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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1372 - (show annotations)
Fri Oct 11 17:05:19 2013 UTC (5 years, 11 months ago) by ph10
File size: 254141 byte(s)
Fix auto-callout with PCRE_UCP, which was compiling rubbish.
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
131
132 Version 8.33 28-May-2013
133 ------------------------
134
135 1. Added 'U' to some constants that are compared to unsigned integers, to
136 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
137 variables that are added to signed variables, to ensure the result is
138 signed and can be negated.
139
140 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
141 CMake config files.
142
143 3. Revise the creation of config.h.generic so that all boolean macros are
144 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
145 overriding via -D on the command line possible.
146
147 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
148 to unsigned int is reported to make a quite noticeable speed difference in
149 a specific Windows environment. Testing on Linux did also appear to show
150 some benefit (and it is clearly not harmful). Also fixed the definition of
151 Xop which should be unsigned.
152
153 5. Related to (4), changing the definition of the intermediate variable cc
154 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
155 improvements.
156
157 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
158 unnecessary spaces.
159
160 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
161 and later.
162
163 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
164
165 9. Optimizing fast_forward_start_bits in JIT.
166
167 10. Adding support for callouts in JIT, and fixing some issues revealed
168 during this work. Namely:
169
170 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
171
172 (b) Minimum length was not checked before the matching is started.
173
174 11. The value of capture_last that is passed to callouts was incorrect in some
175 cases when there was a capture on one path that was subsequently abandoned
176 after a backtrack. Also, the capture_last value is now reset after a
177 recursion, since all captures are also reset in this case.
178
179 12. The interpreter no longer returns the "too many substrings" error in the
180 case when an overflowing capture is in a branch that is subsequently
181 abandoned after a backtrack.
182
183 13. In the pathological case when an offset vector of size 2 is used, pcretest
184 now prints out the matched string after a yield of 0 or 1.
185
186 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
187 Only supported by the JIT compiler at the moment.
188
189 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
190
191 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
192 offset of the starting point of the matching process, provided the offsets
193 vector is large enough.
194
195 17. The \A escape now records a lookbehind value of 1, though its execution
196 does not actually inspect the previous character. This is to ensure that,
197 in partial multi-segment matching, at least one character from the old
198 segment is retained when a new segment is processed. Otherwise, if there
199 are no lookbehinds in the pattern, \A might match incorrectly at the start
200 of a new segment.
201
202 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
203
204 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
205 gives some modest performance improvement in 8-bit mode.
206
207 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
208 be expressed in certain programming languages using Universal Character
209 Names.
210
211 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
212 which points out that "non characters" are not "characters that may not
213 appear in Unicode strings" but rather "characters that are reserved for
214 internal use and have only local meaning".
215
216 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
217 there was a conditional group that depended on an assertion, if the
218 assertion was false, the callout that immediately followed the alternation
219 in the condition was skipped when pcre_exec() was used for matching.
220
221 23. Allow an explicit callout to be inserted before an assertion that is the
222 condition for a conditional group, for compatibility with automatic
223 callouts, which always insert a callout at this point.
224
225 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
226 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
227
228 25. (*PRUNE) is now supported by the JIT compiler.
229
230 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
231
232 27. Fix the case where there are two or more SKIPs with arguments that may be
233 ignored.
234
235 28. (*SKIP) is now supported by the JIT compiler.
236
237 29. (*THEN) is now supported by the JIT compiler.
238
239 30. Update RunTest with additional test selector options.
240
241 31. The way PCRE handles backtracking verbs has been changed in two ways.
242
243 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
244 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
245 backtracking. In some cases this makes it more Perl-compatible, but Perl's
246 rather obscure rules do not always do the same thing.
247
248 (2) Previously, backtracking verbs were confined within assertions. This is
249 no longer the case for positive assertions, except for (*ACCEPT). Again,
250 this sometimes improves Perl compatibility, and sometimes does not.
251
252 32. A number of tests that were in test 2 because Perl did things differently
253 have been moved to test 1, because either Perl or PCRE has changed, and
254 these tests are now compatible.
255
256 32. Backtracking control verbs are now handled in the same way in JIT and
257 interpreter.
258
259 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
260 contained a forward subroutine reference caused a compile error.
261
262 34. Auto-detect and optimize limited repetitions in JIT.
263
264 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
265 blocking (*UTF) etc.
266
267 36. In the interpreter, maximizing pattern repetitions for characters and
268 character types now use tail recursion, which reduces stack usage.
269
270 37. The value of the max lookbehind was not correctly preserved if a compiled
271 and saved regex was reloaded on a host of different endianness.
272
273 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
274 of the compiled pattern block, expand the flags field from 16 to 32 bits
275 because it was almost full.
276
277 39. Try madvise first before posix_madvise.
278
279 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
280 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
281 original change locked out all matches of empty strings. This has been
282 changed so that one match of an empty string per line is recognized.
283 Subsequent searches on the same line (for colouring or for --only-matching,
284 for example) do not recognize empty strings.
285
286 41. Applied a user patch to fix a number of spelling mistakes in comments.
287
288 42. Data lines longer than 65536 caused pcretest to crash.
289
290 43. Clarified the data type for length and startoffset arguments for pcre_exec
291 and pcre_dfa_exec in the function-specific man pages, where they were
292 explicitly stated to be in bytes, never having been updated. I also added
293 some clarification to the pcreapi man page.
294
295 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
296 a segmentation fault.
297
298
299 Version 8.32 30-November-2012
300 -----------------------------
301
302 1. Improved JIT compiler optimizations for first character search and single
303 character iterators.
304
305 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
306 Patch by Daniel Richard G.
307
308 3. Single character iterator optimizations in the JIT compiler.
309
310 4. Improved JIT compiler optimizations for character ranges.
311
312 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
313 Reported by Giuseppe D'Angelo.
314
315 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
316 of a line, was being set incorrectly in cases where .* appeared inside
317 atomic brackets at the start of a pattern, or where there was a subsequent
318 *PRUNE or *SKIP.
319
320 7. Improved instruction cache flush for POWER/PowerPC.
321 Patch by Daniel Richard G.
322
323 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
324 grep:
325
326 (a) There is now no limit to the number of patterns to be matched.
327
328 (b) An error is given if a pattern is too long.
329
330 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
331 are now supported.
332
333 (d) --exclude-from and --include-from (multiple use) have been added.
334
335 (e) Exclusions and inclusions now apply to all files and directories, not
336 just to those obtained from scanning a directory recursively.
337
338 (f) Multiple uses of -f and --file-list are now supported.
339
340 (g) In a Windows environment, the default for -d has been changed from
341 "read" (the GNU grep default) to "skip", because otherwise the presence
342 of a directory in the file list provokes an error.
343
344 (h) The documentation has been revised and clarified in places.
345
346 9. Improve the matching speed of capturing brackets.
347
348 10. Changed the meaning of \X so that it now matches a Unicode extended
349 grapheme cluster.
350
351 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
352 out POSIX threads when JIT support is configured.
353
354 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
355
356 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
357 structure could go wrong in environments where size_t is not the same size
358 as int.
359
360 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
361 set.
362
363 15. The EBCDIC support had decayed; later updates to the code had included
364 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
365 general tidy up of EBCDIC-related issues, and the documentation was also
366 not quite right. There is now a test that can be run on ASCII systems to
367 check some of the EBCDIC-related things (but is it not a full test).
368
369 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
370 in a small tidy to the code.
371
372 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
373
374 18. If the --only-matching (-o) option in pcregrep is specified multiple
375 times, each one causes appropriate output. For example, -o1 -o2 outputs the
376 substrings matched by the 1st and 2nd capturing parentheses. A separating
377 string can be specified by --om-separator (default empty).
378
379 19. Improving the first n character searches.
380
381 20. Turn case lists for horizontal and vertical white space into macros so that
382 they are defined only once.
383
384 21. This set of changes together give more compatible Unicode case-folding
385 behaviour for characters that have more than one other case when UCP
386 support is available.
387
388 (a) The Unicode property table now has offsets into a new table of sets of
389 three or more characters that are case-equivalent. The MultiStage2.py
390 script that generates these tables (the pcre_ucd.c file) now scans
391 CaseFolding.txt instead of UnicodeData.txt for character case
392 information.
393
394 (b) The code for adding characters or ranges of characters to a character
395 class has been abstracted into a generalized function that also handles
396 case-independence. In UTF-mode with UCP support, this uses the new data
397 to handle characters with more than one other case.
398
399 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
400 whose other case is greater than 256 are now correctly matched
401 caselessly. Previously, the high codepoint matched the low one, but not
402 vice versa.
403
404 (d) The processing of \h, \H, \v, and \ in character classes now makes use
405 of the new class addition function, using character lists defined as
406 macros alongside the case definitions of 20 above.
407
408 (e) Caseless back references now work with characters that have more than
409 one other case.
410
411 (f) General caseless matching of characters with more than one other case
412 is supported.
413
414 22. Unicode character properties were updated from Unicode 6.2.0
415
416 23. Improved CMake support under Windows. Patch by Daniel Richard G.
417
418 24. Add support for 32-bit character strings, and UTF-32
419
420 25. Major JIT compiler update (code refactoring and bugfixing).
421 Experimental Sparc 32 support is added.
422
423 26. Applied a modified version of Daniel Richard G's patch to create
424 pcre.h.generic and config.h.generic by "make" instead of in the
425 PrepareRelease script.
426
427 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
428 pcre_compile.c when checking for a zero character.
429
430 28. Introducing a native interface for JIT. Through this interface, the compiled
431 machine code can be directly executed. The purpose of this interface is to
432 provide fast pattern matching, so several sanity checks are not performed.
433 However, feature tests are still performed. The new interface provides
434 1.4x speedup compared to the old one.
435
436 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
437 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
438 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
439
440 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
441 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
442 the "old" RFC 2279). Instead, it ended up passing a negative length to
443 pcre_exec().
444
445 31. Add support for GCC's visibility feature to hide internal functions.
446
447 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
448 "unknown -C option" after outputting 0 or 1.
449
450 33. There is now support for generating a code coverage report for the test
451 suite in environments where gcc is the compiler and lcov is installed. This
452 is mainly for the benefit of the developers.
453
454 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
455 unaddressable using valgrind annotations, allowing valgrind to detect
456 invalid memory accesses. This is mainly for the benefit of the developers.
457
458 25. (*UTF) can now be used to start a pattern in any of the three libraries.
459
460 26. Give configure error if --enable-cpp but no C++ compiler found.
461
462
463 Version 8.31 06-July-2012
464 -------------------------
465
466 1. Fixing a wrong JIT test case and some compiler warnings.
467
468 2. Removed a bashism from the RunTest script.
469
470 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
471 to unsigned type, result still unsigned" that was given by an MS compiler
472 on encountering the code "-sizeof(xxx)".
473
474 4. Partial matching support is added to the JIT compiler.
475
476 5. Fixed several bugs concerned with partial matching of items that consist
477 of more than one character:
478
479 (a) /^(..)\1/ did not partially match "aba" because checking references was
480 done on an "all or nothing" basis. This also applied to repeated
481 references.
482
483 (b) \R did not give a hard partial match if \r was found at the end of the
484 subject.
485
486 (c) \X did not give a hard partial match after matching one or more
487 characters at the end of the subject.
488
489 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
490 a partial match for the string "\r".
491
492 (e) When newline was set to CRLF, the metacharacter "." did not recognize
493 a partial match for a CR character at the end of the subject string.
494
495 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
496 running pcretest, the text "(JIT)" added to the output whenever JIT is
497 actually used to run the match.
498
499 7. Individual JIT compile options can be set in pcretest by following -s+[+]
500 or /S+[+] with a digit between 1 and 7.
501
502 8. OP_NOT now supports any UTF character not just single-byte ones.
503
504 9. (*MARK) control verb is now supported by the JIT compiler.
505
506 10. The command "./RunTest list" lists the available tests without actually
507 running any of them. (Because I keep forgetting what they all are.)
508
509 11. Add PCRE_INFO_MAXLOOKBEHIND.
510
511 12. Applied a (slightly modified) user-supplied patch that improves performance
512 when the heap is used for recursion (compiled with --disable-stack-for-
513 recursion). Instead of malloc and free for each heap frame each time a
514 logical recursion happens, frames are retained on a chain and re-used where
515 possible. This sometimes gives as much as 30% improvement.
516
517 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
518 call.
519
520 14. As documented, (*COMMIT) is now confined to within a positive assertion.
521
522 15. It is now possible to link pcretest with libedit as an alternative to
523 libreadline.
524
525 16. (*COMMIT) control verb is now supported by the JIT compiler.
526
527 17. The Unicode data tables have been updated to Unicode 6.1.0.
528
529 18. Added --file-list option to pcregrep.
530
531 19. Added binary file support to pcregrep, including the -a, --binary-files,
532 -I, and --text options.
533
534 20. The madvise function is renamed for posix_madvise for QNX compatibility
535 reasons. Fixed by Giuseppe D'Angelo.
536
537 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
538
539 22. Changed the option for creating symbolic links for 16-bit man pages from
540 -s to -sf so that re-installing does not cause issues.
541
542 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
543
544 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
545 matches in certain environments (the workspace was not being correctly
546 retained). Also added to pcre_dfa_exec() a simple plausibility check on
547 some of the workspace data at the beginning of a restart.
548
549 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
550 was not doing so when it should - probably a typo introduced by SVN 528
551 (change 8.10/14).
552
553 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
554 \w+ when the character tables indicated that \x{c4} was a word character.
555 There were several related cases, all because the tests for doing a table
556 lookup were testing for characters less than 127 instead of 255.
557
558 27. If a pattern contains capturing parentheses that are not used in a match,
559 their slots in the ovector are set to -1. For those that are higher than
560 any matched groups, this happens at the end of processing. In the case when
561 there were back references that the ovector was too small to contain
562 (causing temporary malloc'd memory to be used during matching), and the
563 highest capturing number was not used, memory off the end of the ovector
564 was incorrectly being set to -1. (It was using the size of the temporary
565 memory instead of the true size.)
566
567 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
568 ovector size, it uses memory at the end of the block that it has got.
569
570 29. Check for an overlong MARK name and give an error at compile time. The
571 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
572
573 30. JIT compiler update.
574
575 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
576 Rill for the patch.
577
578 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
579
580 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
581
582 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
583 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
584 was enabled.
585
586 35. Improve JIT code generation for greedy plus quantifier.
587
588 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
589 1 to "aa" instead of to an empty string. The bug affected repeated groups
590 that could potentially match an empty string.
591
592 37. Optimizing single character iterators in JIT.
593
594 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
595 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
596 codepoints that are too big for the mode are faulted, and in a UTF mode,
597 disallowed codepoints are also faulted.
598
599 39. If PCRE was compiled with UTF support, in three places in the DFA
600 matcher there was code that should only have been obeyed in UTF mode, but
601 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
602 processing when bytes with values greater than 127 were present. In 16-bit
603 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
604 both cases the values are those that cannot be the first data item in a UTF
605 character. The three items that might have provoked this were recursions,
606 possessively repeated groups, and atomic groups.
607
608 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
609 and pcregrep, because some OS require shared objects to be explicitly
610 passed to ld, causing the link step to fail if they are not.
611
612 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
613 mode, patterns that started with \h* or \R* might be incorrectly matched.
614
615
616 Version 8.30 04-February-2012
617 -----------------------------
618
619 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
620 name is defined in ctype.h.
621
622 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
623 only in quite long subpatterns.
624
625 3. Removed the function pcre_info(), which has been obsolete and deprecated
626 since it was replaced by pcre_fullinfo() in February 2000.
627
628 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
629 match a (*MARK), and the match failed at the start of the subject, a
630 reference to memory before the start of the subject could occur. This bug
631 was introduced by fix 17 of release 8.21.
632
633 5. A reference to an unset group with zero minimum repetition was giving
634 totally wrong answers (in non-JavaScript-compatibility mode). For example,
635 /(another)?(\1?)test/ matched against "hello world test". This bug was
636 introduced in release 8.13.
637
638 6. Add support for 16-bit character strings (a large amount of work involving
639 many changes and refactorings).
640
641 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
642 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
643 from a file.
644
645 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
646 rounding is not applied in this particular case).
647
648 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
649 if they appear, or are escaped, in patterns.
650
651 10. Get rid of a number of -Wunused-but-set-variable warnings.
652
653 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
654 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
655 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
656 also returns the mark "x". This bug applied to capturing parentheses,
657 non-capturing parentheses, and atomic parentheses. It also applied to some
658 assertions.
659
660 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
661 information out of configure.ac instead of relying on pcre.h.generic, which
662 is not stored in the repository.
663
664 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
665 -lreadline.
666
667 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
668
669 15. Applied Graycode's patch to put the top-level frame on the stack rather
670 than the heap when not using the stack for recursion. This gives a
671 performance improvement in many cases when recursion is not deep.
672
673 16. Experimental code added to "pcretest -C" to output the stack frame size.
674
675
676 Version 8.21 12-Dec-2011
677 ------------------------
678
679 1. Updating the JIT compiler.
680
681 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
682 are added as well.
683
684 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
685 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
686 calling _pcre_jit_exec. Some extra comments are added.
687
688 4. (*MARK) settings inside atomic groups that do not contain any capturing
689 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
690 was introduced by change 18 for 8.20.
691
692 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
693 ECMA-262 standard.
694
695 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
696 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
697 This bug was probably introduced by change 9 of 8.13.
698
699 7. While fixing 6 above, I noticed that a number of other items were being
700 incorrectly rejected as "not fixed length". This arose partly because newer
701 opcodes had not been added to the fixed-length checking code. I have (a)
702 corrected the bug and added tests for these items, and (b) arranged for an
703 error to occur if an unknown opcode is encountered while checking for fixed
704 length instead of just assuming "not fixed length". The items that were
705 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
706 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
707 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
708
709 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
710 being incorrectly compiled and would have given unpredicatble results.
711
712 9. A possessively repeated subpattern with minimum repeat count greater than
713 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
714 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
715 the first (A) could occur when it should not.
716
717 10. Add a cast and remove a redundant test from the code.
718
719 11. JIT should use pcre_malloc/pcre_free for allocation.
720
721 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
722 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
723 is anything other than /usr, -L is still shown).
724
725 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
726
727 14. Perl does not support \N without a following name in a [] class; PCRE now
728 also gives an error.
729
730 15. If a forward reference was repeated with an upper limit of around 2000,
731 it caused the error "internal error: overran compiling workspace". The
732 maximum number of forward references (including repeats) was limited by the
733 internal workspace, and dependent on the LINK_SIZE. The code has been
734 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
735 the default depends on LINK_SIZE. There is a new upper limit (for safety)
736 of around 200,000 forward references. While doing this, I also speeded up
737 the filling in of repeated forward references.
738
739 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
740 incorrectly expecting the subject to contain another "a" after the start.
741
742 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
743 in the match, the SKIP should be ignored. This was not happening; instead
744 the SKIP was being treated as NOMATCH. For patterns such as
745 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
746 tested.
747
748 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
749 now much more compatible with Perl, in particular in cases where the result
750 is a non-match for a non-anchored pattern. For example, if
751 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
752 "m", where previously it did not return a name. A side effect of this
753 change is that for partial matches, the last encountered mark name is
754 returned, as for non matches. A number of tests that were previously not
755 Perl-compatible have been moved into the Perl-compatible test files. The
756 refactoring has had the pleasing side effect of removing one argument from
757 the match() function, thus reducing its stack requirements.
758
759 19. If the /S+ option was used in pcretest to study a pattern using JIT,
760 subsequent uses of /S (without +) incorrectly behaved like /S+.
761
762 21. Retrieve executable code size support for the JIT compiler and fixing
763 some warnings.
764
765 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
766 not work when the shorter character appeared right at the end of the
767 subject string.
768
769 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
770 systems.
771
772 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
773 output it when the /M option is used in pcretest.
774
775 25. The CheckMan script was not being included in the distribution. Also, added
776 an explicit "perl" to run Perl scripts from the PrepareRelease script
777 because this is reportedly needed in Windows.
778
779 26. If study data was being save in a file and studying had not found a set of
780 "starts with" bytes for the pattern, the data written to the file (though
781 never used) was taken from uninitialized memory and so caused valgrind to
782 complain.
783
784 27. Updated RunTest.bat as provided by Sheri Pierce.
785
786 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
787
788 29. Computation of memory usage for the table of capturing group names was
789 giving an unnecessarily large value.
790
791
792 Version 8.20 21-Oct-2011
793 ------------------------
794
795 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
796 a POSIX class. After further experiments with Perl, which convinced me that
797 Perl has bugs and confusions, a closing square bracket is no longer allowed
798 in a POSIX name. This bug also affected patterns with classes that started
799 with full stops.
800
801 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
802 captured substring, but while checking the failing first alternative,
803 substring 1 is temporarily captured. If the output vector supplied to
804 pcre_exec() was not big enough for this capture, the yield of the function
805 was still zero ("insufficient space for captured substrings"). This cannot
806 be totally fixed without adding another stack variable, which seems a lot
807 of expense for a edge case. However, I have improved the situation in cases
808 such as /(a)(b)x|abc/ matched against "abc", where the return code
809 indicates that fewer than the maximum number of slots in the ovector have
810 been set.
811
812 3. Related to (2) above: when there are more back references in a pattern than
813 slots in the output vector, pcre_exec() uses temporary memory during
814 matching, and copies in the captures as far as possible afterwards. It was
815 using the entire output vector, but this conflicts with the specification
816 that only 2/3 is used for passing back captured substrings. Now it uses
817 only the first 2/3, for compatibility. This is, of course, another edge
818 case.
819
820 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
821 main code base, and can be used by building with --enable-jit. When this is
822 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
823 runtime --no-jit option is given.
824
825 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
826 ovector, the return from the function was zero, implying that there were
827 other matches that did not fit. The correct "exactly full" value is now
828 returned.
829
830 6. If a subpattern that was called recursively or as a subroutine contained
831 (*PRUNE) or any other control that caused it to give a non-standard return,
832 invalid errors such as "Error -26 (nested recursion at the same subject
833 position)" or even infinite loops could occur.
834
835 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
836 computing the minimum length on reaching *ACCEPT, and so ended up with the
837 wrong value of 1 rather than 0. Further investigation indicates that
838 computing a minimum subject length in the presence of *ACCEPT is difficult
839 (think back references, subroutine calls), and so I have changed the code
840 so that no minimum is registered for a pattern that contains *ACCEPT.
841
842 8. If (*THEN) was present in the first (true) branch of a conditional group,
843 it was not handled as intended. [But see 16 below.]
844
845 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
846 Sheri Pierce.
847
848 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
849 the first byte in a match must be "a".
850
851 11. Change 17 for 8.13 increased the recursion depth for patterns like
852 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
853 pattern contains any instances of (*THEN). If it does not, the old
854 optimizations are restored. It would be nice to do this on a per-group
855 basis, but at the moment that is not feasible.
856
857 12. In some environments, the output of pcretest -C is CRLF terminated. This
858 broke RunTest's code that checks for the link size. A single white space
859 character after the value is now allowed for.
860
861 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
862 For "fr", it uses the Windows-specific input and output files.
863
864 14. If (*THEN) appeared in a group that was called recursively or as a
865 subroutine, it did not work as intended. [But see next item.]
866
867 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
868 pattern fragments (but not containing any | characters). If A and B are
869 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
870 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
871 D. In other words, Perl considers parentheses that do not contain any |
872 characters to be part of a surrounding alternative, whereas PCRE was
873 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
874 differently. PCRE now behaves in the same way as Perl, except in the case
875 of subroutine/recursion calls such as (?1) which have in any case always
876 been different (but PCRE had them first :-).
877
878 16. Related to 15 above: Perl does not treat the | in a conditional group as
879 creating alternatives. Such a group is treated in the same way as an
880 ordinary group without any | characters when processing (*THEN). PCRE has
881 been changed to match Perl's behaviour.
882
883 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
884 RunGrepTest script failed.
885
886 18. Change 22 for version 13 caused atomic groups to use more stack. This is
887 inevitable for groups that contain captures, but it can lead to a lot of
888 stack use in large patterns. The old behaviour has been restored for atomic
889 groups that do not contain any capturing parentheses.
890
891 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
892 suppress the check for a minimum subject length at run time. (If it was
893 given to pcre_exec() or pcre_dfa_exec() it did work.)
894
895 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
896 fail to work when decoding hex characters in data strings in EBCDIC
897 environments.
898
899 21. It appears that in at least one Mac OS environment, the isxdigit() function
900 is implemented as a macro that evaluates to its argument more than once,
901 contravening the C 90 Standard (I haven't checked a later standard). There
902 was an instance in pcretest which caused it to go wrong when processing
903 \x{...} escapes in subject strings. The has been rewritten to avoid using
904 things like p++ in the argument of isxdigit().
905
906
907 Version 8.13 16-Aug-2011
908 ------------------------
909
910 1. The Unicode data tables have been updated to Unicode 6.0.0.
911
912 2. Two minor typos in pcre_internal.h have been fixed.
913
914 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
915 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
916 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
917
918 4. There were a number of related bugs in the code for matching backrefences
919 caselessly in UTF-8 mode when codes for the characters concerned were
920 different numbers of bytes. For example, U+023A and U+2C65 are an upper
921 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
922 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
923 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
924 2-byte code at the end of the subject (it thought there wasn't enough data
925 left).
926
927 5. Comprehensive information about what went wrong is now returned by
928 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
929 as the output vector has at least 2 elements. The offset of the start of
930 the failing character and a reason code are placed in the vector.
931
932 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
933 now returned is for the first byte of the failing character, instead of the
934 last byte inspected. This is an incompatible change, but I hope it is small
935 enough not to be a problem. It makes the returned offset consistent with
936 pcre_exec() and pcre_dfa_exec().
937
938 7. pcretest now gives a text phrase as well as the error number when
939 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
940 failure, the offset and reason code are output.
941
942 8. When \R was used with a maximizing quantifier it failed to skip backwards
943 over a \r\n pair if the subsequent match failed. Instead, it just skipped
944 back over a single character (\n). This seems wrong (because it treated the
945 two characters as a single entity when going forwards), conflicts with the
946 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
947 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
948 has been changed.
949
950 9. Some internal refactoring has changed the processing so that the handling
951 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
952 time (the PCRE_DOTALL option was changed this way some time ago: version
953 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
954 which was always a bit of a fudge. It also means that there is one less
955 argument for the match() function, which reduces its stack requirements
956 slightly. This change also fixes an incompatibility with Perl: the pattern
957 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
958
959 10. More internal refactoring has drastically reduced the number of recursive
960 calls to match() for possessively repeated groups such as (abc)++ when
961 using pcre_exec().
962
963 11. While implementing 10, a number of bugs in the handling of groups were
964 discovered and fixed:
965
966 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
967 (a|)*(?1) gave a compile-time internal error.
968 ((a|)+)+ did not notice that the outer group could match an empty string.
969 (^a|^)+ was not marked as anchored.
970 (.*a|.*)+ was not marked as matching at start or after a newline.
971
972 12. Yet more internal refactoring has removed another argument from the match()
973 function. Special calls to this function are now indicated by setting a
974 value in a variable in the "match data" data block.
975
976 13. Be more explicit in pcre_study() instead of relying on "default" for
977 opcodes that mean there is no starting character; this means that when new
978 ones are added and accidentally left out of pcre_study(), testing should
979 pick them up.
980
981 14. The -s option of pcretest has been documented for ages as being an old
982 synonym of -m (show memory usage). I have changed it to mean "force study
983 for every regex", that is, assume /S for every regex. This is similar to -i
984 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
985 using it. It makes it easier to run collections of tests with and without
986 study enabled, and thereby test pcre_study() more easily. All the standard
987 tests are now run with and without -s (but some patterns can be marked as
988 "never study" - see 20 below).
989
990 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
991 restoration of the capturing data to the outer values was not happening
992 correctly.
993
994 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
995 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
996 pattern had matched an empty string, and so incorrectly returned a no
997 match.
998
999 17. There was optimizing code for the last branch of non-capturing parentheses,
1000 and also for the obeyed branch of a conditional subexpression, which used
1001 tail recursion to cut down on stack usage. Unfortunately, now that there is
1002 the possibility of (*THEN) occurring in these branches, tail recursion is
1003 no longer possible because the return has to be checked for (*THEN). These
1004 two optimizations have therefore been removed. [But see 8.20/11 above.]
1005
1006 18. If a pattern containing \R was studied, it was assumed that \R always
1007 matched two bytes, thus causing the minimum subject length to be
1008 incorrectly computed because \R can also match just one byte.
1009
1010 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1011 was incorrectly computed.
1012
1013 20. If /S is present twice on a test pattern in pcretest input, it now
1014 *disables* studying, thereby overriding the use of -s on the command line
1015 (see 14 above). This is necessary for one or two tests to keep the output
1016 identical in both cases.
1017
1018 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1019 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1020
1021 22. When an atomic group that contained a capturing parenthesis was
1022 successfully matched, but the branch in which it appeared failed, the
1023 capturing was not being forgotten if a higher numbered group was later
1024 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1025 group 1 to "a", when in fact it should be unset. This applied to multi-
1026 branched capturing and non-capturing groups, repeated or not, and also to
1027 positive assertions (capturing in negative assertions does not happen
1028 in PCRE) and also to nested atomic groups.
1029
1030 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1031 subject after a captured substring, to make it easier to tell which of a
1032 number of identical substrings has been captured.
1033
1034 24. The way atomic groups are processed by pcre_exec() has been changed so that
1035 if they are repeated, backtracking one repetition now resets captured
1036 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1037 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1038 "aaa". Previously, it would have been "a". As part of this code
1039 refactoring, the way recursive calls are handled has also been changed.
1040
1041 25. If an assertion condition captured any substrings, they were not passed
1042 back unless some other capturing happened later. For example, if
1043 (?(?=(a))a) was matched against "a", no capturing was returned.
1044
1045 26. When studying a pattern that contained subroutine calls or assertions,
1046 the code for finding the minimum length of a possible match was handling
1047 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1048 group 1 called group 2 while simultaneously a separate group 2 called group
1049 1). A stack overflow occurred in this case. I have fixed this by limiting
1050 the recursion depth to 10.
1051
1052 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1053 Fortmann. This supports explicit test numbers on the command line, and has
1054 argument validation and error reporting.
1055
1056 28. An instance of \X with an unlimited repeat could fail if at any point the
1057 first character it looked at was a mark character.
1058
1059 29. Some minor code refactoring concerning Unicode properties and scripts
1060 should reduce the stack requirement of match() slightly.
1061
1062 30. Added the '=' option to pcretest to check the setting of unused capturing
1063 slots at the end of the pattern, which are documented as being -1, but are
1064 not included in the return count.
1065
1066 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1067 compiled something random. Now it gives a compile-time error (as does
1068 Perl).
1069
1070 32. A *MARK encountered during the processing of a positive assertion is now
1071 recorded and passed back (compatible with Perl).
1072
1073 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1074 had alternative anchored branches, the search for a second match in a line
1075 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1076 matched the line "0102" twice. The same bug affected patterns that started
1077 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1078 twice.
1079
1080 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1081 does, and because of capturing effects, quantifying parenthesized
1082 assertions may at times be useful. Quantifiers are now allowed for
1083 parenthesized assertions.
1084
1085 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1086
1087 36. \g was being checked for fancy things in a character class, when it should
1088 just be a literal "g".
1089
1090 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1091 appearance of a nested POSIX class supersedes an apparent external class.
1092 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1093 unescaped square brackets may also appear as part of class names. For
1094 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1095 more like Perl. (But see 8.20/1 above.)
1096
1097 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1098 was because it thought it was \N{name}, which is not supported).
1099
1100 39. Add minix to OS list not supporting the -S option in pcretest.
1101
1102 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1103 cannot analyze patterns in sufficient detail to catch mutual recursions
1104 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1105 subgroup is called recursively as a subpattern for a second time at the
1106 same position in the subject string. In previous releases this might have
1107 been caught by the recursion limit, or it might have run out of stack.
1108
1109 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1110 happen only once. PCRE was, however incorrectly giving a compile time error
1111 "recursive call could loop indefinitely" because it cannot analyze the
1112 pattern in sufficient detail. The compile time test no longer happens when
1113 PCRE is compiling a conditional subpattern, but actual runaway loops are
1114 now caught at runtime (see 40 above).
1115
1116 42. It seems that Perl allows any characters other than a closing parenthesis
1117 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1118 has been changed to be the same.
1119
1120 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1121 as not to get warnings when autogen.sh is called. Also changed
1122 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1123
1124 44. To help people who use pcregrep to scan files containing exceedingly long
1125 lines, the following changes have been made:
1126
1127 (a) The default value of the buffer size parameter has been increased from
1128 8K to 20K. (The actual buffer used is three times this size.)
1129
1130 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1131 PCRE is built.
1132
1133 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1134 to be set at run time.
1135
1136 (d) Numerical values in pcregrep options can be followed by K or M, for
1137 example --buffer-size=50K.
1138
1139 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1140 given and the return code is set to 2.
1141
1142 45. Add a pointer to the latest mark to the callout data block.
1143
1144 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1145 partial match of an empty string instead of no match. This was specific to
1146 the use of ".".
1147
1148 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1149 complete match instead of a partial match. This bug was dependent on both
1150 the PCRE_UTF8 and PCRE_DOTALL options being set.
1151
1152 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1153 starting byte set, because \b was not being ignored.
1154
1155
1156 Version 8.12 15-Jan-2011
1157 ------------------------
1158
1159 1. Fixed some typos in the markup of the man pages, and wrote a script that
1160 checks for such things as part of the documentation building process.
1161
1162 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1163 --match-limit and --recursion-limit options (added for 8.11). In
1164 particular, this made one of the standard tests fail. (The integer value
1165 went into the wrong half of a long int.)
1166
1167 3. If the --colour option was given to pcregrep with -v (invert match), it
1168 did strange things, either producing crazy output, or crashing. It should,
1169 of course, ignore a request for colour when reporting lines that do not
1170 match.
1171
1172 4. Another pcregrep bug caused similar problems if --colour was specified with
1173 -M (multiline) and the pattern match finished with a line ending.
1174
1175 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1176 matched in multiline mode, the following line was shown as part of the
1177 match. This seems wrong, so I have changed it.
1178
1179 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1180 the check for further matches in the same line (so they could be coloured)
1181 to overrun the end of the current line. If another match was found, it was
1182 incorrectly shown (and then shown again when found in the next line).
1183
1184 7. If pcregrep was compiled under Windows, there was a reference to the
1185 function pcregrep_exit() before it was defined. I am assuming this was
1186 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1187 reported by a user. I've moved the definition above the reference.
1188
1189
1190 Version 8.11 10-Dec-2010
1191 ------------------------
1192
1193 1. (*THEN) was not working properly if there were untried alternatives prior
1194 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1195 backtracked to try for "b" instead of moving to the next alternative branch
1196 at the same level (in this case, to look for "c"). The Perl documentation
1197 is clear that when (*THEN) is backtracked onto, it goes to the "next
1198 alternative in the innermost enclosing group".
1199
1200 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1201 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1202 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1203 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1204 (*THEN).
1205
1206 3. If \s appeared in a character class, it removed the VT character from
1207 the class, even if it had been included by some previous item, for example
1208 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1209 of \s, but is part of the POSIX "space" class.)
1210
1211 4. A partial match never returns an empty string (because you can always
1212 match an empty string at the end of the subject); however the checking for
1213 an empty string was starting at the "start of match" point. This has been
1214 changed to the "earliest inspected character" point, because the returned
1215 data for a partial match starts at this character. This means that, for
1216 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1217 (previously it gave "no match").
1218
1219 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1220 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1221 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1222 has an implication that the given string is incomplete (because a partial
1223 match is preferred over a full match). For this reason, these items now
1224 give a partial match in this situation. [Aside: previously, the one case
1225 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1226 match rather than a full match, which was wrong by the old rules, but is
1227 now correct.]
1228
1229 6. There was a bug in the handling of #-introduced comments, recognized when
1230 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1231 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1232 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1233 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1234 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1235 places in pcre_compile().
1236
1237 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1238 comments when looking ahead for named forward references to subpatterns,
1239 the only newline sequence it recognized was NL. It now handles newlines
1240 according to the set newline convention.
1241
1242 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1243 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1244 cater for a lack of strerror(). These oversights have been fixed.
1245
1246 9. Added --match-limit and --recursion-limit to pcregrep.
1247
1248 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1249
1250 11. When the -o option was used, pcregrep was setting a return code of 1, even
1251 when matches were found, and --line-buffered was not being honoured.
1252
1253 12. Added an optional parentheses number to the -o and --only-matching options
1254 of pcregrep.
1255
1256 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1257 can match an empty string. The code to do it in pcretest and pcredemo
1258 needed fixing:
1259
1260 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1261 only one byte after an empty string match just before CRLF (this case
1262 just got forgotten; "any" and "anycrlf" were OK).
1263
1264 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1265 mode when an empty string match preceded an ASCII character followed by
1266 a non-ASCII character. (The code for advancing by one character rather
1267 than one byte was nonsense.)
1268
1269 (c) The pcredemo.c sample program did not have any code at all to handle
1270 the cases when CRLF is a valid newline sequence.
1271
1272 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1273 as a starting offset was within the subject string. There is now a new
1274 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1275 negative or greater than the length of the string. In order to test this,
1276 pcretest is extended to allow the setting of negative starting offsets.
1277
1278 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1279 starting offset points to the beginning of a UTF-8 character was
1280 unnecessarily clumsy. I tidied it up.
1281
1282 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1283 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1284
1285 17. Nobody had reported that the --include_dir option, which was added in
1286 release 7.7 should have been called --include-dir (hyphen, not underscore)
1287 for compatibility with GNU grep. I have changed it to --include-dir, but
1288 left --include_dir as an undocumented synonym, and the same for
1289 --exclude-dir, though that is not available in GNU grep, at least as of
1290 release 2.5.4.
1291
1292 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1293 characters from a string of bytes) have been redefined so as not to use
1294 loops, in order to improve performance in some environments. At the same
1295 time, I abstracted some of the common code into auxiliary macros to save
1296 repetition (this should not affect the compiled code).
1297
1298 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1299 compile-time error is now given if \c is not followed by an ASCII
1300 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1301 different, and any byte value is allowed.)
1302
1303 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1304 START_OPTIMIZE option, which is now allowed at compile time - but just
1305 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1306 to pcregrep and other applications that have no direct access to PCRE
1307 options. The new /Y option in pcretest sets this option when calling
1308 pcre_compile().
1309
1310 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1311 back references. Groups containing recursive back references were forced to
1312 be atomic by that change, but in the case of named groups, the amount of
1313 memory required was incorrectly computed, leading to "Failed: internal
1314 error: code overflow". This has been fixed.
1315
1316 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1317 pcretest.c, to avoid build problems in some Borland environments.
1318
1319
1320 Version 8.10 25-Jun-2010
1321 ------------------------
1322
1323 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1324 THEN.
1325
1326 2. (*ACCEPT) was not working when inside an atomic group.
1327
1328 3. Inside a character class, \B is treated as a literal by default, but
1329 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1330 causes the error). The code is unchanged, but I tidied the documentation.
1331
1332 4. Inside a character class, PCRE always treated \R and \X as literals,
1333 whereas Perl faults them if its -w option is set. I have changed PCRE so
1334 that it faults them when PCRE_EXTRA is set.
1335
1336 5. Added support for \N, which always matches any character other than
1337 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1338
1339 6. When compiling pcregrep with newer versions of gcc which may have
1340 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1341 declared with attribute warn_unused_result" were given. Just casting the
1342 result to (void) does not stop the warnings; a more elaborate fudge is
1343 needed. I've used a macro to implement this.
1344
1345 7. Minor change to pcretest.c to avoid a compiler warning.
1346
1347 8. Added four artifical Unicode properties to help with an option to make
1348 \s etc use properties (see next item). The new properties are: Xan
1349 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1350
1351 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1352 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1353 this option. Modified pcretest to add /W to test this facility. Added
1354 REG_UCP to make it available via the POSIX interface.
1355
1356 10. Added --line-buffered to pcregrep.
1357
1358 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1359 studied, and the match started with a letter with a code point greater than
1360 127 whose first byte was different to the first byte of the other case of
1361 the letter, the other case of this starting letter was not recognized
1362 (#976).
1363
1364 12. If a pattern that was studied started with a repeated Unicode property
1365 test, for example, \p{Nd}+, there was the theoretical possibility of
1366 setting up an incorrect bitmap of starting bytes, but fortunately it could
1367 not have actually happened in practice until change 8 above was made (it
1368 added property types that matched character-matching opcodes).
1369
1370 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1371 possible starting bytes for non-anchored patterns.
1372
1373 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1374 \R, and also a number of cases that involve Unicode properties, both
1375 explicit and implicit when PCRE_UCP is set.
1376
1377 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1378 input, it could crash or give wrong results if characters with values
1379 greater than 0xc0 were present in the subject string. (Detail: it assumed
1380 UTF-8 input when processing these items.)
1381
1382 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1383 size_t is 64-bit (#991).
1384
1385 17. Added a check for running out of memory when PCRE is compiled with
1386 --disable-stack-for-recursion (#990).
1387
1388 18. If the last data line in a file for pcretest does not have a newline on
1389 the end, a newline was missing in the output.
1390
1391 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1392 less than 128) in its various bitmaps. However, there is a facility for
1393 generating tables according to the current locale when PCRE is compiled. It
1394 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1395 characters, are recognized by isspace() and therefore were getting set in
1396 these tables, and indeed these tables seem to approximate to ISO 8859. This
1397 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1398 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1399 which of course cannot start UTF-8 characters. I have changed the code so
1400 that only real ASCII characters (less than 128) and the correct starting
1401 bytes for UTF-8 encodings are set for characters greater than 127 when in
1402 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1403 altogether.)
1404
1405 20. Added the /T option to pcretest so as to be able to run tests with non-
1406 standard character tables, thus making it possible to include the tests
1407 used for 19 above in the standard set of tests.
1408
1409 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1410 reference to a subpattern the other side of a comment that contains an
1411 opening parenthesis caused either an internal compiling error, or a
1412 reference to the wrong subpattern.
1413
1414
1415 Version 8.02 19-Mar-2010
1416 ------------------------
1417
1418 1. The Unicode data tables have been updated to Unicode 5.2.0.
1419
1420 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1421 configured.
1422
1423 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1424 original author of that file, following a query about its status.
1425
1426 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1427 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1428
1429 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1430 quantifier applied to a forward-referencing subroutine call, could compile
1431 incorrect code or give the error "internal error: previously-checked
1432 referenced subpattern not found".
1433
1434 6. Both MS Visual Studio and Symbian OS have problems with initializing
1435 variables to point to external functions. For these systems, therefore,
1436 pcre_malloc etc. are now initialized to local functions that call the
1437 relevant global functions.
1438
1439 7. There were two entries missing in the vectors called coptable and poptable
1440 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1441 I've fixed the data, and added a kludgy way of testing at compile time that
1442 the lengths are correct (equal to the number of opcodes).
1443
1444 8. Following on from 7, I added a similar kludge to check the length of the
1445 eint vector in pcreposix.c.
1446
1447 9. Error texts for pcre_compile() are held as one long string to avoid too
1448 much relocation at load time. To find a text, the string is searched,
1449 counting zeros. There was no check for running off the end of the string,
1450 which could happen if a new error number was added without updating the
1451 string.
1452
1453 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1454
1455 11. \K was not working if it appeared in an atomic group or in a group that
1456 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1457 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1458 the assertion is positive, but not if it is negative.
1459
1460 12. Change 11 fortuitously reduced the size of the stack frame used in the
1461 "match()" function of pcre_exec.c by one pointer. Forthcoming
1462 implementation of support for (*MARK) will need an extra pointer on the
1463 stack; I have reserved it now, so that the stack frame size does not
1464 decrease.
1465
1466 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1467 item in branch that calls a recursion is a subroutine call - as in the
1468 second branch in the above example - was incorrectly given the compile-
1469 time error "recursive call could loop indefinitely" because pcre_compile()
1470 was not correctly checking the subroutine for matching a non-empty string.
1471
1472 14. The checks for overrunning compiling workspace could trigger after an
1473 overrun had occurred. This is a "should never occur" error, but it can be
1474 triggered by pathological patterns such as hundreds of nested parentheses.
1475 The checks now trigger 100 bytes before the end of the workspace.
1476
1477 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1478
1479
1480 Version 8.01 19-Jan-2010
1481 ------------------------
1482
1483 1. If a pattern contained a conditional subpattern with only one branch (in
1484 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1485 computed the wrong minimum data length (which is of course zero for such
1486 subpatterns). This could cause incorrect "no match" results.
1487
1488 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1489 the pattern is reset in the first branch, pcre_compile() failed with
1490 "internal error: code overflow at offset...". This happened only when
1491 the reset was to the original external option setting. (An optimization
1492 abstracts leading options settings into an external setting, which was the
1493 cause of this.)
1494
1495 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1496 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1497 assertion pattern did not match (meaning that the assertion was true), it
1498 was incorrectly treated as false if the SKIP had been reached during the
1499 matching. This also applied to assertions used as conditions.
1500
1501 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1502 assertion subpattern, including such a pattern used as a condition,
1503 unpredictable results occurred, instead of the error return
1504 PCRE_ERROR_DFA_UITEM.
1505
1506 5. The C++ GlobalReplace function was not working like Perl for the special
1507 situation when an empty string is matched. It now does the fancy magic
1508 stuff that is necessary.
1509
1510 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1511 removed. (These were left over from very, very early versions of PCRE.)
1512
1513 7. Some cosmetic changes to the code to make life easier when compiling it
1514 as part of something else:
1515
1516 (a) Change DEBUG to PCRE_DEBUG.
1517
1518 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1519 called "current" as "current_branch", to prevent a collision with the
1520 Linux macro when compiled as a kernel module.
1521
1522 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1523 prevent a collision with the Linux macro when compiled as a kernel
1524 module.
1525
1526 8. In pcre_compile() there are some checks for integer overflows that used to
1527 cast potentially large values to (double). This has been changed to that
1528 when building, a check for int64_t is made, and if it is found, it is used
1529 instead, thus avoiding the use of floating point arithmetic. (There is no
1530 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1531 double.
1532
1533 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1534 2005 (difference between two addresses compared to an unsigned value).
1535
1536 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1537 custom one, because of the following reported problem in Windows:
1538
1539 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1540 under Win32.
1541 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1542 therefore missing the function definition.
1543 - The compiler thus generates a "C" signature for the test function.
1544 - The linker fails to find the "C" function.
1545 - PCRE fails to configure if asked to do so against libbz2.
1546
1547 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1548 messages were output:
1549
1550 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1551 rerunning libtoolize, to keep the correct libtool macros in-tree.
1552 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1553
1554 I have done both of these things.
1555
1556 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1557 most of the time, it *can* run out if it is given a pattern that contains a
1558 runaway infinite recursion. I updated the discussion in the pcrestack man
1559 page.
1560
1561 13. Now that we have gone to the x.xx style of version numbers, the minor
1562 version may start with zero. Using 08 or 09 is a bad idea because users
1563 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1564 interpreted as invalid octal numbers. I've updated the previous comment in
1565 configure.ac, and also added a check that gives an error if 08 or 09 are
1566 used.
1567
1568 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1569 causing partial matching to fail when the end of the subject matched \W
1570 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1571
1572 15. There were some discrepancies between the declarations in pcre_internal.h
1573 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1574 their definitions. The declarations used "const uschar *" and the
1575 definitions used USPTR. Even though USPTR is normally defined as "const
1576 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1577 reported that: "This difference in casting confuses some C++ compilers, for
1578 example, SunCC recognizes above declarations as different functions and
1579 generates broken code for hbpcre." I have changed the declarations to use
1580 USPTR.
1581
1582 16. GNU libtool is named differently on some systems. The autogen.sh script now
1583 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1584 (FreeBSD).
1585
1586 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1587 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1588 comment: "Figure out how to create a longlong from a string: strtoll and
1589 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1590 instance, but it only takes 2 args instead of 3!"
1591
1592 18. A subtle bug concerned with back references has been fixed by a change of
1593 specification, with a corresponding code fix. A pattern such as
1594 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1595 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1596 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1597 same bug. Such groups have to be quantified to be useful, or contained
1598 inside another quantified group. (If there's no repetition, the reference
1599 can never match.) The problem arises because, having left the group and
1600 moved on to the rest of the pattern, a later failure that backtracks into
1601 the group uses the captured value from the final iteration of the group
1602 rather than the correct earlier one. I have fixed this in PCRE by forcing
1603 any group that contains a reference to itself to be an atomic group; that
1604 is, there cannot be any backtracking into it once it has completed. This is
1605 similar to recursive and subroutine calls.
1606
1607
1608 Version 8.00 19-Oct-09
1609 ----------------------
1610
1611 1. The table for translating pcre_compile() error codes into POSIX error codes
1612 was out-of-date, and there was no check on the pcre_compile() error code
1613 being within the table. This could lead to an OK return being given in
1614 error.
1615
1616 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1617 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1618 in a Windows environment.
1619
1620 3. The pcregrep --count option prints the count for each file even when it is
1621 zero, as does GNU grep. However, pcregrep was also printing all files when
1622 --files-with-matches was added. Now, when both options are given, it prints
1623 counts only for those files that have at least one match. (GNU grep just
1624 prints the file name in this circumstance, but including the count seems
1625 more useful - otherwise, why use --count?) Also ensured that the
1626 combination -clh just lists non-zero counts, with no names.
1627
1628 4. The long form of the pcregrep -F option was incorrectly implemented as
1629 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1630 but it seems right to fix it, and I didn't think it was worth preserving
1631 the old behaviour.
1632
1633 5. The command line items --regex=pattern and --regexp=pattern were not
1634 recognized by pcregrep, which required --regex pattern or --regexp pattern
1635 (with a space rather than an '='). The man page documented the '=' forms,
1636 which are compatible with GNU grep; these now work.
1637
1638 6. No libpcreposix.pc file was created for pkg-config; there was just
1639 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1640
1641 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1642 when UCP support is not needed, by modifying the Python script that
1643 generates it from Unicode data files. This should not matter if the module
1644 is correctly used as a library, but I received one complaint about 50K of
1645 unwanted data. My guess is that the person linked everything into his
1646 program rather than using a library. Anyway, it does no harm.
1647
1648 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1649 was a minimum greater than 1 for a wide character in a possessive
1650 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1651 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1652 character. Chaos in the form of incorrect output or a compiling loop could
1653 result.
1654
1655 9. The restrictions on what a pattern can contain when partial matching is
1656 requested for pcre_exec() have been removed. All patterns can now be
1657 partially matched by this function. In addition, if there are at least two
1658 slots in the offset vector, the offset of the earliest inspected character
1659 for the match and the offset of the end of the subject are set in them when
1660 PCRE_ERROR_PARTIAL is returned.
1661
1662 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1663 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1664 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1665 and may be more useful for multi-segment matching.
1666
1667 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1668 used to be given if ever the end of the subject was reached; now it is
1669 given only if matching could not proceed because another character was
1670 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1671 string "Z", which now yields "no match" instead of "partial match". In the
1672 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1673 final character ended with (*FAIL).
1674
1675 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1676 if the pattern had a "must contain" character that was already found in the
1677 earlier partial match, unless partial matching was again requested. For
1678 example, with the pattern /dog.(body)?/, the "must contain" character is
1679 "g". If the first part-match was for the string "dog", restarting with
1680 "sbody" failed. This bug has been fixed.
1681
1682 13. The string returned by pcre_dfa_exec() after a partial match has been
1683 changed so that it starts at the first inspected character rather than the
1684 first character of the match. This makes a difference only if the pattern
1685 starts with a lookbehind assertion or \b or \B (\K is not supported by
1686 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1687 matching functions compatible, and I think it's the right thing to do.
1688
1689 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1690 so that the demonstration program is easily available in environments where
1691 PCRE has not been installed from source.
1692
1693 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1694 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1695 library.
1696
1697 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1698 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1699 is not the first non-POSIX option to be added. Clearly some people find
1700 these options useful.
1701
1702 17. If a caller to the POSIX matching function regexec() passes a non-zero
1703 value for nmatch with a NULL value for pmatch, the value of
1704 nmatch is forced to zero.
1705
1706 18. RunGrepTest did not have a test for the availability of the -u option of
1707 the diff command, as RunTest does. It now checks in the same way as
1708 RunTest, and also checks for the -b option.
1709
1710 19. If an odd number of negated classes containing just a single character
1711 interposed, within parentheses, between a forward reference to a named
1712 subpattern and the definition of the subpattern, compilation crashed with
1713 an internal error, complaining that it could not find the referenced
1714 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1715 [The bug was that it was starting one character too far in when skipping
1716 over the character class, thus treating the ] as data rather than
1717 terminating the class. This meant it could skip too much.]
1718
1719 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1720 /g option in pcretest when the pattern contains \K, which makes it possible
1721 to have an empty string match not at the start, even when the pattern is
1722 anchored. Updated pcretest and pcredemo to use this option.
1723
1724 21. If the maximum number of capturing subpatterns in a recursion was greater
1725 than the maximum at the outer level, the higher number was returned, but
1726 with unset values at the outer level. The correct (outer level) value is
1727 now given.
1728
1729 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1730 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1731 make it do so. The string so far is captured, making this feature
1732 compatible with Perl.
1733
1734 23. The tests have been re-organized, adding tests 11 and 12, to make it
1735 possible to check the Perl 5.10 features against Perl 5.10.
1736
1737 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1738 pattern matches a fixed length string. PCRE did not allow this; now it
1739 does. Neither allows recursion.
1740
1741 25. I finally figured out how to implement a request to provide the minimum
1742 length of subject string that was needed in order to match a given pattern.
1743 (It was back references and recursion that I had previously got hung up
1744 on.) This code has now been added to pcre_study(); it finds a lower bound
1745 to the length of subject needed. It is not necessarily the greatest lower
1746 bound, but using it to avoid searching strings that are too short does give
1747 some useful speed-ups. The value is available to calling programs via
1748 pcre_fullinfo().
1749
1750 26. While implementing 25, I discovered to my embarrassment that pcretest had
1751 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1752 study optimizations had never been tested with that matching function.
1753 Oops. What is worse, even when it was passed study data, there was a bug in
1754 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1755 were also very few tests of studied patterns with pcre_dfa_exec().
1756
1757 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1758 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1759 on the other side of the coin, they are no longer allowed to have different
1760 names, because these cannot be distinguished in PCRE, and this has caused
1761 confusion. (This is a difference from Perl.)
1762
1763 28. When duplicate subpattern names are present (necessarily with different
1764 numbers, as required by 27 above), and a test is made by name in a
1765 conditional pattern, either for a subpattern having been matched, or for
1766 recursion in such a pattern, all the associated numbered subpatterns are
1767 tested, and the overall condition is true if the condition is true for any
1768 one of them. This is the way Perl works, and is also more like the way
1769 testing by number works.
1770
1771
1772 Version 7.9 11-Apr-09
1773 ---------------------
1774
1775 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1776 (pcretest), all targets were linked against these libraries. This included
1777 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1778 libraries. This caused unwanted dependencies to be created. This problem
1779 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1780 pcretest is linked with readline.
1781
1782 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1783 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1784 moved outside it again, because FALSE and TRUE are already defined in AIX,
1785 but BOOL is not.
1786
1787 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1788 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1789
1790 4. The pcregrep documentation said spaces were inserted as well as colons (or
1791 hyphens) following file names and line numbers when outputting matching
1792 lines. This is not true; no spaces are inserted. I have also clarified the
1793 wording for the --colour (or --color) option.
1794
1795 5. In pcregrep, when --colour was used with -o, the list of matching strings
1796 was not coloured; this is different to GNU grep, so I have changed it to be
1797 the same.
1798
1799 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1800 each matching line was coloured. Now it goes on to look for further matches
1801 of any of the test patterns, which is the same behaviour as GNU grep.
1802
1803 7. A pattern that could match an empty string could cause pcregrep to loop; it
1804 doesn't make sense to accept an empty string match in pcregrep, so I have
1805 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1806 seems to be how GNU grep behaves. [But see later change 40 for release
1807 8.33.]
1808
1809 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1810 start or after a newline", because the conditional assertion was not being
1811 correctly handled. The rule now is that both the assertion and what follows
1812 in the first alternative must satisfy the test.
1813
1814 9. If auto-callout was enabled in a pattern with a conditional group whose
1815 condition was an assertion, PCRE could crash during matching, both with
1816 pcre_exec() and pcre_dfa_exec().
1817
1818 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1819 used for matching.
1820
1821 11. Unicode property support in character classes was not working for
1822 characters (bytes) greater than 127 when not in UTF-8 mode.
1823
1824 12. Added the -M command line option to pcretest.
1825
1826 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1827
1828 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1829
1830 16. Added comments and documentation about mis-use of no_arg in the C++
1831 wrapper.
1832
1833 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1834 from Martin Jerabek that uses macro names for all relevant character and
1835 string constants.
1836
1837 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1838 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1839 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1840 these, but not everybody uses configure.
1841
1842 19. A conditional group that had only one branch was not being correctly
1843 recognized as an item that could match an empty string. This meant that an
1844 enclosing group might also not be so recognized, causing infinite looping
1845 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1846 with the subject "ab", where knowledge that the repeated group can match
1847 nothing is needed in order to break the loop.
1848
1849 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1850 exec(), but without supplying a callout function, matching went wrong.
1851
1852 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1853 leak if the size of the offset vector was greater than 30. When the vector
1854 is smaller, the saved offsets during recursion go onto a local stack
1855 vector, but for larger vectors malloc() is used. It was failing to free
1856 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1857 error, in fact).
1858
1859 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1860 heapframe that is used only when UTF-8 support is enabled. This caused no
1861 problem, but was untidy.
1862
1863 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1864 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1865 included within another project.
1866
1867 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1868 slightly modified by me:
1869
1870 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1871 not building pcregrep.
1872
1873 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1874 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1875
1876 25. Forward references, both numeric and by name, in patterns that made use of
1877 duplicate group numbers, could behave incorrectly or give incorrect errors,
1878 because when scanning forward to find the reference group, PCRE was not
1879 taking into account the duplicate group numbers. A pattern such as
1880 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1881
1882 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1883 the feature of a custom pointer more persuasive (as requested by a user).
1884
1885 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1886 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1887 pre-defined.
1888
1889 28. Added support for (*UTF8) at the start of a pattern.
1890
1891 29. Arrange for flags added by the "release type" setting in CMake to be shown
1892 in the configuration summary.
1893
1894
1895 Version 7.8 05-Sep-08
1896 ---------------------
1897
1898 1. Replaced UCP searching code with optimized version as implemented for Ad
1899 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1900 stage table and inline lookup instead of a function, giving speed ups of 2
1901 to 5 times on some simple patterns that I tested. Permission was given to
1902 distribute the MultiStage2.py script that generates the tables (it's not in
1903 the tarball, but is in the Subversion repository).
1904
1905 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1906 scripts.
1907
1908 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1909 a group with a zero qualifier. The result of the study could be incorrect,
1910 or the function might crash, depending on the pattern.
1911
1912 4. Caseless matching was not working for non-ASCII characters in back
1913 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1914 It now works when Unicode Property Support is available.
1915
1916 5. In pcretest, an escape such as \x{de} in the data was always generating
1917 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1918 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1919 truncation.
1920
1921 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1922
1923 7. Added two (int) casts to pcregrep when printing the difference of two
1924 pointers, in case they are 64-bit values.
1925
1926 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1927 test 2 if it fails.
1928
1929 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1930 and a #define of that name to empty if it is not externally set. This is to
1931 allow users of MSVC to set it if necessary.
1932
1933 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1934 the convenience functions in the pcre_get.c source file.
1935
1936 11. An option change at the start of a pattern that had top-level alternatives
1937 could cause overwriting and/or a crash. This command provoked a crash in
1938 some environments:
1939
1940 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1941
1942 This potential security problem was recorded as CVE-2008-2371.
1943
1944 12. For a pattern where the match had to start at the beginning or immediately
1945 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1946 pcre_dfa_exec() could read past the end of the passed subject if there was
1947 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1948 pcretest so that it places the subject at the end of its malloc-ed buffer.
1949
1950 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1951 exec() might read past the end of the data buffer in UTF-8 mode.
1952
1953 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1954 the data contained the byte 0x85 as part of a UTF-8 character within its
1955 first line. This applied both to normal and DFA matching.
1956
1957 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1958 /^[^d]*?$/8 failed to match "abc".
1959
1960 16. Added a missing copyright notice to pcrecpp_internal.h.
1961
1962 17. Make it more clear in the documentation that values returned from
1963 pcre_exec() in ovector are byte offsets, not character counts.
1964
1965 18. Tidied a few places to stop certain compilers from issuing warnings.
1966
1967 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1968 supplied by Stefan Weber. I made a further small update for 7.8 because
1969 there is a change of source arrangements: the pcre_searchfuncs.c module is
1970 replaced by pcre_ucd.c.
1971
1972
1973 Version 7.7 07-May-08
1974 ---------------------
1975
1976 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1977 a string to a long long, pretend we don't even have a long long." This is
1978 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1979
1980 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1981 pre-7.6 versions, which defined a global no_arg variable instead of putting
1982 it in the RE class. (See also #8 below.)
1983
1984 3. Remove a line of dead code, identified by coverity and reported by Nuno
1985 Lopes.
1986
1987 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1988
1989 (1) The include/exclude patterns were being applied to the whole pathnames
1990 of files, instead of just to the final components.
1991
1992 (2) If there was more than one level of directory, the subdirectories were
1993 skipped unless they satisfied the include/exclude conditions. This is
1994 inconsistent with GNU grep (and could even be seen as contrary to the
1995 pcregrep specification - which I improved to make it absolutely clear).
1996 The action now is always to scan all levels of directory, and just
1997 apply the include/exclude patterns to regular files.
1998
1999 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2000 --exclude_dir in the tests to avoid scanning .svn directories.
2001
2002 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2003 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2004 doesn't support NULs in patterns.
2005
2006 7. Added some missing "const"s to declarations of static tables in
2007 pcre_compile.c and pcre_dfa_exec.c.
2008
2009 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2010 caused by fix #2 above. (Subsequently also a second patch to fix the
2011 first patch. And a third patch - this was a messy problem.)
2012
2013 9. Applied Craig's patch to remove the use of push_back().
2014
2015 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2016 matching function regexec().
2017
2018 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2019 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2020 references. PCRE supports relative numbers with this syntax (I don't think
2021 Oniguruma does).
2022
2023 12. Previously, a group with a zero repeat such as (...){0} was completely
2024 omitted from the compiled regex. However, this means that if the group
2025 was called as a subroutine from elsewhere in the pattern, things went wrong
2026 (an internal error was given). Such groups are now left in the compiled
2027 pattern, with a new opcode that causes them to be skipped at execution
2028 time.
2029
2030 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2031 to the way PCRE behaves:
2032
2033 (a) A lone ] character is dis-allowed (Perl treats it as data).
2034
2035 (b) A back reference to an unmatched subpattern matches an empty string
2036 (Perl fails the current match path).
2037
2038 (c) A data ] in a character class must be notated as \] because if the
2039 first data character in a class is ], it defines an empty class. (In
2040 Perl it is not possible to have an empty class.) The empty class []
2041 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2042 The negative empty class [^] matches any one character, independently
2043 of the DOTALL setting.
2044
2045 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2046 non-existent subpattern following a character class starting with ']' and
2047 containing () gave an internal compiling error instead of "reference to
2048 non-existent subpattern". Fortunately, when the pattern did exist, the
2049 compiled code was correct. (When scanning forwards to check for the
2050 existence of the subpattern, it was treating the data ']' as terminating
2051 the class, so got the count wrong. When actually compiling, the reference
2052 was subsequently set up correctly.)
2053
2054 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2055 it was being rejected as not supported by pcre_dfa_exec(), even though
2056 other assertions are supported. I have made pcre_dfa_exec() support
2057 (*FAIL).
2058
2059 16. The implementation of 13c above involved the invention of a new opcode,
2060 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2061 cannot be changed at match time, I realized I could make a small
2062 improvement to matching performance by compiling OP_ALLANY instead of
2063 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2064 on the OP_ANY path.
2065
2066 17. Compiling pcretest on Windows with readline support failed without the
2067 following two fixes: (1) Make the unistd.h include conditional on
2068 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2069
2070 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2071 ncurses library to be included for pcretest when ReadLine support is
2072 requested, but also to allow for it to be overridden. This patch came from
2073 Daniel Bergström.
2074
2075 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2076 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2077 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2078 spotting this.
2079
2080
2081 Version 7.6 28-Jan-08
2082 ---------------------
2083
2084 1. A character class containing a very large number of characters with
2085 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2086 overflow.
2087
2088 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2089 HAVE_LONG_LONG is not defined.
2090
2091 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2092 bring them up to date and include new features. This patch includes:
2093
2094 - Fixed PH's badly added libz and libbz2 support.
2095 - Fixed a problem with static linking.
2096 - Added pcredemo. [But later removed - see 7 below.]
2097 - Fixed dftables problem and added an option.
2098 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2099 HAVE_LONG_LONG.
2100 - Added readline support for pcretest.
2101 - Added an listing of the option settings after cmake has run.
2102
2103 4. A user submitted a patch to Makefile that makes it easy to create
2104 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2105 Makefile.am that cause it to include this special target, without
2106 affecting anything else. Note that the same mingw target plus all
2107 the other distribution libraries and programs are now supported
2108 when configuring with CMake (see 6 below) instead of with
2109 Configure/Make.
2110
2111 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2112 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2113 exported in the Windows port". It has not yet been confirmed that the patch
2114 solves the problem, but it does no harm.
2115
2116 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2117 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2118 with CMake, and also correct the comment about stack recursion.
2119
2120 7. Remove the automatic building of pcredemo from the ./configure system and
2121 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2122 of a program that users should build themselves after PCRE is installed, so
2123 building it automatically is not really right. What is more, it gave
2124 trouble in some build environments.
2125
2126 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2127
2128
2129 Version 7.5 10-Jan-08
2130 ---------------------
2131
2132 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2133 values in parens when parsing an RE using the C++ wrapper."
2134
2135 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2136 Characters greater than 255 were excluded from the class instead of being
2137 included.
2138
2139 3. The same bug as (2) above applied to negated POSIX classes such as
2140 [:^space:].
2141
2142 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2143 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2144 I have changed it.
2145
2146 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2147 first named subpattern) and a construct such as (?&a) would reference the
2148 first named subpattern whose name started with "a" (in other words, the
2149 length check was missing). Both these problems are fixed. "Subpattern name
2150 expected" is now given for (?&) (a zero-length name), and this patch also
2151 makes it give the same error for \k'' (previously it complained that that
2152 was a reference to a non-existent subpattern).
2153
2154 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2155 this is right because (?- can be followed by option settings as well as by
2156 digits. I have, however, made the messages clearer.
2157
2158 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2159 than the number used in the conditional) now cause a compile-time error.
2160 This is actually not compatible with Perl, which accepts such patterns, but
2161 treats the conditional as always being FALSE (as PCRE used to), but it
2162 seems to me that giving a diagnostic is better.
2163
2164 8. Change "alphameric" to the more common word "alphanumeric" in comments
2165 and messages.
2166
2167 9. Fix two occurrences of "backslash" in comments that should have been
2168 "backspace".
2169
2170 10. Remove two redundant lines of code that can never be obeyed (their function
2171 was moved elsewhere).
2172
2173 11. The program that makes PCRE's Unicode character property table had a bug
2174 which caused it to generate incorrect table entries for sequences of
2175 characters that have the same character type, but are in different scripts.
2176 It amalgamated them into a single range, with the script of the first of
2177 them. In other words, some characters were in the wrong script. There were
2178 thirteen such cases, affecting characters in the following ranges:
2179
2180 U+002b0 - U+002c1
2181 U+0060c - U+0060d
2182 U+0061e - U+00612
2183 U+0064b - U+0065e
2184 U+0074d - U+0076d
2185 U+01800 - U+01805
2186 U+01d00 - U+01d77
2187 U+01d9b - U+01dbf
2188 U+0200b - U+0200f
2189 U+030fc - U+030fe
2190 U+03260 - U+0327f
2191 U+0fb46 - U+0fbb1
2192 U+10450 - U+1049d
2193
2194 12. The -o option (show only the matching part of a line) for pcregrep was not
2195 compatible with GNU grep in that, if there was more than one match in a
2196 line, it showed only the first of them. It now behaves in the same way as
2197 GNU grep.
2198
2199 13. If the -o and -v options were combined for pcregrep, it printed a blank
2200 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2201 does the same. The return code can be used to tell if there were any
2202 non-matching lines.
2203
2204 14. Added --file-offsets and --line-offsets to pcregrep.
2205
2206 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2207 infinitely looping recursion. The bug was that positive lookaheads were not
2208 being skipped when checking for a possible empty match (negative lookaheads
2209 and both kinds of lookbehind were skipped).
2210
2211 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2212 inclusion of <windows.h> to before rather than after the definition of
2213 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2214
2215 17. Specifying a possessive quantifier with a specific limit for a Unicode
2216 character property caused pcre_compile() to compile bad code, which led at
2217 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2218 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2219 caused the error; without that there was no problem.
2220
2221 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2222
2223 19. Added --enable-pcretest-libreadline.
2224
2225 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2226 RE::GlobalReplace(). As a result, the number of replacements returned was
2227 double what it should be. I removed one of the increments, but Craig sent a
2228 later patch that removed the other one (the right fix) and added unit tests
2229 that check the return values (which was not done before).
2230
2231 21. Several CMake things:
2232
2233 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2234 the names libpcre and libpcreposix, not just pcre and pcreposix.
2235
2236 (2) The above change means that pcretest and pcregrep are now correctly
2237 linked with the newly-built libraries, not previously installed ones.
2238
2239 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2240
2241 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2242 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2243 UTF-8 newline character). The key issue is that the pattern starts .*;
2244 this means that the match must be either at the beginning, or after a
2245 newline. The bug was in the code for advancing after a failed match and
2246 checking that the new position followed a newline. It was not taking
2247 account of UTF-8 characters correctly.
2248
2249 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2250 character classes. PCRE was not treating the sequence [:...:] as a
2251 character class unless the ... were all letters. Perl, however, seems to
2252 allow any characters between [: and :], though of course it rejects as
2253 unknown any "names" that contain non-letters, because all the known class
2254 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2255 for example, whereas PCRE did not - it did not recognize a POSIX character
2256 class. This seemed a bit dangerous, so the code has been changed to be
2257 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2258 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2259 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2260 Perl does, and where it didn't before.
2261
2262 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2263 Windows environments %n is disabled by default.
2264
2265
2266 Version 7.4 21-Sep-07
2267 ---------------------
2268
2269 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2270 means that a class such as [\s] counted as "explicit reference to CR or
2271 LF". That isn't really right - the whole point of the change was to try to
2272 help when there was an actual mention of one of the two characters. So now
2273 the change happens only if \r or \n (or a literal CR or LF) character is
2274 encountered.
2275
2276 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2277 of both had grown to the point where there were only 3 bits left.
2278 Fortunately, there was spare space in the data structure, and so I have
2279 moved the internal flags into a new 16-bit field to free up more option
2280 bits.
2281
2282 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2283 but did not set the internal JCHANGED flag - either of these is enough to
2284 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2285 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2286 start sets both bits.
2287
2288 4. Added options (at build time, compile time, exec time) to change \R from
2289 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2290
2291 5. doc/pcresyntax.html was missing from the distribution.
2292
2293 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2294 compatibility, even though it is no longer used.
2295
2296 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2297 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2298 windows.h file is present (where different names are used). [This was
2299 reversed later after testing - see 16 below.]
2300
2301 8. Changed all #include <config.h> to #include "config.h". There were also
2302 some further <pcre.h> cases that I changed to "pcre.h".
2303
2304 9. When pcregrep was used with the --colour option, it missed the line ending
2305 sequence off the lines that it output.
2306
2307 10. It was pointed out to me that arrays of string pointers cause lots of
2308 relocations when a shared library is dynamically loaded. A technique of
2309 using a single long string with a table of offsets can drastically reduce
2310 these. I have refactored PCRE in four places to do this. The result is
2311 dramatic:
2312
2313 Originally: 290
2314 After changing UCP table: 187
2315 After changing error message table: 43
2316 After changing table of "verbs" 36
2317 After changing table of Posix names 22
2318
2319 Thanks to the folks working on Gregex for glib for this insight.
2320
2321 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2322 unicode-properties was also set.
2323
2324 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2325
2326 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2327 checked only for CRLF.
2328
2329 14. Added casts to pcretest.c to avoid compiler warnings.
2330
2331 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2332
2333 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2334 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2335 entirely. This removes changes made in 7 above.
2336
2337 17. The CMake files have been updated, and there is now more information about
2338 building with CMake in the NON-UNIX-USE document.
2339
2340
2341 Version 7.3 28-Aug-07
2342 ---------------------
2343
2344 1. In the rejigging of the build system that eventually resulted in 7.1, the
2345 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2346 brackets there is not right, since it causes compilers to look for an
2347 installed pcre.h, not the version that is in the source that is being
2348 compiled (which of course may be different). I have changed it back to:
2349
2350 #include "pcre.h"
2351
2352 I have a vague recollection that the change was concerned with compiling in
2353 different directories, but in the new build system, that is taken care of
2354 by the VPATH setting the Makefile.
2355
2356 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2357 when the subject happened to end in the byte 0x85 (e.g. if the last
2358 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2359 characters but of course it shouldn't be taken as a newline when it is part
2360 of another character. The bug was that, for an unlimited repeat of . in
2361 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2362 characters when looking for a newline.
2363
2364 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2365
2366 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2367 in debug output.
2368
2369 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2370 long printing in the pcrecpp unittest when running under MinGW.
2371
2372 6. ESC_K was left out of the EBCDIC table.
2373
2374 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2375 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2376 limit also applies to "virtual nesting" when a pattern is recursive, and in
2377 this case 1000 isn't so big. I have been able to remove this limit at the
2378 expense of backing off one optimization in certain circumstances. Normally,
2379 when pcre_exec() would call its internal match() function recursively and
2380 immediately return the result unconditionally, it uses a "tail recursion"
2381 feature to save stack. However, when a subpattern that can match an empty
2382 string has an unlimited repetition quantifier, it no longer makes this
2383 optimization. That gives it a stack frame in which to save the data for
2384 checking that an empty string has been matched. Previously this was taken
2385 from the 1000-entry workspace that had been reserved. So now there is no
2386 explicit limit, but more stack is used.
2387
2388 8. Applied Daniel's patches to solve problems with the import/export magic
2389 syntax that is required for Windows, and which was going wrong for the
2390 pcreposix and pcrecpp parts of the library. These were overlooked when this
2391 problem was solved for the main library.
2392
2393 9. There were some crude static tests to avoid integer overflow when computing
2394 the size of patterns that contain repeated groups with explicit upper
2395 limits. As the maximum quantifier is 65535, the maximum group length was
2396 set at 30,000 so that the product of these two numbers did not overflow a
2397 32-bit integer. However, it turns out that people want to use groups that
2398 are longer than 30,000 bytes (though not repeat them that many times).
2399 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2400 made it possible to implement the integer overflow checks in a much more
2401 dynamic way, which I have now done. The artificial limitation on group
2402 length has been removed - we now have only the limit on the total length of
2403 the compiled pattern, which depends on the LINK_SIZE setting.
2404
2405 10. Fixed a bug in the documentation for get/copy named substring when
2406 duplicate names are permitted. If none of the named substrings are set, the
2407 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2408 empty string.
2409
2410 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2411 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2412 because the ] is interpreted as the first data character and the
2413 terminating ] is not found. PCRE has been made compatible with Perl in this
2414 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2415 cause memory overwriting.
2416
2417 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2418 string has been matched (to stop an infinite loop). It was not recognizing
2419 a conditional subpattern that could match an empty string if that
2420 subpattern was within another subpattern. For example, it looped when
2421 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2422 condition was not nested. This bug has been fixed.
2423
2424 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2425 past the start of the subject in the presence of bytes with the top bit
2426 set, for example "\x8aBCD".
2427
2428 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2429 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2430
2431 14. Optimized (?!) to (*FAIL).
2432
2433 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2434 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2435 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2436 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2437 does: it's just the validity check that is more restrictive.
2438
2439 16. Inserted checks for integer overflows during escape sequence (backslash)
2440 processing, and also fixed erroneous offset values for syntax errors during
2441 backslash processing.
2442
2443 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2444 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2445
2446 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2447 caused an overrun.
2448
2449 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2450 something other than just ASCII characters) inside a group that had an
2451 unlimited repeat caused a loop at compile time (while checking to see
2452 whether the group could match an empty string).
2453
2454 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2455 [\P{Any}] did so. (Error in the code for printing property names.)
2456
2457 21. An orphan \E inside a character class could cause a crash.
2458
2459 22. A repeated capturing bracket such as (A)? could cause a wild memory
2460 reference during compilation.
2461
2462 23. There are several functions in pcre_compile() that scan along a compiled
2463 expression for various reasons (e.g. to see if it's fixed length for look
2464 behind). There were bugs in these functions when a repeated \p or \P was
2465 present in the pattern. These operators have additional parameters compared
2466 with \d, etc, and these were not being taken into account when moving along
2467 the compiled data. Specifically:
2468
2469 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2470 length.
2471
2472 (b) An item such as \pL+ within a repeated group could cause crashes or
2473 loops.
2474
2475 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2476 "reference to non-existent subpattern" error.
2477
2478 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2479
2480 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2481 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2482
2483 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2484
2485 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2486 character were causing crashes (broken optimization).
2487
2488 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2489 \p or \P) caused a compile-time loop.
2490
2491 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2492 break. For example, the unstudied pattern [\r\n]A does not match the string
2493 "\r\nA" because change 7.0/46 below moves the current point on by two
2494 characters after failing to match at the start. However, the pattern \nA
2495 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2496 the same is true. There doesn't seem any very clean way out of this, but
2497 what I have chosen to do makes the common cases work: PCRE now takes note
2498 of whether there can be an explicit match for \r or \n anywhere in the
2499 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2500 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2501 pattern has explicit CR or LF references.
2502
2503 29. Added (*CR) etc for changing newline setting at start of pattern.
2504
2505
2506 Version 7.2 19-Jun-07
2507 ---------------------
2508
2509 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2510 which is apparently normally available under Windows.
2511
2512 2. Re-jig the pcregrep tests with different newline settings in an attempt
2513 to make them independent of the local environment's newline setting.
2514
2515 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2516
2517 4. Some of the "internals" tests were previously cut out when the link size
2518 was not 2, because the output contained actual offsets. The recent new
2519 "Z" feature of pcretest means that these can be cut out, making the tests
2520 usable with all link sizes.
2521
2522 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2523 stack recursion. This gives a massive performance boost under BSD, but just
2524 a small improvement under Linux. However, it saves one field in the frame
2525 in all cases.
2526
2527 6. Added more features from the forthcoming Perl 5.10:
2528
2529 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2530 recursion call. It refers to the nth most recently opened parentheses.
2531
2532 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2533 to be opened parentheses.
2534
2535 (c) Conditions that refer to capturing parentheses can be specified
2536 relatively, for example, (?(-2)... or (?(+3)...
2537
2538 (d) \K resets the start of the current match so that everything before
2539 is not part of it.
2540
2541 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2542
2543 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2544 reference syntax.
2545
2546 (g) (?| introduces a group in which the numbering of parentheses in each
2547 alternative starts with the same number.
2548
2549 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2550
2551 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2552 PCRE_INFO_JCHANGED.
2553
2554 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2555 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2556 for detecting groups that can match an empty string.
2557
2558 9. A pattern with a very large number of alternatives (more than several
2559 hundred) was running out of internal workspace during the pre-compile
2560 phase, where pcre_compile() figures out how much memory will be needed. A
2561 bit of new cunning has reduced the workspace needed for groups with
2562 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2563 workspace instead of running out of the 4096 that are available.
2564
2565 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2566
2567 11. Applied patch from Google to remove an optimization that didn't quite work.
2568 The report of the bug said:
2569
2570 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2571 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2572 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2573
2574 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2575 it matched the wrong number of bytes.
2576
2577
2578 Version 7.1 24-Apr-07
2579 ---------------------
2580
2581 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2582 that is more "standard", making use of automake and other Autotools. There
2583 is some re-arrangement of the files and adjustment of comments consequent
2584 on this.
2585
2586 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2587 for recursive directory scanning broke on some systems because the files
2588 are not scanned in any specific order and on different systems the order
2589 was different. A call to "sort" has been inserted into RunGrepTest for the
2590 approprate test as a short-term fix. In the longer term there may be an
2591 alternative.
2592
2593 3. I had an email from Eric Raymond about problems translating some of PCRE's
2594 man pages to HTML (despite the fact that I distribute HTML pages, some
2595 people do their own conversions for various reasons). The problems
2596 concerned the use of low-level troff macros .br and .in. I have therefore
2597 removed all such uses from the man pages (some were redundant, some could
2598 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2599 HTML has been updated to handle .nf/.fi and to complain if it encounters
2600 .br or .in.
2601
2602 4. Updated comments in configure.ac that get placed in config.h.in and also
2603 arranged for config.h to be included in the distribution, with the name
2604 config.h.generic, for the benefit of those who have to compile without
2605 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2606
2607 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2608 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2609 makevp.bat for the current PCRE, using the additional files
2610 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2611
2612 6. A Windows user reported a minor discrepancy with test 2, which turned out
2613 to be caused by a trailing space on an input line that had got lost in his
2614 copy. The trailing space was an accident, so I've just removed it.
2615
2616 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2617 that is needed.
2618
2619 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2620 as "const" (a) because they are and (b) because it helps the PHP
2621 maintainers who have recently made a script to detect big data structures
2622 in the php code that should be moved to the .rodata section. I remembered
2623 to update Builducptable as well, so it won't revert if ucptable.h is ever
2624 re-created.
2625
2626 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2627 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2628 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2629 support is not required. This saves 1.5-2K of code, which is important in
2630 some applications.
2631
2632 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2633 so as not to refer to the tables, even though these functions will never be
2634 called when UTF-8 support is disabled. Otherwise there are problems with a
2635 shared library.
2636
2637 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2638
2639 (a) It was defining its arguments as char * instead of void *.
2640
2641 (b) It was assuming that all moves were upwards in memory; this was true
2642 a long time ago when I wrote it, but is no longer the case.
2643
2644 The emulated memove() is provided for those environments that have neither
2645 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2646 is clearly not the case, as these two bugs were recently reported.
2647
2648 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2649 and Detrail to create the HTML documentation, the .txt form of the man
2650 pages, and it removes trailing spaces from listed files. It also creates
2651 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2652 case, it wraps all the #defines with #ifndefs. This script should be run
2653 before "make dist".
2654
2655 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2656 with Unicode property support.
2657
2658 (a) For a maximizing quantifier, if the two different cases of the
2659 character were of different lengths in their UTF-8 codings (there are
2660 some cases like this - I found 11), and the matching function had to
2661 back up over a mixture of the two cases, it incorrectly assumed they
2662 were both the same length.
2663
2664 (b) When PCRE was configured to use the heap rather than the stack for
2665 recursion during matching, it was not correctly preserving the data for
2666 the other case of a UTF-8 character when checking ahead for a match
2667 while processing a minimizing repeat. If the check also involved
2668 matching a wide character, but failed, corruption could cause an
2669 erroneous result when trying to check for a repeat of the original
2670 character.
2671
2672 13. Some tidying changes to the testing mechanism:
2673
2674 (a) The RunTest script now detects the internal link size and whether there
2675 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2676 values substituted by "configure". (The RunGrepTest script already did
2677 this for UTF-8.) The configure.ac script no longer substitutes the
2678 relevant variables.
2679
2680 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2681 with length and offset values. This means that the output is different
2682 for different internal link sizes. Test 2 is skipped for link sizes
2683 other than 2 because of this, bypassing the problem. Unfortunately,
2684 there was also a test in test 3 (the locale tests) that used /B and
2685 failed for link sizes other than 2. Rather than cut the whole test out,
2686 I have added a new /Z option to pcretest that replaces the length and
2687 offset values with spaces. This is now used to make test 3 independent
2688 of link size. (Test 2 will be tidied up later.)
2689
2690 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2691 segmentation fault instead of returning the appropriate error message.
2692
2693 15. In multiline mode when the newline sequence was set to "any", the pattern
2694 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2695 This doesn't seem right; it now treats the CRLF combination as the line
2696 ending, and so does not match in that case. It's only a pattern such as ^$
2697 that would hit this one: something like ^ABC$ would have failed after \r
2698 and then tried again after \r\n.
2699
2700 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2701 in an attempt to make files that differ only in their line terminators
2702 compare equal. This works on Linux.
2703
2704 17. Under certain error circumstances pcregrep might try to free random memory
2705 as it exited. This is now fixed, thanks to valgrind.
2706
2707 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2708 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2709 was because its rules for how to advance for /g after matching an empty
2710 string at the end of a line did not allow for this case. They now check for
2711 it specially.
2712
2713 20. pcretest is supposed to handle patterns and data of any length, by
2714 extending its buffers when necessary. It was getting this wrong when the
2715 buffer for a data line had to be extended.
2716
2717 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2718 CRLF as a newline sequence.
2719
2720 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2721 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2722 I have nevertheless tidied it up.
2723
2724 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2725
2726 24. Added a man page for pcre-config.
2727
2728
2729 Version 7.0 19-Dec-06
2730 ---------------------
2731
2732 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2733 moving to gcc 4.1.1.
2734
2735 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2736 sys/time.h, which is documented as needed for this function. It doesn't
2737 seem to matter on Linux, but it showed up on some releases of OS X.
2738
2739 3. It seems that there are systems where bytes whose values are greater than
2740 127 match isprint() in the "C" locale. The "C" locale should be the
2741 default when a C program starts up. In most systems, only ASCII printing
2742 characters match isprint(). This difference caused the output from pcretest
2743 to vary, making some of the tests fail. I have changed pcretest so that:
2744
2745 (a) When it is outputting text in the compiled version of a pattern, bytes
2746 other than 32-126 are always shown as hex escapes.
2747
2748 (b) When it is outputting text that is a matched part of a subject string,
2749 it does the same, unless a different locale has been set for the match
2750 (using the /L modifier). In this case, it uses isprint() to decide.
2751
2752 4. Fixed a major bug that caused incorrect computation of the amount of memory
2753 required for a compiled pattern when options that changed within the
2754 pattern affected the logic of the preliminary scan that determines the
2755 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2756 that the computed length was too small. The symptoms of this bug were
2757 either the PCRE error "internal error: code overflow" from pcre_compile(),
2758 or a glibc crash with a message such as "pcretest: free(): invalid next
2759 size (fast)". Examples of patterns that provoked this bug (shown in
2760 pcretest format) are:
2761
2762 /(?-x: )/x
2763 /(?x)(?-x: \s*#\s*)/
2764 /((?i)[\x{c0}])/8
2765 /(?i:[\x{c0}])/8
2766
2767 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2768 is now done differently.
2769
2770 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2771 wrapper classes; (b) implement a new function in the C++ scanner that is
2772 more efficient than the old way of doing things because it avoids levels of
2773 recursion in the regex matching; (c) add a paragraph to the documentation
2774 for the FullMatch() function.
2775
2776 6. The escape sequence \n was being treated as whatever was defined as
2777 "newline". Not only was this contrary to the documentation, which states
2778 that \n is character 10 (hex 0A), but it also went horribly wrong when
2779 "newline" was defined as CRLF. This has been fixed.
2780
2781 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2782 was being set to -1 for the "end of line" case (supposedly a value that no
2783 character can have). Though this value is never used (the check for end of
2784 line is "zero bytes in current character"), it caused compiler complaints.
2785 I've changed it to 0xffffffff.
2786
2787 8. In pcre_version.c, the version string was being built by a sequence of
2788 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2789 string (as it is for production releases) called a macro with an empty
2790 argument. The C standard says the result of this is undefined. The gcc
2791 compiler treats it as an empty string (which was what was wanted) but it is
2792 reported that Visual C gives an error. The source has been hacked around to
2793 avoid this problem.
2794
2795 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2796 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2797 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2798 of them did).
2799
2800 10. Originally, pcretest opened its input and output without "b"; then I was
2801 told that "b" was needed in some environments, so it was added for release
2802 5.0 to both the input and output. (It makes no difference on Unix-like
2803 systems.) Later I was told that it is wrong for the input on Windows. I've
2804 now abstracted the modes into two macros, to make it easier to fiddle with
2805 them, and removed "b" from the input mode under Windows.
2806
2807 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2808
2809 12. Added -help and --help to pcretest as an official way of being reminded
2810 of the options.
2811
2812 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2813 and pcrecpp.cc because they annoy compilers at high warning levels.
2814
2815 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2816
2817 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2818 scripts are not C programs :-) and which was not noticed because it works
2819 on Linux.
2820
2821 16. pcretest is supposed to handle any length of pattern and data line (as one
2822 line or as a continued sequence of lines) by extending its input buffer if
2823 necessary. This feature was broken for very long pattern lines, leading to
2824 a string of junk being passed to pcre_compile() if the pattern was longer
2825 than about 50K.
2826
2827 17. I have done a major re-factoring of the way pcre_compile() computes the
2828 amount of memory needed for a compiled pattern. Previously, there was code
2829 that made a preliminary scan of the pattern in order to do this. That was
2830 OK when PCRE was new, but as the facilities have expanded, it has become
2831 harder and harder to keep it in step with the real compile phase, and there
2832 have been a number of bugs (see for example, 4 above). I have now found a
2833 cunning way of running the real compile function in a "fake" mode that
2834 enables it to compute how much memory it would need, while actually only
2835 ever using a few hundred bytes of working memory and without too many
2836 tests of the mode. This should make future maintenance and development
2837 easier. A side effect of this work is that the limit of 200 on the nesting
2838 depth of parentheses has been removed (though this was never a serious
2839 limitation, I suspect). However, there is a downside: pcre_compile() now
2840 runs more slowly than before (30% or more, depending on the pattern). I
2841 hope this isn't a big issue. There is no effect on runtime performance.
2842
2843 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2844 newline (only possible for the last line of a file) and it was a
2845 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2846
2847 19. Added additional timing features to pcretest. (1) The -tm option now times
2848 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2849 separate command line item, by a number that specifies the number of
2850 repeats to use when timing. The default is 50000; this gives better
2851 precision, but takes uncomfortably long for very large patterns.
2852
2853 20. Extended pcre_study() to be more clever in cases where a branch of a
2854 subpattern has no definite first character. For example, (a*|b*)[cd] would
2855 previously give no result from pcre_study(). Now it recognizes that the
2856 first character must be a, b, c, or d.
2857
2858 21. There was an incorrect error "recursive call could loop indefinitely" if
2859 a subpattern (or the entire pattern) that was being tested for matching an
2860 empty string contained only one non-empty item after a nested subpattern.
2861 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2862 incorrectly, because the \d was being skipped in the check.
2863
2864 22. The pcretest program now has a new pattern option /B and a command line
2865 option -b, which is equivalent to adding /B to every pattern. This causes
2866 it to show the compiled bytecode, without the additional information that
2867 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2868 is the same as /B/I).
2869
2870 23. A new optimization is now able automatically to treat some sequences such
2871 as a*b as a*+b. More specifically, if something simple (such as a character
2872 or a simple class like \d) has an unlimited quantifier, and is followed by
2873 something that cannot possibly match the quantified thing, the quantifier
2874 is automatically "possessified".
2875
2876 24. A recursive reference to a subpattern whose number was greater than 39
2877 went wrong under certain circumstances in UTF-8 mode. This bug could also
2878 have affected the operation of pcre_study().
2879
2880 25. Realized that a little bit of performance could be had by replacing
2881 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2882
2883 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2884
2885 27. Possessive quantifiers such as a++ were previously implemented by turning
2886 them into atomic groups such as ($>a+). Now they have their own opcodes,
2887 which improves performance. This includes the automatically created ones
2888 from 23 above.
2889
2890 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2891 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2892 the first matched character to be a colon. This applied both to named and
2893 numbered groups.
2894
2895 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2896
2897 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2898 building PCRE on Windows easier, so I have included it in the distribution.
2899
2900 31. There is now a check in pcretest against a ridiculously large number being
2901 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2902 loop, the loop is abandoned.
2903
2904 32. Forward references to subpatterns in conditions such as (?(2)...) where
2905 subpattern 2 is defined later cause pcre_compile() to search forwards in
2906 the pattern for the relevant set of parentheses. This search went wrong
2907 when there were unescaped parentheses in a character class, parentheses
2908 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2909
2910 33. "Subroutine" calls and backreferences were previously restricted to
2911 referencing subpatterns earlier in the regex. This restriction has now
2912 been removed.
2913
2914 34. Added a number of extra features that are going to be in Perl 5.10. On the
2915 whole, these are just syntactic alternatives for features that PCRE had
2916 previously implemented using the Python syntax or my own invention. The
2917 other formats are all retained for compatibility.
2918
2919 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2920 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2921 also .NET compatible.
2922
2923 (b) A recursion or subroutine call to a named group can now be defined as
2924 (?&name) as well as (?P>name).
2925
2926 (c) A backreference to a named group can now be defined as \k<name> or
2927 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2928 5.10, are also .NET compatible.
2929
2930 (d) A conditional reference to a named group can now use the syntax
2931 (?(<name>) or (?('name') as well as (?(name).
2932
2933 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2934 groups (named and numbered) that are never evaluated inline, but can be
2935 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2936 is always false. There may be only one alternative in such a group.
2937
2938 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2939 as the simple (?(R). The condition is true only if the most recent
2940 recursion is that of the given number or name. It does not search out
2941 through the entire recursion stack.
2942
2943 (g) The escape \gN or \g{N} has been added, where N is a positive or
2944 negative number, specifying an absolute or relative reference.
2945
2946 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2947 some "unreachable code" warnings.
2948
2949 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2950 things, this adds five new scripts.
2951
2952 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2953 There were also incompatibilities regarding the handling of \Q..\E inside
2954 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2955 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2956
2957 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2958 matches an empty string, and forcibly breaks the loop. There were bugs in
2959 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2960 against aaaa the result was just "a" rather than "aaaa", for example. Two
2961 separate and independent bugs (that affected different cases) have been
2962 fixed.
2963
2964 39. Refactored the code to abolish the use of different opcodes for small
2965 capturing bracket numbers. This is a tidy that I avoided doing when I
2966 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2967 The new approach is not only tidier, it makes it possible to reduce the
2968 memory needed to fix the previous bug (38).
2969
2970 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2971 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2972 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2973 mode.
2974
2975 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2976 report.
2977
2978 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2979 copy construction and assignment in the C++ wrapper.
2980
2981 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2982 couple of bugs that could have given wrong results in the "--newline=crlf"
2983 case.
2984
2985 44. Added a number of casts and did some reorganization of signed/unsigned int
2986 variables following suggestions from Dair Grant. Also renamed the variable
2987 "this" as "item" because it is a C++ keyword.
2988
2989 45. Arranged for dftables to add
2990
2991 #include "pcre_internal.h"
2992
2993 to pcre_chartables.c because without it, gcc 4.x may remove the array
2994 definition from the final binary if PCRE is built into a static library and
2995 dead code stripping is activated.
2996
2997 46. For an unanchored pattern, if a match attempt fails at the start of a
2998 newline sequence, and the newline setting is CRLF or ANY, and the next two
2999 characters are CRLF, advance by two characters instead of one.
3000
3001
3002 Version 6.7 04-Jul-06
3003 ---------------------
3004
3005 1. In order to handle tests when input lines are enormously long, pcretest has
3006 been re-factored so that it automatically extends its buffers when
3007 necessary. The code is crude, but this _is_ just a test program. The
3008 default size has been increased from 32K to 50K.
3009
3010 2. The code in pcre_study() was using the value of the re argument before
3011 testing it for NULL. (Of course, in any sensible call of the function, it
3012 won't be NULL.)
3013
3014 3. The memmove() emulation function in pcre_internal.h, which is used on
3015 systems that lack both memmove() and bcopy() - that is, hardly ever -
3016 was missing a "static" storage class specifier.
3017
3018 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3019 containing an extended class (one that cannot be represented by a bitmap
3020 because it contains high-valued characters or Unicode property items, e.g.
3021 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3022 pattern, but PCRE should not loop if you do not (it no longer does).
3023 [Detail: two cases were found: (a) a repeated subpattern containing an
3024 extended class; (b) a recursive reference to a subpattern that followed a
3025 previous extended class. It wasn't skipping over the extended class
3026 correctly when UTF-8 mode was not set.]
3027
3028 5. A negated single-character class was not being recognized as fixed-length
3029 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3030 compile error "lookbehind assertion is not fixed length".
3031
3032 6. The RunPerlTest auxiliary script was showing an unexpected difference
3033 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3034 write a Perl script that can interpret lines of an input file either as
3035 byte characters or as UTF-8, which is what "perltest" was being required to
3036 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3037 can't do is switch easily at run time between having the "use utf8;" pragma
3038 or not. In the end, I fudged it by using the RunPerlTest script to insert
3039 "use utf8;" explicitly for the UTF-8 tests.
3040
3041 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3042 the end of the subject string, contrary to the documentation and to what
3043 Perl does. This was true of both matching functions. Now it matches only at
3044 the start of the subject and immediately after *internal* newlines.
3045
3046 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3047 a pointer to an int instead of a pointer to an unsigned long int. This
3048 caused problems on 64-bit systems.
3049
3050 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3051 instance of the 'standard' template library not being so standard".
3052
3053 10. There was no check on the number of named subpatterns nor the maximum
3054 length of a subpattern name. The product of these values is used to compute
3055 the size of the memory block for a compiled pattern. By supplying a very
3056 long subpattern name and a large number of named subpatterns, the size
3057 computation could be caused to overflow. This is now prevented by limiting
3058 the length of names to 32 characters, and the number of named subpatterns
3059 to 10,000.
3060
3061 11. Subpatterns that are repeated with specific counts have to be replicated in
3062 the compiled pattern. The size of memory for this was computed from the
3063 length of the subpattern and the repeat count. The latter is limited to
3064 65535, but there was no limit on the former, meaning that integer overflow
3065 could in principle occur. The compiled length of a repeated subpattern is
3066 now limited to 30,000 bytes in order to prevent this.
3067
3068 12. Added the optional facility to have named substrings with the same name.
3069
3070 13. Added the ability to use a named substring as a condition, using the
3071 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3072 are numbers (not recommended). Forward references are permitted.
3073
3074 14. Added forward references in named backreferences (if you see what I mean).
3075
3076 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3077 pattern could run off the end of the subject. For example, the pattern
3078 "(?s)(.{1,5})"8 did this with the subject "ab".
3079
3080 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3081 PCRE_CASELESS was set when matching characters that were quantified with ?
3082 or *.
3083
3084 17. A character class other than a single negated character that had a minimum
3085 but no maximum quantifier - for example [ab]{6,} - was not handled
3086 correctly by pce_dfa_exec(). It would match only one character.
3087
3088 18. A valid (though odd) pattern that looked like a POSIX character
3089 class but used an invalid character after [ (for example [[,abc,]]) caused
3090 pcre_compile() to give the error "Failed: internal error: code overflow" or
3091 in some cases to crash with a glibc free() error. This could even happen if
3092 the pattern terminated after [[ but there just happened to be a sequence of
3093 letters, a binary zero, and a closing ] in the memory that followed.
3094
3095 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3096 over the years. Originally (before any Unicode support), just the bottom 8
3097 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3098 output from "man perlunicode" includes this:
3099
3100 The regular expression compiler produces polymorphic opcodes. That
3101 is, the pattern adapts to the data and automatically switches to
3102 the Unicode character scheme when presented with Unicode data--or
3103 instead uses a traditional byte scheme when presented with byte
3104 data.
3105
3106 Sadly, a wide octal escape does not cause a switch, and in a string with
3107 no other multibyte characters, these octal escapes are treated as before.
3108 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3109 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3110 Unicode string.
3111
3112 I have not perpetrated such confusion in PCRE. Up till now, it took just
3113 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3114 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3115 translate to the appropriate multibyte character.
3116
3117 29. Applied some refactoring to reduce the number of warnings from Microsoft
3118 and Borland compilers. This has included removing the fudge introduced
3119 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3120 a warning about an unused variable.
3121
3122 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3123 characters since release 4.0, because Perl (from release 5.004) does not.
3124 [Or at least, is documented not to: some releases seem to be in conflict
3125 with the documentation.] However, when a pattern was studied with
3126 pcre_study() and all its branches started with \s, PCRE still included VT
3127 as a possible starting character. Of course, this did no harm; it just
3128 caused an unnecessary match attempt.
3129
3130 22. Removed a now-redundant internal flag bit that recorded the fact that case
3131 dependency changed within the pattern. This was once needed for "required
3132 byte" processing, but is no longer used. This recovers a now-scarce options
3133 bit. Also moved the least significant internal flag bit to the most-
3134 significant bit of the word, which was not previously used (hangover from
3135 the days when it was an int rather than a uint) to free up another bit for
3136 the future.
3137
3138 23. Added support for CRLF line endings as well as CR and LF. As well as the
3139 default being selectable at build time, it can now be changed at runtime
3140 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3141 specify that it is scanning data with non-default line endings.
3142
3143 24. Changed the definition of CXXLINK to make it agree with the definition of
3144 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3145
3146 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3147 recursions. This makes a big different to stack usage for some patterns.
3148
3149 26. If a subpattern containing a named recursion or subroutine reference such
3150 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3151 the space required for the compiled pattern went wrong and gave too small a
3152 value. Depending on the environment, this could lead to "Failed: internal
3153 error: code overflow at offset 49" or "glibc detected double free or
3154 corruption" errors.
3155
3156 27. Applied patches from Google (a) to support the new newline modes and (b) to
3157 advance over multibyte UTF-8 characters in GlobalReplace.
3158
3159 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3160 difference for some implementation of PCRE in some Windows version.
3161
3162 29. Added some extra testing facilities to pcretest:
3163
3164 \q<number> in a data line sets the "match limit" value
3165 \Q<number> in a data line sets the "match recursion limt" value
3166 -S <number> sets the stack size, where <number> is in megabytes
3167
3168 The -S option isn't available for Windows.
3169
3170
3171 Version 6.6 06-Feb-06
3172 ---------------------
3173
3174 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3175 in pcreposix.h. I have copied the definition from pcre.h.
3176
3177 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3178 because pcre.h is no longer a built file.
3179
3180 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3181 not normally included in the compiled code.
3182
3183
3184 Version 6.5 01-Feb-06
3185 ---------------------
3186
3187 1. When using the partial match feature with pcre_dfa_exec(), it was not
3188 anchoring the second and subsequent partial matches at the new starting
3189 point. This could lead to incorrect results. For example, with the pattern
3190 /1234/, partially matching against "123" and then "a4" gave a match.
3191
3192 2. Changes to pcregrep:
3193
3194 (a) All non-match returns from pcre_exec() were being treated as failures
3195 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3196 error message is output. Some extra information is given for the
3197 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3198 probably the only errors that are likely to be caused by users (by
3199 specifying a regex that has nested indefinite repeats, for instance).
3200 If there are more than 20 of these errors, pcregrep is abandoned.
3201
3202 (b) A binary zero was treated as data while matching, but terminated the
3203 output line if it was written out. This has been fixed: binary zeroes
3204 are now no different to any other data bytes.
3205
3206 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3207 used to set a locale for matching. The --locale=xxxx long option has
3208 been added (no short equivalent) to specify a locale explicitly on the
3209 pcregrep command, overriding the environment variables.
3210
3211 (d) When -B was used with -n, some line numbers in the output were one less
3212 than they should have been.
3213
3214 (e) Added the -o (--only-matching) option.
3215
3216 (f) If -A or -C was used with -c (count only), some lines of context were
3217 accidentally printed for the final match.
3218
3219 (g) Added the -H (--with-filename) option.
3220
3221 (h) The combination of options -rh failed to suppress file names for files
3222 that were found from directory arguments.
3223
3224 (i) Added the -D (--devices) and -d (--directories) options.
3225
3226 (j) Added the -F (--fixed-strings) option.
3227
3228 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3229
3230 (l) Added the --colo(u)r option.
3231
3232 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3233 is not present by default.
3234
3235 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3236 items such as (?R) or (?1), when the recursion could match a number of
3237 alternatives. If it matched one of the alternatives, but subsequently,
3238 outside the recursion, there was a failure, the code tried to back up into
3239 the recursion. However, because of the way PCRE is implemented, this is not
3240 possible, and the result was an incorrect result from the match.
3241
3242 In order to prevent this happening, the specification of recursion has
3243 been changed so that all such subpatterns are automatically treated as
3244 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3245
3246 4. I had overlooked the fact that, in some locales, there are characters for
3247 which isalpha() is true but neither isupper() nor islower() are true. In
3248 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3249 and ordfeminine) are like this. This affected the treatment of \w and \W
3250 when they appeared in character classes, but not when they appeared outside
3251 a character class. The bit map for "word" characters is now created
3252 separately from the results of isalnum() instead of just taking it from the
3253 upper, lower, and digit maps. (Plus the underscore character, of course.)
3254
3255 5. The above bug also affected the handling of POSIX character classes such as
3256 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3257 permanent tables. Instead, the bit maps for such a class were previously
3258 created as the appropriate unions of the upper, lower, and digit bitmaps.
3259 Now they are created by subtraction from the [[:word:]] class, which has
3260 its own bitmap.
3261
3262 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3263 It is created by subtracting the vertical space characters (\x09, \x0a,
3264 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3265 subtraction was done in the overall bitmap for a character class, meaning
3266 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3267 be recognized. This bug has been fixed.
3268
3269 7. Patches from the folks at Google:
3270
3271 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3272 real life, but is still worth protecting against".
3273
3274 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3275 regular expressions".
3276
3277 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3278 have it.
3279
3280 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3281 "configure" and the latter not, in order to fix a problem somebody had
3282 with compiling the Arg class on HP-UX.
3283
3284 (e) Improve the error-handling of the C++ wrapper a little bit.
3285
3286 (f) New tests for checking recursion limiting.
3287
3288 8. The pcre_memmove() function, which is used only if the environment does not
3289 have a standard memmove() function (and is therefore rarely compiled),
3290 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3291 returning a result (though PCRE never actually uses the result).
3292
3293 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3294 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3295 returned instead of calling malloc() with an overflowing number that would
3296 most likely cause subsequent chaos.
3297
3298 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3299
3300 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3301 with this option is matched, the nmatch and pmatch options of regexec() are
3302 ignored.
3303
3304 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3305 provided in case anyone wants to the the POSIX interface with UTF-8
3306 strings.
3307
3308 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3309 C++ linking (needed for some HP-UX environments).
3310
3311 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3312 (unused parameter) and in the pcre_printint() function (omitted "default"
3313 switch label when the default is to do nothing).
3314
3315 15. Added some code to make it possible, when PCRE is compiled as a C++
3316 library, to replace subject pointers for pcre_exec() with a smart pointer
3317 class, thus making it possible to process discontinuous strings.
3318
3319 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3320 much the same function. They were added by different people who were trying
3321 to make PCRE easy to compile on non-Unix systems. It has been suggested
3322 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3323 for compiling on Windows systems. I have therefore replaced it with
3324 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3325 defaults to "extern" for C or "extern C" for C++, which works fine on
3326 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3327 SCOPE with something explicit in config.h. In addition:
3328
3329 (a) pcreposix.h still had just "extern" instead of either of these macros;
3330 I have replaced it with PCRE_DATA_SCOPE.
3331
3332 (b) Functions such as _pcre_xclass(), which are internal to the library,
3333 but external in the C sense, all had PCRE_EXPORT in their definitions.
3334 This is apparently wrong for the Windows case, so I have removed it.
3335 (It makes no difference on Unix-like systems.)
3336
3337 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3338 of recursive calls to match(). This is different to MATCH_LIMIT because
3339 that limits the total number of calls to match(), not all of which increase
3340 the depth of recursion. Limiting the recursion depth limits the amount of
3341 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3342 when PCRE is compiled, and changed at run time. A patch from Google adds
3343 this functionality to the C++ interface.
3344
3345 18. Changes to the handling of Unicode character properties:
3346
3347 (a) Updated the table to Unicode 4.1.0.
3348
3349 (b) Recognize characters that are not in the table as "Cn" (undefined).
3350
3351 (c) I revised the way the table is implemented to a much improved format
3352 which includes recognition of ranges. It now supports the ranges that
3353 are defined in UnicodeData.txt, and it also amalgamates other
3354 characters into ranges. This has reduced the number of entries in the
3355 table from around 16,000 to around 3,000, thus reducing its size
3356 considerably. I realized I did not need to use a tree structure after
3357 all - a binary chop search is just as efficient. Having reduced the
3358 number of entries, I extended their size from 6 bytes to 8 bytes to
3359 allow for more data.
3360
3361 (d) Added support for Unicode script names via properties such as \p{Han}.
3362
3363 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3364 matching that character.
3365
3366 20. When matching a repeated Unicode property with a minimum greater than zero,
3367 (for example \pL{2,}), PCRE could look past the end of the subject if it
3368 reached it while seeking the minimum number of characters. This could
3369 happen only if some of the characters were more than one byte long, because
3370 there is a check for at least the minimum number of bytes.
3371
3372 21. Refactored the implementation of \p and \P so as to be more general, to
3373 allow for more different types of property in future. This has changed the
3374 compiled form incompatibly. Anybody with saved compiled patterns that use
3375 \p or \P will have to recompile them.
3376
3377 22. Added "Any" and "L&" to the supported property types.
3378
3379 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3380 but give a compile time error if the value is greater than 0xff.
3381
3382 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3383 accidentally not being installed or uninstalled.
3384
3385 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3386 made were to insert the current release number. This seemed silly, because
3387 it made things harder for people building PCRE on systems that don't run
3388 "configure". I have turned pcre.h into a distributed file, no longer built
3389 by "configure", with the version identification directly included. There is
3390 no longer a pcre.h.in file.
3391
3392 However, this change necessitated a change to the pcre-config script as
3393 well. It is built from pcre-config.in, and one of the substitutions was the
3394 release number. I have updated configure.ac so that ./configure now finds
3395 the release number by grepping pcre.h.
3396
3397 26. Added the ability to run the tests under valgrind.
3398
3399
3400 Version 6.4 05-Sep-05
3401 ---------------------
3402
3403 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3404 "--" to be printed when multiple files were scanned, even when none of the
3405 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3406 consider it to be a bug, and have restored the previous behaviour.
3407
3408 2. A couple of code tidies to get rid of compiler warnings.
3409
3410 3. The pcretest program used to cheat by referring to symbols in the library
3411 whose names begin with _pcre_. These are internal symbols that are not
3412 really supposed to be visible externally, and in some environments it is
3413 possible to suppress them. The cheating is now confined to including
3414 certain files from the library's source, which is a bit cleaner.
3415
3416 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3417 file's purpose clearer.
3418
3419 5. Reorganized pcre_ucp_findchar().
3420
3421
3422 Version 6.3 15-Aug-05
3423 ---------------------
3424
3425 1. The file libpcre.pc.in did not have general read permission in the tarball.
3426
3427 2. There were some problems when building without C++ support:
3428
3429 (a) If C++ support was not built, "make install" and "make test" still
3430 tried to test it.
3431
3432 (b) There were problems when the value of CXX was explicitly set. Some
3433 changes have been made to try to fix these, and ...
3434
3435 (c) --disable-cpp can now be used to explicitly disable C++ support.
3436
3437 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3438 backslash in a target when C++ was disabled. This confuses some
3439 versions of "make", apparently. Using an intermediate variable solves
3440 this. (Same for CPP_LOBJ.)
3441
3442 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3443 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3444 necessary on certain architectures.
3445
3446 4. Added a setting of -export-symbols-regex to the link command to remove
3447 those symbols that are exported in the C sense, but actually are local
3448 within the library, and not documented. Their names all begin with
3449 "_pcre_". This is not a perfect job, because (a) we have to except some
3450 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3451 available (and never for static libraries). I have made a note to try to
3452 find a way round (a) in the future.
3453
3454
3455 Version 6.2 01-Aug-05
3456 ---------------------
3457
3458 1. There was no test for integer overflow of quantifier values. A construction
3459 such as {1111111111111111} would give undefined results. What is worse, if
3460 a minimum quantifier for a parenthesized subpattern overflowed and became
3461 negative, the calculation of the memory size went wrong. This could have
3462 led to memory overwriting.
3463
3464 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3465
3466 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3467 operating environments where this matters.
3468
3469 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3470 PCRE options from within the C++ wrapper.
3471
3472 5. Named capturing subpatterns were not being correctly counted when a pattern
3473 was compiled. This caused two problems: (a) If there were more than 100
3474 such subpatterns, the calculation of the memory needed for the whole
3475 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3476 back references of the form \12, where the number was greater than 9, were
3477 not recognized as back references, even though there were sufficient
3478 previous subpatterns.
3479
3480 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3481 versions of gcc, e.g. 2.95.4.
3482
3483
3484 Version 6.1 21-Jun-05
3485 ---------------------
3486
3487 1. There was one reference to the variable "posix" in pcretest.c that was not
3488 surrounded by "#if !defined NOPOSIX".
3489
3490 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3491 the cross-check on the old pcre_info() function, for the benefit of the
3492 cut-down version of PCRE that is currently imported into Exim.
3493
3494 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3495 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3496 patterns that start (?i) but I don't think that matters. The use of (?i) is
3497 just an example; this all applies to the other options as well.
3498
3499 4. Since libtool seems to echo the compile commands it is issuing, the output
3500 from "make" can be reduced a bit by putting "@" in front of each libtool
3501 compile command.
3502
3503 5. Patch from the folks at Google for configure.in to be a bit more thorough
3504 in checking for a suitable C++ installation before trying to compile the
3505 C++ stuff. This should fix a reported problem when a compiler was present,
3506 but no suitable headers.
3507
3508 6. The man pages all had just "PCRE" as their title. I have changed them to
3509 be the relevant file name. I have also arranged that these names are
3510 retained in the file doc/pcre.txt, which is a concatenation in text format
3511 of all the man pages except the little individual ones for each function.
3512
3513 7. The NON-UNIX-USE file had not been updated for the different set of source
3514 files that come with release 6. I also added a few comments about the C++
3515 wrapper.
3516
3517
3518 Version 6.0 07-Jun-05
3519 ---------------------
3520
3521 1. Some minor internal re-organization to help with my DFA experiments.
3522
3523 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3524 didn't matter for the library itself when fully configured, but did matter
3525 when compiling without UCP support, or within Exim, where the ucp files are
3526 not imported.
3527
3528 3. Refactoring of the library code to split up the various functions into
3529 different source modules. The addition of the new DFA matching code (see
3530 below) to a single monolithic source would have made it really too
3531 unwieldy, quite apart from causing all the code to be include in a
3532 statically linked application, when only some functions are used. This is
3533 relevant even without the DFA addition now that patterns can be compiled in
3534 one application and matched in another.
3535
3536 The downside of splitting up is that there have to be some external
3537 functions and data tables that are used internally in different modules of
3538 the library but which are not part of the API. These have all had their
3539 names changed to start with "_pcre_" so that they are unlikely to clash
3540 with other external names.
3541
3542 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3543 a different (DFA) algorithm. Although it is slower than the original
3544 function, it does have some advantages for certain types of matching
3545 problem.
3546
3547 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3548 including restarting after a partial match.
3549
3550 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3551 defined when compiling for Windows was sent to me. I have put it into the
3552 code, though I have no means of testing or verifying it.
3553
3554 7. Added the pcre_refcount() auxiliary function.
3555
3556 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3557 match before or at the first newline in the subject string. In pcretest,
3558 the /f option on a pattern can be used to set this.
3559
3560 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3561 would behave wrongly. This has been present in PCRE since release 4.0.
3562
3563 10. A number of changes to the pcregrep command:
3564
3565 (a) Refactored how -x works; insert ^(...)$ instead of setting
3566 PCRE_ANCHORED and checking the length, in preparation for adding
3567 something similar for -w.
3568
3569 (b) Added the -w (match as a word) option.
3570
3571 (c) Refactored the way lines are read and buffered so as to have more
3572 than one at a time available.
3573
3574 (d) Implemented a pcregrep test script.
3575
3576 (e) Added the -M (multiline match) option. This allows patterns to match
3577 over several lines of the subject. The buffering ensures that at least
3578 8K, or the rest of the document (whichever is the shorter) is available
3579 for matching (and similarly the previous 8K for lookbehind assertions).
3580
3581 (f) Changed the --help output so that it now says
3582
3583 -w, --word-regex(p)
3584
3585 instead of two lines, one with "regex" and the other with "regexp"
3586 because that confused at least one person since the short forms are the
3587 same. (This required a bit of code, as the output is generated
3588 automatically from a table. It wasn't just a text change.)
3589
3590 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3591 option but starts with a hyphen. Could be a pattern or a path name
3592 starting with a hyphen, for instance.
3593
3594 (h) "-" can be given as a file name to represent stdin.
3595
3596 (i) When file names are being printed, "(standard input)" is used for
3597 the standard input, for compatibility with GNU grep. Previously
3598 "<stdin>" was used.
3599
3600 (j) The option --label=xxx can be used to supply a name to be used for
3601 stdin when file names are being printed. There is no short form.
3602
3603 (k) Re-factored the options decoding logic because we are going to add
3604 two more options that take data. Such options can now be given in four
3605 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3606
3607 (l) Added the -A, -B, and -C options for requesting that lines of context
3608 around matches be printed.
3609
3610 (m) Added the -L option to print the names of files that do not contain
3611 any matching lines, that is, the complement of -l.
3612
3613 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3614 continue to scan other files.
3615
3616 (o) The -s option was incorrectly implemented. For compatibility with other
3617 greps, it now suppresses the error message for a non-existent or non-
3618 accessible file (but not the return code). There is a new option called
3619 -q that suppresses the output of matching lines, which was what -s was
3620 previously doing.
3621
3622 (p) Added --include and --exclude options to specify files for inclusion
3623 and exclusion when recursing.
3624
3625 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3626 Hopefully, it now does.
3627
3628 12. Missing cast in pcre_study().
3629
3630 13. Added an "uninstall" target to the makefile.
3631
3632 14. Replaced "extern" in the function prototypes in Makefile.in with
3633 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3634 world, but is set differently for Windows.
3635
3636 15. Added a second compiling function called pcre_compile2(). The only
3637 difference is that it has an extra argument, which is a pointer to an
3638 integer error code. When there is a compile-time failure, this is set
3639 non-zero, in addition to the error test pointer being set to point to an
3640 error message. The new argument may be NULL if no error number is required
3641 (but then you may as well call pcre_compile(), which is now just a
3642 wrapper). This facility is provided because some applications need a
3643 numeric error indication, but it has also enabled me to tidy up the way
3644 compile-time errors are handled in the POSIX wrapper.
3645
3646 16. Added VPATH=.libs to the makefile; this should help when building with one
3647 prefix path and installing with another. (Or so I'm told by someone who
3648 knows more about this stuff than I do.)
3649
3650 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3651 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3652 match everything, including newlines. This is not POSIX-compatible, but
3653 somebody wanted the feature. From pcretest it can be activated by using
3654 both the P and the s flags.
3655
3656 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3657
3658 19. libpcre.pc was being incorrectly installed as executable.
3659
3660 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3661 it now also looks for '\r' so that it will work unmodified on Windows.
3662
3663 21. Added Google's contributed C++ wrapper to the distribution.
3664
3665 22. Added some untidy missing memory free() calls in pcretest, to keep
3666 Electric Fence happy when testing.
3667
3668
3669
3670 Version 5.0 13-Sep-04
3671 ---------------------
3672
3673 1. Internal change: literal characters are no longer packed up into items
3674 containing multiple characters in a single byte-string. Each character
3675 is now matched using a separate opcode. However, there may be more than one
3676 byte in the character in UTF-8 mode.
3677
3678 2. The pcre_callout_block structure has two new fields: pattern_position and
3679 next_item_length. These contain the offset in the pattern to the next match
3680 item, and its length, respectively.
3681
3682 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3683 insertion of callouts before each pattern item. Added the /C option to
3684 pcretest to make use of this.
3685
3686 4. On the advice of a Windows user, the lines
3687
3688 #if defined(_WIN32) || defined(WIN32)
3689 _setmode( _fileno( stdout ), 0x8000 );
3690 #endif /* defined(_WIN32) || defined(WIN32) */
3691
3692 have been added to the source of pcretest. This apparently does useful
3693 magic in relation to line terminators.
3694
3695 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3696 for the benefit of those environments where the "b" makes a difference.
3697
3698 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3699 to know about it. I have put a hack into configure.in that adds in code
3700 to set GCC=yes if CC=icc. This seems to end up at a point in the
3701 generated configure script that is early enough to affect the setting of
3702 compiler options, which is what is needed, but I have no means of testing
3703 whether it really works. (The user who reported this had patched the
3704 generated configure script, which of course I cannot do.)
3705
3706 LATER: After change 22 below (new libtool files), the configure script
3707 seems to know about icc (and also ecc). Therefore, I have commented out
3708 this hack in configure.in.
3709
3710 7. Added support for pkg-config (2 patches were sent in).
3711
3712 8. Negated POSIX character classes that used a combination of internal tables
3713 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3714 [[:^ascii]]. Typically, they would match almost any characters. The other
3715 POSIX classes were not broken in this way.
3716
3717 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3718 to find the match, as PCRE was deluded into thinking that the match had to
3719 start at the start point or following a newline. The same bug applied to
3720 patterns with negative forward assertions or any backward assertions
3721 preceding ".*" at the start, unless the pattern required a fixed first
3722 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3723
3724 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3725 starting at the last subject character, bytes beyond the end of the subject
3726 string were read.
3727
3728 11. Renamed the variable "class" as "classbits" to make life easier for C++
3729 users. (Previously there was a macro definition, but it apparently wasn't
3730 enough.)
3731
3732 12. Added the new field "tables" to the extra data so that tables can be passed
3733 in at exec time, or the internal tables can be re-selected. This allows
3734 a compiled regex to be saved and re-used at a later time by a different
3735 program that might have everything at different addresses.
3736
3737 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3738 -R library as well as a -L library.
3739
3740 14. The debugging options of pcretest (-d on the command line or D on a
3741 pattern) showed incorrect output for anything following an extended class
3742 that contained multibyte characters and which was followed by a quantifier.
3743
3744 15. Added optional support for general category Unicode character properties
3745 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3746 support. It adds about 90K to the size of the library. The meanings of the
3747 inbuilt class escapes such as \d and \s have NOT been changed.
3748
3749 16. Updated pcredemo.c to include calls to free() to release the memory for the
3750 compiled pattern.
3751
3752 17. The generated file chartables.c was being created in the source directory
3753 instead of in the building directory. This caused the build to fail if the
3754 source directory was different from the building directory, and was
3755 read-only.
3756
3757 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3758 file. No doubt somebody will tell me if they don't make sense... Also added
3759 Dan Mooney's comments about building on OpenVMS.
3760
3761 19. Added support for partial matching via the PCRE_PARTIAL option for
3762 pcre_exec() and the \P data escape in pcretest.
3763
3764 20. Extended pcretest with 3 new pattern features:
3765
3766 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3767 write the compiled pattern to the file whose name is "rest-of-line".
3768 This is a straight binary dump of the data, with the saved pointer to
3769 the character tables forced to be NULL. The study data, if any, is
3770 written too. After writing, pcretest reads a new pattern.
3771
3772 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3773 compiled pattern from the given file. There must not be any
3774 occurrences of "<" in the file name (pretty unlikely); if there are,
3775 pcretest will instead treat the initial "<" as a pattern delimiter.
3776 After reading in the pattern, pcretest goes on to read data lines as
3777 usual.
3778
3779 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3780 and 16-bit fields in a compiled pattern, to simulate a pattern that
3781 was compiled on a host of opposite endianness.
3782
3783 21. The pcre-exec() function can now cope with patterns that were compiled on
3784 hosts of opposite endianness, with this restriction:
3785
3786 As for any compiled expression that is saved and used later, the tables
3787 pointer field cannot be preserved; the extra_data field in the arguments
3788 to pcre_exec() should be used to pass in a tables address if a value
3789 other than the default internal tables were used at compile time.
3790
3791 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3792 now diagnosed as an error. Previously, most of the time, a negative number
3793 would have been treated as zero, but if in addition "ovector" was passed as
3794 NULL, a crash could occur.
3795
3796 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3797 new versions from the libtool 1.5 distribution (the last one is a copy of
3798 a file called libtool.m4). This seems to have fixed the need to patch
3799 "configure" to support Darwin 1.3 (which I used to do). However, I still
3800 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3801 workstation).
3802
3803 24. Changed the PCRE licence to be the more standard "BSD" licence.
3804
3805
3806 Version 4.5 01-Dec-03
3807 ---------------------
3808
3809 1. There has been some re-arrangement of the code for the match() function so
3810 that it can be compiled in a version that does not call itself recursively.
3811 Instead, it keeps those local variables that need separate instances for
3812 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3813 needs to "recurse". Keeping track of where control must go is done by means
3814 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3815 hide most of the details from the main code, and operates only if
3816 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3817 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3818 operating.
3819
3820 To make it easier for callers to provide specially tailored get/free
3821 functions for this usage, two new functions, pcre_stack_malloc, and
3822 pcre_stack_free, are used. They are always called in strict stacking order,
3823 and the size of block requested is always the same.
3824
3825 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3826 PCRE has been compiled to use the stack or the heap for recursion. The
3827 -C option of pcretest uses this to show which version is compiled.
3828
3829 A new data escape \S, is added to pcretest; it causes the amounts of store
3830 obtained and freed by both kinds of malloc/free at match time to be added
3831 to the output.
3832
3833 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3834 what's available on my current Linux desktop machine.
3835
3836 3. When matching a UTF-8 string, the test for a valid string at the start has
3837 been extended. If start_offset is not zero, PCRE now checks that it points
3838 to a byte that is the start of a UTF-8 character. If not, it returns
3839 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3840 this is necessary because there may be backward assertions in the pattern.
3841 When matching the same subject several times, it may save resources to use
3842 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3843
3844 4. The code for checking the validity of UTF-8 strings has been tightened so
3845 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3846 containing "overlong sequences".
3847
3848 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3849 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3850 should have been "&", but it just so happened that all the cases this let
3851 through by mistake were picked up later in the function.
3852
3853 6. I had used a variable called "isblank" - this is a C99 function, causing
3854 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3855
3856 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3857 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3858 so the version is shown; (c) stop "make test" echoing "./RunTest".
3859
3860 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3861
3862 9. The prototype for memmove() for systems that don't have it was using
3863 size_t, but the inclusion of the header that defines size_t was later. I've
3864 moved the #includes for the C headers earlier to avoid this.
3865
3866 10. Added some adjustments to the code to make it easier to compiler on certain
3867 special systems:
3868
3869 (a) Some "const" qualifiers were missing.
3870 (b) Added the macro EXPORT before all exported functions; by default this
3871 is defined to be empty.
3872 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3873 that it reads its output file name as an argument instead of writing
3874 to the standard output and assuming this can be redirected.
3875
3876 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3877 class containing characters with values greater than 255, PCRE compilation
3878 went into a loop.
3879
3880 12. A recursive reference to a subpattern that was within another subpattern
3881 that had a minimum quantifier of zero caused PCRE to crash. For example,
3882 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3883 recursion. If the recursively-called subpattern itself had a zero repeat,
3884 that was OK.
3885
3886 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3887 buffer into which it was copied (for escape processing) was still set at
3888 1024, so long lines caused crashes.
3889
3890 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3891 "internal error: code overflow...". This applied to any character class
3892 that was followed by a possessive quantifier.
3893
3894 15. Modified the Makefile to add libpcre.la as a prerequisite for
3895 libpcreposix.la because I was told this is needed for a parallel build to
3896 work.
3897
3898 16. If a pattern that contained .* following optional items at the start was
3899 studied, the wrong optimizing data was generated, leading to matching
3900 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3901 matching string must start with a or b or c. The correct conclusion for
3902 this pattern is that a match can start with any character.
3903
3904
3905 Version 4.4 13-Aug-03
3906 ---------------------
3907
3908 1. In UTF-8 mode, a character class containing characters with values between
3909 127 and 255 was not handled correctly if the compiled pattern was studied.
3910 In fixing this, I have also improved the studying algorithm for such
3911 classes (slightly).
3912
3913 2. Three internal functions had redundant arguments passed to them. Removal
3914 might give a very teeny performance improvement.
3915
3916 3. Documentation bug: the value of the capture_top field in a callout is *one
3917 more than* the number of the hightest numbered captured substring.
3918
3919 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3920 in incorrectly linking with a previously installed version. They now link
3921 explicitly with libpcre.la.
3922
3923 5. configure.in no longer needs to recognize Cygwin specially.
3924
3925 6. A problem in pcre.in for Windows platforms is fixed.
3926
3927 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3928 pcretest, it used to include the size of the study block as part of its
3929 output. Unfortunately, the structure contains a field that has a different
3930 size on different hardware architectures. This meant that the tests that
3931 showed this size failed. As the block is currently always of a fixed size,
3932 this information isn't actually particularly useful in pcretest output, so
3933 I have just removed it.
3934
3935 8. Three pre-processor statements accidentally did not start in column 1.
3936 Sadly, there are *still* compilers around that complain, even though
3937 standard C has not required this for well over a decade. Sigh.
3938
3939 9. In pcretest, the code for checking callouts passed small integers in the
3940 callout_data field, which is a void * field. However, some picky compilers
3941 complained about the casts involved for this on 64-bit systems. Now
3942 pcretest passes the address of the small integer instead, which should get
3943 rid of the warnings.
3944
3945 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3946 both compile and run time, and gives an error if an invalid UTF-8 sequence
3947 is found. There is a option for disabling this check in cases where the
3948 string is known to be correct and/or the maximum performance is wanted.
3949
3950 11. In response to a bug report, I changed one line in Makefile.in from
3951
3952 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3953 to
3954 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3955
3956 to look similar to other lines, but I have no way of telling whether this
3957 is the right thing to do, as I do not use Windows. No doubt I'll get told
3958 if it's wrong...
3959
3960
3961 Version 4.3 21-May-03
3962 ---------------------
3963
3964 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3965 Makefile.
3966
3967 2. Some refactoring to improve the quality of the code:
3968
3969 (i) The utf8_table... variables are now declared "const".
3970
3971 (ii) The code for \cx, which used the "case flipping" table to upper case
3972 lower case letters, now just substracts 32. This is ASCII-specific,
3973 but the whole concept of \cx is ASCII-specific, so it seems
3974 reasonable.
3975
3976 (iii) PCRE was using its character types table to recognize decimal and
3977 hexadecimal digits in the pattern. This is silly, because it handles
3978 only 0-9, a-f, and A-F, but the character types table is locale-
3979 specific, which means strange things might happen. A private
3980 table is now used for this - though it costs 256 bytes, a table is
3981 much faster than multiple explicit tests. Of course, the standard
3982 character types table is still used for matching digits in subject
3983 strings against \d.
3984
3985 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3986 ending in _t are). So I've renamed it as ESC_tee.
3987
3988 3. The first argument for regexec() in the POSIX wrapper should have been
3989 defined as "const".
3990
3991 4. Changed pcretest to use malloc() for its buffers so that they can be
3992 Electric Fenced for debugging.
3993
3994 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3995 to read one or more bytes before the start of the subject string. Often this
3996 had no effect on PCRE's behaviour, but in some circumstances it could
3997 provoke a segmentation fault.
3998
3999 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4000 to try to read one or more bytes before the start of the subject string.
4001
4002 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4003 UTF-8 support could misbehave in various ways if the subject string
4004 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4005 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4006 back over UTF-8 characters.)
4007
4008
4009 Version 4.2 14-Apr-03
4010 ---------------------
4011
4012 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4013
4014 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4015 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4016 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4017 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4018 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4019 and BUILD_EXEEXT
4020 Note: automatic setting of the BUILD variables is not yet working
4021 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4022 compile-time but not at link-time
4023 [LINK]: use for linking executables only
4024 make different versions for Windows and non-Windows
4025 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4026 libraries
4027 [LINK_FOR_BUILD]: new variable
4028 [OBJEXT]: use throughout
4029 [EXEEXT]: use throughout
4030 <winshared>: new target
4031 <wininstall>: new target
4032 <dftables.o>: use native compiler
4033 <dftables>: use native linker
4034 <install>: handle Windows platform correctly
4035 <clean>: ditto
4036 <check>: ditto
4037 copy DLL to top builddir before testing
4038
4039 As part of these changes, -no-undefined was removed again. This was reported
4040 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4041 in any case.
4042
4043 3. Some tidies to get rid of compiler warnings:
4044
4045 . In the match_data structure, match_limit was an unsigned long int, whereas
4046 match_call_count was an int. I've made them both unsigned long ints.
4047
4048 . In pcretest the fact that a const uschar * doesn't automatically cast to
4049 a void * provoked a warning.
4050
4051 . Turning on some more compiler warnings threw up some "shadow" variables
4052 and a few more missing casts.
4053
4054 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4055 option, a class that contained a single character with a value between 128
4056 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4057
4058 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4059 option, a class that contained several characters, but with at least one
4060 whose value was between 128 and 255 caused PCRE to crash.
4061
4062
4063 Version 4.1 12-Mar-03
4064 ---------------------
4065
4066 1. Compiling with gcc -pedantic found a couple of places where casts were
4067 needed, and a string in dftables.c that was longer than standard compilers are
4068 required to support.
4069
4070 2. Compiling with Sun's compiler found a few more places where the code could
4071 be tidied up in order to avoid warnings.
4072
4073 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4074 first of these names is deprecated in the latest Autoconf in favour of the name
4075 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4076 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4077 analogy I have changed it to CFLAGS_FOR_BUILD.
4078
4079 4. Added -no-undefined to the linking command in the Makefile, because this is
4080 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4081 linking step for the pcreposix library.
4082
4083 5. PCRE was failing to diagnose the case of two named groups with the same
4084 name.
4085
4086 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4087 literal character that is needed in the subject for a match, and scans along to
4088 ensure that it is present before embarking on the full matching process. This
4089 saves time in cases of nested unlimited repeats that are never going to match.
4090 Problem: the scan can take a lot of time if the subject is very long (e.g.
4091 megabytes), thus penalizing straightforward matches. It is now done only if the
4092 amount of subject to be scanned is less than 1000 bytes.
4093
4094 7. A lesser problem with the same optimization is that it was recording the
4095 first character of an anchored pattern as "needed", thus provoking a search
4096 right along the subject, even when the first match of the pattern was going to
4097 fail. The "needed" character is now not set for anchored patterns, unless it
4098 follows something in the pattern that is of non-fixed length. Thus, it still
4099 fulfils its original purpose of finding quick non-matches in cases of nested
4100 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4101
4102
4103 Version 4.0 17-Feb-03
4104 ---------------------
4105
4106 1. If a comment in an extended regex that started immediately after a meta-item
4107 extended to the end of string, PCRE compiled incorrect data. This could lead to
4108 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4109
4110 2. Moved to autoconf 2.53 and libtool 1.4.2.
4111
4112 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4113 the special perltest8 script is no longer needed - all the tests can be run
4114 from a single perltest script.
4115
4116 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4117 by \s. It has now been removed in PCRE. This means it isn't recognized as
4118 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4119 class [:space:] *does* include VT, thereby creating a mess.
4120
4121 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4122 space and tab.
4123
4124 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4125 its new features into the main test script, reducing the number of scripts.
4126
4127 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4128 were backward compatible, and made the (?i) apply to the whole pattern, as if
4129 /i were given. Now it behaves more logically, and applies the option setting
4130 only to what follows. PCRE has been changed to follow suit. However, if it
4131 finds options settings right at the start of the pattern, it extracts them into
4132 the global options, as before. Thus, they show up in the info data.
4133
4134 8. Added support for the \Q...\E escape sequence. Characters in between are
4135 treated as literals. This is slightly different from Perl in that $ and @ are
4136 also handled as literals inside the quotes. In Perl, they will cause variable
4137 interpolation. Note the following examples:
4138
4139 Pattern PCRE matches Perl matches
4140
4141 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4142 \Qabc\$xyz\E abc\$xyz abc\$xyz
4143 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4144
4145 For compatibility with Perl, \Q...\E sequences are recognized inside character
4146 classes as well as outside them.
4147
4148 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4149 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4150 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4151 signed/unsigned warnings.
4152
4153 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4154 option for pcretest, so I've replaced it by a simple function that does just
4155 that job.
4156
4157 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4158 "pcregrep -".
4159
4160 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4161 Java package. This provides some syntactic sugar for simple cases of what my
4162 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4163 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4164 item, you can use this simplified notation. Note that only makes sense with
4165 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4166 greediness, whatever the setting of the PCRE_UNGREEDY option.
4167
4168 13. A change of greediness default within a pattern was not taking effect at
4169 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4170 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4171 was abstracted outside.
4172
4173 14. PCRE now supports the \G assertion. It is true when the current matching
4174 position is at the start point of the match. This differs from \A when the
4175 starting offset is non-zero. Used with the /g option of pcretest (or similar
4176 code), it works in the same way as it does for Perl's /g option. If all
4177 alternatives of a regex begin with \G, the expression is anchored to the start
4178 match position, and the "anchored" flag is set in the compiled expression.
4179
4180 15. Some bugs concerning the handling of certain option changes within patterns
4181 have been fixed. These applied to options other than (?ims). For example,
4182 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4183 been the other way round. Some of this was related to change 7 above.
4184
4185 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4186 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4187 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4188 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4189
4190 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4191 mode. Unlike ".", it always matches newline, whatever the setting of
4192 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4193 assertions. Perl allows it, but it doesn't (in general) work because it can't
4194 calculate the length of the lookbehind. At least, that's the case for Perl
4195 5.8.0 - I've been told they are going to document that it doesn't work in
4196 future.
4197
4198 18. Added an error diagnosis for escapes that PCRE does not support: these are
4199 \L, \l, \N, \P, \p, \U, \u, and \X.
4200
4201 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4202 reading past the end of the pattern in cases such as /[abcd/.
4203
4204 20. PCRE was getting more memory than necessary for patterns with classes that
4205 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4206
4207 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4208 compiling PCRE for use with Virtual Pascal.
4209
4210 22. Small fix to the Makefile to make it work properly if the build is done
4211 outside the source tree.
4212
4213 23. Added a new extension: a condition to go with recursion. If a conditional
4214 subpattern starts with (?(R) the "true" branch is used if recursion has
4215 happened, whereas the "false" branch is used only at the top level.
4216
4217 24. When there was a very long string of literal characters (over 255 bytes
4218 without UTF support, over 250 bytes with UTF support), the computation of how
4219 much memory was required could be incorrect, leading to segfaults or other
4220 strange effects.
4221
4222 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4223 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4224 there was a subsequent back reference to those brackets. This meant that, for
4225 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4226 possible to check for precisely this case. All we can do is abandon the
4227 optimization if .* occurs inside capturing brackets when there are any back
4228 references whatsoever. (See below for a better fix that came later.)
4229
4230 26. The handling of the optimization for finding the first character of a
4231 non-anchored pattern, and for finding a character that is required later in the
4232 match were failing in some cases. This didn't break the matching; it just
4233 failed to optimize when it could. The way this is done has been re-implemented.
4234
4235 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4236
4237 28. Added a new feature that provides some of the functionality that Perl
4238 provides with (?{...}). The facility is termed a "callout". The way it is done
4239 in PCRE is for the caller to provide an optional function, by setting
4240 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4241 global variable. By default it is unset, which disables all calling out. To get
4242 the function called, the regex must include (?C) at appropriate points. This
4243 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4244 This provides a means of identifying different callout points. When PCRE
4245 reaches such a point in the regex, if pcre_callout has been set, the external
4246 function is called. It is provided with data in a structure called
4247 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4248 matching continues; if it returns a non-zero value, the match at the current
4249 point fails. However, backtracking will occur if possible. [This was changed
4250 later and other features added - see item 49 below.]
4251
4252 29. pcretest is upgraded to test the callout functionality. It provides a
4253 callout function that displays information. By default, it shows the start of
4254 the match and the current position in the text. There are some new data escapes
4255 to vary what happens:
4256
4257 \C+ in addition, show current contents of captured substrings
4258 \C- do not supply a callout function
4259 \C!n return 1 when callout number n is reached
4260 \C!n!m return 1 when callout number n is reached for the mth time
4261
4262 30. If pcregrep was called with the -l option and just a single file name, it
4263 output "<stdin>" if a match was found, instead of the file name.
4264
4265 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4266 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4267 pcre_exec(). This saves a malloc/free per call. The default value of
4268 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4269 when configuring.
4270
4271 32. The default maximum size of a compiled pattern is 64K. There have been a
4272 few cases of people hitting this limit. The code now uses macros to handle the
4273 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4274 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4275 configuring. Tests 2 and 5 work only with 2-byte links because they output
4276 debugging information about compiled patterns.
4277
4278 33. Internal code re-arrangements:
4279
4280 (a) Moved the debugging function for printing out a compiled regex into
4281 its own source file (printint.c) and used #include to pull it into
4282 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4283 separate copies.
4284
4285 (b) Defined the list of op-code names for debugging as a macro in
4286 internal.h so that it is next to the definition of the opcodes.
4287
4288 (c) Defined a table of op-code lengths for simpler skipping along compiled
4289 code. This is again a macro in internal.h so that it is next to the
4290 definition of the opcodes.
4291
4292 34. Added support for recursive calls to individual subpatterns, along the
4293 lines of Robin Houston's patch (but implemented somewhat differently).
4294
4295 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4296 allow it to read and process whole directories in Win32. This code was
4297 contributed by Lionel Fourquaux; it has not been tested by me.
4298
4299 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4300 used to name a group. Names consist of alphanumerics and underscores, and must
4301 be unique. Back references use the syntax (?P=name) and recursive calls use
4302 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4303 numbers. The function pcre_fullinfo() can be used after compilation to extract
4304 a name/number map. There are three relevant calls:
4305
4306 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4307 PCRE_INFO_NAMECOUNT yields the number of entries
4308 PCRE_INFO_NAMETABLE yields a pointer to the map.
4309
4310 The map is a vector of fixed-size entries. The size of each entry depends on
4311 the length of the longest name used. The first two bytes of each entry are the
4312 group number, most significant byte first. There follows the corresponding
4313 name, zero terminated. The names are in alphabetical order.
4314
4315 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4316 case instead of 255. Making it the same both with and without UTF-8 support
4317 means that the same test output works with both.
4318
4319 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4320 calling malloc() with a zero argument.
4321
4322 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4323 optimization. I've improved things by keeping a bitmap of backreferences with
4324 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4325 fact referenced, the optimization can be applied. It is unlikely that a
4326 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4327 the match to follow \n) will appear inside brackets with a number greater than
4328 31, but if it does, any back reference > 31 suppresses the optimization.
4329
4330 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4331 of disabling numbered capturing parentheses. Any opening parenthesis that is
4332 not followed by ? behaves as if it were followed by ?: but named parentheses
4333 can still be used for capturing (and they will acquire numbers in the usual
4334 way).
4335
4336 41. Redesigned the return codes from the match() function into yes/no/error so
4337 that errors can be passed back from deep inside the nested calls. A malloc
4338 failure while inside a recursive subpattern call now causes the
4339 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4340
4341 42. It is now possible to set a limit on the number of times the match()
4342 function is called in a call to pcre_exec(). This facility makes it possible to
4343 limit the amount of recursion and backtracking, though not in a directly
4344 obvious way, because the match() function is used in a number of different
4345 circumstances. The count starts from zero for each position in the subject
4346 string (for non-anchored patterns). The default limit is, for compatibility, a
4347 large number, namely 10 000 000. You can change this in two ways:
4348
4349 (a) When configuring PCRE before making, you can use --with-match-limit=n
4350 to set a default value for the compiled library.
4351
4352 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4353 a different value is set. See 45 below.
4354
4355 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4356
4357 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4358 of things that can be changed at compile time. The first argument specifies
4359 what is wanted and the second points to where the information is to be placed.
4360 The current list of available information is:
4361
4362 PCRE_CONFIG_UTF8
4363
4364 The output is an integer that is set to one if UTF-8 support is available;
4365 otherwise it is set to zero.
4366
4367 PCRE_CONFIG_NEWLINE
4368
4369 The output is an integer that it set to the value of the code that is used for
4370 newline. It is either LF (10) or CR (13).
4371
4372 PCRE_CONFIG_LINK_SIZE
4373
4374 The output is an integer that contains the number of bytes used for internal
4375 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4376
4377 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4378
4379 The output is an integer that contains the threshold above which the POSIX
4380 interface uses malloc() for output vectors. See item 31 above.
4381
4382 PCRE_CONFIG_MATCH_LIMIT
4383
4384 The output is an unsigned integer that contains the default limit of the number
4385 of match() calls in a pcre_exec() execution. See 42 above.
4386
4387 44. pcretest has been upgraded by the addition of the -C option. This causes it
4388 to extract all the available output from the new pcre_config() function, and to
4389 output it. The program then exits immediately.
4390
4391 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4392 order to support additional features. One way would have been to define
4393 pcre_exec2() (for example) with extra arguments, but this would not have been
4394 extensible, and would also have required all calls to the original function to
4395 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4396 is used for passing in "extra" data from pcre_study().
4397
4398 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4399 contains the following fields:
4400
4401 flags a bitmap indicating which of the following fields are set
4402 study_data opaque data from pcre_study()
4403 match_limit a way of specifying a limit on match() calls for a specific
4404 call to pcre_exec()
4405 callout_data data for callouts (see 49 below)
4406
4407 The flag bits are also defined in pcre.h, and are
4408
4409 PCRE_EXTRA_STUDY_DATA
4410 PCRE_EXTRA_MATCH_LIMIT
4411 PCRE_EXTRA_CALLOUT_DATA
4412
4413 The pcre_study() function now returns one of these new pcre_extra blocks, with
4414 the actual study data pointed to by the study_data field, and the
4415 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4416 before. That is, this change is entirely upwards-compatible and requires no
4417 change to existing code.
4418
4419 If you want to pass in additional data to pcre_exec(), you can either place it
4420 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4421 block.
4422
4423 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4424 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4425 times with different match limits, until it finds the minimum value needed for
4426 pcre_exec() to complete. The value is then output. This can be instructive; for
4427 most simple matches the number is quite small, but for pathological cases it
4428 gets very large very quickly.
4429
4430 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4431 returns the size of the data block pointed to by the study_data field in a
4432 pcre_extra block, that is, the value that was passed as the argument to
4433 pcre_malloc() when PCRE was getting memory in which to place the information
4434 created by pcre_study(). The fourth argument should point to a size_t variable.
4435 pcretest has been extended so that this information is shown after a successful
4436 pcre_study() call when information about the compiled regex is being displayed.
4437
4438 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4439 because what follows is always an absolute path. (Later: it turns out that this
4440 is more than cosmetic for MinGW, because it doesn't like empty path
4441 components.)
4442
4443 49. Some changes have been made to the callout feature (see 28 above):
4444
4445 (i) A callout function now has three choices for what it returns:
4446
4447 0 => success, carry on matching
4448 > 0 => failure at this point, but backtrack if possible
4449 < 0 => serious error, return this value from pcre_exec()
4450
4451 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4452 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4453 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4454 use by callout functions. It will never be used by PCRE itself.
4455
4456 (ii) The pcre_extra structure (see 45 above) has a void * field called
4457 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4458 pcre_callout_block structure has a field of the same name. The contents of
4459 the field passed in the pcre_extra structure are passed to the callout
4460 function in the corresponding field in the callout block. This makes it
4461 easier to use the same callout-containing regex from multiple threads. For
4462 testing, the pcretest program has a new data escape
4463
4464 \C*n pass the number n (may be negative) as callout_data
4465
4466 If the callout function in pcretest receives a non-zero value as
4467 callout_data, it returns that value.
4468
4469 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4470 there were some redundant $(CFLAGS) in commands that are now specified as
4471 $(LINK), which already includes $(CFLAGS).
4472
4473 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4474 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4475 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4476 one-byte characters throughout. Note that case-insensitive matching applies
4477 only to characters whose values are less than 256. PCRE doesn't support the
4478 notion of cases for higher-valued characters.
4479
4480 (i) A character class whose characters are all within 0-255 is handled as
4481 a bit map, and the map is inverted for negative classes. Previously, a
4482 character > 255 always failed to match such a class; however it should
4483 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4484
4485 (ii) A negated character class with a single character < 255 is coded as
4486 "not this character" (OP_NOT). This wasn't working properly when the test
4487 character was multibyte, either singly or repeated.
4488
4489 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4490 mode, for example: \x{100}{2,3}.
4491
4492 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4493 singly or repeated) now correctly test multibyte characters. However,
4494 PCRE doesn't recognize any characters with values greater than 255 as
4495 digits, spaces, or word characters. Such characters always match \D, \S,
4496 and \W, and never match \d, \s, or \w.
4497
4498 (v) Classes may now contain characters and character ranges with values
4499 greater than 255. For example: [ab\x{100}-\x{400}].
4500
4501 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4502 PCRE in UTF-8 mode.
4503
4504 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4505 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4506 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4507 value.)
4508
4509 53. The single man page has become too large. I have therefore split it up into
4510 a number of separate man pages. These also give rise to individual HTML pages;
4511 these are now put in a separate directory, and there is an index.html page that
4512 lists them all. Some hyperlinking between the pages has been installed.
4513
4514 54. Added convenience functions for handling named capturing parentheses.
4515
4516 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4517 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4518 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4519 are faulted.
4520
4521 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4522 calling configure. These values are used when compiling the dftables.c program
4523 which is run to generate the source of the default character tables. They
4524 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4525 you will need to set these values.
4526
4527 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4528
4529
4530 Version 3.9 02-Jan-02
4531 ---------------------
4532
4533 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4534
4535 2. If --disable-static was given, the building process failed when trying to
4536 build pcretest and pcregrep. (For some reason it was using libtool to compile
4537 them, which is not right, as they aren't part of the library.)
4538
4539
4540 Version 3.8 18-Dec-01
4541 ---------------------
4542
4543 1. The experimental UTF-8 code was completely screwed up. It was packing the
4544 bytes in the wrong order. How dumb can you get?
4545
4546
4547 Version 3.7 29-Oct-01
4548 ---------------------
4549
4550 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4551 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4552 this didn't happen under Solaris 8, where I normally test things.
4553
4554 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4555 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4556 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4557 this fix an hour or so after the initial 3.7 release.)
4558
4559
4560 Version 3.6 23-Oct-01
4561 ---------------------
4562
4563 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4564 offsets passed as NULL with zero offset count.
4565
4566 2. The config.guess and config.sub files had not been updated when I moved to
4567 the latest autoconf.
4568
4569
4570 Version 3.5 15-Aug-01
4571 ---------------------
4572
4573 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4574 had been forgotten.
4575
4576 2. By using declared but undefined structures, we can avoid using "void"
4577 definitions in pcre.h while keeping the internal definitions of the structures
4578 private.
4579
4580 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4581 user point of view, this means that both static and shared libraries are built
4582 by default, but this can be individually controlled. More of the work of
4583 handling this static/shared cases is now inside libtool instead of PCRE's make
4584 file.
4585
4586 4. The pcretest utility is now installed along with pcregrep because it is
4587 useful for users (to test regexs) and by doing this, it automatically gets
4588 relinked by libtool. The documentation has been turned into a man page, so
4589 there are now .1, .txt, and .html versions in /doc.
4590
4591 5. Upgrades to pcregrep:
4592 (i) Added long-form option names like gnu grep.
4593 (ii) Added --help to list all options with an explanatory phrase.
4594 (iii) Added -r, --recursive to recurse into sub-directories.
4595 (iv) Added -f, --file to read patterns from a file.
4596
4597 6. pcre_exec() was referring to its "code" argument before testing that
4598 argument for NULL (and giving an error if it was NULL).
4599
4600 7. Upgraded Makefile.in to allow for compiling in a different directory from
4601 the source directory.
4602
4603 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4604 options bits, the pointer it was passed was to an int instead of to an unsigned
4605 long int. This mattered only on 64-bit systems.
4606
4607 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4608 generated) instead of pcre.in, which it its source. Also made the same change
4609 in several of the .c files.
4610
4611 10. A new release of gcc defines printf() as a macro, which broke pcretest
4612 because it had an ifdef in the middle of a string argument for printf(). Fixed
4613 by using separate calls to printf().
4614
4615 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4616 script, to force use of CR or LF instead of \n in the source. On non-Unix
4617 systems, the value can be set in config.h.
4618
4619 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4620 absolute limit. Changed the text of the error message to make this clear, and
4621 likewise updated the man page.
4622
4623 13. The limit of 99 on the number of capturing subpatterns has been removed.
4624 The new limit is 65535, which I hope will not be a "real" limit.
4625
4626
4627 Version 3.4 22-Aug-00
4628 ---------------------
4629
4630 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4631
4632 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4633
4634
4635 Version 3.3 01-Aug-00
4636 ---------------------
4637
4638 1. If an octal character was given, but the value was greater than \377, it
4639 was not getting masked to the least significant bits, as documented. This could
4640 lead to crashes in some systems.
4641
4642 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4643 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4644
4645 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4646 These just pass their arguments on to (pcre_free)(), but they are provided
4647 because some uses of PCRE bind it to non-C systems that can call its functions,
4648 but cannot call free() or pcre_free() directly.
4649
4650 4. Add "make test" as a synonym for "make check". Corrected some comments in
4651 the Makefile.
4652
4653 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4654 Makefile.
4655
4656 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4657 command called pgrep for grepping around the active processes.
4658
4659 7. Added the beginnings of support for UTF-8 character strings.
4660
4661 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4662 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4663 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4664 out for the ar command.)
4665
4666
4667 Version 3.2 12-May-00
4668 ---------------------
4669
4670 This is purely a bug fixing release.
4671
4672 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4673 of ZA. This was just one example of several cases that could provoke this bug,
4674 which was introduced by change 9 of version 2.00. The code for breaking
4675 infinite loops after an iteration that matches an empty string was't working
4676 correctly.
4677
4678 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4679 when matched against abbab (for example). After matching an empty string, it
4680 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4681 caused it to match further down the string than it should.
4682
4683 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4684 was there because it doesn't seem to be needed, and it causes trouble on some
4685 systems, as it is not a Standard C header. It has been removed.
4686
4687 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4688 were reported on the Macintosh. The changes were from
4689
4690 while ((c = *(++ptr)) != 0 && c != '\n');
4691 to
4692 while ((c = *(++ptr)) != 0 && c != '\n') ;
4693
4694 Totally extraordinary, but if that's what it takes...
4695
4696 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4697 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4698 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4699 assumes the way PCRE uses memmove() (always moving upwards).
4700
4701 6. PCRE is being used in one environment where strchr() is not available. There
4702 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4703 faster code anyway.
4704
4705
4706 Version 3.1 09-Feb-00
4707 ---------------------
4708
4709 The only change in this release is the fixing of some bugs in Makefile.in for
4710 the "install" target:
4711
4712 (1) It was failing to install pcreposix.h.
4713
4714 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4715
4716
4717 Version 3.0 01-Feb-00
4718 ---------------------
4719
4720 1. Add support for the /+ modifier to perltest (to output $` like it does in
4721 pcretest).
4722
4723 2. Add support for the /g modifier to perltest.
4724
4725 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4726 matches null strings.
4727
4728 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4729 pattern. Perl treats empty patterns specially - it reuses the most recent
4730 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4731 effect.
4732
4733 5. The POSIX interface was broken in that it was just handing over the POSIX
4734 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4735 required a bigger vector, with some working space on the end. This means that
4736 the POSIX wrapper now has to get and free some memory, and copy the results.
4737
4738 6. Added some simple autoconf support, placing the test data and the
4739 documentation in separate directories, re-organizing some of the
4740 information files, and making it build pcre-config (a GNU standard). Also added
4741 libtool support for building PCRE as a shared library, which is now the
4742 default.
4743
4744 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4745 09 are not valid octal constants. Single digits will be used for minor values
4746 less than 10.
4747
4748 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4749 existing programs that set these in the POSIX interface can use PCRE without
4750 modification.
4751
4752 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4753 return all that pcre_info() returns, plus additional data. The pcre_info()
4754 function is retained for compatibility, but is considered to be obsolete.
4755
4756 10. Added experimental recursion feature (?R) to handle one common case that
4757 Perl 5.6 will be able to do with (?p{...}).
4758
4759 11. Added support for POSIX character classes like [:alpha:], which Perl is
4760 adopting.
4761
4762
4763 Version 2.08 31-Aug-99
4764 ----------------------
4765
4766 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4767 trying to match at the startoffset position, but instead was moving forward to
4768 the next newline as if a previous match had failed.
4769
4770 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4771 and could get into a loop if a null string was matched other than at the start
4772 of the subject.
4773
4774 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4775 be distinguished at compile time, and for completeness also added PCRE_DATE.
4776
4777 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4778 in GnuWin32 environments.
4779
4780
4781 Version 2.07 29-Jul-99
4782 ----------------------
4783
4784 1. The documentation is now supplied in plain text form and HTML as well as in
4785 the form of man page sources.
4786
4787 2. C++ compilers don't like assigning (void *) values to other pointer types.
4788 In particular this affects malloc(). Although there is no problem in Standard
4789 C, I've put in casts to keep C++ compilers happy.
4790
4791 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4792 should be (const char *).
4793
4794 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4795 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4796 However, I haven't made this a standard facility. The documentation doesn't
4797 mention it, and the Makefile doesn't support it.
4798
4799 5. The Makefile now contains an "install" target, with editable destinations at
4800 the top of the file. The pcretest program is not installed.
4801
4802 6. pgrep -V now gives the PCRE version number and date.
4803
4804 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4805 causing the entire string to be ignored, instead of just the last character.
4806
4807 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4808 non-matching string, it can take a very, very long time, even for strings of
4809 quite modest length, because of the nested recursion. PCRE now does better in
4810 some of these cases. It does this by remembering the last required literal
4811 character in the pattern, and pre-searching the subject to ensure it is present
4812 before running the real match. In other words, it applies a heuristic to detect
4813 some types of certain failure quickly, and in the above example, if presented
4814 with a string that has no trailing " it gives "no match" very quickly.
4815
4816 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4817 other alternatives are tried instead.
4818
4819
4820 Version 2.06 09-Jun-99
4821 ----------------------
4822
4823 1. Change pcretest's output for amount of store used to show just the code
4824 space, because the remainder (the data block) varies in size between 32-bit and
4825 64-bit systems.
4826
4827 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4828 start matching at. This allows lookbehinds to work when searching for multiple
4829 occurrences in a string.
4830
4831 3. Added additional options to pcretest for testing multiple occurrences:
4832
4833 /+ outputs the rest of the string that follows a match
4834 /g loops for multiple occurrences, using the new startoffset argument
4835 /G loops for multiple occurrences by passing an incremented pointer
4836
4837 4. PCRE wasn't doing the "first character" optimization for patterns starting
4838 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4839 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4840 the letter 'x'. On long subject strings, this gives a significant speed-up.
4841
4842
4843 Version 2.05 21-Apr-99
4844 ----------------------
4845
4846 1. Changed the type of magic_number from int to long int so that it works
4847 properly on 16-bit systems.
4848
4849 2. Fixed a bug which caused patterns starting with .* not to work correctly
4850 when the subject string contained newline characters. PCRE was assuming
4851 anchoring for such patterns in all cases, which is not correct because .* will
4852 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4853 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4854 must be retried after every newline in the subject.
4855
4856
4857 Version 2.04 18-Feb-99
4858 ----------------------
4859
4860 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4861 computation of the store needed to hold the pattern was incorrect (too large).
4862 If such patterns were nested a few deep, this could multiply and become a real
4863 problem.
4864
4865 2. Added /M option to pcretest to show the memory requirement of a specific
4866 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4867
4868 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4869 compiled in such a way that the backtracking after subsequent failure was
4870 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4871 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4872
4873
4874 Version 2.03 02-Feb-99
4875 ----------------------
4876
4877 1. Fixed typo and small mistake in man page.
4878
4879 2. Added 4th condition (GPL supersedes if conflict) and created separate
4880 LICENCE file containing the conditions.
4881
4882 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4883 Perl, that is the internal \ allows the delimiter to be included in the
4884 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4885 the final delimiter, add \ to the end of the pattern (to test the error).
4886
4887 4. Added the convenience functions for extracting substrings after a successful
4888 match. Updated pcretest to make it able to test these functions.
4889
4890
4891 Version 2.02 14-Jan-99
4892 ----------------------
4893
4894 1. Initialized the working variables associated with each extraction so that
4895 their saving and restoring doesn't refer to uninitialized store.
4896
4897 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4898 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4899 fix the problem.
4900
4901 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4902 calls, and wasn't printing the correct value for compiling calls. Increased the
4903 default value of LOOPREPEAT, and the number of significant figures in the
4904 times.
4905
4906 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4907
4908 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4909 a building problem on Windows NT with a FAT file system.
4910
4911
4912 Version 2.01 21-Oct-98
4913 ----------------------
4914
4915 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4916 to character tables built by pcre_maketables() in the current locale. If NULL
4917 is passed, the default tables are used.
4918
4919
4920 Version 2.00 24-Sep-98
4921 ----------------------
4922
4923 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4924 it any more.
4925
4926 2. Allow quantification of (?>) groups, and make it work correctly.
4927
4928 3. The first character computation wasn't working for (?>) groups.
4929
4930 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4931 end of the subject) and add 5.005's \z, which really does match only at the
4932 very end of the subject.
4933
4934 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4935
4936 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4937 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4938 localized options. All options to pcre_study() were also removed.
4939
4940 7. Add other new features from 5.005:
4941
4942 $(?<= positive lookbehind
4943 $(?<! negative lookbehind
4944 (?imsx-imsx) added the unsetting capability
4945 such a setting is global if at outer level; local otherwise
4946 (?imsx-imsx:) non-capturing groups with option setting
4947 (?(cond)re|re) conditional pattern matching
4948
4949 A backreference to itself in a repeated group matches the previous
4950 captured string.
4951
4952 8. General tidying up of studying (both automatic and via "study")
4953 consequential on the addition of new assertions.
4954
4955 9. As in 5.005, unlimited repeated groups that could match an empty substring
4956 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4957 runtime if any iteration does actually match an empty substring.
4958
4959 10. Include the RunTest script in the distribution.
4960
4961 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4962 discrepancies, some of which were old and were also with respect to 5.004. They
4963 have now been fixed.
4964
4965
4966 Version 1.09 28-Apr-98
4967 ----------------------
4968
4969 1. A negated single character class followed by a quantifier with a minimum
4970 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4971 program crashes, or just wrong answers. This did not apply to negated classes
4972 containing more than one character, or to minima other than one.
4973
4974
4975 Version 1.08 27-Mar-98
4976 ----------------------
4977
4978 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4979
4980 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4981 latter must appear before anything that relies on it in the pattern.
4982
4983
4984 Version 1.07 16-Feb-98
4985 ----------------------
4986
4987 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4988 repeat of a potentially empty string).
4989
4990
4991 Version 1.06 23-Jan-98
4992 ----------------------
4993
4994 1. Added Markus Oberhumer's little patches for C++.
4995
4996 2. Literal strings longer than 255 characters were broken.
4997
4998
4999 Version 1.05 23-Dec-97
5000 ----------------------
5001
5002 1. Negated character classes containing more than one character were failing if
5003 PCRE_CASELESS was set at run time.
5004
5005
5006 Version 1.04 19-Dec-97
5007 ----------------------
5008
5009 1. Corrected the man page, where some "const" qualifiers had been omitted.
5010
5011 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5012 input syntax.
5013
5014 3. Fixed memory leak which occurred when a regex with back references was
5015 matched with an offsets vector that wasn't big enough. The temporary memory
5016 that is used in this case wasn't being freed if the match failed.
5017
5018 4. Tidied pcretest to ensure it frees memory that it gets.
5019
5020 5. Temporary memory was being obtained in the case where the passed offsets
5021 vector was exactly big enough.
5022
5023 6. Corrected definition of offsetof() from change 5 below.
5024
5025 7. I had screwed up change 6 below and broken the rules for the use of
5026 setjmp(). Now fixed.
5027
5028
5029 Version 1.03 18-Dec-97
5030 ----------------------
5031
5032 1. A erroneous regex with a missing opening parenthesis was correctly
5033 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5034 on some systems.
5035
5036 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5037 it was reported that one broken compiler failed on the former because "code" is
5038 also an independent variable.
5039
5040 3. The erroneous regex a[]b caused an array overrun reference.
5041
5042 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5043 fail on data ending with that character. (It was going on too far, and checking
5044 the next character, typically a binary zero.) This was specific to the
5045 optimized code for single-character negative classes.
5046
5047 5. Added a contributed patch from the TIN world which does the following:
5048
5049 + Add an undef for memmove, in case the the system defines a macro for it.
5050
5051 + Add a definition of offsetof(), in case there isn't one. (I don't know
5052 the reason behind this - offsetof() is part of the ANSI standard - but
5053 it does no harm).
5054
5055 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5056 most of the places where whitespace preceded '#'. I have given up and
5057 allowed the remaining 2 cases to be at the margin.
5058
5059 + Rename some variables in pcre to eliminate shadowing. This seems very
5060 pedantic, but does no harm, of course.
5061
5062 6. Moved the call to setjmp() into its own function, to get rid of warnings
5063 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5064
5065 7. Constructs such as \d{8,} were compiling into the equivalent of
5066 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5067 outcome, but in this particular case used more store than had been allocated,
5068 which caused the bug to be discovered because it threw up an internal error.
5069
5070 8. The debugging code in both pcre and pcretest for outputting the compiled
5071 form of a regex was going wrong in the case of back references followed by
5072 curly-bracketed repeats.
5073
5074
5075 Version 1.02 12-Dec-97
5076 ----------------------
5077
5078 1. Typos in pcre.3 and comments in the source fixed.
5079
5080 2. Applied a contributed patch to get rid of places where it used to remove
5081 'const' from variables, and fixed some signed/unsigned and uninitialized
5082 variable warnings.
5083
5084 3. Added the "runtest" target to Makefile.
5085
5086 4. Set default compiler flag to -O2 rather than just -O.
5087
5088
5089 Version 1.01 19-Nov-97
5090 ----------------------
5091
5092 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5093 like /([ab]*)*/, that is, for classes with more than one character in them.
5094
5095 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5096 as /((?>a*))*/ (a PCRE_EXTRA facility).
5097
5098
5099 Version 1.00 18-Nov-97
5100 ----------------------
5101
5102 1. Added compile-time macros to support systems such as SunOS4 which don't have
5103 memmove() or strerror() but have other things that can be used instead.
5104
5105 2. Arranged that "make clean" removes the executables.
5106
5107
5108 Version 0.99 27-Oct-97
5109 ----------------------
5110
5111 1. Fixed bug in code for optimizing classes with only one character. It was
5112 initializing a 32-byte map regardless, which could cause it to run off the end
5113 of the memory it had got.
5114
5115 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5116
5117
5118 Version 0.98 22-Oct-97
5119 ----------------------
5120
5121 1. Fixed bug in code for handling temporary memory usage when there are more
5122 back references than supplied space in the ovector. This could cause segfaults.
5123
5124
5125 Version 0.97 21-Oct-97
5126 ----------------------
5127
5128 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5129
5130 2. Optimized negated single characters not to use a bit map.
5131
5132 3. Brought error texts together as macro definitions; clarified some of them;
5133 fixed one that was wrong - it said "range out of order" when it meant "invalid
5134 escape sequence".
5135
5136 4. Changed some char * arguments to const char *.
5137
5138 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5139
5140 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5141 pcretest.
5142
5143
5144 Version 0.96 16-Oct-97
5145 ----------------------
5146
5147 1. Added a simple "pgrep" utility to the distribution.
5148
5149 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5150 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5151 where "ddd" means "one or more decimal digits".
5152
5153 3. Fixed serious bug. If a pattern had a back reference, but the call to
5154 pcre_exec() didn't supply a large enough ovector to record the related
5155 identifying subpattern, the match always failed. PCRE now remembers the number
5156 of the largest back reference, and gets some temporary memory in which to save
5157 the offsets during matching if necessary, in order to ensure that
5158 backreferences always work.
5159
5160 4. Increased the compatibility with Perl in a number of ways:
5161
5162 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5163 to request this handling. The option can be set at compile or exec time.
5164
5165 (b) $ matches before a terminating newline by default; an option
5166 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5167 mode). The option can be set at compile or exec time.
5168
5169 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5170 the same as Perl's. If the decimal number it represents is less than 10
5171 or there aren't that many previous left capturing parentheses, an octal
5172 escape is read. Inside a character class, it's always an octal escape,
5173 even if it is a single digit.
5174
5175 (d) An escaped but undefined alphabetic character is taken as a literal,
5176 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5177 escapes.
5178
5179 (e) {0} is now permitted. (The previous item is removed from the compiled
5180 pattern).
5181
5182 5. Changed all the names of code files so that the basic parts are no longer
5183 than 10 characters, and abolished the teeny "globals.c" file.
5184
5185 6. Changed the handling of character classes; they are now done with a 32-byte
5186 bit map always.
5187
5188 7. Added the -d and /D options to pcretest to make it possible to look at the
5189 internals of compilation without having to recompile pcre.
5190
5191
5192 Version 0.95 23-Sep-97
5193 ----------------------
5194
5195 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5196 \x20 at the start of a run of normal characters. These were being treated as
5197 real characters, instead of the source characters being re-checked.
5198
5199
5200 Version 0.94 18-Sep-97
5201 ----------------------
5202
5203 1. The functions are now thread-safe, with the caveat that the global variables
5204 containing pointers to malloc() and free() or alternative functions are the
5205 same for all threads.
5206
5207 2. Get pcre_study() to generate a bitmap of initial characters for non-
5208 anchored patterns when this is possible, and use it if passed to pcre_exec().
5209
5210
5211 Version 0.93 15-Sep-97
5212 ----------------------
5213
5214 1. /(b)|(:+)/ was computing an incorrect first character.
5215
5216 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5217 but not actually doing anything yet.
5218
5219 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5220 as Perl does (e.g. [-az] or [az-]).
5221
5222 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5223 all possible positions.
5224
5225 5. Split up into different modules to avoid including unneeded functions in a
5226 compiled binary. However, compile and exec are still in one module. The "study"
5227 function is split off.
5228
5229 6. The character tables are now in a separate module whose source is generated
5230 by an auxiliary program - but can then be edited by hand if required. There are
5231 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5232 toupper() in the code.
5233
5234 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5235 make them global. Abolish the function for setting them, as the caller can now
5236 set them directly.
5237
5238
5239 Version 0.92 11-Sep-97
5240 ----------------------
5241
5242 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5243 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5244
5245 2. Caseless matching was not working in character classes if the characters in
5246 the pattern were in upper case.
5247
5248 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5249
5250 4. Make PCRE_ANCHORED public and accept as a compile option.
5251
5252 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5253 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5254 pass them.
5255
5256 6. Give an error if bad option bits passed at compile or run time.
5257
5258 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5259 pcretest to cause it to pass that flag.
5260
5261 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5262 options, and the first character, if set.
5263
5264 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5265
5266
5267 Version 0.91 10-Sep-97
5268 ----------------------
5269
5270 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5271 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5272
5273 2. PCRE was looping on encountering an indefinitely repeated back reference to
5274 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5275 Perl does - treats the match as successful.
5276
5277 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5