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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1359 - (show annotations)
Tue Sep 3 10:10:59 2013 UTC (6 years, 3 months ago) by ph10
File size: 251402 byte(s)
Error occurred while calculating annotation data.
Refactor the code for creating the name/number table.
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
79 Version 8.33 28-May-2013
80 ------------------------
81
82 1. Added 'U' to some constants that are compared to unsigned integers, to
83 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
84 variables that are added to signed variables, to ensure the result is
85 signed and can be negated.
86
87 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
88 CMake config files.
89
90 3. Revise the creation of config.h.generic so that all boolean macros are
91 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
92 overriding via -D on the command line possible.
93
94 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
95 to unsigned int is reported to make a quite noticeable speed difference in
96 a specific Windows environment. Testing on Linux did also appear to show
97 some benefit (and it is clearly not harmful). Also fixed the definition of
98 Xop which should be unsigned.
99
100 5. Related to (4), changing the definition of the intermediate variable cc
101 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
102 improvements.
103
104 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
105 unnecessary spaces.
106
107 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
108 and later.
109
110 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
111
112 9. Optimizing fast_forward_start_bits in JIT.
113
114 10. Adding support for callouts in JIT, and fixing some issues revealed
115 during this work. Namely:
116
117 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
118
119 (b) Minimum length was not checked before the matching is started.
120
121 11. The value of capture_last that is passed to callouts was incorrect in some
122 cases when there was a capture on one path that was subsequently abandoned
123 after a backtrack. Also, the capture_last value is now reset after a
124 recursion, since all captures are also reset in this case.
125
126 12. The interpreter no longer returns the "too many substrings" error in the
127 case when an overflowing capture is in a branch that is subsequently
128 abandoned after a backtrack.
129
130 13. In the pathological case when an offset vector of size 2 is used, pcretest
131 now prints out the matched string after a yield of 0 or 1.
132
133 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
134 Only supported by the JIT compiler at the moment.
135
136 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
137
138 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
139 offset of the starting point of the matching process, provided the offsets
140 vector is large enough.
141
142 17. The \A escape now records a lookbehind value of 1, though its execution
143 does not actually inspect the previous character. This is to ensure that,
144 in partial multi-segment matching, at least one character from the old
145 segment is retained when a new segment is processed. Otherwise, if there
146 are no lookbehinds in the pattern, \A might match incorrectly at the start
147 of a new segment.
148
149 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
150
151 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
152 gives some modest performance improvement in 8-bit mode.
153
154 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
155 be expressed in certain programming languages using Universal Character
156 Names.
157
158 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
159 which points out that "non characters" are not "characters that may not
160 appear in Unicode strings" but rather "characters that are reserved for
161 internal use and have only local meaning".
162
163 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
164 there was a conditional group that depended on an assertion, if the
165 assertion was false, the callout that immediately followed the alternation
166 in the condition was skipped when pcre_exec() was used for matching.
167
168 23. Allow an explicit callout to be inserted before an assertion that is the
169 condition for a conditional group, for compatibility with automatic
170 callouts, which always insert a callout at this point.
171
172 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
173 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
174
175 25. (*PRUNE) is now supported by the JIT compiler.
176
177 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
178
179 27. Fix the case where there are two or more SKIPs with arguments that may be
180 ignored.
181
182 28. (*SKIP) is now supported by the JIT compiler.
183
184 29. (*THEN) is now supported by the JIT compiler.
185
186 30. Update RunTest with additional test selector options.
187
188 31. The way PCRE handles backtracking verbs has been changed in two ways.
189
190 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
191 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
192 backtracking. In some cases this makes it more Perl-compatible, but Perl's
193 rather obscure rules do not always do the same thing.
194
195 (2) Previously, backtracking verbs were confined within assertions. This is
196 no longer the case for positive assertions, except for (*ACCEPT). Again,
197 this sometimes improves Perl compatibility, and sometimes does not.
198
199 32. A number of tests that were in test 2 because Perl did things differently
200 have been moved to test 1, because either Perl or PCRE has changed, and
201 these tests are now compatible.
202
203 32. Backtracking control verbs are now handled in the same way in JIT and
204 interpreter.
205
206 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
207 contained a forward subroutine reference caused a compile error.
208
209 34. Auto-detect and optimize limited repetitions in JIT.
210
211 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
212 blocking (*UTF) etc.
213
214 36. In the interpreter, maximizing pattern repetitions for characters and
215 character types now use tail recursion, which reduces stack usage.
216
217 37. The value of the max lookbehind was not correctly preserved if a compiled
218 and saved regex was reloaded on a host of different endianness.
219
220 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
221 of the compiled pattern block, expand the flags field from 16 to 32 bits
222 because it was almost full.
223
224 39. Try madvise first before posix_madvise.
225
226 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
227 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
228 original change locked out all matches of empty strings. This has been
229 changed so that one match of an empty string per line is recognized.
230 Subsequent searches on the same line (for colouring or for --only-matching,
231 for example) do not recognize empty strings.
232
233 41. Applied a user patch to fix a number of spelling mistakes in comments.
234
235 42. Data lines longer than 65536 caused pcretest to crash.
236
237 43. Clarified the data type for length and startoffset arguments for pcre_exec
238 and pcre_dfa_exec in the function-specific man pages, where they were
239 explicitly stated to be in bytes, never having been updated. I also added
240 some clarification to the pcreapi man page.
241
242 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
243 a segmentation fault.
244
245
246 Version 8.32 30-November-2012
247 -----------------------------
248
249 1. Improved JIT compiler optimizations for first character search and single
250 character iterators.
251
252 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
253 Patch by Daniel Richard G.
254
255 3. Single character iterator optimizations in the JIT compiler.
256
257 4. Improved JIT compiler optimizations for character ranges.
258
259 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
260 Reported by Giuseppe D'Angelo.
261
262 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
263 of a line, was being set incorrectly in cases where .* appeared inside
264 atomic brackets at the start of a pattern, or where there was a subsequent
265 *PRUNE or *SKIP.
266
267 7. Improved instruction cache flush for POWER/PowerPC.
268 Patch by Daniel Richard G.
269
270 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
271 grep:
272
273 (a) There is now no limit to the number of patterns to be matched.
274
275 (b) An error is given if a pattern is too long.
276
277 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
278 are now supported.
279
280 (d) --exclude-from and --include-from (multiple use) have been added.
281
282 (e) Exclusions and inclusions now apply to all files and directories, not
283 just to those obtained from scanning a directory recursively.
284
285 (f) Multiple uses of -f and --file-list are now supported.
286
287 (g) In a Windows environment, the default for -d has been changed from
288 "read" (the GNU grep default) to "skip", because otherwise the presence
289 of a directory in the file list provokes an error.
290
291 (h) The documentation has been revised and clarified in places.
292
293 9. Improve the matching speed of capturing brackets.
294
295 10. Changed the meaning of \X so that it now matches a Unicode extended
296 grapheme cluster.
297
298 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
299 out POSIX threads when JIT support is configured.
300
301 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
302
303 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
304 structure could go wrong in environments where size_t is not the same size
305 as int.
306
307 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
308 set.
309
310 15. The EBCDIC support had decayed; later updates to the code had included
311 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
312 general tidy up of EBCDIC-related issues, and the documentation was also
313 not quite right. There is now a test that can be run on ASCII systems to
314 check some of the EBCDIC-related things (but is it not a full test).
315
316 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
317 in a small tidy to the code.
318
319 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
320
321 18. If the --only-matching (-o) option in pcregrep is specified multiple
322 times, each one causes appropriate output. For example, -o1 -o2 outputs the
323 substrings matched by the 1st and 2nd capturing parentheses. A separating
324 string can be specified by --om-separator (default empty).
325
326 19. Improving the first n character searches.
327
328 20. Turn case lists for horizontal and vertical white space into macros so that
329 they are defined only once.
330
331 21. This set of changes together give more compatible Unicode case-folding
332 behaviour for characters that have more than one other case when UCP
333 support is available.
334
335 (a) The Unicode property table now has offsets into a new table of sets of
336 three or more characters that are case-equivalent. The MultiStage2.py
337 script that generates these tables (the pcre_ucd.c file) now scans
338 CaseFolding.txt instead of UnicodeData.txt for character case
339 information.
340
341 (b) The code for adding characters or ranges of characters to a character
342 class has been abstracted into a generalized function that also handles
343 case-independence. In UTF-mode with UCP support, this uses the new data
344 to handle characters with more than one other case.
345
346 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
347 whose other case is greater than 256 are now correctly matched
348 caselessly. Previously, the high codepoint matched the low one, but not
349 vice versa.
350
351 (d) The processing of \h, \H, \v, and \ in character classes now makes use
352 of the new class addition function, using character lists defined as
353 macros alongside the case definitions of 20 above.
354
355 (e) Caseless back references now work with characters that have more than
356 one other case.
357
358 (f) General caseless matching of characters with more than one other case
359 is supported.
360
361 22. Unicode character properties were updated from Unicode 6.2.0
362
363 23. Improved CMake support under Windows. Patch by Daniel Richard G.
364
365 24. Add support for 32-bit character strings, and UTF-32
366
367 25. Major JIT compiler update (code refactoring and bugfixing).
368 Experimental Sparc 32 support is added.
369
370 26. Applied a modified version of Daniel Richard G's patch to create
371 pcre.h.generic and config.h.generic by "make" instead of in the
372 PrepareRelease script.
373
374 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
375 pcre_compile.c when checking for a zero character.
376
377 28. Introducing a native interface for JIT. Through this interface, the compiled
378 machine code can be directly executed. The purpose of this interface is to
379 provide fast pattern matching, so several sanity checks are not performed.
380 However, feature tests are still performed. The new interface provides
381 1.4x speedup compared to the old one.
382
383 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
384 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
385 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
386
387 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
388 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
389 the "old" RFC 2279). Instead, it ended up passing a negative length to
390 pcre_exec().
391
392 31. Add support for GCC's visibility feature to hide internal functions.
393
394 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
395 "unknown -C option" after outputting 0 or 1.
396
397 33. There is now support for generating a code coverage report for the test
398 suite in environments where gcc is the compiler and lcov is installed. This
399 is mainly for the benefit of the developers.
400
401 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
402 unaddressable using valgrind annotations, allowing valgrind to detect
403 invalid memory accesses. This is mainly for the benefit of the developers.
404
405 25. (*UTF) can now be used to start a pattern in any of the three libraries.
406
407 26. Give configure error if --enable-cpp but no C++ compiler found.
408
409
410 Version 8.31 06-July-2012
411 -------------------------
412
413 1. Fixing a wrong JIT test case and some compiler warnings.
414
415 2. Removed a bashism from the RunTest script.
416
417 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
418 to unsigned type, result still unsigned" that was given by an MS compiler
419 on encountering the code "-sizeof(xxx)".
420
421 4. Partial matching support is added to the JIT compiler.
422
423 5. Fixed several bugs concerned with partial matching of items that consist
424 of more than one character:
425
426 (a) /^(..)\1/ did not partially match "aba" because checking references was
427 done on an "all or nothing" basis. This also applied to repeated
428 references.
429
430 (b) \R did not give a hard partial match if \r was found at the end of the
431 subject.
432
433 (c) \X did not give a hard partial match after matching one or more
434 characters at the end of the subject.
435
436 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
437 a partial match for the string "\r".
438
439 (e) When newline was set to CRLF, the metacharacter "." did not recognize
440 a partial match for a CR character at the end of the subject string.
441
442 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
443 running pcretest, the text "(JIT)" added to the output whenever JIT is
444 actually used to run the match.
445
446 7. Individual JIT compile options can be set in pcretest by following -s+[+]
447 or /S+[+] with a digit between 1 and 7.
448
449 8. OP_NOT now supports any UTF character not just single-byte ones.
450
451 9. (*MARK) control verb is now supported by the JIT compiler.
452
453 10. The command "./RunTest list" lists the available tests without actually
454 running any of them. (Because I keep forgetting what they all are.)
455
456 11. Add PCRE_INFO_MAXLOOKBEHIND.
457
458 12. Applied a (slightly modified) user-supplied patch that improves performance
459 when the heap is used for recursion (compiled with --disable-stack-for-
460 recursion). Instead of malloc and free for each heap frame each time a
461 logical recursion happens, frames are retained on a chain and re-used where
462 possible. This sometimes gives as much as 30% improvement.
463
464 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
465 call.
466
467 14. As documented, (*COMMIT) is now confined to within a positive assertion.
468
469 15. It is now possible to link pcretest with libedit as an alternative to
470 libreadline.
471
472 16. (*COMMIT) control verb is now supported by the JIT compiler.
473
474 17. The Unicode data tables have been updated to Unicode 6.1.0.
475
476 18. Added --file-list option to pcregrep.
477
478 19. Added binary file support to pcregrep, including the -a, --binary-files,
479 -I, and --text options.
480
481 20. The madvise function is renamed for posix_madvise for QNX compatibility
482 reasons. Fixed by Giuseppe D'Angelo.
483
484 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
485
486 22. Changed the option for creating symbolic links for 16-bit man pages from
487 -s to -sf so that re-installing does not cause issues.
488
489 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
490
491 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
492 matches in certain environments (the workspace was not being correctly
493 retained). Also added to pcre_dfa_exec() a simple plausibility check on
494 some of the workspace data at the beginning of a restart.
495
496 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
497 was not doing so when it should - probably a typo introduced by SVN 528
498 (change 8.10/14).
499
500 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
501 \w+ when the character tables indicated that \x{c4} was a word character.
502 There were several related cases, all because the tests for doing a table
503 lookup were testing for characters less than 127 instead of 255.
504
505 27. If a pattern contains capturing parentheses that are not used in a match,
506 their slots in the ovector are set to -1. For those that are higher than
507 any matched groups, this happens at the end of processing. In the case when
508 there were back references that the ovector was too small to contain
509 (causing temporary malloc'd memory to be used during matching), and the
510 highest capturing number was not used, memory off the end of the ovector
511 was incorrectly being set to -1. (It was using the size of the temporary
512 memory instead of the true size.)
513
514 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
515 ovector size, it uses memory at the end of the block that it has got.
516
517 29. Check for an overlong MARK name and give an error at compile time. The
518 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
519
520 30. JIT compiler update.
521
522 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
523 Rill for the patch.
524
525 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
526
527 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
528
529 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
530 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
531 was enabled.
532
533 35. Improve JIT code generation for greedy plus quantifier.
534
535 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
536 1 to "aa" instead of to an empty string. The bug affected repeated groups
537 that could potentially match an empty string.
538
539 37. Optimizing single character iterators in JIT.
540
541 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
542 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
543 codepoints that are too big for the mode are faulted, and in a UTF mode,
544 disallowed codepoints are also faulted.
545
546 39. If PCRE was compiled with UTF support, in three places in the DFA
547 matcher there was code that should only have been obeyed in UTF mode, but
548 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
549 processing when bytes with values greater than 127 were present. In 16-bit
550 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
551 both cases the values are those that cannot be the first data item in a UTF
552 character. The three items that might have provoked this were recursions,
553 possessively repeated groups, and atomic groups.
554
555 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
556 and pcregrep, because some OS require shared objects to be explicitly
557 passed to ld, causing the link step to fail if they are not.
558
559 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
560 mode, patterns that started with \h* or \R* might be incorrectly matched.
561
562
563 Version 8.30 04-February-2012
564 -----------------------------
565
566 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
567 name is defined in ctype.h.
568
569 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
570 only in quite long subpatterns.
571
572 3. Removed the function pcre_info(), which has been obsolete and deprecated
573 since it was replaced by pcre_fullinfo() in February 2000.
574
575 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
576 match a (*MARK), and the match failed at the start of the subject, a
577 reference to memory before the start of the subject could occur. This bug
578 was introduced by fix 17 of release 8.21.
579
580 5. A reference to an unset group with zero minimum repetition was giving
581 totally wrong answers (in non-JavaScript-compatibility mode). For example,
582 /(another)?(\1?)test/ matched against "hello world test". This bug was
583 introduced in release 8.13.
584
585 6. Add support for 16-bit character strings (a large amount of work involving
586 many changes and refactorings).
587
588 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
589 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
590 from a file.
591
592 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
593 rounding is not applied in this particular case).
594
595 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
596 if they appear, or are escaped, in patterns.
597
598 10. Get rid of a number of -Wunused-but-set-variable warnings.
599
600 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
601 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
602 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
603 also returns the mark "x". This bug applied to capturing parentheses,
604 non-capturing parentheses, and atomic parentheses. It also applied to some
605 assertions.
606
607 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
608 information out of configure.ac instead of relying on pcre.h.generic, which
609 is not stored in the repository.
610
611 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
612 -lreadline.
613
614 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
615
616 15. Applied Graycode's patch to put the top-level frame on the stack rather
617 than the heap when not using the stack for recursion. This gives a
618 performance improvement in many cases when recursion is not deep.
619
620 16. Experimental code added to "pcretest -C" to output the stack frame size.
621
622
623 Version 8.21 12-Dec-2011
624 ------------------------
625
626 1. Updating the JIT compiler.
627
628 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
629 are added as well.
630
631 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
632 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
633 calling _pcre_jit_exec. Some extra comments are added.
634
635 4. (*MARK) settings inside atomic groups that do not contain any capturing
636 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
637 was introduced by change 18 for 8.20.
638
639 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
640 ECMA-262 standard.
641
642 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
643 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
644 This bug was probably introduced by change 9 of 8.13.
645
646 7. While fixing 6 above, I noticed that a number of other items were being
647 incorrectly rejected as "not fixed length". This arose partly because newer
648 opcodes had not been added to the fixed-length checking code. I have (a)
649 corrected the bug and added tests for these items, and (b) arranged for an
650 error to occur if an unknown opcode is encountered while checking for fixed
651 length instead of just assuming "not fixed length". The items that were
652 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
653 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
654 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
655
656 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
657 being incorrectly compiled and would have given unpredicatble results.
658
659 9. A possessively repeated subpattern with minimum repeat count greater than
660 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
661 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
662 the first (A) could occur when it should not.
663
664 10. Add a cast and remove a redundant test from the code.
665
666 11. JIT should use pcre_malloc/pcre_free for allocation.
667
668 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
669 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
670 is anything other than /usr, -L is still shown).
671
672 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
673
674 14. Perl does not support \N without a following name in a [] class; PCRE now
675 also gives an error.
676
677 15. If a forward reference was repeated with an upper limit of around 2000,
678 it caused the error "internal error: overran compiling workspace". The
679 maximum number of forward references (including repeats) was limited by the
680 internal workspace, and dependent on the LINK_SIZE. The code has been
681 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
682 the default depends on LINK_SIZE. There is a new upper limit (for safety)
683 of around 200,000 forward references. While doing this, I also speeded up
684 the filling in of repeated forward references.
685
686 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
687 incorrectly expecting the subject to contain another "a" after the start.
688
689 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
690 in the match, the SKIP should be ignored. This was not happening; instead
691 the SKIP was being treated as NOMATCH. For patterns such as
692 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
693 tested.
694
695 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
696 now much more compatible with Perl, in particular in cases where the result
697 is a non-match for a non-anchored pattern. For example, if
698 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
699 "m", where previously it did not return a name. A side effect of this
700 change is that for partial matches, the last encountered mark name is
701 returned, as for non matches. A number of tests that were previously not
702 Perl-compatible have been moved into the Perl-compatible test files. The
703 refactoring has had the pleasing side effect of removing one argument from
704 the match() function, thus reducing its stack requirements.
705
706 19. If the /S+ option was used in pcretest to study a pattern using JIT,
707 subsequent uses of /S (without +) incorrectly behaved like /S+.
708
709 21. Retrieve executable code size support for the JIT compiler and fixing
710 some warnings.
711
712 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
713 not work when the shorter character appeared right at the end of the
714 subject string.
715
716 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
717 systems.
718
719 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
720 output it when the /M option is used in pcretest.
721
722 25. The CheckMan script was not being included in the distribution. Also, added
723 an explicit "perl" to run Perl scripts from the PrepareRelease script
724 because this is reportedly needed in Windows.
725
726 26. If study data was being save in a file and studying had not found a set of
727 "starts with" bytes for the pattern, the data written to the file (though
728 never used) was taken from uninitialized memory and so caused valgrind to
729 complain.
730
731 27. Updated RunTest.bat as provided by Sheri Pierce.
732
733 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
734
735 29. Computation of memory usage for the table of capturing group names was
736 giving an unnecessarily large value.
737
738
739 Version 8.20 21-Oct-2011
740 ------------------------
741
742 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
743 a POSIX class. After further experiments with Perl, which convinced me that
744 Perl has bugs and confusions, a closing square bracket is no longer allowed
745 in a POSIX name. This bug also affected patterns with classes that started
746 with full stops.
747
748 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
749 captured substring, but while checking the failing first alternative,
750 substring 1 is temporarily captured. If the output vector supplied to
751 pcre_exec() was not big enough for this capture, the yield of the function
752 was still zero ("insufficient space for captured substrings"). This cannot
753 be totally fixed without adding another stack variable, which seems a lot
754 of expense for a edge case. However, I have improved the situation in cases
755 such as /(a)(b)x|abc/ matched against "abc", where the return code
756 indicates that fewer than the maximum number of slots in the ovector have
757 been set.
758
759 3. Related to (2) above: when there are more back references in a pattern than
760 slots in the output vector, pcre_exec() uses temporary memory during
761 matching, and copies in the captures as far as possible afterwards. It was
762 using the entire output vector, but this conflicts with the specification
763 that only 2/3 is used for passing back captured substrings. Now it uses
764 only the first 2/3, for compatibility. This is, of course, another edge
765 case.
766
767 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
768 main code base, and can be used by building with --enable-jit. When this is
769 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
770 runtime --no-jit option is given.
771
772 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
773 ovector, the return from the function was zero, implying that there were
774 other matches that did not fit. The correct "exactly full" value is now
775 returned.
776
777 6. If a subpattern that was called recursively or as a subroutine contained
778 (*PRUNE) or any other control that caused it to give a non-standard return,
779 invalid errors such as "Error -26 (nested recursion at the same subject
780 position)" or even infinite loops could occur.
781
782 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
783 computing the minimum length on reaching *ACCEPT, and so ended up with the
784 wrong value of 1 rather than 0. Further investigation indicates that
785 computing a minimum subject length in the presence of *ACCEPT is difficult
786 (think back references, subroutine calls), and so I have changed the code
787 so that no minimum is registered for a pattern that contains *ACCEPT.
788
789 8. If (*THEN) was present in the first (true) branch of a conditional group,
790 it was not handled as intended. [But see 16 below.]
791
792 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
793 Sheri Pierce.
794
795 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
796 the first byte in a match must be "a".
797
798 11. Change 17 for 8.13 increased the recursion depth for patterns like
799 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
800 pattern contains any instances of (*THEN). If it does not, the old
801 optimizations are restored. It would be nice to do this on a per-group
802 basis, but at the moment that is not feasible.
803
804 12. In some environments, the output of pcretest -C is CRLF terminated. This
805 broke RunTest's code that checks for the link size. A single white space
806 character after the value is now allowed for.
807
808 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
809 For "fr", it uses the Windows-specific input and output files.
810
811 14. If (*THEN) appeared in a group that was called recursively or as a
812 subroutine, it did not work as intended. [But see next item.]
813
814 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
815 pattern fragments (but not containing any | characters). If A and B are
816 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
817 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
818 D. In other words, Perl considers parentheses that do not contain any |
819 characters to be part of a surrounding alternative, whereas PCRE was
820 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
821 differently. PCRE now behaves in the same way as Perl, except in the case
822 of subroutine/recursion calls such as (?1) which have in any case always
823 been different (but PCRE had them first :-).
824
825 16. Related to 15 above: Perl does not treat the | in a conditional group as
826 creating alternatives. Such a group is treated in the same way as an
827 ordinary group without any | characters when processing (*THEN). PCRE has
828 been changed to match Perl's behaviour.
829
830 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
831 RunGrepTest script failed.
832
833 18. Change 22 for version 13 caused atomic groups to use more stack. This is
834 inevitable for groups that contain captures, but it can lead to a lot of
835 stack use in large patterns. The old behaviour has been restored for atomic
836 groups that do not contain any capturing parentheses.
837
838 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
839 suppress the check for a minimum subject length at run time. (If it was
840 given to pcre_exec() or pcre_dfa_exec() it did work.)
841
842 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
843 fail to work when decoding hex characters in data strings in EBCDIC
844 environments.
845
846 21. It appears that in at least one Mac OS environment, the isxdigit() function
847 is implemented as a macro that evaluates to its argument more than once,
848 contravening the C 90 Standard (I haven't checked a later standard). There
849 was an instance in pcretest which caused it to go wrong when processing
850 \x{...} escapes in subject strings. The has been rewritten to avoid using
851 things like p++ in the argument of isxdigit().
852
853
854 Version 8.13 16-Aug-2011
855 ------------------------
856
857 1. The Unicode data tables have been updated to Unicode 6.0.0.
858
859 2. Two minor typos in pcre_internal.h have been fixed.
860
861 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
862 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
863 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
864
865 4. There were a number of related bugs in the code for matching backrefences
866 caselessly in UTF-8 mode when codes for the characters concerned were
867 different numbers of bytes. For example, U+023A and U+2C65 are an upper
868 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
869 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
870 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
871 2-byte code at the end of the subject (it thought there wasn't enough data
872 left).
873
874 5. Comprehensive information about what went wrong is now returned by
875 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
876 as the output vector has at least 2 elements. The offset of the start of
877 the failing character and a reason code are placed in the vector.
878
879 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
880 now returned is for the first byte of the failing character, instead of the
881 last byte inspected. This is an incompatible change, but I hope it is small
882 enough not to be a problem. It makes the returned offset consistent with
883 pcre_exec() and pcre_dfa_exec().
884
885 7. pcretest now gives a text phrase as well as the error number when
886 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
887 failure, the offset and reason code are output.
888
889 8. When \R was used with a maximizing quantifier it failed to skip backwards
890 over a \r\n pair if the subsequent match failed. Instead, it just skipped
891 back over a single character (\n). This seems wrong (because it treated the
892 two characters as a single entity when going forwards), conflicts with the
893 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
894 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
895 has been changed.
896
897 9. Some internal refactoring has changed the processing so that the handling
898 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
899 time (the PCRE_DOTALL option was changed this way some time ago: version
900 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
901 which was always a bit of a fudge. It also means that there is one less
902 argument for the match() function, which reduces its stack requirements
903 slightly. This change also fixes an incompatibility with Perl: the pattern
904 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
905
906 10. More internal refactoring has drastically reduced the number of recursive
907 calls to match() for possessively repeated groups such as (abc)++ when
908 using pcre_exec().
909
910 11. While implementing 10, a number of bugs in the handling of groups were
911 discovered and fixed:
912
913 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
914 (a|)*(?1) gave a compile-time internal error.
915 ((a|)+)+ did not notice that the outer group could match an empty string.
916 (^a|^)+ was not marked as anchored.
917 (.*a|.*)+ was not marked as matching at start or after a newline.
918
919 12. Yet more internal refactoring has removed another argument from the match()
920 function. Special calls to this function are now indicated by setting a
921 value in a variable in the "match data" data block.
922
923 13. Be more explicit in pcre_study() instead of relying on "default" for
924 opcodes that mean there is no starting character; this means that when new
925 ones are added and accidentally left out of pcre_study(), testing should
926 pick them up.
927
928 14. The -s option of pcretest has been documented for ages as being an old
929 synonym of -m (show memory usage). I have changed it to mean "force study
930 for every regex", that is, assume /S for every regex. This is similar to -i
931 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
932 using it. It makes it easier to run collections of tests with and without
933 study enabled, and thereby test pcre_study() more easily. All the standard
934 tests are now run with and without -s (but some patterns can be marked as
935 "never study" - see 20 below).
936
937 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
938 restoration of the capturing data to the outer values was not happening
939 correctly.
940
941 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
942 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
943 pattern had matched an empty string, and so incorrectly returned a no
944 match.
945
946 17. There was optimizing code for the last branch of non-capturing parentheses,
947 and also for the obeyed branch of a conditional subexpression, which used
948 tail recursion to cut down on stack usage. Unfortunately, now that there is
949 the possibility of (*THEN) occurring in these branches, tail recursion is
950 no longer possible because the return has to be checked for (*THEN). These
951 two optimizations have therefore been removed. [But see 8.20/11 above.]
952
953 18. If a pattern containing \R was studied, it was assumed that \R always
954 matched two bytes, thus causing the minimum subject length to be
955 incorrectly computed because \R can also match just one byte.
956
957 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
958 was incorrectly computed.
959
960 20. If /S is present twice on a test pattern in pcretest input, it now
961 *disables* studying, thereby overriding the use of -s on the command line
962 (see 14 above). This is necessary for one or two tests to keep the output
963 identical in both cases.
964
965 21. When (*ACCEPT) was used in an assertion that matched an empty string and
966 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
967
968 22. When an atomic group that contained a capturing parenthesis was
969 successfully matched, but the branch in which it appeared failed, the
970 capturing was not being forgotten if a higher numbered group was later
971 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
972 group 1 to "a", when in fact it should be unset. This applied to multi-
973 branched capturing and non-capturing groups, repeated or not, and also to
974 positive assertions (capturing in negative assertions does not happen
975 in PCRE) and also to nested atomic groups.
976
977 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
978 subject after a captured substring, to make it easier to tell which of a
979 number of identical substrings has been captured.
980
981 24. The way atomic groups are processed by pcre_exec() has been changed so that
982 if they are repeated, backtracking one repetition now resets captured
983 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
984 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
985 "aaa". Previously, it would have been "a". As part of this code
986 refactoring, the way recursive calls are handled has also been changed.
987
988 25. If an assertion condition captured any substrings, they were not passed
989 back unless some other capturing happened later. For example, if
990 (?(?=(a))a) was matched against "a", no capturing was returned.
991
992 26. When studying a pattern that contained subroutine calls or assertions,
993 the code for finding the minimum length of a possible match was handling
994 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
995 group 1 called group 2 while simultaneously a separate group 2 called group
996 1). A stack overflow occurred in this case. I have fixed this by limiting
997 the recursion depth to 10.
998
999 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1000 Fortmann. This supports explicit test numbers on the command line, and has
1001 argument validation and error reporting.
1002
1003 28. An instance of \X with an unlimited repeat could fail if at any point the
1004 first character it looked at was a mark character.
1005
1006 29. Some minor code refactoring concerning Unicode properties and scripts
1007 should reduce the stack requirement of match() slightly.
1008
1009 30. Added the '=' option to pcretest to check the setting of unused capturing
1010 slots at the end of the pattern, which are documented as being -1, but are
1011 not included in the return count.
1012
1013 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1014 compiled something random. Now it gives a compile-time error (as does
1015 Perl).
1016
1017 32. A *MARK encountered during the processing of a positive assertion is now
1018 recorded and passed back (compatible with Perl).
1019
1020 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1021 had alternative anchored branches, the search for a second match in a line
1022 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1023 matched the line "0102" twice. The same bug affected patterns that started
1024 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1025 twice.
1026
1027 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1028 does, and because of capturing effects, quantifying parenthesized
1029 assertions may at times be useful. Quantifiers are now allowed for
1030 parenthesized assertions.
1031
1032 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1033
1034 36. \g was being checked for fancy things in a character class, when it should
1035 just be a literal "g".
1036
1037 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1038 appearance of a nested POSIX class supersedes an apparent external class.
1039 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1040 unescaped square brackets may also appear as part of class names. For
1041 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1042 more like Perl. (But see 8.20/1 above.)
1043
1044 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1045 was because it thought it was \N{name}, which is not supported).
1046
1047 39. Add minix to OS list not supporting the -S option in pcretest.
1048
1049 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1050 cannot analyze patterns in sufficient detail to catch mutual recursions
1051 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1052 subgroup is called recursively as a subpattern for a second time at the
1053 same position in the subject string. In previous releases this might have
1054 been caught by the recursion limit, or it might have run out of stack.
1055
1056 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1057 happen only once. PCRE was, however incorrectly giving a compile time error
1058 "recursive call could loop indefinitely" because it cannot analyze the
1059 pattern in sufficient detail. The compile time test no longer happens when
1060 PCRE is compiling a conditional subpattern, but actual runaway loops are
1061 now caught at runtime (see 40 above).
1062
1063 42. It seems that Perl allows any characters other than a closing parenthesis
1064 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1065 has been changed to be the same.
1066
1067 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1068 as not to get warnings when autogen.sh is called. Also changed
1069 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1070
1071 44. To help people who use pcregrep to scan files containing exceedingly long
1072 lines, the following changes have been made:
1073
1074 (a) The default value of the buffer size parameter has been increased from
1075 8K to 20K. (The actual buffer used is three times this size.)
1076
1077 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1078 PCRE is built.
1079
1080 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1081 to be set at run time.
1082
1083 (d) Numerical values in pcregrep options can be followed by K or M, for
1084 example --buffer-size=50K.
1085
1086 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1087 given and the return code is set to 2.
1088
1089 45. Add a pointer to the latest mark to the callout data block.
1090
1091 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1092 partial match of an empty string instead of no match. This was specific to
1093 the use of ".".
1094
1095 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1096 complete match instead of a partial match. This bug was dependent on both
1097 the PCRE_UTF8 and PCRE_DOTALL options being set.
1098
1099 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1100 starting byte set, because \b was not being ignored.
1101
1102
1103 Version 8.12 15-Jan-2011
1104 ------------------------
1105
1106 1. Fixed some typos in the markup of the man pages, and wrote a script that
1107 checks for such things as part of the documentation building process.
1108
1109 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1110 --match-limit and --recursion-limit options (added for 8.11). In
1111 particular, this made one of the standard tests fail. (The integer value
1112 went into the wrong half of a long int.)
1113
1114 3. If the --colour option was given to pcregrep with -v (invert match), it
1115 did strange things, either producing crazy output, or crashing. It should,
1116 of course, ignore a request for colour when reporting lines that do not
1117 match.
1118
1119 4. Another pcregrep bug caused similar problems if --colour was specified with
1120 -M (multiline) and the pattern match finished with a line ending.
1121
1122 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1123 matched in multiline mode, the following line was shown as part of the
1124 match. This seems wrong, so I have changed it.
1125
1126 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1127 the check for further matches in the same line (so they could be coloured)
1128 to overrun the end of the current line. If another match was found, it was
1129 incorrectly shown (and then shown again when found in the next line).
1130
1131 7. If pcregrep was compiled under Windows, there was a reference to the
1132 function pcregrep_exit() before it was defined. I am assuming this was
1133 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1134 reported by a user. I've moved the definition above the reference.
1135
1136
1137 Version 8.11 10-Dec-2010
1138 ------------------------
1139
1140 1. (*THEN) was not working properly if there were untried alternatives prior
1141 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1142 backtracked to try for "b" instead of moving to the next alternative branch
1143 at the same level (in this case, to look for "c"). The Perl documentation
1144 is clear that when (*THEN) is backtracked onto, it goes to the "next
1145 alternative in the innermost enclosing group".
1146
1147 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1148 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1149 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1150 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1151 (*THEN).
1152
1153 3. If \s appeared in a character class, it removed the VT character from
1154 the class, even if it had been included by some previous item, for example
1155 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1156 of \s, but is part of the POSIX "space" class.)
1157
1158 4. A partial match never returns an empty string (because you can always
1159 match an empty string at the end of the subject); however the checking for
1160 an empty string was starting at the "start of match" point. This has been
1161 changed to the "earliest inspected character" point, because the returned
1162 data for a partial match starts at this character. This means that, for
1163 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1164 (previously it gave "no match").
1165
1166 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1167 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1168 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1169 has an implication that the given string is incomplete (because a partial
1170 match is preferred over a full match). For this reason, these items now
1171 give a partial match in this situation. [Aside: previously, the one case
1172 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1173 match rather than a full match, which was wrong by the old rules, but is
1174 now correct.]
1175
1176 6. There was a bug in the handling of #-introduced comments, recognized when
1177 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1178 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1179 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1180 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1181 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1182 places in pcre_compile().
1183
1184 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1185 comments when looking ahead for named forward references to subpatterns,
1186 the only newline sequence it recognized was NL. It now handles newlines
1187 according to the set newline convention.
1188
1189 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1190 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1191 cater for a lack of strerror(). These oversights have been fixed.
1192
1193 9. Added --match-limit and --recursion-limit to pcregrep.
1194
1195 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1196
1197 11. When the -o option was used, pcregrep was setting a return code of 1, even
1198 when matches were found, and --line-buffered was not being honoured.
1199
1200 12. Added an optional parentheses number to the -o and --only-matching options
1201 of pcregrep.
1202
1203 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1204 can match an empty string. The code to do it in pcretest and pcredemo
1205 needed fixing:
1206
1207 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1208 only one byte after an empty string match just before CRLF (this case
1209 just got forgotten; "any" and "anycrlf" were OK).
1210
1211 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1212 mode when an empty string match preceded an ASCII character followed by
1213 a non-ASCII character. (The code for advancing by one character rather
1214 than one byte was nonsense.)
1215
1216 (c) The pcredemo.c sample program did not have any code at all to handle
1217 the cases when CRLF is a valid newline sequence.
1218
1219 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1220 as a starting offset was within the subject string. There is now a new
1221 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1222 negative or greater than the length of the string. In order to test this,
1223 pcretest is extended to allow the setting of negative starting offsets.
1224
1225 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1226 starting offset points to the beginning of a UTF-8 character was
1227 unnecessarily clumsy. I tidied it up.
1228
1229 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1230 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1231
1232 17. Nobody had reported that the --include_dir option, which was added in
1233 release 7.7 should have been called --include-dir (hyphen, not underscore)
1234 for compatibility with GNU grep. I have changed it to --include-dir, but
1235 left --include_dir as an undocumented synonym, and the same for
1236 --exclude-dir, though that is not available in GNU grep, at least as of
1237 release 2.5.4.
1238
1239 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1240 characters from a string of bytes) have been redefined so as not to use
1241 loops, in order to improve performance in some environments. At the same
1242 time, I abstracted some of the common code into auxiliary macros to save
1243 repetition (this should not affect the compiled code).
1244
1245 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1246 compile-time error is now given if \c is not followed by an ASCII
1247 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1248 different, and any byte value is allowed.)
1249
1250 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1251 START_OPTIMIZE option, which is now allowed at compile time - but just
1252 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1253 to pcregrep and other applications that have no direct access to PCRE
1254 options. The new /Y option in pcretest sets this option when calling
1255 pcre_compile().
1256
1257 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1258 back references. Groups containing recursive back references were forced to
1259 be atomic by that change, but in the case of named groups, the amount of
1260 memory required was incorrectly computed, leading to "Failed: internal
1261 error: code overflow". This has been fixed.
1262
1263 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1264 pcretest.c, to avoid build problems in some Borland environments.
1265
1266
1267 Version 8.10 25-Jun-2010
1268 ------------------------
1269
1270 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1271 THEN.
1272
1273 2. (*ACCEPT) was not working when inside an atomic group.
1274
1275 3. Inside a character class, \B is treated as a literal by default, but
1276 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1277 causes the error). The code is unchanged, but I tidied the documentation.
1278
1279 4. Inside a character class, PCRE always treated \R and \X as literals,
1280 whereas Perl faults them if its -w option is set. I have changed PCRE so
1281 that it faults them when PCRE_EXTRA is set.
1282
1283 5. Added support for \N, which always matches any character other than
1284 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1285
1286 6. When compiling pcregrep with newer versions of gcc which may have
1287 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1288 declared with attribute warn_unused_result" were given. Just casting the
1289 result to (void) does not stop the warnings; a more elaborate fudge is
1290 needed. I've used a macro to implement this.
1291
1292 7. Minor change to pcretest.c to avoid a compiler warning.
1293
1294 8. Added four artifical Unicode properties to help with an option to make
1295 \s etc use properties (see next item). The new properties are: Xan
1296 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1297
1298 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1299 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1300 this option. Modified pcretest to add /W to test this facility. Added
1301 REG_UCP to make it available via the POSIX interface.
1302
1303 10. Added --line-buffered to pcregrep.
1304
1305 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1306 studied, and the match started with a letter with a code point greater than
1307 127 whose first byte was different to the first byte of the other case of
1308 the letter, the other case of this starting letter was not recognized
1309 (#976).
1310
1311 12. If a pattern that was studied started with a repeated Unicode property
1312 test, for example, \p{Nd}+, there was the theoretical possibility of
1313 setting up an incorrect bitmap of starting bytes, but fortunately it could
1314 not have actually happened in practice until change 8 above was made (it
1315 added property types that matched character-matching opcodes).
1316
1317 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1318 possible starting bytes for non-anchored patterns.
1319
1320 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1321 \R, and also a number of cases that involve Unicode properties, both
1322 explicit and implicit when PCRE_UCP is set.
1323
1324 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1325 input, it could crash or give wrong results if characters with values
1326 greater than 0xc0 were present in the subject string. (Detail: it assumed
1327 UTF-8 input when processing these items.)
1328
1329 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1330 size_t is 64-bit (#991).
1331
1332 17. Added a check for running out of memory when PCRE is compiled with
1333 --disable-stack-for-recursion (#990).
1334
1335 18. If the last data line in a file for pcretest does not have a newline on
1336 the end, a newline was missing in the output.
1337
1338 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1339 less than 128) in its various bitmaps. However, there is a facility for
1340 generating tables according to the current locale when PCRE is compiled. It
1341 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1342 characters, are recognized by isspace() and therefore were getting set in
1343 these tables, and indeed these tables seem to approximate to ISO 8859. This
1344 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1345 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1346 which of course cannot start UTF-8 characters. I have changed the code so
1347 that only real ASCII characters (less than 128) and the correct starting
1348 bytes for UTF-8 encodings are set for characters greater than 127 when in
1349 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1350 altogether.)
1351
1352 20. Added the /T option to pcretest so as to be able to run tests with non-
1353 standard character tables, thus making it possible to include the tests
1354 used for 19 above in the standard set of tests.
1355
1356 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1357 reference to a subpattern the other side of a comment that contains an
1358 opening parenthesis caused either an internal compiling error, or a
1359 reference to the wrong subpattern.
1360
1361
1362 Version 8.02 19-Mar-2010
1363 ------------------------
1364
1365 1. The Unicode data tables have been updated to Unicode 5.2.0.
1366
1367 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1368 configured.
1369
1370 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1371 original author of that file, following a query about its status.
1372
1373 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1374 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1375
1376 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1377 quantifier applied to a forward-referencing subroutine call, could compile
1378 incorrect code or give the error "internal error: previously-checked
1379 referenced subpattern not found".
1380
1381 6. Both MS Visual Studio and Symbian OS have problems with initializing
1382 variables to point to external functions. For these systems, therefore,
1383 pcre_malloc etc. are now initialized to local functions that call the
1384 relevant global functions.
1385
1386 7. There were two entries missing in the vectors called coptable and poptable
1387 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1388 I've fixed the data, and added a kludgy way of testing at compile time that
1389 the lengths are correct (equal to the number of opcodes).
1390
1391 8. Following on from 7, I added a similar kludge to check the length of the
1392 eint vector in pcreposix.c.
1393
1394 9. Error texts for pcre_compile() are held as one long string to avoid too
1395 much relocation at load time. To find a text, the string is searched,
1396 counting zeros. There was no check for running off the end of the string,
1397 which could happen if a new error number was added without updating the
1398 string.
1399
1400 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1401
1402 11. \K was not working if it appeared in an atomic group or in a group that
1403 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1404 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1405 the assertion is positive, but not if it is negative.
1406
1407 12. Change 11 fortuitously reduced the size of the stack frame used in the
1408 "match()" function of pcre_exec.c by one pointer. Forthcoming
1409 implementation of support for (*MARK) will need an extra pointer on the
1410 stack; I have reserved it now, so that the stack frame size does not
1411 decrease.
1412
1413 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1414 item in branch that calls a recursion is a subroutine call - as in the
1415 second branch in the above example - was incorrectly given the compile-
1416 time error "recursive call could loop indefinitely" because pcre_compile()
1417 was not correctly checking the subroutine for matching a non-empty string.
1418
1419 14. The checks for overrunning compiling workspace could trigger after an
1420 overrun had occurred. This is a "should never occur" error, but it can be
1421 triggered by pathological patterns such as hundreds of nested parentheses.
1422 The checks now trigger 100 bytes before the end of the workspace.
1423
1424 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1425
1426
1427 Version 8.01 19-Jan-2010
1428 ------------------------
1429
1430 1. If a pattern contained a conditional subpattern with only one branch (in
1431 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1432 computed the wrong minimum data length (which is of course zero for such
1433 subpatterns). This could cause incorrect "no match" results.
1434
1435 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1436 the pattern is reset in the first branch, pcre_compile() failed with
1437 "internal error: code overflow at offset...". This happened only when
1438 the reset was to the original external option setting. (An optimization
1439 abstracts leading options settings into an external setting, which was the
1440 cause of this.)
1441
1442 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1443 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1444 assertion pattern did not match (meaning that the assertion was true), it
1445 was incorrectly treated as false if the SKIP had been reached during the
1446 matching. This also applied to assertions used as conditions.
1447
1448 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1449 assertion subpattern, including such a pattern used as a condition,
1450 unpredictable results occurred, instead of the error return
1451 PCRE_ERROR_DFA_UITEM.
1452
1453 5. The C++ GlobalReplace function was not working like Perl for the special
1454 situation when an empty string is matched. It now does the fancy magic
1455 stuff that is necessary.
1456
1457 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1458 removed. (These were left over from very, very early versions of PCRE.)
1459
1460 7. Some cosmetic changes to the code to make life easier when compiling it
1461 as part of something else:
1462
1463 (a) Change DEBUG to PCRE_DEBUG.
1464
1465 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1466 called "current" as "current_branch", to prevent a collision with the
1467 Linux macro when compiled as a kernel module.
1468
1469 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1470 prevent a collision with the Linux macro when compiled as a kernel
1471 module.
1472
1473 8. In pcre_compile() there are some checks for integer overflows that used to
1474 cast potentially large values to (double). This has been changed to that
1475 when building, a check for int64_t is made, and if it is found, it is used
1476 instead, thus avoiding the use of floating point arithmetic. (There is no
1477 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1478 double.
1479
1480 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1481 2005 (difference between two addresses compared to an unsigned value).
1482
1483 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1484 custom one, because of the following reported problem in Windows:
1485
1486 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1487 under Win32.
1488 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1489 therefore missing the function definition.
1490 - The compiler thus generates a "C" signature for the test function.
1491 - The linker fails to find the "C" function.
1492 - PCRE fails to configure if asked to do so against libbz2.
1493
1494 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1495 messages were output:
1496
1497 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1498 rerunning libtoolize, to keep the correct libtool macros in-tree.
1499 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1500
1501 I have done both of these things.
1502
1503 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1504 most of the time, it *can* run out if it is given a pattern that contains a
1505 runaway infinite recursion. I updated the discussion in the pcrestack man
1506 page.
1507
1508 13. Now that we have gone to the x.xx style of version numbers, the minor
1509 version may start with zero. Using 08 or 09 is a bad idea because users
1510 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1511 interpreted as invalid octal numbers. I've updated the previous comment in
1512 configure.ac, and also added a check that gives an error if 08 or 09 are
1513 used.
1514
1515 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1516 causing partial matching to fail when the end of the subject matched \W
1517 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1518
1519 15. There were some discrepancies between the declarations in pcre_internal.h
1520 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1521 their definitions. The declarations used "const uschar *" and the
1522 definitions used USPTR. Even though USPTR is normally defined as "const
1523 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1524 reported that: "This difference in casting confuses some C++ compilers, for
1525 example, SunCC recognizes above declarations as different functions and
1526 generates broken code for hbpcre." I have changed the declarations to use
1527 USPTR.
1528
1529 16. GNU libtool is named differently on some systems. The autogen.sh script now
1530 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1531 (FreeBSD).
1532
1533 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1534 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1535 comment: "Figure out how to create a longlong from a string: strtoll and
1536 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1537 instance, but it only takes 2 args instead of 3!"
1538
1539 18. A subtle bug concerned with back references has been fixed by a change of
1540 specification, with a corresponding code fix. A pattern such as
1541 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1542 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1543 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1544 same bug. Such groups have to be quantified to be useful, or contained
1545 inside another quantified group. (If there's no repetition, the reference
1546 can never match.) The problem arises because, having left the group and
1547 moved on to the rest of the pattern, a later failure that backtracks into
1548 the group uses the captured value from the final iteration of the group
1549 rather than the correct earlier one. I have fixed this in PCRE by forcing
1550 any group that contains a reference to itself to be an atomic group; that
1551 is, there cannot be any backtracking into it once it has completed. This is
1552 similar to recursive and subroutine calls.
1553
1554
1555 Version 8.00 19-Oct-09
1556 ----------------------
1557
1558 1. The table for translating pcre_compile() error codes into POSIX error codes
1559 was out-of-date, and there was no check on the pcre_compile() error code
1560 being within the table. This could lead to an OK return being given in
1561 error.
1562
1563 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1564 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1565 in a Windows environment.
1566
1567 3. The pcregrep --count option prints the count for each file even when it is
1568 zero, as does GNU grep. However, pcregrep was also printing all files when
1569 --files-with-matches was added. Now, when both options are given, it prints
1570 counts only for those files that have at least one match. (GNU grep just
1571 prints the file name in this circumstance, but including the count seems
1572 more useful - otherwise, why use --count?) Also ensured that the
1573 combination -clh just lists non-zero counts, with no names.
1574
1575 4. The long form of the pcregrep -F option was incorrectly implemented as
1576 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1577 but it seems right to fix it, and I didn't think it was worth preserving
1578 the old behaviour.
1579
1580 5. The command line items --regex=pattern and --regexp=pattern were not
1581 recognized by pcregrep, which required --regex pattern or --regexp pattern
1582 (with a space rather than an '='). The man page documented the '=' forms,
1583 which are compatible with GNU grep; these now work.
1584
1585 6. No libpcreposix.pc file was created for pkg-config; there was just
1586 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1587
1588 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1589 when UCP support is not needed, by modifying the Python script that
1590 generates it from Unicode data files. This should not matter if the module
1591 is correctly used as a library, but I received one complaint about 50K of
1592 unwanted data. My guess is that the person linked everything into his
1593 program rather than using a library. Anyway, it does no harm.
1594
1595 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1596 was a minimum greater than 1 for a wide character in a possessive
1597 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1598 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1599 character. Chaos in the form of incorrect output or a compiling loop could
1600 result.
1601
1602 9. The restrictions on what a pattern can contain when partial matching is
1603 requested for pcre_exec() have been removed. All patterns can now be
1604 partially matched by this function. In addition, if there are at least two
1605 slots in the offset vector, the offset of the earliest inspected character
1606 for the match and the offset of the end of the subject are set in them when
1607 PCRE_ERROR_PARTIAL is returned.
1608
1609 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1610 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1611 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1612 and may be more useful for multi-segment matching.
1613
1614 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1615 used to be given if ever the end of the subject was reached; now it is
1616 given only if matching could not proceed because another character was
1617 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1618 string "Z", which now yields "no match" instead of "partial match". In the
1619 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1620 final character ended with (*FAIL).
1621
1622 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1623 if the pattern had a "must contain" character that was already found in the
1624 earlier partial match, unless partial matching was again requested. For
1625 example, with the pattern /dog.(body)?/, the "must contain" character is
1626 "g". If the first part-match was for the string "dog", restarting with
1627 "sbody" failed. This bug has been fixed.
1628
1629 13. The string returned by pcre_dfa_exec() after a partial match has been
1630 changed so that it starts at the first inspected character rather than the
1631 first character of the match. This makes a difference only if the pattern
1632 starts with a lookbehind assertion or \b or \B (\K is not supported by
1633 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1634 matching functions compatible, and I think it's the right thing to do.
1635
1636 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1637 so that the demonstration program is easily available in environments where
1638 PCRE has not been installed from source.
1639
1640 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1641 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1642 library.
1643
1644 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1645 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1646 is not the first non-POSIX option to be added. Clearly some people find
1647 these options useful.
1648
1649 17. If a caller to the POSIX matching function regexec() passes a non-zero
1650 value for nmatch with a NULL value for pmatch, the value of
1651 nmatch is forced to zero.
1652
1653 18. RunGrepTest did not have a test for the availability of the -u option of
1654 the diff command, as RunTest does. It now checks in the same way as
1655 RunTest, and also checks for the -b option.
1656
1657 19. If an odd number of negated classes containing just a single character
1658 interposed, within parentheses, between a forward reference to a named
1659 subpattern and the definition of the subpattern, compilation crashed with
1660 an internal error, complaining that it could not find the referenced
1661 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1662 [The bug was that it was starting one character too far in when skipping
1663 over the character class, thus treating the ] as data rather than
1664 terminating the class. This meant it could skip too much.]
1665
1666 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1667 /g option in pcretest when the pattern contains \K, which makes it possible
1668 to have an empty string match not at the start, even when the pattern is
1669 anchored. Updated pcretest and pcredemo to use this option.
1670
1671 21. If the maximum number of capturing subpatterns in a recursion was greater
1672 than the maximum at the outer level, the higher number was returned, but
1673 with unset values at the outer level. The correct (outer level) value is
1674 now given.
1675
1676 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1677 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1678 make it do so. The string so far is captured, making this feature
1679 compatible with Perl.
1680
1681 23. The tests have been re-organized, adding tests 11 and 12, to make it
1682 possible to check the Perl 5.10 features against Perl 5.10.
1683
1684 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1685 pattern matches a fixed length string. PCRE did not allow this; now it
1686 does. Neither allows recursion.
1687
1688 25. I finally figured out how to implement a request to provide the minimum
1689 length of subject string that was needed in order to match a given pattern.
1690 (It was back references and recursion that I had previously got hung up
1691 on.) This code has now been added to pcre_study(); it finds a lower bound
1692 to the length of subject needed. It is not necessarily the greatest lower
1693 bound, but using it to avoid searching strings that are too short does give
1694 some useful speed-ups. The value is available to calling programs via
1695 pcre_fullinfo().
1696
1697 26. While implementing 25, I discovered to my embarrassment that pcretest had
1698 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1699 study optimizations had never been tested with that matching function.
1700 Oops. What is worse, even when it was passed study data, there was a bug in
1701 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1702 were also very few tests of studied patterns with pcre_dfa_exec().
1703
1704 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1705 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1706 on the other side of the coin, they are no longer allowed to have different
1707 names, because these cannot be distinguished in PCRE, and this has caused
1708 confusion. (This is a difference from Perl.)
1709
1710 28. When duplicate subpattern names are present (necessarily with different
1711 numbers, as required by 27 above), and a test is made by name in a
1712 conditional pattern, either for a subpattern having been matched, or for
1713 recursion in such a pattern, all the associated numbered subpatterns are
1714 tested, and the overall condition is true if the condition is true for any
1715 one of them. This is the way Perl works, and is also more like the way
1716 testing by number works.
1717
1718
1719 Version 7.9 11-Apr-09
1720 ---------------------
1721
1722 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1723 (pcretest), all targets were linked against these libraries. This included
1724 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1725 libraries. This caused unwanted dependencies to be created. This problem
1726 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1727 pcretest is linked with readline.
1728
1729 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1730 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1731 moved outside it again, because FALSE and TRUE are already defined in AIX,
1732 but BOOL is not.
1733
1734 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1735 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1736
1737 4. The pcregrep documentation said spaces were inserted as well as colons (or
1738 hyphens) following file names and line numbers when outputting matching
1739 lines. This is not true; no spaces are inserted. I have also clarified the
1740 wording for the --colour (or --color) option.
1741
1742 5. In pcregrep, when --colour was used with -o, the list of matching strings
1743 was not coloured; this is different to GNU grep, so I have changed it to be
1744 the same.
1745
1746 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1747 each matching line was coloured. Now it goes on to look for further matches
1748 of any of the test patterns, which is the same behaviour as GNU grep.
1749
1750 7. A pattern that could match an empty string could cause pcregrep to loop; it
1751 doesn't make sense to accept an empty string match in pcregrep, so I have
1752 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1753 seems to be how GNU grep behaves. [But see later change 40 for release
1754 8.33.]
1755
1756 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1757 start or after a newline", because the conditional assertion was not being
1758 correctly handled. The rule now is that both the assertion and what follows
1759 in the first alternative must satisfy the test.
1760
1761 9. If auto-callout was enabled in a pattern with a conditional group whose
1762 condition was an assertion, PCRE could crash during matching, both with
1763 pcre_exec() and pcre_dfa_exec().
1764
1765 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1766 used for matching.
1767
1768 11. Unicode property support in character classes was not working for
1769 characters (bytes) greater than 127 when not in UTF-8 mode.
1770
1771 12. Added the -M command line option to pcretest.
1772
1773 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1774
1775 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1776
1777 16. Added comments and documentation about mis-use of no_arg in the C++
1778 wrapper.
1779
1780 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1781 from Martin Jerabek that uses macro names for all relevant character and
1782 string constants.
1783
1784 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1785 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1786 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1787 these, but not everybody uses configure.
1788
1789 19. A conditional group that had only one branch was not being correctly
1790 recognized as an item that could match an empty string. This meant that an
1791 enclosing group might also not be so recognized, causing infinite looping
1792 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1793 with the subject "ab", where knowledge that the repeated group can match
1794 nothing is needed in order to break the loop.
1795
1796 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1797 exec(), but without supplying a callout function, matching went wrong.
1798
1799 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1800 leak if the size of the offset vector was greater than 30. When the vector
1801 is smaller, the saved offsets during recursion go onto a local stack
1802 vector, but for larger vectors malloc() is used. It was failing to free
1803 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1804 error, in fact).
1805
1806 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1807 heapframe that is used only when UTF-8 support is enabled. This caused no
1808 problem, but was untidy.
1809
1810 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1811 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1812 included within another project.
1813
1814 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1815 slightly modified by me:
1816
1817 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1818 not building pcregrep.
1819
1820 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1821 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1822
1823 25. Forward references, both numeric and by name, in patterns that made use of
1824 duplicate group numbers, could behave incorrectly or give incorrect errors,
1825 because when scanning forward to find the reference group, PCRE was not
1826 taking into account the duplicate group numbers. A pattern such as
1827 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1828
1829 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1830 the feature of a custom pointer more persuasive (as requested by a user).
1831
1832 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1833 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1834 pre-defined.
1835
1836 28. Added support for (*UTF8) at the start of a pattern.
1837
1838 29. Arrange for flags added by the "release type" setting in CMake to be shown
1839 in the configuration summary.
1840
1841
1842 Version 7.8 05-Sep-08
1843 ---------------------
1844
1845 1. Replaced UCP searching code with optimized version as implemented for Ad
1846 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1847 stage table and inline lookup instead of a function, giving speed ups of 2
1848 to 5 times on some simple patterns that I tested. Permission was given to
1849 distribute the MultiStage2.py script that generates the tables (it's not in
1850 the tarball, but is in the Subversion repository).
1851
1852 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1853 scripts.
1854
1855 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1856 a group with a zero qualifier. The result of the study could be incorrect,
1857 or the function might crash, depending on the pattern.
1858
1859 4. Caseless matching was not working for non-ASCII characters in back
1860 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1861 It now works when Unicode Property Support is available.
1862
1863 5. In pcretest, an escape such as \x{de} in the data was always generating
1864 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1865 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1866 truncation.
1867
1868 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1869
1870 7. Added two (int) casts to pcregrep when printing the difference of two
1871 pointers, in case they are 64-bit values.
1872
1873 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1874 test 2 if it fails.
1875
1876 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1877 and a #define of that name to empty if it is not externally set. This is to
1878 allow users of MSVC to set it if necessary.
1879
1880 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1881 the convenience functions in the pcre_get.c source file.
1882
1883 11. An option change at the start of a pattern that had top-level alternatives
1884 could cause overwriting and/or a crash. This command provoked a crash in
1885 some environments:
1886
1887 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1888
1889 This potential security problem was recorded as CVE-2008-2371.
1890
1891 12. For a pattern where the match had to start at the beginning or immediately
1892 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1893 pcre_dfa_exec() could read past the end of the passed subject if there was
1894 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1895 pcretest so that it places the subject at the end of its malloc-ed buffer.
1896
1897 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1898 exec() might read past the end of the data buffer in UTF-8 mode.
1899
1900 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1901 the data contained the byte 0x85 as part of a UTF-8 character within its
1902 first line. This applied both to normal and DFA matching.
1903
1904 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1905 /^[^d]*?$/8 failed to match "abc".
1906
1907 16. Added a missing copyright notice to pcrecpp_internal.h.
1908
1909 17. Make it more clear in the documentation that values returned from
1910 pcre_exec() in ovector are byte offsets, not character counts.
1911
1912 18. Tidied a few places to stop certain compilers from issuing warnings.
1913
1914 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1915 supplied by Stefan Weber. I made a further small update for 7.8 because
1916 there is a change of source arrangements: the pcre_searchfuncs.c module is
1917 replaced by pcre_ucd.c.
1918
1919
1920 Version 7.7 07-May-08
1921 ---------------------
1922
1923 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1924 a string to a long long, pretend we don't even have a long long." This is
1925 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1926
1927 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1928 pre-7.6 versions, which defined a global no_arg variable instead of putting
1929 it in the RE class. (See also #8 below.)
1930
1931 3. Remove a line of dead code, identified by coverity and reported by Nuno
1932 Lopes.
1933
1934 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1935
1936 (1) The include/exclude patterns were being applied to the whole pathnames
1937 of files, instead of just to the final components.
1938
1939 (2) If there was more than one level of directory, the subdirectories were
1940 skipped unless they satisfied the include/exclude conditions. This is
1941 inconsistent with GNU grep (and could even be seen as contrary to the
1942 pcregrep specification - which I improved to make it absolutely clear).
1943 The action now is always to scan all levels of directory, and just
1944 apply the include/exclude patterns to regular files.
1945
1946 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1947 --exclude_dir in the tests to avoid scanning .svn directories.
1948
1949 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1950 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1951 doesn't support NULs in patterns.
1952
1953 7. Added some missing "const"s to declarations of static tables in
1954 pcre_compile.c and pcre_dfa_exec.c.
1955
1956 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1957 caused by fix #2 above. (Subsequently also a second patch to fix the
1958 first patch. And a third patch - this was a messy problem.)
1959
1960 9. Applied Craig's patch to remove the use of push_back().
1961
1962 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1963 matching function regexec().
1964
1965 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1966 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1967 references. PCRE supports relative numbers with this syntax (I don't think
1968 Oniguruma does).
1969
1970 12. Previously, a group with a zero repeat such as (...){0} was completely
1971 omitted from the compiled regex. However, this means that if the group
1972 was called as a subroutine from elsewhere in the pattern, things went wrong
1973 (an internal error was given). Such groups are now left in the compiled
1974 pattern, with a new opcode that causes them to be skipped at execution
1975 time.
1976
1977 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1978 to the way PCRE behaves:
1979
1980 (a) A lone ] character is dis-allowed (Perl treats it as data).
1981
1982 (b) A back reference to an unmatched subpattern matches an empty string
1983 (Perl fails the current match path).
1984
1985 (c) A data ] in a character class must be notated as \] because if the
1986 first data character in a class is ], it defines an empty class. (In
1987 Perl it is not possible to have an empty class.) The empty class []
1988 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1989 The negative empty class [^] matches any one character, independently
1990 of the DOTALL setting.
1991
1992 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1993 non-existent subpattern following a character class starting with ']' and
1994 containing () gave an internal compiling error instead of "reference to
1995 non-existent subpattern". Fortunately, when the pattern did exist, the
1996 compiled code was correct. (When scanning forwards to check for the
1997 existence of the subpattern, it was treating the data ']' as terminating
1998 the class, so got the count wrong. When actually compiling, the reference
1999 was subsequently set up correctly.)
2000
2001 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2002 it was being rejected as not supported by pcre_dfa_exec(), even though
2003 other assertions are supported. I have made pcre_dfa_exec() support
2004 (*FAIL).
2005
2006 16. The implementation of 13c above involved the invention of a new opcode,
2007 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2008 cannot be changed at match time, I realized I could make a small
2009 improvement to matching performance by compiling OP_ALLANY instead of
2010 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2011 on the OP_ANY path.
2012
2013 17. Compiling pcretest on Windows with readline support failed without the
2014 following two fixes: (1) Make the unistd.h include conditional on
2015 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2016
2017 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2018 ncurses library to be included for pcretest when ReadLine support is
2019 requested, but also to allow for it to be overridden. This patch came from
2020 Daniel Bergström.
2021
2022 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2023 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2024 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2025 spotting this.
2026
2027
2028 Version 7.6 28-Jan-08
2029 ---------------------
2030
2031 1. A character class containing a very large number of characters with
2032 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2033 overflow.
2034
2035 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2036 HAVE_LONG_LONG is not defined.
2037
2038 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2039 bring them up to date and include new features. This patch includes:
2040
2041 - Fixed PH's badly added libz and libbz2 support.
2042 - Fixed a problem with static linking.
2043 - Added pcredemo. [But later removed - see 7 below.]
2044 - Fixed dftables problem and added an option.
2045 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2046 HAVE_LONG_LONG.
2047 - Added readline support for pcretest.
2048 - Added an listing of the option settings after cmake has run.
2049
2050 4. A user submitted a patch to Makefile that makes it easy to create
2051 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2052 Makefile.am that cause it to include this special target, without
2053 affecting anything else. Note that the same mingw target plus all
2054 the other distribution libraries and programs are now supported
2055 when configuring with CMake (see 6 below) instead of with
2056 Configure/Make.
2057
2058 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2059 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2060 exported in the Windows port". It has not yet been confirmed that the patch
2061 solves the problem, but it does no harm.
2062
2063 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2064 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2065 with CMake, and also correct the comment about stack recursion.
2066
2067 7. Remove the automatic building of pcredemo from the ./configure system and
2068 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2069 of a program that users should build themselves after PCRE is installed, so
2070 building it automatically is not really right. What is more, it gave
2071 trouble in some build environments.
2072
2073 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2074
2075
2076 Version 7.5 10-Jan-08
2077 ---------------------
2078
2079 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2080 values in parens when parsing an RE using the C++ wrapper."
2081
2082 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2083 Characters greater than 255 were excluded from the class instead of being
2084 included.
2085
2086 3. The same bug as (2) above applied to negated POSIX classes such as
2087 [:^space:].
2088
2089 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2090 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2091 I have changed it.
2092
2093 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2094 first named subpattern) and a construct such as (?&a) would reference the
2095 first named subpattern whose name started with "a" (in other words, the
2096 length check was missing). Both these problems are fixed. "Subpattern name
2097 expected" is now given for (?&) (a zero-length name), and this patch also
2098 makes it give the same error for \k'' (previously it complained that that
2099 was a reference to a non-existent subpattern).
2100
2101 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2102 this is right because (?- can be followed by option settings as well as by
2103 digits. I have, however, made the messages clearer.
2104
2105 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2106 than the number used in the conditional) now cause a compile-time error.
2107 This is actually not compatible with Perl, which accepts such patterns, but
2108 treats the conditional as always being FALSE (as PCRE used to), but it
2109 seems to me that giving a diagnostic is better.
2110
2111 8. Change "alphameric" to the more common word "alphanumeric" in comments
2112 and messages.
2113
2114 9. Fix two occurrences of "backslash" in comments that should have been
2115 "backspace".
2116
2117 10. Remove two redundant lines of code that can never be obeyed (their function
2118 was moved elsewhere).
2119
2120 11. The program that makes PCRE's Unicode character property table had a bug
2121 which caused it to generate incorrect table entries for sequences of
2122 characters that have the same character type, but are in different scripts.
2123 It amalgamated them into a single range, with the script of the first of
2124 them. In other words, some characters were in the wrong script. There were
2125 thirteen such cases, affecting characters in the following ranges:
2126
2127 U+002b0 - U+002c1
2128 U+0060c - U+0060d
2129 U+0061e - U+00612
2130 U+0064b - U+0065e
2131 U+0074d - U+0076d
2132 U+01800 - U+01805
2133 U+01d00 - U+01d77
2134 U+01d9b - U+01dbf
2135 U+0200b - U+0200f
2136 U+030fc - U+030fe
2137 U+03260 - U+0327f
2138 U+0fb46 - U+0fbb1
2139 U+10450 - U+1049d
2140
2141 12. The -o option (show only the matching part of a line) for pcregrep was not
2142 compatible with GNU grep in that, if there was more than one match in a
2143 line, it showed only the first of them. It now behaves in the same way as
2144 GNU grep.
2145
2146 13. If the -o and -v options were combined for pcregrep, it printed a blank
2147 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2148 does the same. The return code can be used to tell if there were any
2149 non-matching lines.
2150
2151 14. Added --file-offsets and --line-offsets to pcregrep.
2152
2153 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2154 infinitely looping recursion. The bug was that positive lookaheads were not
2155 being skipped when checking for a possible empty match (negative lookaheads
2156 and both kinds of lookbehind were skipped).
2157
2158 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2159 inclusion of <windows.h> to before rather than after the definition of
2160 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2161
2162 17. Specifying a possessive quantifier with a specific limit for a Unicode
2163 character property caused pcre_compile() to compile bad code, which led at
2164 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2165 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2166 caused the error; without that there was no problem.
2167
2168 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2169
2170 19. Added --enable-pcretest-libreadline.
2171
2172 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2173 RE::GlobalReplace(). As a result, the number of replacements returned was
2174 double what it should be. I removed one of the increments, but Craig sent a
2175 later patch that removed the other one (the right fix) and added unit tests
2176 that check the return values (which was not done before).
2177
2178 21. Several CMake things:
2179
2180 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2181 the names libpcre and libpcreposix, not just pcre and pcreposix.
2182
2183 (2) The above change means that pcretest and pcregrep are now correctly
2184 linked with the newly-built libraries, not previously installed ones.
2185
2186 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2187
2188 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2189 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2190 UTF-8 newline character). The key issue is that the pattern starts .*;
2191 this means that the match must be either at the beginning, or after a
2192 newline. The bug was in the code for advancing after a failed match and
2193 checking that the new position followed a newline. It was not taking
2194 account of UTF-8 characters correctly.
2195
2196 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2197 character classes. PCRE was not treating the sequence [:...:] as a
2198 character class unless the ... were all letters. Perl, however, seems to
2199 allow any characters between [: and :], though of course it rejects as
2200 unknown any "names" that contain non-letters, because all the known class
2201 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2202 for example, whereas PCRE did not - it did not recognize a POSIX character
2203 class. This seemed a bit dangerous, so the code has been changed to be
2204 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2205 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2206 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2207 Perl does, and where it didn't before.
2208
2209 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2210 Windows environments %n is disabled by default.
2211
2212
2213 Version 7.4 21-Sep-07
2214 ---------------------
2215
2216 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2217 means that a class such as [\s] counted as "explicit reference to CR or
2218 LF". That isn't really right - the whole point of the change was to try to
2219 help when there was an actual mention of one of the two characters. So now
2220 the change happens only if \r or \n (or a literal CR or LF) character is
2221 encountered.
2222
2223 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2224 of both had grown to the point where there were only 3 bits left.
2225 Fortunately, there was spare space in the data structure, and so I have
2226 moved the internal flags into a new 16-bit field to free up more option
2227 bits.
2228
2229 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2230 but did not set the internal JCHANGED flag - either of these is enough to
2231 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2232 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2233 start sets both bits.
2234
2235 4. Added options (at build time, compile time, exec time) to change \R from
2236 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2237
2238 5. doc/pcresyntax.html was missing from the distribution.
2239
2240 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2241 compatibility, even though it is no longer used.
2242
2243 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2244 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2245 windows.h file is present (where different names are used). [This was
2246 reversed later after testing - see 16 below.]
2247
2248 8. Changed all #include <config.h> to #include "config.h". There were also
2249 some further <pcre.h> cases that I changed to "pcre.h".
2250
2251 9. When pcregrep was used with the --colour option, it missed the line ending
2252 sequence off the lines that it output.
2253
2254 10. It was pointed out to me that arrays of string pointers cause lots of
2255 relocations when a shared library is dynamically loaded. A technique of
2256 using a single long string with a table of offsets can drastically reduce
2257 these. I have refactored PCRE in four places to do this. The result is
2258 dramatic:
2259
2260 Originally: 290
2261 After changing UCP table: 187
2262 After changing error message table: 43
2263 After changing table of "verbs" 36
2264 After changing table of Posix names 22
2265
2266 Thanks to the folks working on Gregex for glib for this insight.
2267
2268 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2269 unicode-properties was also set.
2270
2271 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2272
2273 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2274 checked only for CRLF.
2275
2276 14. Added casts to pcretest.c to avoid compiler warnings.
2277
2278 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2279
2280 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2281 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2282 entirely. This removes changes made in 7 above.
2283
2284 17. The CMake files have been updated, and there is now more information about
2285 building with CMake in the NON-UNIX-USE document.
2286
2287
2288 Version 7.3 28-Aug-07
2289 ---------------------
2290
2291 1. In the rejigging of the build system that eventually resulted in 7.1, the
2292 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2293 brackets there is not right, since it causes compilers to look for an
2294 installed pcre.h, not the version that is in the source that is being
2295 compiled (which of course may be different). I have changed it back to:
2296
2297 #include "pcre.h"
2298
2299 I have a vague recollection that the change was concerned with compiling in
2300 different directories, but in the new build system, that is taken care of
2301 by the VPATH setting the Makefile.
2302
2303 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2304 when the subject happened to end in the byte 0x85 (e.g. if the last
2305 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2306 characters but of course it shouldn't be taken as a newline when it is part
2307 of another character. The bug was that, for an unlimited repeat of . in
2308 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2309 characters when looking for a newline.
2310
2311 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2312
2313 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2314 in debug output.
2315
2316 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2317 long printing in the pcrecpp unittest when running under MinGW.
2318
2319 6. ESC_K was left out of the EBCDIC table.
2320
2321 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2322 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2323 limit also applies to "virtual nesting" when a pattern is recursive, and in
2324 this case 1000 isn't so big. I have been able to remove this limit at the
2325 expense of backing off one optimization in certain circumstances. Normally,
2326 when pcre_exec() would call its internal match() function recursively and
2327 immediately return the result unconditionally, it uses a "tail recursion"
2328 feature to save stack. However, when a subpattern that can match an empty
2329 string has an unlimited repetition quantifier, it no longer makes this
2330 optimization. That gives it a stack frame in which to save the data for
2331 checking that an empty string has been matched. Previously this was taken
2332 from the 1000-entry workspace that had been reserved. So now there is no
2333 explicit limit, but more stack is used.
2334
2335 8. Applied Daniel's patches to solve problems with the import/export magic
2336 syntax that is required for Windows, and which was going wrong for the
2337 pcreposix and pcrecpp parts of the library. These were overlooked when this
2338 problem was solved for the main library.
2339
2340 9. There were some crude static tests to avoid integer overflow when computing
2341 the size of patterns that contain repeated groups with explicit upper
2342 limits. As the maximum quantifier is 65535, the maximum group length was
2343 set at 30,000 so that the product of these two numbers did not overflow a
2344 32-bit integer. However, it turns out that people want to use groups that
2345 are longer than 30,000 bytes (though not repeat them that many times).
2346 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2347 made it possible to implement the integer overflow checks in a much more
2348 dynamic way, which I have now done. The artificial limitation on group
2349 length has been removed - we now have only the limit on the total length of
2350 the compiled pattern, which depends on the LINK_SIZE setting.
2351
2352 10. Fixed a bug in the documentation for get/copy named substring when
2353 duplicate names are permitted. If none of the named substrings are set, the
2354 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2355 empty string.
2356
2357 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2358 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2359 because the ] is interpreted as the first data character and the
2360 terminating ] is not found. PCRE has been made compatible with Perl in this
2361 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2362 cause memory overwriting.
2363
2364 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2365 string has been matched (to stop an infinite loop). It was not recognizing
2366 a conditional subpattern that could match an empty string if that
2367 subpattern was within another subpattern. For example, it looped when
2368 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2369 condition was not nested. This bug has been fixed.
2370
2371 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2372 past the start of the subject in the presence of bytes with the top bit
2373 set, for example "\x8aBCD".
2374
2375 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2376 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2377
2378 14. Optimized (?!) to (*FAIL).
2379
2380 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2381 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2382 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2383 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2384 does: it's just the validity check that is more restrictive.
2385
2386 16. Inserted checks for integer overflows during escape sequence (backslash)
2387 processing, and also fixed erroneous offset values for syntax errors during
2388 backslash processing.
2389
2390 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2391 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2392
2393 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2394 caused an overrun.
2395
2396 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2397 something other than just ASCII characters) inside a group that had an
2398 unlimited repeat caused a loop at compile time (while checking to see
2399 whether the group could match an empty string).
2400
2401 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2402 [\P{Any}] did so. (Error in the code for printing property names.)
2403
2404 21. An orphan \E inside a character class could cause a crash.
2405
2406 22. A repeated capturing bracket such as (A)? could cause a wild memory
2407 reference during compilation.
2408
2409 23. There are several functions in pcre_compile() that scan along a compiled
2410 expression for various reasons (e.g. to see if it's fixed length for look
2411 behind). There were bugs in these functions when a repeated \p or \P was
2412 present in the pattern. These operators have additional parameters compared
2413 with \d, etc, and these were not being taken into account when moving along
2414 the compiled data. Specifically:
2415
2416 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2417 length.
2418
2419 (b) An item such as \pL+ within a repeated group could cause crashes or
2420 loops.
2421
2422 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2423 "reference to non-existent subpattern" error.
2424
2425 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2426
2427 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2428 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2429
2430 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2431
2432 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2433 character were causing crashes (broken optimization).
2434
2435 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2436 \p or \P) caused a compile-time loop.
2437
2438 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2439 break. For example, the unstudied pattern [\r\n]A does not match the string
2440 "\r\nA" because change 7.0/46 below moves the current point on by two
2441 characters after failing to match at the start. However, the pattern \nA
2442 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2443 the same is true. There doesn't seem any very clean way out of this, but
2444 what I have chosen to do makes the common cases work: PCRE now takes note
2445 of whether there can be an explicit match for \r or \n anywhere in the
2446 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2447 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2448 pattern has explicit CR or LF references.
2449
2450 29. Added (*CR) etc for changing newline setting at start of pattern.
2451
2452
2453 Version 7.2 19-Jun-07
2454 ---------------------
2455
2456 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2457 which is apparently normally available under Windows.
2458
2459 2. Re-jig the pcregrep tests with different newline settings in an attempt
2460 to make them independent of the local environment's newline setting.
2461
2462 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2463
2464 4. Some of the "internals" tests were previously cut out when the link size
2465 was not 2, because the output contained actual offsets. The recent new
2466 "Z" feature of pcretest means that these can be cut out, making the tests
2467 usable with all link sizes.
2468
2469 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2470 stack recursion. This gives a massive performance boost under BSD, but just
2471 a small improvement under Linux. However, it saves one field in the frame
2472 in all cases.
2473
2474 6. Added more features from the forthcoming Perl 5.10:
2475
2476 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2477 recursion call. It refers to the nth most recently opened parentheses.
2478
2479 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2480 to be opened parentheses.
2481
2482 (c) Conditions that refer to capturing parentheses can be specified
2483 relatively, for example, (?(-2)... or (?(+3)...
2484
2485 (d) \K resets the start of the current match so that everything before
2486 is not part of it.
2487
2488 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2489
2490 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2491 reference syntax.
2492
2493 (g) (?| introduces a group in which the numbering of parentheses in each
2494 alternative starts with the same number.
2495
2496 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2497
2498 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2499 PCRE_INFO_JCHANGED.
2500
2501 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2502 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2503 for detecting groups that can match an empty string.
2504
2505 9. A pattern with a very large number of alternatives (more than several
2506 hundred) was running out of internal workspace during the pre-compile
2507 phase, where pcre_compile() figures out how much memory will be needed. A
2508 bit of new cunning has reduced the workspace needed for groups with
2509 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2510 workspace instead of running out of the 4096 that are available.
2511
2512 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2513
2514 11. Applied patch from Google to remove an optimization that didn't quite work.
2515 The report of the bug said:
2516
2517 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2518 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2519 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2520
2521 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2522 it matched the wrong number of bytes.
2523
2524
2525 Version 7.1 24-Apr-07
2526 ---------------------
2527
2528 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2529 that is more "standard", making use of automake and other Autotools. There
2530 is some re-arrangement of the files and adjustment of comments consequent
2531 on this.
2532
2533 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2534 for recursive directory scanning broke on some systems because the files
2535 are not scanned in any specific order and on different systems the order
2536 was different. A call to "sort" has been inserted into RunGrepTest for the
2537 approprate test as a short-term fix. In the longer term there may be an
2538 alternative.
2539
2540 3. I had an email from Eric Raymond about problems translating some of PCRE's
2541 man pages to HTML (despite the fact that I distribute HTML pages, some
2542 people do their own conversions for various reasons). The problems
2543 concerned the use of low-level troff macros .br and .in. I have therefore
2544 removed all such uses from the man pages (some were redundant, some could
2545 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2546 HTML has been updated to handle .nf/.fi and to complain if it encounters
2547 .br or .in.
2548
2549 4. Updated comments in configure.ac that get placed in config.h.in and also
2550 arranged for config.h to be included in the distribution, with the name
2551 config.h.generic, for the benefit of those who have to compile without
2552 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2553
2554 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2555 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2556 makevp.bat for the current PCRE, using the additional files
2557 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2558
2559 6. A Windows user reported a minor discrepancy with test 2, which turned out
2560 to be caused by a trailing space on an input line that had got lost in his
2561 copy. The trailing space was an accident, so I've just removed it.
2562
2563 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2564 that is needed.
2565
2566 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2567 as "const" (a) because they are and (b) because it helps the PHP
2568 maintainers who have recently made a script to detect big data structures
2569 in the php code that should be moved to the .rodata section. I remembered
2570 to update Builducptable as well, so it won't revert if ucptable.h is ever
2571 re-created.
2572
2573 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2574 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2575 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2576 support is not required. This saves 1.5-2K of code, which is important in
2577 some applications.
2578
2579 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2580 so as not to refer to the tables, even though these functions will never be
2581 called when UTF-8 support is disabled. Otherwise there are problems with a
2582 shared library.
2583
2584 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2585
2586 (a) It was defining its arguments as char * instead of void *.
2587
2588 (b) It was assuming that all moves were upwards in memory; this was true
2589 a long time ago when I wrote it, but is no longer the case.
2590
2591 The emulated memove() is provided for those environments that have neither
2592 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2593 is clearly not the case, as these two bugs were recently reported.
2594
2595 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2596 and Detrail to create the HTML documentation, the .txt form of the man
2597 pages, and it removes trailing spaces from listed files. It also creates
2598 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2599 case, it wraps all the #defines with #ifndefs. This script should be run
2600 before "make dist".
2601
2602 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2603 with Unicode property support.
2604
2605 (a) For a maximizing quantifier, if the two different cases of the
2606 character were of different lengths in their UTF-8 codings (there are
2607 some cases like this - I found 11), and the matching function had to
2608 back up over a mixture of the two cases, it incorrectly assumed they
2609 were both the same length.
2610
2611 (b) When PCRE was configured to use the heap rather than the stack for
2612 recursion during matching, it was not correctly preserving the data for
2613 the other case of a UTF-8 character when checking ahead for a match
2614 while processing a minimizing repeat. If the check also involved
2615 matching a wide character, but failed, corruption could cause an
2616 erroneous result when trying to check for a repeat of the original
2617 character.
2618
2619 13. Some tidying changes to the testing mechanism:
2620
2621 (a) The RunTest script now detects the internal link size and whether there
2622 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2623 values substituted by "configure". (The RunGrepTest script already did
2624 this for UTF-8.) The configure.ac script no longer substitutes the
2625 relevant variables.
2626
2627 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2628 with length and offset values. This means that the output is different
2629 for different internal link sizes. Test 2 is skipped for link sizes
2630 other than 2 because of this, bypassing the problem. Unfortunately,
2631 there was also a test in test 3 (the locale tests) that used /B and
2632 failed for link sizes other than 2. Rather than cut the whole test out,
2633 I have added a new /Z option to pcretest that replaces the length and
2634 offset values with spaces. This is now used to make test 3 independent
2635 of link size. (Test 2 will be tidied up later.)
2636
2637 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2638 segmentation fault instead of returning the appropriate error message.
2639
2640 15. In multiline mode when the newline sequence was set to "any", the pattern
2641 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2642 This doesn't seem right; it now treats the CRLF combination as the line
2643 ending, and so does not match in that case. It's only a pattern such as ^$
2644 that would hit this one: something like ^ABC$ would have failed after \r
2645 and then tried again after \r\n.
2646
2647 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2648 in an attempt to make files that differ only in their line terminators
2649 compare equal. This works on Linux.
2650
2651 17. Under certain error circumstances pcregrep might try to free random memory
2652 as it exited. This is now fixed, thanks to valgrind.
2653
2654 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2655 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2656 was because its rules for how to advance for /g after matching an empty
2657 string at the end of a line did not allow for this case. They now check for
2658 it specially.
2659
2660 20. pcretest is supposed to handle patterns and data of any length, by
2661 extending its buffers when necessary. It was getting this wrong when the
2662 buffer for a data line had to be extended.
2663
2664 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2665 CRLF as a newline sequence.
2666
2667 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2668 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2669 I have nevertheless tidied it up.
2670
2671 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2672
2673 24. Added a man page for pcre-config.
2674
2675
2676 Version 7.0 19-Dec-06
2677 ---------------------
2678
2679 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2680 moving to gcc 4.1.1.
2681
2682 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2683 sys/time.h, which is documented as needed for this function. It doesn't
2684 seem to matter on Linux, but it showed up on some releases of OS X.
2685
2686 3. It seems that there are systems where bytes whose values are greater than
2687 127 match isprint() in the "C" locale. The "C" locale should be the
2688 default when a C program starts up. In most systems, only ASCII printing
2689 characters match isprint(). This difference caused the output from pcretest
2690 to vary, making some of the tests fail. I have changed pcretest so that:
2691
2692 (a) When it is outputting text in the compiled version of a pattern, bytes
2693 other than 32-126 are always shown as hex escapes.
2694
2695 (b) When it is outputting text that is a matched part of a subject string,
2696 it does the same, unless a different locale has been set for the match
2697 (using the /L modifier). In this case, it uses isprint() to decide.
2698
2699 4. Fixed a major bug that caused incorrect computation of the amount of memory
2700 required for a compiled pattern when options that changed within the
2701 pattern affected the logic of the preliminary scan that determines the
2702 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2703 that the computed length was too small. The symptoms of this bug were
2704 either the PCRE error "internal error: code overflow" from pcre_compile(),
2705 or a glibc crash with a message such as "pcretest: free(): invalid next
2706 size (fast)". Examples of patterns that provoked this bug (shown in
2707 pcretest format) are:
2708
2709 /(?-x: )/x
2710 /(?x)(?-x: \s*#\s*)/
2711 /((?i)[\x{c0}])/8
2712 /(?i:[\x{c0}])/8
2713
2714 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2715 is now done differently.
2716
2717 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2718 wrapper classes; (b) implement a new function in the C++ scanner that is
2719 more efficient than the old way of doing things because it avoids levels of
2720 recursion in the regex matching; (c) add a paragraph to the documentation
2721 for the FullMatch() function.
2722
2723 6. The escape sequence \n was being treated as whatever was defined as
2724 "newline". Not only was this contrary to the documentation, which states
2725 that \n is character 10 (hex 0A), but it also went horribly wrong when
2726 "newline" was defined as CRLF. This has been fixed.
2727
2728 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2729 was being set to -1 for the "end of line" case (supposedly a value that no
2730 character can have). Though this value is never used (the check for end of
2731 line is "zero bytes in current character"), it caused compiler complaints.
2732 I've changed it to 0xffffffff.
2733
2734 8. In pcre_version.c, the version string was being built by a sequence of
2735 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2736 string (as it is for production releases) called a macro with an empty
2737 argument. The C standard says the result of this is undefined. The gcc
2738 compiler treats it as an empty string (which was what was wanted) but it is
2739 reported that Visual C gives an error. The source has been hacked around to
2740 avoid this problem.
2741
2742 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2743 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2744 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2745 of them did).
2746
2747 10. Originally, pcretest opened its input and output without "b"; then I was
2748 told that "b" was needed in some environments, so it was added for release
2749 5.0 to both the input and output. (It makes no difference on Unix-like
2750 systems.) Later I was told that it is wrong for the input on Windows. I've
2751 now abstracted the modes into two macros, to make it easier to fiddle with
2752 them, and removed "b" from the input mode under Windows.
2753
2754 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2755
2756 12. Added -help and --help to pcretest as an official way of being reminded
2757 of the options.
2758
2759 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2760 and pcrecpp.cc because they annoy compilers at high warning levels.
2761
2762 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2763
2764 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2765 scripts are not C programs :-) and which was not noticed because it works
2766 on Linux.
2767
2768 16. pcretest is supposed to handle any length of pattern and data line (as one
2769 line or as a continued sequence of lines) by extending its input buffer if
2770 necessary. This feature was broken for very long pattern lines, leading to
2771 a string of junk being passed to pcre_compile() if the pattern was longer
2772 than about 50K.
2773
2774 17. I have done a major re-factoring of the way pcre_compile() computes the
2775 amount of memory needed for a compiled pattern. Previously, there was code
2776 that made a preliminary scan of the pattern in order to do this. That was
2777 OK when PCRE was new, but as the facilities have expanded, it has become
2778 harder and harder to keep it in step with the real compile phase, and there
2779 have been a number of bugs (see for example, 4 above). I have now found a
2780 cunning way of running the real compile function in a "fake" mode that
2781 enables it to compute how much memory it would need, while actually only
2782 ever using a few hundred bytes of working memory and without too many
2783 tests of the mode. This should make future maintenance and development
2784 easier. A side effect of this work is that the limit of 200 on the nesting
2785 depth of parentheses has been removed (though this was never a serious
2786 limitation, I suspect). However, there is a downside: pcre_compile() now
2787 runs more slowly than before (30% or more, depending on the pattern). I
2788 hope this isn't a big issue. There is no effect on runtime performance.
2789
2790 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2791 newline (only possible for the last line of a file) and it was a
2792 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2793
2794 19. Added additional timing features to pcretest. (1) The -tm option now times
2795 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2796 separate command line item, by a number that specifies the number of
2797 repeats to use when timing. The default is 50000; this gives better
2798 precision, but takes uncomfortably long for very large patterns.
2799
2800 20. Extended pcre_study() to be more clever in cases where a branch of a
2801 subpattern has no definite first character. For example, (a*|b*)[cd] would
2802 previously give no result from pcre_study(). Now it recognizes that the
2803 first character must be a, b, c, or d.
2804
2805 21. There was an incorrect error "recursive call could loop indefinitely" if
2806 a subpattern (or the entire pattern) that was being tested for matching an
2807 empty string contained only one non-empty item after a nested subpattern.
2808 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2809 incorrectly, because the \d was being skipped in the check.
2810
2811 22. The pcretest program now has a new pattern option /B and a command line
2812 option -b, which is equivalent to adding /B to every pattern. This causes
2813 it to show the compiled bytecode, without the additional information that
2814 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2815 is the same as /B/I).
2816
2817 23. A new optimization is now able automatically to treat some sequences such
2818 as a*b as a*+b. More specifically, if something simple (such as a character
2819 or a simple class like \d) has an unlimited quantifier, and is followed by
2820 something that cannot possibly match the quantified thing, the quantifier
2821 is automatically "possessified".
2822
2823 24. A recursive reference to a subpattern whose number was greater than 39
2824 went wrong under certain circumstances in UTF-8 mode. This bug could also
2825 have affected the operation of pcre_study().
2826
2827 25. Realized that a little bit of performance could be had by replacing
2828 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2829
2830 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2831
2832 27. Possessive quantifiers such as a++ were previously implemented by turning
2833 them into atomic groups such as ($>a+). Now they have their own opcodes,
2834 which improves performance. This includes the automatically created ones
2835 from 23 above.
2836
2837 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2838 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2839 the first matched character to be a colon. This applied both to named and
2840 numbered groups.
2841
2842 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2843
2844 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2845 building PCRE on Windows easier, so I have included it in the distribution.
2846
2847 31. There is now a check in pcretest against a ridiculously large number being
2848 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2849 loop, the loop is abandoned.
2850
2851 32. Forward references to subpatterns in conditions such as (?(2)...) where
2852 subpattern 2 is defined later cause pcre_compile() to search forwards in
2853 the pattern for the relevant set of parentheses. This search went wrong
2854 when there were unescaped parentheses in a character class, parentheses
2855 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2856
2857 33. "Subroutine" calls and backreferences were previously restricted to
2858 referencing subpatterns earlier in the regex. This restriction has now
2859 been removed.
2860
2861 34. Added a number of extra features that are going to be in Perl 5.10. On the
2862 whole, these are just syntactic alternatives for features that PCRE had
2863 previously implemented using the Python syntax or my own invention. The
2864 other formats are all retained for compatibility.
2865
2866 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2867 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2868 also .NET compatible.
2869
2870 (b) A recursion or subroutine call to a named group can now be defined as
2871 (?&name) as well as (?P>name).
2872
2873 (c) A backreference to a named group can now be defined as \k<name> or
2874 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2875 5.10, are also .NET compatible.
2876
2877 (d) A conditional reference to a named group can now use the syntax
2878 (?(<name>) or (?('name') as well as (?(name).
2879
2880 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2881 groups (named and numbered) that are never evaluated inline, but can be
2882 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2883 is always false. There may be only one alternative in such a group.
2884
2885 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2886 as the simple (?(R). The condition is true only if the most recent
2887 recursion is that of the given number or name. It does not search out
2888 through the entire recursion stack.
2889
2890 (g) The escape \gN or \g{N} has been added, where N is a positive or
2891 negative number, specifying an absolute or relative reference.
2892
2893 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2894 some "unreachable code" warnings.
2895
2896 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2897 things, this adds five new scripts.
2898
2899 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2900 There were also incompatibilities regarding the handling of \Q..\E inside
2901 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2902 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2903
2904 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2905 matches an empty string, and forcibly breaks the loop. There were bugs in
2906 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2907 against aaaa the result was just "a" rather than "aaaa", for example. Two
2908 separate and independent bugs (that affected different cases) have been
2909 fixed.
2910
2911 39. Refactored the code to abolish the use of different opcodes for small
2912 capturing bracket numbers. This is a tidy that I avoided doing when I
2913 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2914 The new approach is not only tidier, it makes it possible to reduce the
2915 memory needed to fix the previous bug (38).
2916
2917 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2918 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2919 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2920 mode.
2921
2922 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2923 report.
2924
2925 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2926 copy construction and assignment in the C++ wrapper.
2927
2928 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2929 couple of bugs that could have given wrong results in the "--newline=crlf"
2930 case.
2931
2932 44. Added a number of casts and did some reorganization of signed/unsigned int
2933 variables following suggestions from Dair Grant. Also renamed the variable
2934 "this" as "item" because it is a C++ keyword.
2935
2936 45. Arranged for dftables to add
2937
2938 #include "pcre_internal.h"
2939
2940 to pcre_chartables.c because without it, gcc 4.x may remove the array
2941 definition from the final binary if PCRE is built into a static library and
2942 dead code stripping is activated.
2943
2944 46. For an unanchored pattern, if a match attempt fails at the start of a
2945 newline sequence, and the newline setting is CRLF or ANY, and the next two
2946 characters are CRLF, advance by two characters instead of one.
2947
2948
2949 Version 6.7 04-Jul-06
2950 ---------------------
2951
2952 1. In order to handle tests when input lines are enormously long, pcretest has
2953 been re-factored so that it automatically extends its buffers when
2954 necessary. The code is crude, but this _is_ just a test program. The
2955 default size has been increased from 32K to 50K.
2956
2957 2. The code in pcre_study() was using the value of the re argument before
2958 testing it for NULL. (Of course, in any sensible call of the function, it
2959 won't be NULL.)
2960
2961 3. The memmove() emulation function in pcre_internal.h, which is used on
2962 systems that lack both memmove() and bcopy() - that is, hardly ever -
2963 was missing a "static" storage class specifier.
2964
2965 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2966 containing an extended class (one that cannot be represented by a bitmap
2967 because it contains high-valued characters or Unicode property items, e.g.
2968 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2969 pattern, but PCRE should not loop if you do not (it no longer does).
2970 [Detail: two cases were found: (a) a repeated subpattern containing an
2971 extended class; (b) a recursive reference to a subpattern that followed a
2972 previous extended class. It wasn't skipping over the extended class
2973 correctly when UTF-8 mode was not set.]
2974
2975 5. A negated single-character class was not being recognized as fixed-length
2976 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2977 compile error "lookbehind assertion is not fixed length".
2978
2979 6. The RunPerlTest auxiliary script was showing an unexpected difference
2980 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2981 write a Perl script that can interpret lines of an input file either as
2982 byte characters or as UTF-8, which is what "perltest" was being required to
2983 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2984 can't do is switch easily at run time between having the "use utf8;" pragma
2985 or not. In the end, I fudged it by using the RunPerlTest script to insert
2986 "use utf8;" explicitly for the UTF-8 tests.
2987
2988 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2989 the end of the subject string, contrary to the documentation and to what
2990 Perl does. This was true of both matching functions. Now it matches only at
2991 the start of the subject and immediately after *internal* newlines.
2992
2993 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2994 a pointer to an int instead of a pointer to an unsigned long int. This
2995 caused problems on 64-bit systems.
2996
2997 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2998 instance of the 'standard' template library not being so standard".
2999
3000 10. There was no check on the number of named subpatterns nor the maximum
3001 length of a subpattern name. The product of these values is used to compute
3002 the size of the memory block for a compiled pattern. By supplying a very
3003 long subpattern name and a large number of named subpatterns, the size
3004 computation could be caused to overflow. This is now prevented by limiting
3005 the length of names to 32 characters, and the number of named subpatterns
3006 to 10,000.
3007
3008 11. Subpatterns that are repeated with specific counts have to be replicated in
3009 the compiled pattern. The size of memory for this was computed from the
3010 length of the subpattern and the repeat count. The latter is limited to
3011 65535, but there was no limit on the former, meaning that integer overflow
3012 could in principle occur. The compiled length of a repeated subpattern is
3013 now limited to 30,000 bytes in order to prevent this.
3014
3015 12. Added the optional facility to have named substrings with the same name.
3016
3017 13. Added the ability to use a named substring as a condition, using the
3018 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3019 are numbers (not recommended). Forward references are permitted.
3020
3021 14. Added forward references in named backreferences (if you see what I mean).
3022
3023 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3024 pattern could run off the end of the subject. For example, the pattern
3025 "(?s)(.{1,5})"8 did this with the subject "ab".
3026
3027 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3028 PCRE_CASELESS was set when matching characters that were quantified with ?
3029 or *.
3030
3031 17. A character class other than a single negated character that had a minimum
3032 but no maximum quantifier - for example [ab]{6,} - was not handled
3033 correctly by pce_dfa_exec(). It would match only one character.
3034
3035 18. A valid (though odd) pattern that looked like a POSIX character
3036 class but used an invalid character after [ (for example [[,abc,]]) caused
3037 pcre_compile() to give the error "Failed: internal error: code overflow" or
3038 in some cases to crash with a glibc free() error. This could even happen if
3039 the pattern terminated after [[ but there just happened to be a sequence of
3040 letters, a binary zero, and a closing ] in the memory that followed.
3041
3042 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3043 over the years. Originally (before any Unicode support), just the bottom 8
3044 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3045 output from "man perlunicode" includes this:
3046
3047 The regular expression compiler produces polymorphic opcodes. That
3048 is, the pattern adapts to the data and automatically switches to
3049 the Unicode character scheme when presented with Unicode data--or
3050 instead uses a traditional byte scheme when presented with byte
3051 data.
3052
3053 Sadly, a wide octal escape does not cause a switch, and in a string with
3054 no other multibyte characters, these octal escapes are treated as before.
3055 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3056 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3057 Unicode string.
3058
3059 I have not perpetrated such confusion in PCRE. Up till now, it took just
3060 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3061 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3062 translate to the appropriate multibyte character.
3063
3064 29. Applied some refactoring to reduce the number of warnings from Microsoft
3065 and Borland compilers. This has included removing the fudge introduced
3066 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3067 a warning about an unused variable.
3068
3069 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3070 characters since release 4.0, because Perl (from release 5.004) does not.
3071 [Or at least, is documented not to: some releases seem to be in conflict
3072 with the documentation.] However, when a pattern was studied with
3073 pcre_study() and all its branches started with \s, PCRE still included VT
3074 as a possible starting character. Of course, this did no harm; it just
3075 caused an unnecessary match attempt.
3076
3077 22. Removed a now-redundant internal flag bit that recorded the fact that case
3078 dependency changed within the pattern. This was once needed for "required
3079 byte" processing, but is no longer used. This recovers a now-scarce options
3080 bit. Also moved the least significant internal flag bit to the most-
3081 significant bit of the word, which was not previously used (hangover from
3082 the days when it was an int rather than a uint) to free up another bit for
3083 the future.
3084
3085 23. Added support for CRLF line endings as well as CR and LF. As well as the
3086 default being selectable at build time, it can now be changed at runtime
3087 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3088 specify that it is scanning data with non-default line endings.
3089
3090 24. Changed the definition of CXXLINK to make it agree with the definition of
3091 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3092
3093 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3094 recursions. This makes a big different to stack usage for some patterns.
3095
3096 26. If a subpattern containing a named recursion or subroutine reference such
3097 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3098 the space required for the compiled pattern went wrong and gave too small a
3099 value. Depending on the environment, this could lead to "Failed: internal
3100 error: code overflow at offset 49" or "glibc detected double free or
3101 corruption" errors.
3102
3103 27. Applied patches from Google (a) to support the new newline modes and (b) to
3104 advance over multibyte UTF-8 characters in GlobalReplace.
3105
3106 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3107 difference for some implementation of PCRE in some Windows version.
3108
3109 29. Added some extra testing facilities to pcretest:
3110
3111 \q<number> in a data line sets the "match limit" value
3112 \Q<number> in a data line sets the "match recursion limt" value
3113 -S <number> sets the stack size, where <number> is in megabytes
3114
3115 The -S option isn't available for Windows.
3116
3117
3118 Version 6.6 06-Feb-06
3119 ---------------------
3120
3121 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3122 in pcreposix.h. I have copied the definition from pcre.h.
3123
3124 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3125 because pcre.h is no longer a built file.
3126
3127 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3128 not normally included in the compiled code.
3129
3130
3131 Version 6.5 01-Feb-06
3132 ---------------------
3133
3134 1. When using the partial match feature with pcre_dfa_exec(), it was not
3135 anchoring the second and subsequent partial matches at the new starting
3136 point. This could lead to incorrect results. For example, with the pattern
3137 /1234/, partially matching against "123" and then "a4" gave a match.
3138
3139 2. Changes to pcregrep:
3140
3141 (a) All non-match returns from pcre_exec() were being treated as failures
3142 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3143 error message is output. Some extra information is given for the
3144 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3145 probably the only errors that are likely to be caused by users (by
3146 specifying a regex that has nested indefinite repeats, for instance).
3147 If there are more than 20 of these errors, pcregrep is abandoned.
3148
3149 (b) A binary zero was treated as data while matching, but terminated the
3150 output line if it was written out. This has been fixed: binary zeroes
3151 are now no different to any other data bytes.
3152
3153 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3154 used to set a locale for matching. The --locale=xxxx long option has
3155 been added (no short equivalent) to specify a locale explicitly on the
3156 pcregrep command, overriding the environment variables.
3157
3158 (d) When -B was used with -n, some line numbers in the output were one less
3159 than they should have been.
3160
3161 (e) Added the -o (--only-matching) option.
3162
3163 (f) If -A or -C was used with -c (count only), some lines of context were
3164 accidentally printed for the final match.
3165
3166 (g) Added the -H (--with-filename) option.
3167
3168 (h) The combination of options -rh failed to suppress file names for files
3169 that were found from directory arguments.
3170
3171 (i) Added the -D (--devices) and -d (--directories) options.
3172
3173 (j) Added the -F (--fixed-strings) option.
3174
3175 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3176
3177 (l) Added the --colo(u)r option.
3178
3179 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3180 is not present by default.
3181
3182 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3183 items such as (?R) or (?1), when the recursion could match a number of
3184 alternatives. If it matched one of the alternatives, but subsequently,
3185 outside the recursion, there was a failure, the code tried to back up into
3186 the recursion. However, because of the way PCRE is implemented, this is not
3187 possible, and the result was an incorrect result from the match.
3188
3189 In order to prevent this happening, the specification of recursion has
3190 been changed so that all such subpatterns are automatically treated as
3191 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3192
3193 4. I had overlooked the fact that, in some locales, there are characters for
3194 which isalpha() is true but neither isupper() nor islower() are true. In
3195 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3196 and ordfeminine) are like this. This affected the treatment of \w and \W
3197 when they appeared in character classes, but not when they appeared outside
3198 a character class. The bit map for "word" characters is now created
3199 separately from the results of isalnum() instead of just taking it from the
3200 upper, lower, and digit maps. (Plus the underscore character, of course.)
3201
3202 5. The above bug also affected the handling of POSIX character classes such as
3203 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3204 permanent tables. Instead, the bit maps for such a class were previously
3205 created as the appropriate unions of the upper, lower, and digit bitmaps.
3206 Now they are created by subtraction from the [[:word:]] class, which has
3207 its own bitmap.
3208
3209 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3210 It is created by subtracting the vertical space characters (\x09, \x0a,
3211 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3212 subtraction was done in the overall bitmap for a character class, meaning
3213 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3214 be recognized. This bug has been fixed.
3215
3216 7. Patches from the folks at Google:
3217
3218 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3219 real life, but is still worth protecting against".
3220
3221 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3222 regular expressions".
3223
3224 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3225 have it.
3226
3227 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3228 "configure" and the latter not, in order to fix a problem somebody had
3229 with compiling the Arg class on HP-UX.
3230
3231 (e) Improve the error-handling of the C++ wrapper a little bit.
3232
3233 (f) New tests for checking recursion limiting.
3234
3235 8. The pcre_memmove() function, which is used only if the environment does not
3236 have a standard memmove() function (and is therefore rarely compiled),
3237 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3238 returning a result (though PCRE never actually uses the result).
3239
3240 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3241 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3242 returned instead of calling malloc() with an overflowing number that would
3243 most likely cause subsequent chaos.
3244
3245 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3246
3247 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3248 with this option is matched, the nmatch and pmatch options of regexec() are
3249 ignored.
3250
3251 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3252 provided in case anyone wants to the the POSIX interface with UTF-8
3253 strings.
3254
3255 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3256 C++ linking (needed for some HP-UX environments).
3257
3258 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3259 (unused parameter) and in the pcre_printint() function (omitted "default"
3260 switch label when the default is to do nothing).
3261
3262 15. Added some code to make it possible, when PCRE is compiled as a C++
3263 library, to replace subject pointers for pcre_exec() with a smart pointer
3264 class, thus making it possible to process discontinuous strings.
3265
3266 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3267 much the same function. They were added by different people who were trying
3268 to make PCRE easy to compile on non-Unix systems. It has been suggested
3269 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3270 for compiling on Windows systems. I have therefore replaced it with
3271 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3272 defaults to "extern" for C or "extern C" for C++, which works fine on
3273 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3274 SCOPE with something explicit in config.h. In addition:
3275
3276 (a) pcreposix.h still had just "extern" instead of either of these macros;
3277 I have replaced it with PCRE_DATA_SCOPE.
3278
3279 (b) Functions such as _pcre_xclass(), which are internal to the library,
3280 but external in the C sense, all had PCRE_EXPORT in their definitions.
3281 This is apparently wrong for the Windows case, so I have removed it.
3282 (It makes no difference on Unix-like systems.)
3283
3284 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3285 of recursive calls to match(). This is different to MATCH_LIMIT because
3286 that limits the total number of calls to match(), not all of which increase
3287 the depth of recursion. Limiting the recursion depth limits the amount of
3288 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3289 when PCRE is compiled, and changed at run time. A patch from Google adds
3290 this functionality to the C++ interface.
3291
3292 18. Changes to the handling of Unicode character properties:
3293
3294 (a) Updated the table to Unicode 4.1.0.
3295
3296 (b) Recognize characters that are not in the table as "Cn" (undefined).
3297
3298 (c) I revised the way the table is implemented to a much improved format
3299 which includes recognition of ranges. It now supports the ranges that
3300 are defined in UnicodeData.txt, and it also amalgamates other
3301 characters into ranges. This has reduced the number of entries in the
3302 table from around 16,000 to around 3,000, thus reducing its size
3303 considerably. I realized I did not need to use a tree structure after
3304 all - a binary chop search is just as efficient. Having reduced the
3305 number of entries, I extended their size from 6 bytes to 8 bytes to
3306 allow for more data.
3307
3308 (d) Added support for Unicode script names via properties such as \p{Han}.
3309
3310 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3311 matching that character.
3312
3313 20. When matching a repeated Unicode property with a minimum greater than zero,
3314 (for example \pL{2,}), PCRE could look past the end of the subject if it
3315 reached it while seeking the minimum number of characters. This could
3316 happen only if some of the characters were more than one byte long, because
3317 there is a check for at least the minimum number of bytes.
3318
3319 21. Refactored the implementation of \p and \P so as to be more general, to
3320 allow for more different types of property in future. This has changed the
3321 compiled form incompatibly. Anybody with saved compiled patterns that use
3322 \p or \P will have to recompile them.
3323
3324 22. Added "Any" and "L&" to the supported property types.
3325
3326 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3327 but give a compile time error if the value is greater than 0xff.
3328
3329 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3330 accidentally not being installed or uninstalled.
3331
3332 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3333 made were to insert the current release number. This seemed silly, because
3334 it made things harder for people building PCRE on systems that don't run
3335 "configure". I have turned pcre.h into a distributed file, no longer built
3336 by "configure", with the version identification directly included. There is
3337 no longer a pcre.h.in file.
3338
3339 However, this change necessitated a change to the pcre-config script as
3340 well. It is built from pcre-config.in, and one of the substitutions was the
3341 release number. I have updated configure.ac so that ./configure now finds
3342 the release number by grepping pcre.h.
3343
3344 26. Added the ability to run the tests under valgrind.
3345
3346
3347 Version 6.4 05-Sep-05
3348 ---------------------
3349
3350 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3351 "--" to be printed when multiple files were scanned, even when none of the
3352 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3353 consider it to be a bug, and have restored the previous behaviour.
3354
3355 2. A couple of code tidies to get rid of compiler warnings.
3356
3357 3. The pcretest program used to cheat by referring to symbols in the library
3358 whose names begin with _pcre_. These are internal symbols that are not
3359 really supposed to be visible externally, and in some environments it is
3360 possible to suppress them. The cheating is now confined to including
3361 certain files from the library's source, which is a bit cleaner.
3362
3363 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3364 file's purpose clearer.
3365
3366 5. Reorganized pcre_ucp_findchar().
3367
3368
3369 Version 6.3 15-Aug-05
3370 ---------------------
3371
3372 1. The file libpcre.pc.in did not have general read permission in the tarball.
3373
3374 2. There were some problems when building without C++ support:
3375
3376 (a) If C++ support was not built, "make install" and "make test" still
3377 tried to test it.
3378
3379 (b) There were problems when the value of CXX was explicitly set. Some
3380 changes have been made to try to fix these, and ...
3381
3382 (c) --disable-cpp can now be used to explicitly disable C++ support.
3383
3384 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3385 backslash in a target when C++ was disabled. This confuses some
3386 versions of "make", apparently. Using an intermediate variable solves
3387 this. (Same for CPP_LOBJ.)
3388
3389 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3390 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3391 necessary on certain architectures.
3392
3393 4. Added a setting of -export-symbols-regex to the link command to remove
3394 those symbols that are exported in the C sense, but actually are local
3395 within the library, and not documented. Their names all begin with
3396 "_pcre_". This is not a perfect job, because (a) we have to except some
3397 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3398 available (and never for static libraries). I have made a note to try to
3399 find a way round (a) in the future.
3400
3401
3402 Version 6.2 01-Aug-05
3403 ---------------------
3404
3405 1. There was no test for integer overflow of quantifier values. A construction
3406 such as {1111111111111111} would give undefined results. What is worse, if
3407 a minimum quantifier for a parenthesized subpattern overflowed and became
3408 negative, the calculation of the memory size went wrong. This could have
3409 led to memory overwriting.
3410
3411 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3412
3413 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3414 operating environments where this matters.
3415
3416 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3417 PCRE options from within the C++ wrapper.
3418
3419 5. Named capturing subpatterns were not being correctly counted when a pattern
3420 was compiled. This caused two problems: (a) If there were more than 100
3421 such subpatterns, the calculation of the memory needed for the whole
3422 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3423 back references of the form \12, where the number was greater than 9, were
3424 not recognized as back references, even though there were sufficient
3425 previous subpatterns.
3426
3427 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3428 versions of gcc, e.g. 2.95.4.
3429
3430
3431 Version 6.1 21-Jun-05
3432 ---------------------
3433
3434 1. There was one reference to the variable "posix" in pcretest.c that was not
3435 surrounded by "#if !defined NOPOSIX".
3436
3437 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3438 the cross-check on the old pcre_info() function, for the benefit of the
3439 cut-down version of PCRE that is currently imported into Exim.
3440
3441 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3442 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3443 patterns that start (?i) but I don't think that matters. The use of (?i) is
3444 just an example; this all applies to the other options as well.
3445
3446 4. Since libtool seems to echo the compile commands it is issuing, the output
3447 from "make" can be reduced a bit by putting "@" in front of each libtool
3448 compile command.
3449
3450 5. Patch from the folks at Google for configure.in to be a bit more thorough
3451 in checking for a suitable C++ installation before trying to compile the
3452 C++ stuff. This should fix a reported problem when a compiler was present,
3453 but no suitable headers.
3454
3455 6. The man pages all had just "PCRE" as their title. I have changed them to
3456 be the relevant file name. I have also arranged that these names are
3457 retained in the file doc/pcre.txt, which is a concatenation in text format
3458 of all the man pages except the little individual ones for each function.
3459
3460 7. The NON-UNIX-USE file had not been updated for the different set of source
3461 files that come with release 6. I also added a few comments about the C++
3462 wrapper.
3463
3464
3465 Version 6.0 07-Jun-05
3466 ---------------------
3467
3468 1. Some minor internal re-organization to help with my DFA experiments.
3469
3470 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3471 didn't matter for the library itself when fully configured, but did matter
3472 when compiling without UCP support, or within Exim, where the ucp files are
3473 not imported.
3474
3475 3. Refactoring of the library code to split up the various functions into
3476 different source modules. The addition of the new DFA matching code (see
3477 below) to a single monolithic source would have made it really too
3478 unwieldy, quite apart from causing all the code to be include in a
3479 statically linked application, when only some functions are used. This is
3480 relevant even without the DFA addition now that patterns can be compiled in
3481 one application and matched in another.
3482
3483 The downside of splitting up is that there have to be some external
3484 functions and data tables that are used internally in different modules of
3485 the library but which are not part of the API. These have all had their
3486 names changed to start with "_pcre_" so that they are unlikely to clash
3487 with other external names.
3488
3489 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3490 a different (DFA) algorithm. Although it is slower than the original
3491 function, it does have some advantages for certain types of matching
3492 problem.
3493
3494 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3495 including restarting after a partial match.
3496
3497 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3498 defined when compiling for Windows was sent to me. I have put it into the
3499 code, though I have no means of testing or verifying it.
3500
3501 7. Added the pcre_refcount() auxiliary function.
3502
3503 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3504 match before or at the first newline in the subject string. In pcretest,
3505 the /f option on a pattern can be used to set this.
3506
3507 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3508 would behave wrongly. This has been present in PCRE since release 4.0.
3509
3510 10. A number of changes to the pcregrep command:
3511
3512 (a) Refactored how -x works; insert ^(...)$ instead of setting
3513 PCRE_ANCHORED and checking the length, in preparation for adding
3514 something similar for -w.
3515
3516 (b) Added the -w (match as a word) option.
3517
3518 (c) Refactored the way lines are read and buffered so as to have more
3519 than one at a time available.
3520
3521 (d) Implemented a pcregrep test script.
3522
3523 (e) Added the -M (multiline match) option. This allows patterns to match
3524 over several lines of the subject. The buffering ensures that at least
3525 8K, or the rest of the document (whichever is the shorter) is available
3526 for matching (and similarly the previous 8K for lookbehind assertions).
3527
3528 (f) Changed the --help output so that it now says
3529
3530 -w, --word-regex(p)
3531
3532 instead of two lines, one with "regex" and the other with "regexp"
3533 because that confused at least one person since the short forms are the
3534 same. (This required a bit of code, as the output is generated
3535 automatically from a table. It wasn't just a text change.)
3536
3537 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3538 option but starts with a hyphen. Could be a pattern or a path name
3539 starting with a hyphen, for instance.
3540
3541 (h) "-" can be given as a file name to represent stdin.
3542
3543 (i) When file names are being printed, "(standard input)" is used for
3544 the standard input, for compatibility with GNU grep. Previously
3545 "<stdin>" was used.
3546
3547 (j) The option --label=xxx can be used to supply a name to be used for
3548 stdin when file names are being printed. There is no short form.
3549
3550 (k) Re-factored the options decoding logic because we are going to add
3551 two more options that take data. Such options can now be given in four
3552 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3553
3554 (l) Added the -A, -B, and -C options for requesting that lines of context
3555 around matches be printed.
3556
3557 (m) Added the -L option to print the names of files that do not contain
3558 any matching lines, that is, the complement of -l.
3559
3560 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3561 continue to scan other files.
3562
3563 (o) The -s option was incorrectly implemented. For compatibility with other
3564 greps, it now suppresses the error message for a non-existent or non-
3565 accessible file (but not the return code). There is a new option called
3566 -q that suppresses the output of matching lines, which was what -s was
3567 previously doing.
3568
3569 (p) Added --include and --exclude options to specify files for inclusion
3570 and exclusion when recursing.
3571
3572 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3573 Hopefully, it now does.
3574
3575 12. Missing cast in pcre_study().
3576
3577 13. Added an "uninstall" target to the makefile.
3578
3579 14. Replaced "extern" in the function prototypes in Makefile.in with
3580 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3581 world, but is set differently for Windows.
3582
3583 15. Added a second compiling function called pcre_compile2(). The only
3584 difference is that it has an extra argument, which is a pointer to an
3585 integer error code. When there is a compile-time failure, this is set
3586 non-zero, in addition to the error test pointer being set to point to an
3587 error message. The new argument may be NULL if no error number is required
3588 (but then you may as well call pcre_compile(), which is now just a
3589 wrapper). This facility is provided because some applications need a
3590 numeric error indication, but it has also enabled me to tidy up the way
3591 compile-time errors are handled in the POSIX wrapper.
3592
3593 16. Added VPATH=.libs to the makefile; this should help when building with one
3594 prefix path and installing with another. (Or so I'm told by someone who
3595 knows more about this stuff than I do.)
3596
3597 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3598 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3599 match everything, including newlines. This is not POSIX-compatible, but
3600 somebody wanted the feature. From pcretest it can be activated by using
3601 both the P and the s flags.
3602
3603 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3604
3605 19. libpcre.pc was being incorrectly installed as executable.
3606
3607 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3608 it now also looks for '\r' so that it will work unmodified on Windows.
3609
3610 21. Added Google's contributed C++ wrapper to the distribution.
3611
3612 22. Added some untidy missing memory free() calls in pcretest, to keep
3613 Electric Fence happy when testing.
3614
3615
3616
3617 Version 5.0 13-Sep-04
3618 ---------------------
3619
3620 1. Internal change: literal characters are no longer packed up into items
3621 containing multiple characters in a single byte-string. Each character
3622 is now matched using a separate opcode. However, there may be more than one
3623 byte in the character in UTF-8 mode.
3624
3625 2. The pcre_callout_block structure has two new fields: pattern_position and
3626 next_item_length. These contain the offset in the pattern to the next match
3627 item, and its length, respectively.
3628
3629 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3630 insertion of callouts before each pattern item. Added the /C option to
3631 pcretest to make use of this.
3632
3633 4. On the advice of a Windows user, the lines
3634
3635 #if defined(_WIN32) || defined(WIN32)
3636 _setmode( _fileno( stdout ), 0x8000 );
3637 #endif /* defined(_WIN32) || defined(WIN32) */
3638
3639 have been added to the source of pcretest. This apparently does useful
3640 magic in relation to line terminators.
3641
3642 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3643 for the benefit of those environments where the "b" makes a difference.
3644
3645 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3646 to know about it. I have put a hack into configure.in that adds in code
3647 to set GCC=yes if CC=icc. This seems to end up at a point in the
3648 generated configure script that is early enough to affect the setting of
3649 compiler options, which is what is needed, but I have no means of testing
3650 whether it really works. (The user who reported this had patched the
3651 generated configure script, which of course I cannot do.)
3652
3653 LATER: After change 22 below (new libtool files), the configure script
3654 seems to know about icc (and also ecc). Therefore, I have commented out
3655 this hack in configure.in.
3656
3657 7. Added support for pkg-config (2 patches were sent in).
3658
3659 8. Negated POSIX character classes that used a combination of internal tables
3660 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3661 [[:^ascii]]. Typically, they would match almost any characters. The other
3662 POSIX classes were not broken in this way.
3663
3664 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3665 to find the match, as PCRE was deluded into thinking that the match had to
3666 start at the start point or following a newline. The same bug applied to
3667 patterns with negative forward assertions or any backward assertions
3668 preceding ".*" at the start, unless the pattern required a fixed first
3669 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3670
3671 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3672 starting at the last subject character, bytes beyond the end of the subject
3673 string were read.
3674
3675 11. Renamed the variable "class" as "classbits" to make life easier for C++
3676 users. (Previously there was a macro definition, but it apparently wasn't
3677 enough.)
3678
3679 12. Added the new field "tables" to the extra data so that tables can be passed
3680 in at exec time, or the internal tables can be re-selected. This allows
3681 a compiled regex to be saved and re-used at a later time by a different
3682 program that might have everything at different addresses.
3683
3684 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3685 -R library as well as a -L library.
3686
3687 14. The debugging options of pcretest (-d on the command line or D on a
3688 pattern) showed incorrect output for anything following an extended class
3689 that contained multibyte characters and which was followed by a quantifier.
3690
3691 15. Added optional support for general category Unicode character properties
3692 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3693 support. It adds about 90K to the size of the library. The meanings of the
3694 inbuilt class escapes such as \d and \s have NOT been changed.
3695
3696 16. Updated pcredemo.c to include calls to free() to release the memory for the
3697 compiled pattern.
3698
3699 17. The generated file chartables.c was being created in the source directory
3700 instead of in the building directory. This caused the build to fail if the
3701 source directory was different from the building directory, and was
3702 read-only.
3703
3704 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3705 file. No doubt somebody will tell me if they don't make sense... Also added
3706 Dan Mooney's comments about building on OpenVMS.
3707
3708 19. Added support for partial matching via the PCRE_PARTIAL option for
3709 pcre_exec() and the \P data escape in pcretest.
3710
3711 20. Extended pcretest with 3 new pattern features:
3712
3713 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3714 write the compiled pattern to the file whose name is "rest-of-line".
3715 This is a straight binary dump of the data, with the saved pointer to
3716 the character tables forced to be NULL. The study data, if any, is
3717 written too. After writing, pcretest reads a new pattern.
3718
3719 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3720 compiled pattern from the given file. There must not be any
3721 occurrences of "<" in the file name (pretty unlikely); if there are,
3722 pcretest will instead treat the initial "<" as a pattern delimiter.
3723 After reading in the pattern, pcretest goes on to read data lines as
3724 usual.
3725
3726 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3727 and 16-bit fields in a compiled pattern, to simulate a pattern that
3728 was compiled on a host of opposite endianness.
3729
3730 21. The pcre-exec() function can now cope with patterns that were compiled on
3731 hosts of opposite endianness, with this restriction:
3732
3733 As for any compiled expression that is saved and used later, the tables
3734 pointer field cannot be preserved; the extra_data field in the arguments
3735 to pcre_exec() should be used to pass in a tables address if a value
3736 other than the default internal tables were used at compile time.
3737
3738 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3739 now diagnosed as an error. Previously, most of the time, a negative number
3740 would have been treated as zero, but if in addition "ovector" was passed as
3741 NULL, a crash could occur.
3742
3743 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3744 new versions from the libtool 1.5 distribution (the last one is a copy of
3745 a file called libtool.m4). This seems to have fixed the need to patch
3746 "configure" to support Darwin 1.3 (which I used to do). However, I still
3747 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3748 workstation).
3749
3750 24. Changed the PCRE licence to be the more standard "BSD" licence.
3751
3752
3753 Version 4.5 01-Dec-03
3754 ---------------------
3755
3756 1. There has been some re-arrangement of the code for the match() function so
3757 that it can be compiled in a version that does not call itself recursively.
3758 Instead, it keeps those local variables that need separate instances for
3759 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3760 needs to "recurse". Keeping track of where control must go is done by means
3761 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3762 hide most of the details from the main code, and operates only if
3763 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3764 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3765 operating.
3766
3767 To make it easier for callers to provide specially tailored get/free
3768 functions for this usage, two new functions, pcre_stack_malloc, and
3769 pcre_stack_free, are used. They are always called in strict stacking order,
3770 and the size of block requested is always the same.
3771
3772 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3773 PCRE has been compiled to use the stack or the heap for recursion. The
3774 -C option of pcretest uses this to show which version is compiled.
3775
3776 A new data escape \S, is added to pcretest; it causes the amounts of store
3777 obtained and freed by both kinds of malloc/free at match time to be added
3778 to the output.
3779
3780 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3781 what's available on my current Linux desktop machine.
3782
3783 3. When matching a UTF-8 string, the test for a valid string at the start has
3784 been extended. If start_offset is not zero, PCRE now checks that it points
3785 to a byte that is the start of a UTF-8 character. If not, it returns
3786 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3787 this is necessary because there may be backward assertions in the pattern.
3788 When matching the same subject several times, it may save resources to use
3789 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3790
3791 4. The code for checking the validity of UTF-8 strings has been tightened so
3792 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3793 containing "overlong sequences".
3794
3795 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3796 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3797 should have been "&", but it just so happened that all the cases this let
3798 through by mistake were picked up later in the function.
3799
3800 6. I had used a variable called "isblank" - this is a C99 function, causing
3801 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3802
3803 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3804 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3805 so the version is shown; (c) stop "make test" echoing "./RunTest".
3806
3807 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3808
3809 9. The prototype for memmove() for systems that don't have it was using
3810 size_t, but the inclusion of the header that defines size_t was later. I've
3811 moved the #includes for the C headers earlier to avoid this.
3812
3813 10. Added some adjustments to the code to make it easier to compiler on certain
3814 special systems:
3815
3816 (a) Some "const" qualifiers were missing.
3817 (b) Added the macro EXPORT before all exported functions; by default this
3818 is defined to be empty.
3819 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3820 that it reads its output file name as an argument instead of writing
3821 to the standard output and assuming this can be redirected.
3822
3823 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3824 class containing characters with values greater than 255, PCRE compilation
3825 went into a loop.
3826
3827 12. A recursive reference to a subpattern that was within another subpattern
3828 that had a minimum quantifier of zero caused PCRE to crash. For example,
3829 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3830 recursion. If the recursively-called subpattern itself had a zero repeat,
3831 that was OK.
3832
3833 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3834 buffer into which it was copied (for escape processing) was still set at
3835 1024, so long lines caused crashes.
3836
3837 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3838 "internal error: code overflow...". This applied to any character class
3839 that was followed by a possessive quantifier.
3840
3841 15. Modified the Makefile to add libpcre.la as a prerequisite for
3842 libpcreposix.la because I was told this is needed for a parallel build to
3843 work.
3844
3845 16. If a pattern that contained .* following optional items at the start was
3846 studied, the wrong optimizing data was generated, leading to matching
3847 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3848 matching string must start with a or b or c. The correct conclusion for
3849 this pattern is that a match can start with any character.
3850
3851
3852 Version 4.4 13-Aug-03
3853 ---------------------
3854
3855 1. In UTF-8 mode, a character class containing characters with values between
3856 127 and 255 was not handled correctly if the compiled pattern was studied.
3857 In fixing this, I have also improved the studying algorithm for such
3858 classes (slightly).
3859
3860 2. Three internal functions had redundant arguments passed to them. Removal
3861 might give a very teeny performance improvement.
3862
3863 3. Documentation bug: the value of the capture_top field in a callout is *one
3864 more than* the number of the hightest numbered captured substring.
3865
3866 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3867 in incorrectly linking with a previously installed version. They now link
3868 explicitly with libpcre.la.
3869
3870 5. configure.in no longer needs to recognize Cygwin specially.
3871
3872 6. A problem in pcre.in for Windows platforms is fixed.
3873
3874 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3875 pcretest, it used to include the size of the study block as part of its
3876 output. Unfortunately, the structure contains a field that has a different
3877 size on different hardware architectures. This meant that the tests that
3878 showed this size failed. As the block is currently always of a fixed size,
3879 this information isn't actually particularly useful in pcretest output, so
3880 I have just removed it.
3881
3882 8. Three pre-processor statements accidentally did not start in column 1.
3883 Sadly, there are *still* compilers around that complain, even though
3884 standard C has not required this for well over a decade. Sigh.
3885
3886 9. In pcretest, the code for checking callouts passed small integers in the
3887 callout_data field, which is a void * field. However, some picky compilers
3888 complained about the casts involved for this on 64-bit systems. Now
3889 pcretest passes the address of the small integer instead, which should get
3890 rid of the warnings.
3891
3892 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3893 both compile and run time, and gives an error if an invalid UTF-8 sequence
3894 is found. There is a option for disabling this check in cases where the
3895 string is known to be correct and/or the maximum performance is wanted.
3896
3897 11. In response to a bug report, I changed one line in Makefile.in from
3898
3899 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3900 to
3901 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3902
3903 to look similar to other lines, but I have no way of telling whether this
3904 is the right thing to do, as I do not use Windows. No doubt I'll get told
3905 if it's wrong...
3906
3907
3908 Version 4.3 21-May-03
3909 ---------------------
3910
3911 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3912 Makefile.
3913
3914 2. Some refactoring to improve the quality of the code:
3915
3916 (i) The utf8_table... variables are now declared "const".
3917
3918 (ii) The code for \cx, which used the "case flipping" table to upper case
3919 lower case letters, now just substracts 32. This is ASCII-specific,
3920 but the whole concept of \cx is ASCII-specific, so it seems
3921 reasonable.
3922
3923 (iii) PCRE was using its character types table to recognize decimal and
3924 hexadecimal digits in the pattern. This is silly, because it handles
3925 only 0-9, a-f, and A-F, but the character types table is locale-
3926 specific, which means strange things might happen. A private
3927 table is now used for this - though it costs 256 bytes, a table is
3928 much faster than multiple explicit tests. Of course, the standard
3929 character types table is still used for matching digits in subject
3930 strings against \d.
3931
3932 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3933 ending in _t are). So I've renamed it as ESC_tee.
3934
3935 3. The first argument for regexec() in the POSIX wrapper should have been
3936 defined as "const".
3937
3938 4. Changed pcretest to use malloc() for its buffers so that they can be
3939 Electric Fenced for debugging.
3940
3941 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3942 to read one or more bytes before the start of the subject string. Often this
3943 had no effect on PCRE's behaviour, but in some circumstances it could
3944 provoke a segmentation fault.
3945
3946 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3947 to try to read one or more bytes before the start of the subject string.
3948
3949 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3950 UTF-8 support could misbehave in various ways if the subject string
3951 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3952 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3953 back over UTF-8 characters.)
3954
3955
3956 Version 4.2 14-Apr-03
3957 ---------------------
3958
3959 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3960
3961 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3962 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3963 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3964 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3965 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3966 and BUILD_EXEEXT
3967 Note: automatic setting of the BUILD variables is not yet working
3968 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3969 compile-time but not at link-time
3970 [LINK]: use for linking executables only
3971 make different versions for Windows and non-Windows
3972 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3973 libraries
3974 [LINK_FOR_BUILD]: new variable
3975 [OBJEXT]: use throughout
3976 [EXEEXT]: use throughout
3977 <winshared>: new target
3978 <wininstall>: new target
3979 <dftables.o>: use native compiler
3980 <dftables>: use native linker
3981 <install>: handle Windows platform correctly
3982 <clean>: ditto
3983 <check>: ditto
3984 copy DLL to top builddir before testing
3985
3986 As part of these changes, -no-undefined was removed again. This was reported
3987 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3988 in any case.
3989
3990 3. Some tidies to get rid of compiler warnings:
3991
3992 . In the match_data structure, match_limit was an unsigned long int, whereas
3993 match_call_count was an int. I've made them both unsigned long ints.
3994
3995 . In pcretest the fact that a const uschar * doesn't automatically cast to
3996 a void * provoked a warning.
3997
3998 . Turning on some more compiler warnings threw up some "shadow" variables
3999 and a few more missing casts.
4000
4001 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4002 option, a class that contained a single character with a value between 128
4003 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4004
4005 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4006 option, a class that contained several characters, but with at least one
4007 whose value was between 128 and 255 caused PCRE to crash.
4008
4009
4010 Version 4.1 12-Mar-03
4011 ---------------------
4012
4013 1. Compiling with gcc -pedantic found a couple of places where casts were
4014 needed, and a string in dftables.c that was longer than standard compilers are
4015 required to support.
4016
4017 2. Compiling with Sun's compiler found a few more places where the code could
4018 be tidied up in order to avoid warnings.
4019
4020 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4021 first of these names is deprecated in the latest Autoconf in favour of the name
4022 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4023 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4024 analogy I have changed it to CFLAGS_FOR_BUILD.
4025
4026 4. Added -no-undefined to the linking command in the Makefile, because this is
4027 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4028 linking step for the pcreposix library.
4029
4030 5. PCRE was failing to diagnose the case of two named groups with the same
4031 name.
4032
4033 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4034 literal character that is needed in the subject for a match, and scans along to
4035 ensure that it is present before embarking on the full matching process. This
4036 saves time in cases of nested unlimited repeats that are never going to match.
4037 Problem: the scan can take a lot of time if the subject is very long (e.g.
4038 megabytes), thus penalizing straightforward matches. It is now done only if the
4039 amount of subject to be scanned is less than 1000 bytes.
4040
4041 7. A lesser problem with the same optimization is that it was recording the
4042 first character of an anchored pattern as "needed", thus provoking a search
4043 right along the subject, even when the first match of the pattern was going to
4044 fail. The "needed" character is now not set for anchored patterns, unless it
4045 follows something in the pattern that is of non-fixed length. Thus, it still
4046 fulfils its original purpose of finding quick non-matches in cases of nested
4047 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4048
4049
4050 Version 4.0 17-Feb-03
4051 ---------------------
4052
4053 1. If a comment in an extended regex that started immediately after a meta-item
4054 extended to the end of string, PCRE compiled incorrect data. This could lead to
4055 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4056
4057 2. Moved to autoconf 2.53 and libtool 1.4.2.
4058
4059 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4060 the special perltest8 script is no longer needed - all the tests can be run
4061 from a single perltest script.
4062
4063 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4064 by \s. It has now been removed in PCRE. This means it isn't recognized as
4065 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4066 class [:space:] *does* include VT, thereby creating a mess.
4067
4068 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4069 space and tab.
4070
4071 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4072 its new features into the main test script, reducing the number of scripts.
4073
4074 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4075 were backward compatible, and made the (?i) apply to the whole pattern, as if
4076 /i were given. Now it behaves more logically, and applies the option setting
4077 only to what follows. PCRE has been changed to follow suit. However, if it
4078 finds options settings right at the start of the pattern, it extracts them into
4079 the global options, as before. Thus, they show up in the info data.
4080
4081 8. Added support for the \Q...\E escape sequence. Characters in between are
4082 treated as literals. This is slightly different from Perl in that $ and @ are
4083 also handled as literals inside the quotes. In Perl, they will cause variable
4084 interpolation. Note the following examples:
4085
4086 Pattern PCRE matches Perl matches
4087
4088 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4089 \Qabc\$xyz\E abc\$xyz abc\$xyz
4090 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4091
4092 For compatibility with Perl, \Q...\E sequences are recognized inside character
4093 classes as well as outside them.
4094
4095 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4096 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4097 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4098 signed/unsigned warnings.
4099
4100 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4101 option for pcretest, so I've replaced it by a simple function that does just
4102 that job.
4103
4104 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4105 "pcregrep -".
4106
4107 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4108 Java package. This provides some syntactic sugar for simple cases of what my
4109 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4110 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4111 item, you can use this simplified notation. Note that only makes sense with
4112 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4113 greediness, whatever the setting of the PCRE_UNGREEDY option.
4114
4115 13. A change of greediness default within a pattern was not taking effect at
4116 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4117 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4118 was abstracted outside.
4119
4120 14. PCRE now supports the \G assertion. It is true when the current matching
4121 position is at the start point of the match. This differs from \A when the
4122 starting offset is non-zero. Used with the /g option of pcretest (or similar
4123 code), it works in the same way as it does for Perl's /g option. If all
4124 alternatives of a regex begin with \G, the expression is anchored to the start
4125 match position, and the "anchored" flag is set in the compiled expression.
4126
4127 15. Some bugs concerning the handling of certain option changes within patterns
4128 have been fixed. These applied to options other than (?ims). For example,
4129 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4130 been the other way round. Some of this was related to change 7 above.
4131
4132 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4133 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4134 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4135 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4136
4137 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4138 mode. Unlike ".", it always matches newline, whatever the setting of
4139 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4140 assertions. Perl allows it, but it doesn't (in general) work because it can't
4141 calculate the length of the lookbehind. At least, that's the case for Perl
4142 5.8.0 - I've been told they are going to document that it doesn't work in
4143 future.
4144
4145 18. Added an error diagnosis for escapes that PCRE does not support: these are
4146 \L, \l, \N, \P, \p, \U, \u, and \X.
4147
4148 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4149 reading past the end of the pattern in cases such as /[abcd/.
4150
4151 20. PCRE was getting more memory than necessary for patterns with classes that
4152 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4153
4154 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4155 compiling PCRE for use with Virtual Pascal.
4156
4157 22. Small fix to the Makefile to make it work properly if the build is done
4158 outside the source tree.
4159
4160 23. Added a new extension: a condition to go with recursion. If a conditional
4161 subpattern starts with (?(R) the "true" branch is used if recursion has
4162 happened, whereas the "false" branch is used only at the top level.
4163
4164 24. When there was a very long string of literal characters (over 255 bytes
4165 without UTF support, over 250 bytes with UTF support), the computation of how
4166 much memory was required could be incorrect, leading to segfaults or other
4167 strange effects.
4168
4169 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4170 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4171 there was a subsequent back reference to those brackets. This meant that, for
4172 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4173 possible to check for precisely this case. All we can do is abandon the
4174 optimization if .* occurs inside capturing brackets when there are any back
4175 references whatsoever. (See below for a better fix that came later.)
4176
4177 26. The handling of the optimization for finding the first character of a
4178 non-anchored pattern, and for finding a character that is required later in the
4179 match were failing in some cases. This didn't break the matching; it just
4180 failed to optimize when it could. The way this is done has been re-implemented.
4181
4182 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4183
4184 28. Added a new feature that provides some of the functionality that Perl
4185 provides with (?{...}). The facility is termed a "callout". The way it is done
4186 in PCRE is for the caller to provide an optional function, by setting
4187 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4188 global variable. By default it is unset, which disables all calling out. To get
4189 the function called, the regex must include (?C) at appropriate points. This
4190 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4191 This provides a means of identifying different callout points. When PCRE
4192 reaches such a point in the regex, if pcre_callout has been set, the external
4193 function is called. It is provided with data in a structure called
4194 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4195 matching continues; if it returns a non-zero value, the match at the current
4196 point fails. However, backtracking will occur if possible. [This was changed
4197 later and other features added - see item 49 below.]
4198
4199 29. pcretest is upgraded to test the callout functionality. It provides a
4200 callout function that displays information. By default, it shows the start of
4201 the match and the current position in the text. There are some new data escapes
4202 to vary what happens:
4203
4204 \C+ in addition, show current contents of captured substrings
4205 \C- do not supply a callout function
4206 \C!n return 1 when callout number n is reached
4207 \C!n!m return 1 when callout number n is reached for the mth time
4208
4209 30. If pcregrep was called with the -l option and just a single file name, it
4210 output "<stdin>" if a match was found, instead of the file name.
4211
4212 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4213 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4214 pcre_exec(). This saves a malloc/free per call. The default value of
4215 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4216 when configuring.
4217
4218 32. The default maximum size of a compiled pattern is 64K. There have been a
4219 few cases of people hitting this limit. The code now uses macros to handle the
4220 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4221 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4222 configuring. Tests 2 and 5 work only with 2-byte links because they output
4223 debugging information about compiled patterns.
4224
4225 33. Internal code re-arrangements:
4226
4227 (a) Moved the debugging function for printing out a compiled regex into
4228 its own source file (printint.c) and used #include to pull it into
4229 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4230 separate copies.
4231
4232 (b) Defined the list of op-code names for debugging as a macro in
4233 internal.h so that it is next to the definition of the opcodes.
4234
4235 (c) Defined a table of op-code lengths for simpler skipping along compiled
4236 code. This is again a macro in internal.h so that it is next to the
4237 definition of the opcodes.
4238
4239 34. Added support for recursive calls to individual subpatterns, along the
4240 lines of Robin Houston's patch (but implemented somewhat differently).
4241
4242 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4243 allow it to read and process whole directories in Win32. This code was
4244 contributed by Lionel Fourquaux; it has not been tested by me.
4245
4246 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4247 used to name a group. Names consist of alphanumerics and underscores, and must
4248 be unique. Back references use the syntax (?P=name) and recursive calls use
4249 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4250 numbers. The function pcre_fullinfo() can be used after compilation to extract
4251 a name/number map. There are three relevant calls:
4252
4253 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4254 PCRE_INFO_NAMECOUNT yields the number of entries
4255 PCRE_INFO_NAMETABLE yields a pointer to the map.
4256
4257 The map is a vector of fixed-size entries. The size of each entry depends on
4258 the length of the longest name used. The first two bytes of each entry are the
4259 group number, most significant byte first. There follows the corresponding
4260 name, zero terminated. The names are in alphabetical order.
4261
4262 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4263 case instead of 255. Making it the same both with and without UTF-8 support
4264 means that the same test output works with both.
4265
4266 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4267 calling malloc() with a zero argument.
4268
4269 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4270 optimization. I've improved things by keeping a bitmap of backreferences with
4271 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4272 fact referenced, the optimization can be applied. It is unlikely that a
4273 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4274 the match to follow \n) will appear inside brackets with a number greater than
4275 31, but if it does, any back reference > 31 suppresses the optimization.
4276
4277 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4278 of disabling numbered capturing parentheses. Any opening parenthesis that is
4279 not followed by ? behaves as if it were followed by ?: but named parentheses
4280 can still be used for capturing (and they will acquire numbers in the usual
4281 way).
4282
4283 41. Redesigned the return codes from the match() function into yes/no/error so
4284 that errors can be passed back from deep inside the nested calls. A malloc
4285 failure while inside a recursive subpattern call now causes the
4286 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4287
4288 42. It is now possible to set a limit on the number of times the match()
4289 function is called in a call to pcre_exec(). This facility makes it possible to
4290 limit the amount of recursion and backtracking, though not in a directly
4291 obvious way, because the match() function is used in a number of different
4292 circumstances. The count starts from zero for each position in the subject
4293 string (for non-anchored patterns). The default limit is, for compatibility, a
4294 large number, namely 10 000 000. You can change this in two ways:
4295
4296 (a) When configuring PCRE before making, you can use --with-match-limit=n
4297 to set a default value for the compiled library.
4298
4299 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4300 a different value is set. See 45 below.
4301
4302 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4303
4304 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4305 of things that can be changed at compile time. The first argument specifies
4306 what is wanted and the second points to where the information is to be placed.
4307 The current list of available information is:
4308
4309 PCRE_CONFIG_UTF8
4310
4311 The output is an integer that is set to one if UTF-8 support is available;
4312 otherwise it is set to zero.
4313
4314 PCRE_CONFIG_NEWLINE
4315
4316 The output is an integer that it set to the value of the code that is used for
4317 newline. It is either LF (10) or CR (13).
4318
4319 PCRE_CONFIG_LINK_SIZE
4320
4321 The output is an integer that contains the number of bytes used for internal
4322 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4323
4324 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4325
4326 The output is an integer that contains the threshold above which the POSIX
4327 interface uses malloc() for output vectors. See item 31 above.
4328
4329 PCRE_CONFIG_MATCH_LIMIT
4330
4331 The output is an unsigned integer that contains the default limit of the number
4332 of match() calls in a pcre_exec() execution. See 42 above.
4333
4334 44. pcretest has been upgraded by the addition of the -C option. This causes it
4335 to extract all the available output from the new pcre_config() function, and to
4336 output it. The program then exits immediately.
4337
4338 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4339 order to support additional features. One way would have been to define
4340 pcre_exec2() (for example) with extra arguments, but this would not have been
4341 extensible, and would also have required all calls to the original function to
4342 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4343 is used for passing in "extra" data from pcre_study().
4344
4345 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4346 contains the following fields:
4347
4348 flags a bitmap indicating which of the following fields are set
4349 study_data opaque data from pcre_study()
4350 match_limit a way of specifying a limit on match() calls for a specific
4351 call to pcre_exec()
4352 callout_data data for callouts (see 49 below)
4353
4354 The flag bits are also defined in pcre.h, and are
4355
4356 PCRE_EXTRA_STUDY_DATA
4357 PCRE_EXTRA_MATCH_LIMIT
4358 PCRE_EXTRA_CALLOUT_DATA
4359
4360 The pcre_study() function now returns one of these new pcre_extra blocks, with
4361 the actual study data pointed to by the study_data field, and the
4362 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4363 before. That is, this change is entirely upwards-compatible and requires no
4364 change to existing code.
4365
4366 If you want to pass in additional data to pcre_exec(), you can either place it
4367 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4368 block.
4369
4370 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4371 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4372 times with different match limits, until it finds the minimum value needed for
4373 pcre_exec() to complete. The value is then output. This can be instructive; for
4374 most simple matches the number is quite small, but for pathological cases it
4375 gets very large very quickly.
4376
4377 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4378 returns the size of the data block pointed to by the study_data field in a
4379 pcre_extra block, that is, the value that was passed as the argument to
4380 pcre_malloc() when PCRE was getting memory in which to place the information
4381 created by pcre_study(). The fourth argument should point to a size_t variable.
4382 pcretest has been extended so that this information is shown after a successful
4383 pcre_study() call when information about the compiled regex is being displayed.
4384
4385 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4386 because what follows is always an absolute path. (Later: it turns out that this
4387 is more than cosmetic for MinGW, because it doesn't like empty path
4388 components.)
4389
4390 49. Some changes have been made to the callout feature (see 28 above):
4391
4392 (i) A callout function now has three choices for what it returns:
4393
4394 0 => success, carry on matching
4395 > 0 => failure at this point, but backtrack if possible
4396 < 0 => serious error, return this value from pcre_exec()
4397
4398 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4399 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4400 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4401 use by callout functions. It will never be used by PCRE itself.
4402
4403 (ii) The pcre_extra structure (see 45 above) has a void * field called
4404 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4405 pcre_callout_block structure has a field of the same name. The contents of
4406 the field passed in the pcre_extra structure are passed to the callout
4407 function in the corresponding field in the callout block. This makes it
4408 easier to use the same callout-containing regex from multiple threads. For
4409 testing, the pcretest program has a new data escape
4410
4411 \C*n pass the number n (may be negative) as callout_data
4412
4413 If the callout function in pcretest receives a non-zero value as
4414 callout_data, it returns that value.
4415
4416 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4417 there were some redundant $(CFLAGS) in commands that are now specified as
4418 $(LINK), which already includes $(CFLAGS).
4419
4420 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4421 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4422 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4423 one-byte characters throughout. Note that case-insensitive matching applies
4424 only to characters whose values are less than 256. PCRE doesn't support the
4425 notion of cases for higher-valued characters.
4426
4427 (i) A character class whose characters are all within 0-255 is handled as
4428 a bit map, and the map is inverted for negative classes. Previously, a
4429 character > 255 always failed to match such a class; however it should
4430 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4431
4432 (ii) A negated character class with a single character < 255 is coded as
4433 "not this character" (OP_NOT). This wasn't working properly when the test
4434 character was multibyte, either singly or repeated.
4435
4436 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4437 mode, for example: \x{100}{2,3}.
4438
4439 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4440 singly or repeated) now correctly test multibyte characters. However,
4441 PCRE doesn't recognize any characters with values greater than 255 as
4442 digits, spaces, or word characters. Such characters always match \D, \S,
4443 and \W, and never match \d, \s, or \w.
4444
4445 (v) Classes may now contain characters and character ranges with values
4446 greater than 255. For example: [ab\x{100}-\x{400}].
4447
4448 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4449 PCRE in UTF-8 mode.
4450
4451 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4452 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4453 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4454 value.)
4455
4456 53. The single man page has become too large. I have therefore split it up into
4457 a number of separate man pages. These also give rise to individual HTML pages;
4458 these are now put in a separate directory, and there is an index.html page that
4459 lists them all. Some hyperlinking between the pages has been installed.
4460
4461 54. Added convenience functions for handling named capturing parentheses.
4462
4463 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4464 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4465 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4466 are faulted.
4467
4468 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4469 calling configure. These values are used when compiling the dftables.c program
4470 which is run to generate the source of the default character tables. They
4471 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4472 you will need to set these values.
4473
4474 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4475
4476
4477 Version 3.9 02-Jan-02
4478 ---------------------
4479
4480 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4481
4482 2. If --disable-static was given, the building process failed when trying to
4483 build pcretest and pcregrep. (For some reason it was using libtool to compile
4484 them, which is not right, as they aren't part of the library.)
4485
4486
4487 Version 3.8 18-Dec-01
4488 ---------------------
4489
4490 1. The experimental UTF-8 code was completely screwed up. It was packing the
4491 bytes in the wrong order. How dumb can you get?
4492
4493
4494 Version 3.7 29-Oct-01
4495 ---------------------
4496
4497 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4498 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4499 this didn't happen under Solaris 8, where I normally test things.
4500
4501 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4502 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4503 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4504 this fix an hour or so after the initial 3.7 release.)
4505
4506
4507 Version 3.6 23-Oct-01
4508 ---------------------
4509
4510 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4511 offsets passed as NULL with zero offset count.
4512
4513 2. The config.guess and config.sub files had not been updated when I moved to
4514 the latest autoconf.
4515
4516
4517 Version 3.5 15-Aug-01
4518 ---------------------
4519
4520 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4521 had been forgotten.
4522
4523 2. By using declared but undefined structures, we can avoid using "void"
4524 definitions in pcre.h while keeping the internal definitions of the structures
4525 private.
4526
4527 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4528 user point of view, this means that both static and shared libraries are built
4529 by default, but this can be individually controlled. More of the work of
4530 handling this static/shared cases is now inside libtool instead of PCRE's make
4531 file.
4532
4533 4. The pcretest utility is now installed along with pcregrep because it is
4534 useful for users (to test regexs) and by doing this, it automatically gets
4535 relinked by libtool. The documentation has been turned into a man page, so
4536 there are now .1, .txt, and .html versions in /doc.
4537
4538 5. Upgrades to pcregrep:
4539 (i) Added long-form option names like gnu grep.
4540 (ii) Added --help to list all options with an explanatory phrase.
4541 (iii) Added -r, --recursive to recurse into sub-directories.
4542 (iv) Added -f, --file to read patterns from a file.
4543
4544 6. pcre_exec() was referring to its "code" argument before testing that
4545 argument for NULL (and giving an error if it was NULL).
4546
4547 7. Upgraded Makefile.in to allow for compiling in a different directory from
4548 the source directory.
4549
4550 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4551 options bits, the pointer it was passed was to an int instead of to an unsigned
4552 long int. This mattered only on 64-bit systems.
4553
4554 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4555 generated) instead of pcre.in, which it its source. Also made the same change
4556 in several of the .c files.
4557
4558 10. A new release of gcc defines printf() as a macro, which broke pcretest
4559 because it had an ifdef in the middle of a string argument for printf(). Fixed
4560 by using separate calls to printf().
4561
4562 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4563 script, to force use of CR or LF instead of \n in the source. On non-Unix
4564 systems, the value can be set in config.h.
4565
4566 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4567 absolute limit. Changed the text of the error message to make this clear, and
4568 likewise updated the man page.
4569
4570 13. The limit of 99 on the number of capturing subpatterns has been removed.
4571 The new limit is 65535, which I hope will not be a "real" limit.
4572
4573
4574 Version 3.4 22-Aug-00
4575 ---------------------
4576
4577 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4578
4579 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4580
4581
4582 Version 3.3 01-Aug-00
4583 ---------------------
4584
4585 1. If an octal character was given, but the value was greater than \377, it
4586 was not getting masked to the least significant bits, as documented. This could
4587 lead to crashes in some systems.
4588
4589 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4590 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4591
4592 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4593 These just pass their arguments on to (pcre_free)(), but they are provided
4594 because some uses of PCRE bind it to non-C systems that can call its functions,
4595 but cannot call free() or pcre_free() directly.
4596
4597 4. Add "make test" as a synonym for "make check". Corrected some comments in
4598 the Makefile.
4599
4600 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4601 Makefile.
4602
4603 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4604 command called pgrep for grepping around the active processes.
4605
4606 7. Added the beginnings of support for UTF-8 character strings.
4607
4608 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4609 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4610 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4611 out for the ar command.)
4612
4613
4614 Version 3.2 12-May-00
4615 ---------------------
4616
4617 This is purely a bug fixing release.
4618
4619 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4620 of ZA. This was just one example of several cases that could provoke this bug,
4621 which was introduced by change 9 of version 2.00. The code for breaking
4622 infinite loops after an iteration that matches an empty string was't working
4623 correctly.
4624
4625 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4626 when matched against abbab (for example). After matching an empty string, it
4627 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4628 caused it to match further down the string than it should.
4629
4630 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4631 was there because it doesn't seem to be needed, and it causes trouble on some
4632 systems, as it is not a Standard C header. It has been removed.
4633
4634 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4635 were reported on the Macintosh. The changes were from
4636
4637 while ((c = *(++ptr)) != 0 && c != '\n');
4638 to
4639 while ((c = *(++ptr)) != 0 && c != '\n') ;
4640
4641 Totally extraordinary, but if that's what it takes...
4642
4643 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4644 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4645 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4646 assumes the way PCRE uses memmove() (always moving upwards).
4647
4648 6. PCRE is being used in one environment where strchr() is not available. There
4649 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4650 faster code anyway.
4651
4652
4653 Version 3.1 09-Feb-00
4654 ---------------------
4655
4656 The only change in this release is the fixing of some bugs in Makefile.in for
4657 the "install" target:
4658
4659 (1) It was failing to install pcreposix.h.
4660
4661 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4662
4663
4664 Version 3.0 01-Feb-00
4665 ---------------------
4666
4667 1. Add support for the /+ modifier to perltest (to output $` like it does in
4668 pcretest).
4669
4670 2. Add support for the /g modifier to perltest.
4671
4672 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4673 matches null strings.
4674
4675 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4676 pattern. Perl treats empty patterns specially - it reuses the most recent
4677 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4678 effect.
4679
4680 5. The POSIX interface was broken in that it was just handing over the POSIX
4681 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4682 required a bigger vector, with some working space on the end. This means that
4683 the POSIX wrapper now has to get and free some memory, and copy the results.
4684
4685 6. Added some simple autoconf support, placing the test data and the
4686 documentation in separate directories, re-organizing some of the
4687 information files, and making it build pcre-config (a GNU standard). Also added
4688 libtool support for building PCRE as a shared library, which is now the
4689 default.
4690
4691 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4692 09 are not valid octal constants. Single digits will be used for minor values
4693 less than 10.
4694
4695 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4696 existing programs that set these in the POSIX interface can use PCRE without
4697 modification.
4698
4699 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4700 return all that pcre_info() returns, plus additional data. The pcre_info()
4701 function is retained for compatibility, but is considered to be obsolete.
4702
4703 10. Added experimental recursion feature (?R) to handle one common case that
4704 Perl 5.6 will be able to do with (?p{...}).
4705
4706 11. Added support for POSIX character classes like [:alpha:], which Perl is
4707 adopting.
4708
4709
4710 Version 2.08 31-Aug-99
4711 ----------------------
4712
4713 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4714 trying to match at the startoffset position, but instead was moving forward to
4715 the next newline as if a previous match had failed.
4716
4717 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4718 and could get into a loop if a null string was matched other than at the start
4719 of the subject.
4720
4721 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4722 be distinguished at compile time, and for completeness also added PCRE_DATE.
4723
4724 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4725 in GnuWin32 environments.
4726
4727
4728 Version 2.07 29-Jul-99
4729 ----------------------
4730
4731 1. The documentation is now supplied in plain text form and HTML as well as in
4732 the form of man page sources.
4733
4734 2. C++ compilers don't like assigning (void *) values to other pointer types.
4735 In particular this affects malloc(). Although there is no problem in Standard
4736 C, I've put in casts to keep C++ compilers happy.
4737
4738 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4739 should be (const char *).
4740
4741 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4742 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4743 However, I haven't made this a standard facility. The documentation doesn't
4744 mention it, and the Makefile doesn't support it.
4745
4746 5. The Makefile now contains an "install" target, with editable destinations at
4747 the top of the file. The pcretest program is not installed.
4748
4749 6. pgrep -V now gives the PCRE version number and date.
4750
4751 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4752 causing the entire string to be ignored, instead of just the last character.
4753
4754 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4755 non-matching string, it can take a very, very long time, even for strings of
4756 quite modest length, because of the nested recursion. PCRE now does better in
4757 some of these cases. It does this by remembering the last required literal
4758 character in the pattern, and pre-searching the subject to ensure it is present
4759 before running the real match. In other words, it applies a heuristic to detect
4760 some types of certain failure quickly, and in the above example, if presented
4761 with a string that has no trailing " it gives "no match" very quickly.
4762
4763 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4764 other alternatives are tried instead.
4765
4766
4767 Version 2.06 09-Jun-99
4768 ----------------------
4769
4770 1. Change pcretest's output for amount of store used to show just the code
4771 space, because the remainder (the data block) varies in size between 32-bit and
4772 64-bit systems.
4773
4774 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4775 start matching at. This allows lookbehinds to work when searching for multiple
4776 occurrences in a string.
4777
4778 3. Added additional options to pcretest for testing multiple occurrences:
4779
4780 /+ outputs the rest of the string that follows a match
4781 /g loops for multiple occurrences, using the new startoffset argument
4782 /G loops for multiple occurrences by passing an incremented pointer
4783
4784 4. PCRE wasn't doing the "first character" optimization for patterns starting
4785 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4786 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4787 the letter 'x'. On long subject strings, this gives a significant speed-up.
4788
4789
4790 Version 2.05 21-Apr-99
4791 ----------------------
4792
4793 1. Changed the type of magic_number from int to long int so that it works
4794 properly on 16-bit systems.
4795
4796 2. Fixed a bug which caused patterns starting with .* not to work correctly
4797 when the subject string contained newline characters. PCRE was assuming
4798 anchoring for such patterns in all cases, which is not correct because .* will
4799 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4800 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4801 must be retried after every newline in the subject.
4802
4803
4804 Version 2.04 18-Feb-99
4805 ----------------------
4806
4807 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4808 computation of the store needed to hold the pattern was incorrect (too large).
4809 If such patterns were nested a few deep, this could multiply and become a real
4810 problem.
4811
4812 2. Added /M option to pcretest to show the memory requirement of a specific
4813 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4814
4815 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4816 compiled in such a way that the backtracking after subsequent failure was
4817 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4818 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4819
4820
4821 Version 2.03 02-Feb-99
4822 ----------------------
4823
4824 1. Fixed typo and small mistake in man page.
4825
4826 2. Added 4th condition (GPL supersedes if conflict) and created separate
4827 LICENCE file containing the conditions.
4828
4829 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4830 Perl, that is the internal \ allows the delimiter to be included in the
4831 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4832 the final delimiter, add \ to the end of the pattern (to test the error).
4833
4834 4. Added the convenience functions for extracting substrings after a successful
4835 match. Updated pcretest to make it able to test these functions.
4836
4837
4838 Version 2.02 14-Jan-99
4839 ----------------------
4840
4841 1. Initialized the working variables associated with each extraction so that
4842 their saving and restoring doesn't refer to uninitialized store.
4843
4844 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4845 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4846 fix the problem.
4847
4848 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4849 calls, and wasn't printing the correct value for compiling calls. Increased the
4850 default value of LOOPREPEAT, and the number of significant figures in the
4851 times.
4852
4853 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4854
4855 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4856 a building problem on Windows NT with a FAT file system.
4857
4858
4859 Version 2.01 21-Oct-98
4860 ----------------------
4861
4862 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4863 to character tables built by pcre_maketables() in the current locale. If NULL
4864 is passed, the default tables are used.
4865
4866
4867 Version 2.00 24-Sep-98
4868 ----------------------
4869
4870 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4871 it any more.
4872
4873 2. Allow quantification of (?>) groups, and make it work correctly.
4874
4875 3. The first character computation wasn't working for (?>) groups.
4876
4877 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4878 end of the subject) and add 5.005's \z, which really does match only at the
4879 very end of the subject.
4880
4881 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4882
4883 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4884 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4885 localized options. All options to pcre_study() were also removed.
4886
4887 7. Add other new features from 5.005:
4888
4889 $(?<= positive lookbehind
4890 $(?<! negative lookbehind
4891 (?imsx-imsx) added the unsetting capability
4892 such a setting is global if at outer level; local otherwise
4893 (?imsx-imsx:) non-capturing groups with option setting
4894 (?(cond)re|re) conditional pattern matching
4895
4896 A backreference to itself in a repeated group matches the previous
4897 captured string.
4898
4899 8. General tidying up of studying (both automatic and via "study")
4900 consequential on the addition of new assertions.
4901
4902 9. As in 5.005, unlimited repeated groups that could match an empty substring
4903 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4904 runtime if any iteration does actually match an empty substring.
4905
4906 10. Include the RunTest script in the distribution.
4907
4908 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4909 discrepancies, some of which were old and were also with respect to 5.004. They
4910 have now been fixed.
4911
4912
4913 Version 1.09 28-Apr-98
4914 ----------------------
4915
4916 1. A negated single character class followed by a quantifier with a minimum
4917 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4918 program crashes, or just wrong answers. This did not apply to negated classes
4919 containing more than one character, or to minima other than one.
4920
4921
4922 Version 1.08 27-Mar-98
4923 ----------------------
4924
4925 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4926
4927 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4928 latter must appear before anything that relies on it in the pattern.
4929
4930
4931 Version 1.07 16-Feb-98
4932 ----------------------
4933
4934 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4935 repeat of a potentially empty string).
4936
4937
4938 Version 1.06 23-Jan-98
4939 ----------------------
4940
4941 1. Added Markus Oberhumer's little patches for C++.
4942
4943 2. Literal strings longer than 255 characters were broken.
4944
4945
4946 Version 1.05 23-Dec-97
4947 ----------------------
4948
4949 1. Negated character classes containing more than one character were failing if
4950 PCRE_CASELESS was set at run time.
4951
4952
4953 Version 1.04 19-Dec-97
4954 ----------------------
4955
4956 1. Corrected the man page, where some "const" qualifiers had been omitted.
4957
4958 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4959 input syntax.
4960
4961 3. Fixed memory leak which occurred when a regex with back references was
4962 matched with an offsets vector that wasn't big enough. The temporary memory
4963 that is used in this case wasn't being freed if the match failed.
4964
4965 4. Tidied pcretest to ensure it frees memory that it gets.
4966
4967 5. Temporary memory was being obtained in the case where the passed offsets
4968 vector was exactly big enough.
4969
4970 6. Corrected definition of offsetof() from change 5 below.
4971
4972 7. I had screwed up change 6 below and broken the rules for the use of
4973 setjmp(). Now fixed.
4974
4975
4976 Version 1.03 18-Dec-97
4977 ----------------------
4978
4979 1. A erroneous regex with a missing opening parenthesis was correctly
4980 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4981 on some systems.
4982
4983 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4984 it was reported that one broken compiler failed on the former because "code" is
4985 also an independent variable.
4986
4987 3. The erroneous regex a[]b caused an array overrun reference.
4988
4989 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4990 fail on data ending with that character. (It was going on too far, and checking
4991 the next character, typically a binary zero.) This was specific to the
4992 optimized code for single-character negative classes.
4993
4994 5. Added a contributed patch from the TIN world which does the following:
4995
4996 + Add an undef for memmove, in case the the system defines a macro for it.
4997
4998 + Add a definition of offsetof(), in case there isn't one. (I don't know
4999 the reason behind this - offsetof() is part of the ANSI standard - but
5000 it does no harm).
5001
5002 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5003 most of the places where whitespace preceded '#'. I have given up and
5004 allowed the remaining 2 cases to be at the margin.
5005
5006 + Rename some variables in pcre to eliminate shadowing. This seems very
5007 pedantic, but does no harm, of course.
5008
5009 6. Moved the call to setjmp() into its own function, to get rid of warnings
5010 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5011
5012 7. Constructs such as \d{8,} were compiling into the equivalent of
5013 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5014 outcome, but in this particular case used more store than had been allocated,
5015 which caused the bug to be discovered because it threw up an internal error.
5016
5017 8. The debugging code in both pcre and pcretest for outputting the compiled
5018 form of a regex was going wrong in the case of back references followed by
5019 curly-bracketed repeats.
5020
5021
5022 Version 1.02 12-Dec-97
5023 ----------------------
5024
5025 1. Typos in pcre.3 and comments in the source fixed.
5026
5027 2. Applied a contributed patch to get rid of places where it used to remove
5028 'const' from variables, and fixed some signed/unsigned and uninitialized
5029 variable warnings.
5030
5031 3. Added the "runtest" target to Makefile.
5032
5033 4. Set default compiler flag to -O2 rather than just -O.
5034
5035
5036 Version 1.01 19-Nov-97
5037 ----------------------
5038
5039 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5040 like /([ab]*)*/, that is, for classes with more than one character in them.
5041
5042 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5043 as /((?>a*))*/ (a PCRE_EXTRA facility).
5044
5045
5046 Version 1.00 18-Nov-97
5047 ----------------------
5048
5049 1. Added compile-time macros to support systems such as SunOS4 which don't have
5050 memmove() or strerror() but have other things that can be used instead.
5051
5052 2. Arranged that "make clean" removes the executables.
5053
5054
5055 Version 0.99 27-Oct-97
5056 ----------------------
5057
5058 1. Fixed bug in code for optimizing classes with only one character. It was
5059 initializing a 32-byte map regardless, which could cause it to run off the end
5060 of the memory it had got.
5061
5062 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5063
5064
5065 Version 0.98 22-Oct-97
5066 ----------------------
5067
5068 1. Fixed bug in code for handling temporary memory usage when there are more
5069 back references than supplied space in the ovector. This could cause segfaults.
5070
5071
5072 Version 0.97 21-Oct-97
5073 ----------------------
5074
5075 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5076
5077 2. Optimized negated single characters not to use a bit map.
5078
5079 3. Brought error texts together as macro definitions; clarified some of them;
5080 fixed one that was wrong - it said "range out of order" when it meant "invalid
5081 escape sequence".
5082
5083 4. Changed some char * arguments to const char *.
5084
5085 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5086
5087 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5088 pcretest.
5089
5090
5091 Version 0.96 16-Oct-97
5092 ----------------------
5093
5094 1. Added a simple "pgrep" utility to the distribution.
5095
5096 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5097 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5098 where "ddd" means "one or more decimal digits".
5099
5100 3. Fixed serious bug. If a pattern had a back reference, but the call to
5101 pcre_exec() didn't supply a large enough ovector to record the related
5102 identifying subpattern, the match always failed. PCRE now remembers the number
5103 of the largest back reference, and gets some temporary memory in which to save
5104 the offsets during matching if necessary, in order to ensure that
5105 backreferences always work.
5106
5107 4. Increased the compatibility with Perl in a number of ways:
5108
5109 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5110 to request this handling. The option can be set at compile or exec time.
5111
5112 (b) $ matches before a terminating newline by default; an option
5113 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5114 mode). The option can be set at compile or exec time.
5115
5116 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5117 the same as Perl's. If the decimal number it represents is less than 10
5118 or there aren't that many previous left capturing parentheses, an octal
5119 escape is read. Inside a character class, it's always an octal escape,
5120 even if it is a single digit.
5121
5122 (d) An escaped but undefined alphabetic character is taken as a literal,
5123 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5124 escapes.
5125
5126 (e) {0} is now permitted. (The previous item is removed from the compiled
5127 pattern).
5128
5129 5. Changed all the names of code files so that the basic parts are no longer
5130 than 10 characters, and abolished the teeny "globals.c" file.
5131
5132 6. Changed the handling of character classes; they are now done with a 32-byte
5133 bit map always.
5134
5135 7. Added the -d and /D options to pcretest to make it possible to look at the
5136 internals of compilation without having to recompile pcre.
5137
5138
5139 Version 0.95 23-Sep-97
5140 ----------------------
5141
5142 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5143 \x20 at the start of a run of normal characters. These were being treated as
5144 real characters, instead of the source characters being re-checked.
5145
5146
5147 Version 0.94 18-Sep-97
5148 ----------------------
5149
5150 1. The functions are now thread-safe, with the caveat that the global variables
5151 containing pointers to malloc() and free() or alternative functions are the
5152 same for all threads.
5153
5154 2. Get pcre_study() to generate a bitmap of initial characters for non-
5155 anchored patterns when this is possible, and use it if passed to pcre_exec().
5156
5157
5158 Version 0.93 15-Sep-97
5159 ----------------------
5160
5161 1. /(b)|(:+)/ was computing an incorrect first character.
5162
5163 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5164 but not actually doing anything yet.
5165
5166 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5167 as Perl does (e.g. [-az] or [az-]).
5168
5169 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5170 all possible positions.
5171
5172 5. Split up into different modules to avoid including unneeded functions in a
5173 compiled binary. However, compile and exec are still in one module. The "study"
5174 function is split off.
5175
5176 6. The character tables are now in a separate module whose source is generated
5177 by an auxiliary program - but can then be edited by hand if required. There are
5178 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5179 toupper() in the code.
5180
5181 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5182 make them global. Abolish the function for setting them, as the caller can now
5183 set them directly.
5184
5185
5186 Version 0.92 11-Sep-97
5187 ----------------------
5188
5189 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5190 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5191
5192 2. Caseless matching was not working in character classes if the characters in
5193 the pattern were in upper case.
5194
5195 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5196
5197 4. Make PCRE_ANCHORED public and accept as a compile option.
5198
5199 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5200 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5201 pass them.
5202
5203 6. Give an error if bad option bits passed at compile or run time.
5204
5205 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5206 pcretest to cause it to pass that flag.
5207
5208 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5209 options, and the first character, if set.
5210
5211 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5212
5213
5214 Version 0.91 10-Sep-97
5215 ----------------------
5216
5217 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5218 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5219
5220 2. PCRE was looping on encountering an indefinitely repeated back reference to
5221 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5222 Perl does - treats the match as successful.
5223
5224 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5