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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1261 - (show annotations)
Wed Feb 27 16:27:01 2013 UTC (6 years, 5 months ago) by ph10
File size: 243685 byte(s)
Correct Unicode string checking in the light of corrigendum #9.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80 be expressed in certain programming languages using Universal Character
81 Names.
82
83 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84 which points out that "non characters" are not "characters that may not
85 appear in Unicode strings" but rather "characters that are reserved for
86 internal use and have only local meaning".
87
88
89 Version 8.32 30-November-2012
90 -----------------------------
91
92 1. Improved JIT compiler optimizations for first character search and single
93 character iterators.
94
95 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
96 Patch by Daniel Richard G.
97
98 3. Single character iterator optimizations in the JIT compiler.
99
100 4. Improved JIT compiler optimizations for character ranges.
101
102 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
103 Reported by Giuseppe D'Angelo.
104
105 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
106 of a line, was being set incorrectly in cases where .* appeared inside
107 atomic brackets at the start of a pattern, or where there was a subsequent
108 *PRUNE or *SKIP.
109
110 7. Improved instruction cache flush for POWER/PowerPC.
111 Patch by Daniel Richard G.
112
113 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
114 grep:
115
116 (a) There is now no limit to the number of patterns to be matched.
117
118 (b) An error is given if a pattern is too long.
119
120 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
121 are now supported.
122
123 (d) --exclude-from and --include-from (multiple use) have been added.
124
125 (e) Exclusions and inclusions now apply to all files and directories, not
126 just to those obtained from scanning a directory recursively.
127
128 (f) Multiple uses of -f and --file-list are now supported.
129
130 (g) In a Windows environment, the default for -d has been changed from
131 "read" (the GNU grep default) to "skip", because otherwise the presence
132 of a directory in the file list provokes an error.
133
134 (h) The documentation has been revised and clarified in places.
135
136 9. Improve the matching speed of capturing brackets.
137
138 10. Changed the meaning of \X so that it now matches a Unicode extended
139 grapheme cluster.
140
141 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
142 out POSIX threads when JIT support is configured.
143
144 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
145
146 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
147 structure could go wrong in environments where size_t is not the same size
148 as int.
149
150 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
151 set.
152
153 15. The EBCDIC support had decayed; later updates to the code had included
154 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
155 general tidy up of EBCDIC-related issues, and the documentation was also
156 not quite right. There is now a test that can be run on ASCII systems to
157 check some of the EBCDIC-related things (but is it not a full test).
158
159 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
160 in a small tidy to the code.
161
162 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
163
164 18. If the --only-matching (-o) option in pcregrep is specified multiple
165 times, each one causes appropriate output. For example, -o1 -o2 outputs the
166 substrings matched by the 1st and 2nd capturing parentheses. A separating
167 string can be specified by --om-separator (default empty).
168
169 19. Improving the first n character searches.
170
171 20. Turn case lists for horizontal and vertical white space into macros so that
172 they are defined only once.
173
174 21. This set of changes together give more compatible Unicode case-folding
175 behaviour for characters that have more than one other case when UCP
176 support is available.
177
178 (a) The Unicode property table now has offsets into a new table of sets of
179 three or more characters that are case-equivalent. The MultiStage2.py
180 script that generates these tables (the pcre_ucd.c file) now scans
181 CaseFolding.txt instead of UnicodeData.txt for character case
182 information.
183
184 (b) The code for adding characters or ranges of characters to a character
185 class has been abstracted into a generalized function that also handles
186 case-independence. In UTF-mode with UCP support, this uses the new data
187 to handle characters with more than one other case.
188
189 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
190 whose other case is greater than 256 are now correctly matched
191 caselessly. Previously, the high codepoint matched the low one, but not
192 vice versa.
193
194 (d) The processing of \h, \H, \v, and \ in character classes now makes use
195 of the new class addition function, using character lists defined as
196 macros alongside the case definitions of 20 above.
197
198 (e) Caseless back references now work with characters that have more than
199 one other case.
200
201 (f) General caseless matching of characters with more than one other case
202 is supported.
203
204 22. Unicode character properties were updated from Unicode 6.2.0
205
206 23. Improved CMake support under Windows. Patch by Daniel Richard G.
207
208 24. Add support for 32-bit character strings, and UTF-32
209
210 25. Major JIT compiler update (code refactoring and bugfixing).
211 Experimental Sparc 32 support is added.
212
213 26. Applied a modified version of Daniel Richard G's patch to create
214 pcre.h.generic and config.h.generic by "make" instead of in the
215 PrepareRelease script.
216
217 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
218 pcre_compile.c when checking for a zero character.
219
220 28. Introducing a native interface for JIT. Through this interface, the compiled
221 machine code can be directly executed. The purpose of this interface is to
222 provide fast pattern matching, so several sanity checks are not performed.
223 However, feature tests are still performed. The new interface provides
224 1.4x speedup compared to the old one.
225
226 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
227 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
228 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
229
230 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
231 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
232 the "old" RFC 2279). Instead, it ended up passing a negative length to
233 pcre_exec().
234
235 31. Add support for GCC's visibility feature to hide internal functions.
236
237 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
238 "unknown -C option" after outputting 0 or 1.
239
240 33. There is now support for generating a code coverage report for the test
241 suite in environments where gcc is the compiler and lcov is installed. This
242 is mainly for the benefit of the developers.
243
244 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
245 unaddressable using valgrind annotations, allowing valgrind to detect
246 invalid memory accesses. This is mainly for the benefit of the developers.
247
248 25. (*UTF) can now be used to start a pattern in any of the three libraries.
249
250 26. Give configure error if --enable-cpp but no C++ compiler found.
251
252
253 Version 8.31 06-July-2012
254 -------------------------
255
256 1. Fixing a wrong JIT test case and some compiler warnings.
257
258 2. Removed a bashism from the RunTest script.
259
260 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
261 to unsigned type, result still unsigned" that was given by an MS compiler
262 on encountering the code "-sizeof(xxx)".
263
264 4. Partial matching support is added to the JIT compiler.
265
266 5. Fixed several bugs concerned with partial matching of items that consist
267 of more than one character:
268
269 (a) /^(..)\1/ did not partially match "aba" because checking references was
270 done on an "all or nothing" basis. This also applied to repeated
271 references.
272
273 (b) \R did not give a hard partial match if \r was found at the end of the
274 subject.
275
276 (c) \X did not give a hard partial match after matching one or more
277 characters at the end of the subject.
278
279 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
280 a partial match for the string "\r".
281
282 (e) When newline was set to CRLF, the metacharacter "." did not recognize
283 a partial match for a CR character at the end of the subject string.
284
285 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
286 running pcretest, the text "(JIT)" added to the output whenever JIT is
287 actually used to run the match.
288
289 7. Individual JIT compile options can be set in pcretest by following -s+[+]
290 or /S+[+] with a digit between 1 and 7.
291
292 8. OP_NOT now supports any UTF character not just single-byte ones.
293
294 9. (*MARK) control verb is now supported by the JIT compiler.
295
296 10. The command "./RunTest list" lists the available tests without actually
297 running any of them. (Because I keep forgetting what they all are.)
298
299 11. Add PCRE_INFO_MAXLOOKBEHIND.
300
301 12. Applied a (slightly modified) user-supplied patch that improves performance
302 when the heap is used for recursion (compiled with --disable-stack-for-
303 recursion). Instead of malloc and free for each heap frame each time a
304 logical recursion happens, frames are retained on a chain and re-used where
305 possible. This sometimes gives as much as 30% improvement.
306
307 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
308 call.
309
310 14. As documented, (*COMMIT) is now confined to within a positive assertion.
311
312 15. It is now possible to link pcretest with libedit as an alternative to
313 libreadline.
314
315 16. (*COMMIT) control verb is now supported by the JIT compiler.
316
317 17. The Unicode data tables have been updated to Unicode 6.1.0.
318
319 18. Added --file-list option to pcregrep.
320
321 19. Added binary file support to pcregrep, including the -a, --binary-files,
322 -I, and --text options.
323
324 20. The madvise function is renamed for posix_madvise for QNX compatibility
325 reasons. Fixed by Giuseppe D'Angelo.
326
327 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
328
329 22. Changed the option for creating symbolic links for 16-bit man pages from
330 -s to -sf so that re-installing does not cause issues.
331
332 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
333
334 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
335 matches in certain environments (the workspace was not being correctly
336 retained). Also added to pcre_dfa_exec() a simple plausibility check on
337 some of the workspace data at the beginning of a restart.
338
339 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
340 was not doing so when it should - probably a typo introduced by SVN 528
341 (change 8.10/14).
342
343 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
344 \w+ when the character tables indicated that \x{c4} was a word character.
345 There were several related cases, all because the tests for doing a table
346 lookup were testing for characters less than 127 instead of 255.
347
348 27. If a pattern contains capturing parentheses that are not used in a match,
349 their slots in the ovector are set to -1. For those that are higher than
350 any matched groups, this happens at the end of processing. In the case when
351 there were back references that the ovector was too small to contain
352 (causing temporary malloc'd memory to be used during matching), and the
353 highest capturing number was not used, memory off the end of the ovector
354 was incorrectly being set to -1. (It was using the size of the temporary
355 memory instead of the true size.)
356
357 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
358 ovector size, it uses memory at the end of the block that it has got.
359
360 29. Check for an overlong MARK name and give an error at compile time. The
361 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
362
363 30. JIT compiler update.
364
365 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
366 Rill for the patch.
367
368 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
369
370 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
371
372 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
373 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
374 was enabled.
375
376 35. Improve JIT code generation for greedy plus quantifier.
377
378 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
379 1 to "aa" instead of to an empty string. The bug affected repeated groups
380 that could potentially match an empty string.
381
382 37. Optimizing single character iterators in JIT.
383
384 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
385 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
386 codepoints that are too big for the mode are faulted, and in a UTF mode,
387 disallowed codepoints are also faulted.
388
389 39. If PCRE was compiled with UTF support, in three places in the DFA
390 matcher there was code that should only have been obeyed in UTF mode, but
391 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
392 processing when bytes with values greater than 127 were present. In 16-bit
393 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
394 both cases the values are those that cannot be the first data item in a UTF
395 character. The three items that might have provoked this were recursions,
396 possessively repeated groups, and atomic groups.
397
398 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
399 and pcregrep, because some OS require shared objects to be explicitly
400 passed to ld, causing the link step to fail if they are not.
401
402 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
403 mode, patterns that started with \h* or \R* might be incorrectly matched.
404
405
406 Version 8.30 04-February-2012
407 -----------------------------
408
409 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
410 name is defined in ctype.h.
411
412 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
413 only in quite long subpatterns.
414
415 3. Removed the function pcre_info(), which has been obsolete and deprecated
416 since it was replaced by pcre_fullinfo() in February 2000.
417
418 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
419 match a (*MARK), and the match failed at the start of the subject, a
420 reference to memory before the start of the subject could occur. This bug
421 was introduced by fix 17 of release 8.21.
422
423 5. A reference to an unset group with zero minimum repetition was giving
424 totally wrong answers (in non-JavaScript-compatibility mode). For example,
425 /(another)?(\1?)test/ matched against "hello world test". This bug was
426 introduced in release 8.13.
427
428 6. Add support for 16-bit character strings (a large amount of work involving
429 many changes and refactorings).
430
431 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
432 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
433 from a file.
434
435 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
436 rounding is not applied in this particular case).
437
438 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
439 if they appear, or are escaped, in patterns.
440
441 10. Get rid of a number of -Wunused-but-set-variable warnings.
442
443 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
444 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
445 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
446 also returns the mark "x". This bug applied to capturing parentheses,
447 non-capturing parentheses, and atomic parentheses. It also applied to some
448 assertions.
449
450 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
451 information out of configure.ac instead of relying on pcre.h.generic, which
452 is not stored in the repository.
453
454 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
455 -lreadline.
456
457 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
458
459 15. Applied Graycode's patch to put the top-level frame on the stack rather
460 than the heap when not using the stack for recursion. This gives a
461 performance improvement in many cases when recursion is not deep.
462
463 16. Experimental code added to "pcretest -C" to output the stack frame size.
464
465
466 Version 8.21 12-Dec-2011
467 ------------------------
468
469 1. Updating the JIT compiler.
470
471 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
472 are added as well.
473
474 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
475 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
476 calling _pcre_jit_exec. Some extra comments are added.
477
478 4. (*MARK) settings inside atomic groups that do not contain any capturing
479 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
480 was introduced by change 18 for 8.20.
481
482 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
483 ECMA-262 standard.
484
485 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
486 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
487 This bug was probably introduced by change 9 of 8.13.
488
489 7. While fixing 6 above, I noticed that a number of other items were being
490 incorrectly rejected as "not fixed length". This arose partly because newer
491 opcodes had not been added to the fixed-length checking code. I have (a)
492 corrected the bug and added tests for these items, and (b) arranged for an
493 error to occur if an unknown opcode is encountered while checking for fixed
494 length instead of just assuming "not fixed length". The items that were
495 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
496 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
497 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
498
499 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
500 being incorrectly compiled and would have given unpredicatble results.
501
502 9. A possessively repeated subpattern with minimum repeat count greater than
503 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
504 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
505 the first (A) could occur when it should not.
506
507 10. Add a cast and remove a redundant test from the code.
508
509 11. JIT should use pcre_malloc/pcre_free for allocation.
510
511 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
512 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
513 is anything other than /usr, -L is still shown).
514
515 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
516
517 14. Perl does not support \N without a following name in a [] class; PCRE now
518 also gives an error.
519
520 15. If a forward reference was repeated with an upper limit of around 2000,
521 it caused the error "internal error: overran compiling workspace". The
522 maximum number of forward references (including repeats) was limited by the
523 internal workspace, and dependent on the LINK_SIZE. The code has been
524 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
525 the default depends on LINK_SIZE. There is a new upper limit (for safety)
526 of around 200,000 forward references. While doing this, I also speeded up
527 the filling in of repeated forward references.
528
529 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
530 incorrectly expecting the subject to contain another "a" after the start.
531
532 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
533 in the match, the SKIP should be ignored. This was not happening; instead
534 the SKIP was being treated as NOMATCH. For patterns such as
535 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
536 tested.
537
538 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
539 now much more compatible with Perl, in particular in cases where the result
540 is a non-match for a non-anchored pattern. For example, if
541 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
542 "m", where previously it did not return a name. A side effect of this
543 change is that for partial matches, the last encountered mark name is
544 returned, as for non matches. A number of tests that were previously not
545 Perl-compatible have been moved into the Perl-compatible test files. The
546 refactoring has had the pleasing side effect of removing one argument from
547 the match() function, thus reducing its stack requirements.
548
549 19. If the /S+ option was used in pcretest to study a pattern using JIT,
550 subsequent uses of /S (without +) incorrectly behaved like /S+.
551
552 21. Retrieve executable code size support for the JIT compiler and fixing
553 some warnings.
554
555 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
556 not work when the shorter character appeared right at the end of the
557 subject string.
558
559 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
560 systems.
561
562 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
563 output it when the /M option is used in pcretest.
564
565 25. The CheckMan script was not being included in the distribution. Also, added
566 an explicit "perl" to run Perl scripts from the PrepareRelease script
567 because this is reportedly needed in Windows.
568
569 26. If study data was being save in a file and studying had not found a set of
570 "starts with" bytes for the pattern, the data written to the file (though
571 never used) was taken from uninitialized memory and so caused valgrind to
572 complain.
573
574 27. Updated RunTest.bat as provided by Sheri Pierce.
575
576 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
577
578 29. Computation of memory usage for the table of capturing group names was
579 giving an unnecessarily large value.
580
581
582 Version 8.20 21-Oct-2011
583 ------------------------
584
585 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
586 a POSIX class. After further experiments with Perl, which convinced me that
587 Perl has bugs and confusions, a closing square bracket is no longer allowed
588 in a POSIX name. This bug also affected patterns with classes that started
589 with full stops.
590
591 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
592 captured substring, but while checking the failing first alternative,
593 substring 1 is temporarily captured. If the output vector supplied to
594 pcre_exec() was not big enough for this capture, the yield of the function
595 was still zero ("insufficient space for captured substrings"). This cannot
596 be totally fixed without adding another stack variable, which seems a lot
597 of expense for a edge case. However, I have improved the situation in cases
598 such as /(a)(b)x|abc/ matched against "abc", where the return code
599 indicates that fewer than the maximum number of slots in the ovector have
600 been set.
601
602 3. Related to (2) above: when there are more back references in a pattern than
603 slots in the output vector, pcre_exec() uses temporary memory during
604 matching, and copies in the captures as far as possible afterwards. It was
605 using the entire output vector, but this conflicts with the specification
606 that only 2/3 is used for passing back captured substrings. Now it uses
607 only the first 2/3, for compatibility. This is, of course, another edge
608 case.
609
610 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
611 main code base, and can be used by building with --enable-jit. When this is
612 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
613 runtime --no-jit option is given.
614
615 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
616 ovector, the return from the function was zero, implying that there were
617 other matches that did not fit. The correct "exactly full" value is now
618 returned.
619
620 6. If a subpattern that was called recursively or as a subroutine contained
621 (*PRUNE) or any other control that caused it to give a non-standard return,
622 invalid errors such as "Error -26 (nested recursion at the same subject
623 position)" or even infinite loops could occur.
624
625 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
626 computing the minimum length on reaching *ACCEPT, and so ended up with the
627 wrong value of 1 rather than 0. Further investigation indicates that
628 computing a minimum subject length in the presence of *ACCEPT is difficult
629 (think back references, subroutine calls), and so I have changed the code
630 so that no minimum is registered for a pattern that contains *ACCEPT.
631
632 8. If (*THEN) was present in the first (true) branch of a conditional group,
633 it was not handled as intended. [But see 16 below.]
634
635 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
636 Sheri Pierce.
637
638 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
639 the first byte in a match must be "a".
640
641 11. Change 17 for 8.13 increased the recursion depth for patterns like
642 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
643 pattern contains any instances of (*THEN). If it does not, the old
644 optimizations are restored. It would be nice to do this on a per-group
645 basis, but at the moment that is not feasible.
646
647 12. In some environments, the output of pcretest -C is CRLF terminated. This
648 broke RunTest's code that checks for the link size. A single white space
649 character after the value is now allowed for.
650
651 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
652 For "fr", it uses the Windows-specific input and output files.
653
654 14. If (*THEN) appeared in a group that was called recursively or as a
655 subroutine, it did not work as intended. [But see next item.]
656
657 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
658 pattern fragments (but not containing any | characters). If A and B are
659 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
660 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
661 D. In other words, Perl considers parentheses that do not contain any |
662 characters to be part of a surrounding alternative, whereas PCRE was
663 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
664 differently. PCRE now behaves in the same way as Perl, except in the case
665 of subroutine/recursion calls such as (?1) which have in any case always
666 been different (but PCRE had them first :-).
667
668 16. Related to 15 above: Perl does not treat the | in a conditional group as
669 creating alternatives. Such a group is treated in the same way as an
670 ordinary group without any | characters when processing (*THEN). PCRE has
671 been changed to match Perl's behaviour.
672
673 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
674 RunGrepTest script failed.
675
676 18. Change 22 for version 13 caused atomic groups to use more stack. This is
677 inevitable for groups that contain captures, but it can lead to a lot of
678 stack use in large patterns. The old behaviour has been restored for atomic
679 groups that do not contain any capturing parentheses.
680
681 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
682 suppress the check for a minimum subject length at run time. (If it was
683 given to pcre_exec() or pcre_dfa_exec() it did work.)
684
685 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
686 fail to work when decoding hex characters in data strings in EBCDIC
687 environments.
688
689 21. It appears that in at least one Mac OS environment, the isxdigit() function
690 is implemented as a macro that evaluates to its argument more than once,
691 contravening the C 90 Standard (I haven't checked a later standard). There
692 was an instance in pcretest which caused it to go wrong when processing
693 \x{...} escapes in subject strings. The has been rewritten to avoid using
694 things like p++ in the argument of isxdigit().
695
696
697 Version 8.13 16-Aug-2011
698 ------------------------
699
700 1. The Unicode data tables have been updated to Unicode 6.0.0.
701
702 2. Two minor typos in pcre_internal.h have been fixed.
703
704 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
705 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
706 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
707
708 4. There were a number of related bugs in the code for matching backrefences
709 caselessly in UTF-8 mode when codes for the characters concerned were
710 different numbers of bytes. For example, U+023A and U+2C65 are an upper
711 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
712 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
713 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
714 2-byte code at the end of the subject (it thought there wasn't enough data
715 left).
716
717 5. Comprehensive information about what went wrong is now returned by
718 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
719 as the output vector has at least 2 elements. The offset of the start of
720 the failing character and a reason code are placed in the vector.
721
722 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
723 now returned is for the first byte of the failing character, instead of the
724 last byte inspected. This is an incompatible change, but I hope it is small
725 enough not to be a problem. It makes the returned offset consistent with
726 pcre_exec() and pcre_dfa_exec().
727
728 7. pcretest now gives a text phrase as well as the error number when
729 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
730 failure, the offset and reason code are output.
731
732 8. When \R was used with a maximizing quantifier it failed to skip backwards
733 over a \r\n pair if the subsequent match failed. Instead, it just skipped
734 back over a single character (\n). This seems wrong (because it treated the
735 two characters as a single entity when going forwards), conflicts with the
736 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
737 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
738 has been changed.
739
740 9. Some internal refactoring has changed the processing so that the handling
741 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
742 time (the PCRE_DOTALL option was changed this way some time ago: version
743 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
744 which was always a bit of a fudge. It also means that there is one less
745 argument for the match() function, which reduces its stack requirements
746 slightly. This change also fixes an incompatibility with Perl: the pattern
747 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
748
749 10. More internal refactoring has drastically reduced the number of recursive
750 calls to match() for possessively repeated groups such as (abc)++ when
751 using pcre_exec().
752
753 11. While implementing 10, a number of bugs in the handling of groups were
754 discovered and fixed:
755
756 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
757 (a|)*(?1) gave a compile-time internal error.
758 ((a|)+)+ did not notice that the outer group could match an empty string.
759 (^a|^)+ was not marked as anchored.
760 (.*a|.*)+ was not marked as matching at start or after a newline.
761
762 12. Yet more internal refactoring has removed another argument from the match()
763 function. Special calls to this function are now indicated by setting a
764 value in a variable in the "match data" data block.
765
766 13. Be more explicit in pcre_study() instead of relying on "default" for
767 opcodes that mean there is no starting character; this means that when new
768 ones are added and accidentally left out of pcre_study(), testing should
769 pick them up.
770
771 14. The -s option of pcretest has been documented for ages as being an old
772 synonym of -m (show memory usage). I have changed it to mean "force study
773 for every regex", that is, assume /S for every regex. This is similar to -i
774 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
775 using it. It makes it easier to run collections of tests with and without
776 study enabled, and thereby test pcre_study() more easily. All the standard
777 tests are now run with and without -s (but some patterns can be marked as
778 "never study" - see 20 below).
779
780 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
781 restoration of the capturing data to the outer values was not happening
782 correctly.
783
784 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
785 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
786 pattern had matched an empty string, and so incorrectly returned a no
787 match.
788
789 17. There was optimizing code for the last branch of non-capturing parentheses,
790 and also for the obeyed branch of a conditional subexpression, which used
791 tail recursion to cut down on stack usage. Unfortunately, now that there is
792 the possibility of (*THEN) occurring in these branches, tail recursion is
793 no longer possible because the return has to be checked for (*THEN). These
794 two optimizations have therefore been removed. [But see 8.20/11 above.]
795
796 18. If a pattern containing \R was studied, it was assumed that \R always
797 matched two bytes, thus causing the minimum subject length to be
798 incorrectly computed because \R can also match just one byte.
799
800 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
801 was incorrectly computed.
802
803 20. If /S is present twice on a test pattern in pcretest input, it now
804 *disables* studying, thereby overriding the use of -s on the command line
805 (see 14 above). This is necessary for one or two tests to keep the output
806 identical in both cases.
807
808 21. When (*ACCEPT) was used in an assertion that matched an empty string and
809 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
810
811 22. When an atomic group that contained a capturing parenthesis was
812 successfully matched, but the branch in which it appeared failed, the
813 capturing was not being forgotten if a higher numbered group was later
814 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
815 group 1 to "a", when in fact it should be unset. This applied to multi-
816 branched capturing and non-capturing groups, repeated or not, and also to
817 positive assertions (capturing in negative assertions does not happen
818 in PCRE) and also to nested atomic groups.
819
820 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
821 subject after a captured substring, to make it easier to tell which of a
822 number of identical substrings has been captured.
823
824 24. The way atomic groups are processed by pcre_exec() has been changed so that
825 if they are repeated, backtracking one repetition now resets captured
826 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
827 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
828 "aaa". Previously, it would have been "a". As part of this code
829 refactoring, the way recursive calls are handled has also been changed.
830
831 25. If an assertion condition captured any substrings, they were not passed
832 back unless some other capturing happened later. For example, if
833 (?(?=(a))a) was matched against "a", no capturing was returned.
834
835 26. When studying a pattern that contained subroutine calls or assertions,
836 the code for finding the minimum length of a possible match was handling
837 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
838 group 1 called group 2 while simultaneously a separate group 2 called group
839 1). A stack overflow occurred in this case. I have fixed this by limiting
840 the recursion depth to 10.
841
842 27. Updated RunTest.bat in the distribution to the version supplied by Tom
843 Fortmann. This supports explicit test numbers on the command line, and has
844 argument validation and error reporting.
845
846 28. An instance of \X with an unlimited repeat could fail if at any point the
847 first character it looked at was a mark character.
848
849 29. Some minor code refactoring concerning Unicode properties and scripts
850 should reduce the stack requirement of match() slightly.
851
852 30. Added the '=' option to pcretest to check the setting of unused capturing
853 slots at the end of the pattern, which are documented as being -1, but are
854 not included in the return count.
855
856 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
857 compiled something random. Now it gives a compile-time error (as does
858 Perl).
859
860 32. A *MARK encountered during the processing of a positive assertion is now
861 recorded and passed back (compatible with Perl).
862
863 33. If --only-matching or --colour was set on a pcregrep call whose pattern
864 had alternative anchored branches, the search for a second match in a line
865 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
866 matched the line "0102" twice. The same bug affected patterns that started
867 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
868 twice.
869
870 34. Previously, PCRE did not allow quantification of assertions. However, Perl
871 does, and because of capturing effects, quantifying parenthesized
872 assertions may at times be useful. Quantifiers are now allowed for
873 parenthesized assertions.
874
875 35. A minor code tidy in pcre_compile() when checking options for \R usage.
876
877 36. \g was being checked for fancy things in a character class, when it should
878 just be a literal "g".
879
880 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
881 appearance of a nested POSIX class supersedes an apparent external class.
882 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
883 unescaped square brackets may also appear as part of class names. For
884 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
885 more like Perl. (But see 8.20/1 above.)
886
887 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
888 was because it thought it was \N{name}, which is not supported).
889
890 39. Add minix to OS list not supporting the -S option in pcretest.
891
892 40. PCRE tries to detect cases of infinite recursion at compile time, but it
893 cannot analyze patterns in sufficient detail to catch mutual recursions
894 such as ((?1))((?2)). There is now a runtime test that gives an error if a
895 subgroup is called recursively as a subpattern for a second time at the
896 same position in the subject string. In previous releases this might have
897 been caught by the recursion limit, or it might have run out of stack.
898
899 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
900 happen only once. PCRE was, however incorrectly giving a compile time error
901 "recursive call could loop indefinitely" because it cannot analyze the
902 pattern in sufficient detail. The compile time test no longer happens when
903 PCRE is compiling a conditional subpattern, but actual runaway loops are
904 now caught at runtime (see 40 above).
905
906 42. It seems that Perl allows any characters other than a closing parenthesis
907 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
908 has been changed to be the same.
909
910 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
911 as not to get warnings when autogen.sh is called. Also changed
912 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
913
914 44. To help people who use pcregrep to scan files containing exceedingly long
915 lines, the following changes have been made:
916
917 (a) The default value of the buffer size parameter has been increased from
918 8K to 20K. (The actual buffer used is three times this size.)
919
920 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
921 PCRE is built.
922
923 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
924 to be set at run time.
925
926 (d) Numerical values in pcregrep options can be followed by K or M, for
927 example --buffer-size=50K.
928
929 (e) If a line being scanned overflows pcregrep's buffer, an error is now
930 given and the return code is set to 2.
931
932 45. Add a pointer to the latest mark to the callout data block.
933
934 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
935 partial match of an empty string instead of no match. This was specific to
936 the use of ".".
937
938 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
939 complete match instead of a partial match. This bug was dependent on both
940 the PCRE_UTF8 and PCRE_DOTALL options being set.
941
942 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
943 starting byte set, because \b was not being ignored.
944
945
946 Version 8.12 15-Jan-2011
947 ------------------------
948
949 1. Fixed some typos in the markup of the man pages, and wrote a script that
950 checks for such things as part of the documentation building process.
951
952 2. On a big-endian 64-bit system, pcregrep did not correctly process the
953 --match-limit and --recursion-limit options (added for 8.11). In
954 particular, this made one of the standard tests fail. (The integer value
955 went into the wrong half of a long int.)
956
957 3. If the --colour option was given to pcregrep with -v (invert match), it
958 did strange things, either producing crazy output, or crashing. It should,
959 of course, ignore a request for colour when reporting lines that do not
960 match.
961
962 4. Another pcregrep bug caused similar problems if --colour was specified with
963 -M (multiline) and the pattern match finished with a line ending.
964
965 5. In pcregrep, when a pattern that ended with a literal newline sequence was
966 matched in multiline mode, the following line was shown as part of the
967 match. This seems wrong, so I have changed it.
968
969 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
970 the check for further matches in the same line (so they could be coloured)
971 to overrun the end of the current line. If another match was found, it was
972 incorrectly shown (and then shown again when found in the next line).
973
974 7. If pcregrep was compiled under Windows, there was a reference to the
975 function pcregrep_exit() before it was defined. I am assuming this was
976 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
977 reported by a user. I've moved the definition above the reference.
978
979
980 Version 8.11 10-Dec-2010
981 ------------------------
982
983 1. (*THEN) was not working properly if there were untried alternatives prior
984 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
985 backtracked to try for "b" instead of moving to the next alternative branch
986 at the same level (in this case, to look for "c"). The Perl documentation
987 is clear that when (*THEN) is backtracked onto, it goes to the "next
988 alternative in the innermost enclosing group".
989
990 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
991 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
992 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
993 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
994 (*THEN).
995
996 3. If \s appeared in a character class, it removed the VT character from
997 the class, even if it had been included by some previous item, for example
998 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
999 of \s, but is part of the POSIX "space" class.)
1000
1001 4. A partial match never returns an empty string (because you can always
1002 match an empty string at the end of the subject); however the checking for
1003 an empty string was starting at the "start of match" point. This has been
1004 changed to the "earliest inspected character" point, because the returned
1005 data for a partial match starts at this character. This means that, for
1006 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1007 (previously it gave "no match").
1008
1009 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1010 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1011 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1012 has an implication that the given string is incomplete (because a partial
1013 match is preferred over a full match). For this reason, these items now
1014 give a partial match in this situation. [Aside: previously, the one case
1015 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1016 match rather than a full match, which was wrong by the old rules, but is
1017 now correct.]
1018
1019 6. There was a bug in the handling of #-introduced comments, recognized when
1020 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1021 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1022 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1023 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1024 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1025 places in pcre_compile().
1026
1027 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1028 comments when looking ahead for named forward references to subpatterns,
1029 the only newline sequence it recognized was NL. It now handles newlines
1030 according to the set newline convention.
1031
1032 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1033 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1034 cater for a lack of strerror(). These oversights have been fixed.
1035
1036 9. Added --match-limit and --recursion-limit to pcregrep.
1037
1038 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1039
1040 11. When the -o option was used, pcregrep was setting a return code of 1, even
1041 when matches were found, and --line-buffered was not being honoured.
1042
1043 12. Added an optional parentheses number to the -o and --only-matching options
1044 of pcregrep.
1045
1046 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1047 can match an empty string. The code to do it in pcretest and pcredemo
1048 needed fixing:
1049
1050 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1051 only one byte after an empty string match just before CRLF (this case
1052 just got forgotten; "any" and "anycrlf" were OK).
1053
1054 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1055 mode when an empty string match preceded an ASCII character followed by
1056 a non-ASCII character. (The code for advancing by one character rather
1057 than one byte was nonsense.)
1058
1059 (c) The pcredemo.c sample program did not have any code at all to handle
1060 the cases when CRLF is a valid newline sequence.
1061
1062 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1063 as a starting offset was within the subject string. There is now a new
1064 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1065 negative or greater than the length of the string. In order to test this,
1066 pcretest is extended to allow the setting of negative starting offsets.
1067
1068 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1069 starting offset points to the beginning of a UTF-8 character was
1070 unnecessarily clumsy. I tidied it up.
1071
1072 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1073 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1074
1075 17. Nobody had reported that the --include_dir option, which was added in
1076 release 7.7 should have been called --include-dir (hyphen, not underscore)
1077 for compatibility with GNU grep. I have changed it to --include-dir, but
1078 left --include_dir as an undocumented synonym, and the same for
1079 --exclude-dir, though that is not available in GNU grep, at least as of
1080 release 2.5.4.
1081
1082 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1083 characters from a string of bytes) have been redefined so as not to use
1084 loops, in order to improve performance in some environments. At the same
1085 time, I abstracted some of the common code into auxiliary macros to save
1086 repetition (this should not affect the compiled code).
1087
1088 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1089 compile-time error is now given if \c is not followed by an ASCII
1090 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1091 different, and any byte value is allowed.)
1092
1093 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1094 START_OPTIMIZE option, which is now allowed at compile time - but just
1095 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1096 to pcregrep and other applications that have no direct access to PCRE
1097 options. The new /Y option in pcretest sets this option when calling
1098 pcre_compile().
1099
1100 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1101 back references. Groups containing recursive back references were forced to
1102 be atomic by that change, but in the case of named groups, the amount of
1103 memory required was incorrectly computed, leading to "Failed: internal
1104 error: code overflow". This has been fixed.
1105
1106 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1107 pcretest.c, to avoid build problems in some Borland environments.
1108
1109
1110 Version 8.10 25-Jun-2010
1111 ------------------------
1112
1113 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1114 THEN.
1115
1116 2. (*ACCEPT) was not working when inside an atomic group.
1117
1118 3. Inside a character class, \B is treated as a literal by default, but
1119 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1120 causes the error). The code is unchanged, but I tidied the documentation.
1121
1122 4. Inside a character class, PCRE always treated \R and \X as literals,
1123 whereas Perl faults them if its -w option is set. I have changed PCRE so
1124 that it faults them when PCRE_EXTRA is set.
1125
1126 5. Added support for \N, which always matches any character other than
1127 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1128
1129 6. When compiling pcregrep with newer versions of gcc which may have
1130 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1131 declared with attribute warn_unused_result" were given. Just casting the
1132 result to (void) does not stop the warnings; a more elaborate fudge is
1133 needed. I've used a macro to implement this.
1134
1135 7. Minor change to pcretest.c to avoid a compiler warning.
1136
1137 8. Added four artifical Unicode properties to help with an option to make
1138 \s etc use properties (see next item). The new properties are: Xan
1139 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1140
1141 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1142 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1143 this option. Modified pcretest to add /W to test this facility. Added
1144 REG_UCP to make it available via the POSIX interface.
1145
1146 10. Added --line-buffered to pcregrep.
1147
1148 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1149 studied, and the match started with a letter with a code point greater than
1150 127 whose first byte was different to the first byte of the other case of
1151 the letter, the other case of this starting letter was not recognized
1152 (#976).
1153
1154 12. If a pattern that was studied started with a repeated Unicode property
1155 test, for example, \p{Nd}+, there was the theoretical possibility of
1156 setting up an incorrect bitmap of starting bytes, but fortunately it could
1157 not have actually happened in practice until change 8 above was made (it
1158 added property types that matched character-matching opcodes).
1159
1160 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1161 possible starting bytes for non-anchored patterns.
1162
1163 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1164 \R, and also a number of cases that involve Unicode properties, both
1165 explicit and implicit when PCRE_UCP is set.
1166
1167 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1168 input, it could crash or give wrong results if characters with values
1169 greater than 0xc0 were present in the subject string. (Detail: it assumed
1170 UTF-8 input when processing these items.)
1171
1172 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1173 size_t is 64-bit (#991).
1174
1175 17. Added a check for running out of memory when PCRE is compiled with
1176 --disable-stack-for-recursion (#990).
1177
1178 18. If the last data line in a file for pcretest does not have a newline on
1179 the end, a newline was missing in the output.
1180
1181 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1182 less than 128) in its various bitmaps. However, there is a facility for
1183 generating tables according to the current locale when PCRE is compiled. It
1184 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1185 characters, are recognized by isspace() and therefore were getting set in
1186 these tables, and indeed these tables seem to approximate to ISO 8859. This
1187 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1188 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1189 which of course cannot start UTF-8 characters. I have changed the code so
1190 that only real ASCII characters (less than 128) and the correct starting
1191 bytes for UTF-8 encodings are set for characters greater than 127 when in
1192 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1193 altogether.)
1194
1195 20. Added the /T option to pcretest so as to be able to run tests with non-
1196 standard character tables, thus making it possible to include the tests
1197 used for 19 above in the standard set of tests.
1198
1199 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1200 reference to a subpattern the other side of a comment that contains an
1201 opening parenthesis caused either an internal compiling error, or a
1202 reference to the wrong subpattern.
1203
1204
1205 Version 8.02 19-Mar-2010
1206 ------------------------
1207
1208 1. The Unicode data tables have been updated to Unicode 5.2.0.
1209
1210 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1211 configured.
1212
1213 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1214 original author of that file, following a query about its status.
1215
1216 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1217 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1218
1219 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1220 quantifier applied to a forward-referencing subroutine call, could compile
1221 incorrect code or give the error "internal error: previously-checked
1222 referenced subpattern not found".
1223
1224 6. Both MS Visual Studio and Symbian OS have problems with initializing
1225 variables to point to external functions. For these systems, therefore,
1226 pcre_malloc etc. are now initialized to local functions that call the
1227 relevant global functions.
1228
1229 7. There were two entries missing in the vectors called coptable and poptable
1230 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1231 I've fixed the data, and added a kludgy way of testing at compile time that
1232 the lengths are correct (equal to the number of opcodes).
1233
1234 8. Following on from 7, I added a similar kludge to check the length of the
1235 eint vector in pcreposix.c.
1236
1237 9. Error texts for pcre_compile() are held as one long string to avoid too
1238 much relocation at load time. To find a text, the string is searched,
1239 counting zeros. There was no check for running off the end of the string,
1240 which could happen if a new error number was added without updating the
1241 string.
1242
1243 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1244
1245 11. \K was not working if it appeared in an atomic group or in a group that
1246 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1247 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1248 the assertion is positive, but not if it is negative.
1249
1250 12. Change 11 fortuitously reduced the size of the stack frame used in the
1251 "match()" function of pcre_exec.c by one pointer. Forthcoming
1252 implementation of support for (*MARK) will need an extra pointer on the
1253 stack; I have reserved it now, so that the stack frame size does not
1254 decrease.
1255
1256 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1257 item in branch that calls a recursion is a subroutine call - as in the
1258 second branch in the above example - was incorrectly given the compile-
1259 time error "recursive call could loop indefinitely" because pcre_compile()
1260 was not correctly checking the subroutine for matching a non-empty string.
1261
1262 14. The checks for overrunning compiling workspace could trigger after an
1263 overrun had occurred. This is a "should never occur" error, but it can be
1264 triggered by pathological patterns such as hundreds of nested parentheses.
1265 The checks now trigger 100 bytes before the end of the workspace.
1266
1267 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1268
1269
1270 Version 8.01 19-Jan-2010
1271 ------------------------
1272
1273 1. If a pattern contained a conditional subpattern with only one branch (in
1274 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1275 computed the wrong minimum data length (which is of course zero for such
1276 subpatterns). This could cause incorrect "no match" results.
1277
1278 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1279 the pattern is reset in the first branch, pcre_compile() failed with
1280 "internal error: code overflow at offset...". This happened only when
1281 the reset was to the original external option setting. (An optimization
1282 abstracts leading options settings into an external setting, which was the
1283 cause of this.)
1284
1285 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1286 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1287 assertion pattern did not match (meaning that the assertion was true), it
1288 was incorrectly treated as false if the SKIP had been reached during the
1289 matching. This also applied to assertions used as conditions.
1290
1291 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1292 assertion subpattern, including such a pattern used as a condition,
1293 unpredictable results occurred, instead of the error return
1294 PCRE_ERROR_DFA_UITEM.
1295
1296 5. The C++ GlobalReplace function was not working like Perl for the special
1297 situation when an empty string is matched. It now does the fancy magic
1298 stuff that is necessary.
1299
1300 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1301 removed. (These were left over from very, very early versions of PCRE.)
1302
1303 7. Some cosmetic changes to the code to make life easier when compiling it
1304 as part of something else:
1305
1306 (a) Change DEBUG to PCRE_DEBUG.
1307
1308 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1309 called "current" as "current_branch", to prevent a collision with the
1310 Linux macro when compiled as a kernel module.
1311
1312 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1313 prevent a collision with the Linux macro when compiled as a kernel
1314 module.
1315
1316 8. In pcre_compile() there are some checks for integer overflows that used to
1317 cast potentially large values to (double). This has been changed to that
1318 when building, a check for int64_t is made, and if it is found, it is used
1319 instead, thus avoiding the use of floating point arithmetic. (There is no
1320 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1321 double.
1322
1323 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1324 2005 (difference between two addresses compared to an unsigned value).
1325
1326 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1327 custom one, because of the following reported problem in Windows:
1328
1329 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1330 under Win32.
1331 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1332 therefore missing the function definition.
1333 - The compiler thus generates a "C" signature for the test function.
1334 - The linker fails to find the "C" function.
1335 - PCRE fails to configure if asked to do so against libbz2.
1336
1337 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1338 messages were output:
1339
1340 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1341 rerunning libtoolize, to keep the correct libtool macros in-tree.
1342 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1343
1344 I have done both of these things.
1345
1346 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1347 most of the time, it *can* run out if it is given a pattern that contains a
1348 runaway infinite recursion. I updated the discussion in the pcrestack man
1349 page.
1350
1351 13. Now that we have gone to the x.xx style of version numbers, the minor
1352 version may start with zero. Using 08 or 09 is a bad idea because users
1353 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1354 interpreted as invalid octal numbers. I've updated the previous comment in
1355 configure.ac, and also added a check that gives an error if 08 or 09 are
1356 used.
1357
1358 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1359 causing partial matching to fail when the end of the subject matched \W
1360 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1361
1362 15. There were some discrepancies between the declarations in pcre_internal.h
1363 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1364 their definitions. The declarations used "const uschar *" and the
1365 definitions used USPTR. Even though USPTR is normally defined as "const
1366 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1367 reported that: "This difference in casting confuses some C++ compilers, for
1368 example, SunCC recognizes above declarations as different functions and
1369 generates broken code for hbpcre." I have changed the declarations to use
1370 USPTR.
1371
1372 16. GNU libtool is named differently on some systems. The autogen.sh script now
1373 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1374 (FreeBSD).
1375
1376 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1377 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1378 comment: "Figure out how to create a longlong from a string: strtoll and
1379 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1380 instance, but it only takes 2 args instead of 3!"
1381
1382 18. A subtle bug concerned with back references has been fixed by a change of
1383 specification, with a corresponding code fix. A pattern such as
1384 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1385 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1386 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1387 same bug. Such groups have to be quantified to be useful, or contained
1388 inside another quantified group. (If there's no repetition, the reference
1389 can never match.) The problem arises because, having left the group and
1390 moved on to the rest of the pattern, a later failure that backtracks into
1391 the group uses the captured value from the final iteration of the group
1392 rather than the correct earlier one. I have fixed this in PCRE by forcing
1393 any group that contains a reference to itself to be an atomic group; that
1394 is, there cannot be any backtracking into it once it has completed. This is
1395 similar to recursive and subroutine calls.
1396
1397
1398 Version 8.00 19-Oct-09
1399 ----------------------
1400
1401 1. The table for translating pcre_compile() error codes into POSIX error codes
1402 was out-of-date, and there was no check on the pcre_compile() error code
1403 being within the table. This could lead to an OK return being given in
1404 error.
1405
1406 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1407 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1408 in a Windows environment.
1409
1410 3. The pcregrep --count option prints the count for each file even when it is
1411 zero, as does GNU grep. However, pcregrep was also printing all files when
1412 --files-with-matches was added. Now, when both options are given, it prints
1413 counts only for those files that have at least one match. (GNU grep just
1414 prints the file name in this circumstance, but including the count seems
1415 more useful - otherwise, why use --count?) Also ensured that the
1416 combination -clh just lists non-zero counts, with no names.
1417
1418 4. The long form of the pcregrep -F option was incorrectly implemented as
1419 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1420 but it seems right to fix it, and I didn't think it was worth preserving
1421 the old behaviour.
1422
1423 5. The command line items --regex=pattern and --regexp=pattern were not
1424 recognized by pcregrep, which required --regex pattern or --regexp pattern
1425 (with a space rather than an '='). The man page documented the '=' forms,
1426 which are compatible with GNU grep; these now work.
1427
1428 6. No libpcreposix.pc file was created for pkg-config; there was just
1429 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1430
1431 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1432 when UCP support is not needed, by modifying the Python script that
1433 generates it from Unicode data files. This should not matter if the module
1434 is correctly used as a library, but I received one complaint about 50K of
1435 unwanted data. My guess is that the person linked everything into his
1436 program rather than using a library. Anyway, it does no harm.
1437
1438 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1439 was a minimum greater than 1 for a wide character in a possessive
1440 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1441 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1442 character. Chaos in the form of incorrect output or a compiling loop could
1443 result.
1444
1445 9. The restrictions on what a pattern can contain when partial matching is
1446 requested for pcre_exec() have been removed. All patterns can now be
1447 partially matched by this function. In addition, if there are at least two
1448 slots in the offset vector, the offset of the earliest inspected character
1449 for the match and the offset of the end of the subject are set in them when
1450 PCRE_ERROR_PARTIAL is returned.
1451
1452 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1453 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1454 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1455 and may be more useful for multi-segment matching.
1456
1457 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1458 used to be given if ever the end of the subject was reached; now it is
1459 given only if matching could not proceed because another character was
1460 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1461 string "Z", which now yields "no match" instead of "partial match". In the
1462 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1463 final character ended with (*FAIL).
1464
1465 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1466 if the pattern had a "must contain" character that was already found in the
1467 earlier partial match, unless partial matching was again requested. For
1468 example, with the pattern /dog.(body)?/, the "must contain" character is
1469 "g". If the first part-match was for the string "dog", restarting with
1470 "sbody" failed. This bug has been fixed.
1471
1472 13. The string returned by pcre_dfa_exec() after a partial match has been
1473 changed so that it starts at the first inspected character rather than the
1474 first character of the match. This makes a difference only if the pattern
1475 starts with a lookbehind assertion or \b or \B (\K is not supported by
1476 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1477 matching functions compatible, and I think it's the right thing to do.
1478
1479 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1480 so that the demonstration program is easily available in environments where
1481 PCRE has not been installed from source.
1482
1483 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1484 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1485 library.
1486
1487 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1488 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1489 is not the first non-POSIX option to be added. Clearly some people find
1490 these options useful.
1491
1492 17. If a caller to the POSIX matching function regexec() passes a non-zero
1493 value for nmatch with a NULL value for pmatch, the value of
1494 nmatch is forced to zero.
1495
1496 18. RunGrepTest did not have a test for the availability of the -u option of
1497 the diff command, as RunTest does. It now checks in the same way as
1498 RunTest, and also checks for the -b option.
1499
1500 19. If an odd number of negated classes containing just a single character
1501 interposed, within parentheses, between a forward reference to a named
1502 subpattern and the definition of the subpattern, compilation crashed with
1503 an internal error, complaining that it could not find the referenced
1504 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1505 [The bug was that it was starting one character too far in when skipping
1506 over the character class, thus treating the ] as data rather than
1507 terminating the class. This meant it could skip too much.]
1508
1509 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1510 /g option in pcretest when the pattern contains \K, which makes it possible
1511 to have an empty string match not at the start, even when the pattern is
1512 anchored. Updated pcretest and pcredemo to use this option.
1513
1514 21. If the maximum number of capturing subpatterns in a recursion was greater
1515 than the maximum at the outer level, the higher number was returned, but
1516 with unset values at the outer level. The correct (outer level) value is
1517 now given.
1518
1519 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1520 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1521 make it do so. The string so far is captured, making this feature
1522 compatible with Perl.
1523
1524 23. The tests have been re-organized, adding tests 11 and 12, to make it
1525 possible to check the Perl 5.10 features against Perl 5.10.
1526
1527 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1528 pattern matches a fixed length string. PCRE did not allow this; now it
1529 does. Neither allows recursion.
1530
1531 25. I finally figured out how to implement a request to provide the minimum
1532 length of subject string that was needed in order to match a given pattern.
1533 (It was back references and recursion that I had previously got hung up
1534 on.) This code has now been added to pcre_study(); it finds a lower bound
1535 to the length of subject needed. It is not necessarily the greatest lower
1536 bound, but using it to avoid searching strings that are too short does give
1537 some useful speed-ups. The value is available to calling programs via
1538 pcre_fullinfo().
1539
1540 26. While implementing 25, I discovered to my embarrassment that pcretest had
1541 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1542 study optimizations had never been tested with that matching function.
1543 Oops. What is worse, even when it was passed study data, there was a bug in
1544 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1545 were also very few tests of studied patterns with pcre_dfa_exec().
1546
1547 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1548 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1549 on the other side of the coin, they are no longer allowed to have different
1550 names, because these cannot be distinguished in PCRE, and this has caused
1551 confusion. (This is a difference from Perl.)
1552
1553 28. When duplicate subpattern names are present (necessarily with different
1554 numbers, as required by 27 above), and a test is made by name in a
1555 conditional pattern, either for a subpattern having been matched, or for
1556 recursion in such a pattern, all the associated numbered subpatterns are
1557 tested, and the overall condition is true if the condition is true for any
1558 one of them. This is the way Perl works, and is also more like the way
1559 testing by number works.
1560
1561
1562 Version 7.9 11-Apr-09
1563 ---------------------
1564
1565 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1566 (pcretest), all targets were linked against these libraries. This included
1567 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1568 libraries. This caused unwanted dependencies to be created. This problem
1569 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1570 pcretest is linked with readline.
1571
1572 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1573 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1574 moved outside it again, because FALSE and TRUE are already defined in AIX,
1575 but BOOL is not.
1576
1577 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1578 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1579
1580 4. The pcregrep documentation said spaces were inserted as well as colons (or
1581 hyphens) following file names and line numbers when outputting matching
1582 lines. This is not true; no spaces are inserted. I have also clarified the
1583 wording for the --colour (or --color) option.
1584
1585 5. In pcregrep, when --colour was used with -o, the list of matching strings
1586 was not coloured; this is different to GNU grep, so I have changed it to be
1587 the same.
1588
1589 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1590 each matching line was coloured. Now it goes on to look for further matches
1591 of any of the test patterns, which is the same behaviour as GNU grep.
1592
1593 7. A pattern that could match an empty string could cause pcregrep to loop; it
1594 doesn't make sense to accept an empty string match in pcregrep, so I have
1595 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1596 seems to be how GNU grep behaves.
1597
1598 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1599 start or after a newline", because the conditional assertion was not being
1600 correctly handled. The rule now is that both the assertion and what follows
1601 in the first alternative must satisfy the test.
1602
1603 9. If auto-callout was enabled in a pattern with a conditional group whose
1604 condition was an assertion, PCRE could crash during matching, both with
1605 pcre_exec() and pcre_dfa_exec().
1606
1607 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1608 used for matching.
1609
1610 11. Unicode property support in character classes was not working for
1611 characters (bytes) greater than 127 when not in UTF-8 mode.
1612
1613 12. Added the -M command line option to pcretest.
1614
1615 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1616
1617 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1618
1619 16. Added comments and documentation about mis-use of no_arg in the C++
1620 wrapper.
1621
1622 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1623 from Martin Jerabek that uses macro names for all relevant character and
1624 string constants.
1625
1626 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1627 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1628 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1629 these, but not everybody uses configure.
1630
1631 19. A conditional group that had only one branch was not being correctly
1632 recognized as an item that could match an empty string. This meant that an
1633 enclosing group might also not be so recognized, causing infinite looping
1634 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1635 with the subject "ab", where knowledge that the repeated group can match
1636 nothing is needed in order to break the loop.
1637
1638 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1639 exec(), but without supplying a callout function, matching went wrong.
1640
1641 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1642 leak if the size of the offset vector was greater than 30. When the vector
1643 is smaller, the saved offsets during recursion go onto a local stack
1644 vector, but for larger vectors malloc() is used. It was failing to free
1645 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1646 error, in fact).
1647
1648 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1649 heapframe that is used only when UTF-8 support is enabled. This caused no
1650 problem, but was untidy.
1651
1652 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1653 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1654 included within another project.
1655
1656 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1657 slightly modified by me:
1658
1659 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1660 not building pcregrep.
1661
1662 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1663 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1664
1665 25. Forward references, both numeric and by name, in patterns that made use of
1666 duplicate group numbers, could behave incorrectly or give incorrect errors,
1667 because when scanning forward to find the reference group, PCRE was not
1668 taking into account the duplicate group numbers. A pattern such as
1669 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1670
1671 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1672 the feature of a custom pointer more persuasive (as requested by a user).
1673
1674 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1675 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1676 pre-defined.
1677
1678 28. Added support for (*UTF8) at the start of a pattern.
1679
1680 29. Arrange for flags added by the "release type" setting in CMake to be shown
1681 in the configuration summary.
1682
1683
1684 Version 7.8 05-Sep-08
1685 ---------------------
1686
1687 1. Replaced UCP searching code with optimized version as implemented for Ad
1688 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1689 stage table and inline lookup instead of a function, giving speed ups of 2
1690 to 5 times on some simple patterns that I tested. Permission was given to
1691 distribute the MultiStage2.py script that generates the tables (it's not in
1692 the tarball, but is in the Subversion repository).
1693
1694 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1695 scripts.
1696
1697 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1698 a group with a zero qualifier. The result of the study could be incorrect,
1699 or the function might crash, depending on the pattern.
1700
1701 4. Caseless matching was not working for non-ASCII characters in back
1702 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1703 It now works when Unicode Property Support is available.
1704
1705 5. In pcretest, an escape such as \x{de} in the data was always generating
1706 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1707 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1708 truncation.
1709
1710 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1711
1712 7. Added two (int) casts to pcregrep when printing the difference of two
1713 pointers, in case they are 64-bit values.
1714
1715 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1716 test 2 if it fails.
1717
1718 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1719 and a #define of that name to empty if it is not externally set. This is to
1720 allow users of MSVC to set it if necessary.
1721
1722 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1723 the convenience functions in the pcre_get.c source file.
1724
1725 11. An option change at the start of a pattern that had top-level alternatives
1726 could cause overwriting and/or a crash. This command provoked a crash in
1727 some environments:
1728
1729 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1730
1731 This potential security problem was recorded as CVE-2008-2371.
1732
1733 12. For a pattern where the match had to start at the beginning or immediately
1734 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1735 pcre_dfa_exec() could read past the end of the passed subject if there was
1736 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1737 pcretest so that it places the subject at the end of its malloc-ed buffer.
1738
1739 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1740 exec() might read past the end of the data buffer in UTF-8 mode.
1741
1742 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1743 the data contained the byte 0x85 as part of a UTF-8 character within its
1744 first line. This applied both to normal and DFA matching.
1745
1746 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1747 /^[^d]*?$/8 failed to match "abc".
1748
1749 16. Added a missing copyright notice to pcrecpp_internal.h.
1750
1751 17. Make it more clear in the documentation that values returned from
1752 pcre_exec() in ovector are byte offsets, not character counts.
1753
1754 18. Tidied a few places to stop certain compilers from issuing warnings.
1755
1756 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1757 supplied by Stefan Weber. I made a further small update for 7.8 because
1758 there is a change of source arrangements: the pcre_searchfuncs.c module is
1759 replaced by pcre_ucd.c.
1760
1761
1762 Version 7.7 07-May-08
1763 ---------------------
1764
1765 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1766 a string to a long long, pretend we don't even have a long long." This is
1767 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1768
1769 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1770 pre-7.6 versions, which defined a global no_arg variable instead of putting
1771 it in the RE class. (See also #8 below.)
1772
1773 3. Remove a line of dead code, identified by coverity and reported by Nuno
1774 Lopes.
1775
1776 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1777
1778 (1) The include/exclude patterns were being applied to the whole pathnames
1779 of files, instead of just to the final components.
1780
1781 (2) If there was more than one level of directory, the subdirectories were
1782 skipped unless they satisfied the include/exclude conditions. This is
1783 inconsistent with GNU grep (and could even be seen as contrary to the
1784 pcregrep specification - which I improved to make it absolutely clear).
1785 The action now is always to scan all levels of directory, and just
1786 apply the include/exclude patterns to regular files.
1787
1788 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1789 --exclude_dir in the tests to avoid scanning .svn directories.
1790
1791 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1792 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1793 doesn't support NULs in patterns.
1794
1795 7. Added some missing "const"s to declarations of static tables in
1796 pcre_compile.c and pcre_dfa_exec.c.
1797
1798 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1799 caused by fix #2 above. (Subsequently also a second patch to fix the
1800 first patch. And a third patch - this was a messy problem.)
1801
1802 9. Applied Craig's patch to remove the use of push_back().
1803
1804 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1805 matching function regexec().
1806
1807 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1808 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1809 references. PCRE supports relative numbers with this syntax (I don't think
1810 Oniguruma does).
1811
1812 12. Previously, a group with a zero repeat such as (...){0} was completely
1813 omitted from the compiled regex. However, this means that if the group
1814 was called as a subroutine from elsewhere in the pattern, things went wrong
1815 (an internal error was given). Such groups are now left in the compiled
1816 pattern, with a new opcode that causes them to be skipped at execution
1817 time.
1818
1819 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1820 to the way PCRE behaves:
1821
1822 (a) A lone ] character is dis-allowed (Perl treats it as data).
1823
1824 (b) A back reference to an unmatched subpattern matches an empty string
1825 (Perl fails the current match path).
1826
1827 (c) A data ] in a character class must be notated as \] because if the
1828 first data character in a class is ], it defines an empty class. (In
1829 Perl it is not possible to have an empty class.) The empty class []
1830 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1831 The negative empty class [^] matches any one character, independently
1832 of the DOTALL setting.
1833
1834 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1835 non-existent subpattern following a character class starting with ']' and
1836 containing () gave an internal compiling error instead of "reference to
1837 non-existent subpattern". Fortunately, when the pattern did exist, the
1838 compiled code was correct. (When scanning forwards to check for the
1839 existencd of the subpattern, it was treating the data ']' as terminating
1840 the class, so got the count wrong. When actually compiling, the reference
1841 was subsequently set up correctly.)
1842
1843 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1844 it was being rejected as not supported by pcre_dfa_exec(), even though
1845 other assertions are supported. I have made pcre_dfa_exec() support
1846 (*FAIL).
1847
1848 16. The implementation of 13c above involved the invention of a new opcode,
1849 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1850 cannot be changed at match time, I realized I could make a small
1851 improvement to matching performance by compiling OP_ALLANY instead of
1852 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1853 on the OP_ANY path.
1854
1855 17. Compiling pcretest on Windows with readline support failed without the
1856 following two fixes: (1) Make the unistd.h include conditional on
1857 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1858
1859 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1860 ncurses library to be included for pcretest when ReadLine support is
1861 requested, but also to allow for it to be overridden. This patch came from
1862 Daniel Bergström.
1863
1864 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1865 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1866 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1867 spotting this.
1868
1869
1870 Version 7.6 28-Jan-08
1871 ---------------------
1872
1873 1. A character class containing a very large number of characters with
1874 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1875 overflow.
1876
1877 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1878 HAVE_LONG_LONG is not defined.
1879
1880 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1881 bring them up to date and include new features. This patch includes:
1882
1883 - Fixed PH's badly added libz and libbz2 support.
1884 - Fixed a problem with static linking.
1885 - Added pcredemo. [But later removed - see 7 below.]
1886 - Fixed dftables problem and added an option.
1887 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1888 HAVE_LONG_LONG.
1889 - Added readline support for pcretest.
1890 - Added an listing of the option settings after cmake has run.
1891
1892 4. A user submitted a patch to Makefile that makes it easy to create
1893 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1894 Makefile.am that cause it to include this special target, without
1895 affecting anything else. Note that the same mingw target plus all
1896 the other distribution libraries and programs are now supported
1897 when configuring with CMake (see 6 below) instead of with
1898 Configure/Make.
1899
1900 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1901 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1902 exported in the Windows port". It has not yet been confirmed that the patch
1903 solves the problem, but it does no harm.
1904
1905 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1906 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1907 with CMake, and also correct the comment about stack recursion.
1908
1909 7. Remove the automatic building of pcredemo from the ./configure system and
1910 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1911 of a program that users should build themselves after PCRE is installed, so
1912 building it automatically is not really right. What is more, it gave
1913 trouble in some build environments.
1914
1915 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1916
1917
1918 Version 7.5 10-Jan-08
1919 ---------------------
1920
1921 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1922 values in parens when parsing an RE using the C++ wrapper."
1923
1924 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1925 Characters greater than 255 were excluded from the class instead of being
1926 included.
1927
1928 3. The same bug as (2) above applied to negated POSIX classes such as
1929 [:^space:].
1930
1931 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1932 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1933 I have changed it.
1934
1935 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1936 first named subpattern) and a construct such as (?&a) would reference the
1937 first named subpattern whose name started with "a" (in other words, the
1938 length check was missing). Both these problems are fixed. "Subpattern name
1939 expected" is now given for (?&) (a zero-length name), and this patch also
1940 makes it give the same error for \k'' (previously it complained that that
1941 was a reference to a non-existent subpattern).
1942
1943 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1944 this is right because (?- can be followed by option settings as well as by
1945 digits. I have, however, made the messages clearer.
1946
1947 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1948 than the number used in the conditional) now cause a compile-time error.
1949 This is actually not compatible with Perl, which accepts such patterns, but
1950 treats the conditional as always being FALSE (as PCRE used to), but it
1951 seems to me that giving a diagnostic is better.
1952
1953 8. Change "alphameric" to the more common word "alphanumeric" in comments
1954 and messages.
1955
1956 9. Fix two occurrences of "backslash" in comments that should have been
1957 "backspace".
1958
1959 10. Remove two redundant lines of code that can never be obeyed (their function
1960 was moved elsewhere).
1961
1962 11. The program that makes PCRE's Unicode character property table had a bug
1963 which caused it to generate incorrect table entries for sequences of
1964 characters that have the same character type, but are in different scripts.
1965 It amalgamated them into a single range, with the script of the first of
1966 them. In other words, some characters were in the wrong script. There were
1967 thirteen such cases, affecting characters in the following ranges:
1968
1969 U+002b0 - U+002c1
1970 U+0060c - U+0060d
1971 U+0061e - U+00612
1972 U+0064b - U+0065e
1973 U+0074d - U+0076d
1974 U+01800 - U+01805
1975 U+01d00 - U+01d77
1976 U+01d9b - U+01dbf
1977 U+0200b - U+0200f
1978 U+030fc - U+030fe
1979 U+03260 - U+0327f
1980 U+0fb46 - U+0fbb1
1981 U+10450 - U+1049d
1982
1983 12. The -o option (show only the matching part of a line) for pcregrep was not
1984 compatible with GNU grep in that, if there was more than one match in a
1985 line, it showed only the first of them. It now behaves in the same way as
1986 GNU grep.
1987
1988 13. If the -o and -v options were combined for pcregrep, it printed a blank
1989 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1990 does the same. The return code can be used to tell if there were any
1991 non-matching lines.
1992
1993 14. Added --file-offsets and --line-offsets to pcregrep.
1994
1995 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1996 infinitely looping recursion. The bug was that positive lookaheads were not
1997 being skipped when checking for a possible empty match (negative lookaheads
1998 and both kinds of lookbehind were skipped).
1999
2000 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2001 inclusion of <windows.h> to before rather than after the definition of
2002 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2003
2004 17. Specifying a possessive quantifier with a specific limit for a Unicode
2005 character property caused pcre_compile() to compile bad code, which led at
2006 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2007 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2008 caused the error; without that there was no problem.
2009
2010 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2011
2012 19. Added --enable-pcretest-libreadline.
2013
2014 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2015 RE::GlobalReplace(). As a result, the number of replacements returned was
2016 double what it should be. I removed one of the increments, but Craig sent a
2017 later patch that removed the other one (the right fix) and added unit tests
2018 that check the return values (which was not done before).
2019
2020 21. Several CMake things:
2021
2022 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2023 the names libpcre and libpcreposix, not just pcre and pcreposix.
2024
2025 (2) The above change means that pcretest and pcregrep are now correctly
2026 linked with the newly-built libraries, not previously installed ones.
2027
2028 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2029
2030 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2031 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2032 UTF-8 newline character). The key issue is that the pattern starts .*;
2033 this means that the match must be either at the beginning, or after a
2034 newline. The bug was in the code for advancing after a failed match and
2035 checking that the new position followed a newline. It was not taking
2036 account of UTF-8 characters correctly.
2037
2038 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2039 character classes. PCRE was not treating the sequence [:...:] as a
2040 character class unless the ... were all letters. Perl, however, seems to
2041 allow any characters between [: and :], though of course it rejects as
2042 unknown any "names" that contain non-letters, because all the known class
2043 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2044 for example, whereas PCRE did not - it did not recognize a POSIX character
2045 class. This seemed a bit dangerous, so the code has been changed to be
2046 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2047 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2048 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2049 Perl does, and where it didn't before.
2050
2051 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2052 Windows environments %n is disabled by default.
2053
2054
2055 Version 7.4 21-Sep-07
2056 ---------------------
2057
2058 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2059 means that a class such as [\s] counted as "explicit reference to CR or
2060 LF". That isn't really right - the whole point of the change was to try to
2061 help when there was an actual mention of one of the two characters. So now
2062 the change happens only if \r or \n (or a literal CR or LF) character is
2063 encountered.
2064
2065 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2066 of both had grown to the point where there were only 3 bits left.
2067 Fortunately, there was spare space in the data structure, and so I have
2068 moved the internal flags into a new 16-bit field to free up more option
2069 bits.
2070
2071 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2072 but did not set the internal JCHANGED flag - either of these is enough to
2073 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2074 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2075 start sets both bits.
2076
2077 4. Added options (at build time, compile time, exec time) to change \R from
2078 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2079
2080 5. doc/pcresyntax.html was missing from the distribution.
2081
2082 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2083 compatibility, even though it is no longer used.
2084
2085 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2086 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2087 windows.h file is present (where different names are used). [This was
2088 reversed later after testing - see 16 below.]
2089
2090 8. Changed all #include <config.h> to #include "config.h". There were also
2091 some further <pcre.h> cases that I changed to "pcre.h".
2092
2093 9. When pcregrep was used with the --colour option, it missed the line ending
2094 sequence off the lines that it output.
2095
2096 10. It was pointed out to me that arrays of string pointers cause lots of
2097 relocations when a shared library is dynamically loaded. A technique of
2098 using a single long string with a table of offsets can drastically reduce
2099 these. I have refactored PCRE in four places to do this. The result is
2100 dramatic:
2101
2102 Originally: 290
2103 After changing UCP table: 187
2104 After changing error message table: 43
2105 After changing table of "verbs" 36
2106 After changing table of Posix names 22
2107
2108 Thanks to the folks working on Gregex for glib for this insight.
2109
2110 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2111 unicode-properties was also set.
2112
2113 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2114
2115 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2116 checked only for CRLF.
2117
2118 14. Added casts to pcretest.c to avoid compiler warnings.
2119
2120 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2121
2122 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2123 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2124 entirely. This removes changes made in 7 above.
2125
2126 17. The CMake files have been updated, and there is now more information about
2127 building with CMake in the NON-UNIX-USE document.
2128
2129
2130 Version 7.3 28-Aug-07
2131 ---------------------
2132
2133 1. In the rejigging of the build system that eventually resulted in 7.1, the
2134 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2135 brackets there is not right, since it causes compilers to look for an
2136 installed pcre.h, not the version that is in the source that is being
2137 compiled (which of course may be different). I have changed it back to:
2138
2139 #include "pcre.h"
2140
2141 I have a vague recollection that the change was concerned with compiling in
2142 different directories, but in the new build system, that is taken care of
2143 by the VPATH setting the Makefile.
2144
2145 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2146 when the subject happened to end in the byte 0x85 (e.g. if the last
2147 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2148 characters but of course it shouldn't be taken as a newline when it is part
2149 of another character. The bug was that, for an unlimited repeat of . in
2150 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2151 characters when looking for a newline.
2152
2153 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2154
2155 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2156 in debug output.
2157
2158 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2159 long printing in the pcrecpp unittest when running under MinGW.
2160
2161 6. ESC_K was left out of the EBCDIC table.
2162
2163 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2164 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2165 limit also applies to "virtual nesting" when a pattern is recursive, and in
2166 this case 1000 isn't so big. I have been able to remove this limit at the
2167 expense of backing off one optimization in certain circumstances. Normally,
2168 when pcre_exec() would call its internal match() function recursively and
2169 immediately return the result unconditionally, it uses a "tail recursion"
2170 feature to save stack. However, when a subpattern that can match an empty
2171 string has an unlimited repetition quantifier, it no longer makes this
2172 optimization. That gives it a stack frame in which to save the data for
2173 checking that an empty string has been matched. Previously this was taken
2174 from the 1000-entry workspace that had been reserved. So now there is no
2175 explicit limit, but more stack is used.
2176
2177 8. Applied Daniel's patches to solve problems with the import/export magic
2178 syntax that is required for Windows, and which was going wrong for the
2179 pcreposix and pcrecpp parts of the library. These were overlooked when this
2180 problem was solved for the main library.
2181
2182 9. There were some crude static tests to avoid integer overflow when computing
2183 the size of patterns that contain repeated groups with explicit upper
2184 limits. As the maximum quantifier is 65535, the maximum group length was
2185 set at 30,000 so that the product of these two numbers did not overflow a
2186 32-bit integer. However, it turns out that people want to use groups that
2187 are longer than 30,000 bytes (though not repeat them that many times).
2188 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2189 made it possible to implement the integer overflow checks in a much more
2190 dynamic way, which I have now done. The artificial limitation on group
2191 length has been removed - we now have only the limit on the total length of
2192 the compiled pattern, which depends on the LINK_SIZE setting.
2193
2194 10. Fixed a bug in the documentation for get/copy named substring when
2195 duplicate names are permitted. If none of the named substrings are set, the
2196 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2197 empty string.
2198
2199 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2200 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2201 because the ] is interpreted as the first data character and the
2202 terminating ] is not found. PCRE has been made compatible with Perl in this
2203 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2204 cause memory overwriting.
2205
2206 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2207 string has been matched (to stop an infinite loop). It was not recognizing
2208 a conditional subpattern that could match an empty string if that
2209 subpattern was within another subpattern. For example, it looped when
2210 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2211 condition was not nested. This bug has been fixed.
2212
2213 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2214 past the start of the subject in the presence of bytes with the top bit
2215 set, for example "\x8aBCD".
2216
2217 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2218 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2219
2220 14. Optimized (?!) to (*FAIL).
2221
2222 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2223 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2224 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2225 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2226 does: it's just the validity check that is more restrictive.
2227
2228 16. Inserted checks for integer overflows during escape sequence (backslash)
2229 processing, and also fixed erroneous offset values for syntax errors during
2230 backslash processing.
2231
2232 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2233 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2234
2235 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2236 caused an overrun.
2237
2238 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2239 something other than just ASCII characters) inside a group that had an
2240 unlimited repeat caused a loop at compile time (while checking to see
2241 whether the group could match an empty string).
2242
2243 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2244 [\P{Any}] did so. (Error in the code for printing property names.)
2245
2246 21. An orphan \E inside a character class could cause a crash.
2247
2248 22. A repeated capturing bracket such as (A)? could cause a wild memory
2249 reference during compilation.
2250
2251 23. There are several functions in pcre_compile() that scan along a compiled
2252 expression for various reasons (e.g. to see if it's fixed length for look
2253 behind). There were bugs in these functions when a repeated \p or \P was
2254 present in the pattern. These operators have additional parameters compared
2255 with \d, etc, and these were not being taken into account when moving along
2256 the compiled data. Specifically:
2257
2258 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2259 length.
2260
2261 (b) An item such as \pL+ within a repeated group could cause crashes or
2262 loops.
2263
2264 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2265 "reference to non-existent subpattern" error.
2266
2267 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2268
2269 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2270 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2271
2272 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2273
2274 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2275 character were causing crashes (broken optimization).
2276
2277 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2278 \p or \P) caused a compile-time loop.
2279
2280 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2281 break. For example, the unstudied pattern [\r\n]A does not match the string
2282 "\r\nA" because change 7.0/46 below moves the current point on by two
2283 characters after failing to match at the start. However, the pattern \nA
2284 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2285 the same is true. There doesn't seem any very clean way out of this, but
2286 what I have chosen to do makes the common cases work: PCRE now takes note
2287 of whether there can be an explicit match for \r or \n anywhere in the
2288 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2289 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2290 pattern has explicit CR or LF references.
2291
2292 29. Added (*CR) etc for changing newline setting at start of pattern.
2293
2294
2295 Version 7.2 19-Jun-07
2296 ---------------------
2297
2298 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2299 which is apparently normally available under Windows.
2300
2301 2. Re-jig the pcregrep tests with different newline settings in an attempt
2302 to make them independent of the local environment's newline setting.
2303
2304 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2305
2306 4. Some of the "internals" tests were previously cut out when the link size
2307 was not 2, because the output contained actual offsets. The recent new
2308 "Z" feature of pcretest means that these can be cut out, making the tests
2309 usable with all link sizes.
2310
2311 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2312 stack recursion. This gives a massive performance boost under BSD, but just
2313 a small improvement under Linux. However, it saves one field in the frame
2314 in all cases.
2315
2316 6. Added more features from the forthcoming Perl 5.10:
2317
2318 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2319 recursion call. It refers to the nth most recently opened parentheses.
2320
2321 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2322 to be opened parentheses.
2323
2324 (c) Conditions that refer to capturing parentheses can be specified
2325 relatively, for example, (?(-2)... or (?(+3)...
2326
2327 (d) \K resets the start of the current match so that everything before
2328 is not part of it.
2329
2330 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2331
2332 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2333 reference syntax.
2334
2335 (g) (?| introduces a group in which the numbering of parentheses in each
2336 alternative starts with the same number.
2337
2338 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2339
2340 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2341 PCRE_INFO_JCHANGED.
2342
2343 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2344 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2345 for detecting groups that can match an empty string.
2346
2347 9. A pattern with a very large number of alternatives (more than several
2348 hundred) was running out of internal workspace during the pre-compile
2349 phase, where pcre_compile() figures out how much memory will be needed. A
2350 bit of new cunning has reduced the workspace needed for groups with
2351 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2352 workspace instead of running out of the 4096 that are available.
2353
2354 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2355
2356 11. Applied patch from Google to remove an optimization that didn't quite work.
2357 The report of the bug said:
2358
2359 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2360 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2361 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2362
2363 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2364 it matched the wrong number of bytes.
2365
2366
2367 Version 7.1 24-Apr-07
2368 ---------------------
2369
2370 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2371 that is more "standard", making use of automake and other Autotools. There
2372 is some re-arrangement of the files and adjustment of comments consequent
2373 on this.
2374
2375 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2376 for recursive directory scanning broke on some systems because the files
2377 are not scanned in any specific order and on different systems the order
2378 was different. A call to "sort" has been inserted into RunGrepTest for the
2379 approprate test as a short-term fix. In the longer term there may be an
2380 alternative.
2381
2382 3. I had an email from Eric Raymond about problems translating some of PCRE's
2383 man pages to HTML (despite the fact that I distribute HTML pages, some
2384 people do their own conversions for various reasons). The problems
2385 concerned the use of low-level troff macros .br and .in. I have therefore
2386 removed all such uses from the man pages (some were redundant, some could
2387 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2388 HTML has been updated to handle .nf/.fi and to complain if it encounters
2389 .br or .in.
2390
2391 4. Updated comments in configure.ac that get placed in config.h.in and also
2392 arranged for config.h to be included in the distribution, with the name
2393 config.h.generic, for the benefit of those who have to compile without
2394 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2395
2396 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2397 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2398 makevp.bat for the current PCRE, using the additional files
2399 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2400
2401 6. A Windows user reported a minor discrepancy with test 2, which turned out
2402 to be caused by a trailing space on an input line that had got lost in his
2403 copy. The trailing space was an accident, so I've just removed it.
2404
2405 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2406 that is needed.
2407
2408 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2409 as "const" (a) because they are and (b) because it helps the PHP
2410 maintainers who have recently made a script to detect big data structures
2411 in the php code that should be moved to the .rodata section. I remembered
2412 to update Builducptable as well, so it won't revert if ucptable.h is ever
2413 re-created.
2414
2415 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2416 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2417 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2418 support is not required. This saves 1.5-2K of code, which is important in
2419 some applications.
2420
2421 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2422 so as not to refer to the tables, even though these functions will never be
2423 called when UTF-8 support is disabled. Otherwise there are problems with a
2424 shared library.
2425
2426 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2427
2428 (a) It was defining its arguments as char * instead of void *.
2429
2430 (b) It was assuming that all moves were upwards in memory; this was true
2431 a long time ago when I wrote it, but is no longer the case.
2432
2433 The emulated memove() is provided for those environments that have neither
2434 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2435 is clearly not the case, as these two bugs were recently reported.
2436
2437 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2438 and Detrail to create the HTML documentation, the .txt form of the man
2439 pages, and it removes trailing spaces from listed files. It also creates
2440 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2441 case, it wraps all the #defines with #ifndefs. This script should be run
2442 before "make dist".
2443
2444 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2445 with Unicode property support.
2446
2447 (a) For a maximizing quantifier, if the two different cases of the
2448 character were of different lengths in their UTF-8 codings (there are
2449 some cases like this - I found 11), and the matching function had to
2450 back up over a mixture of the two cases, it incorrectly assumed they
2451 were both the same length.
2452
2453 (b) When PCRE was configured to use the heap rather than the stack for
2454 recursion during matching, it was not correctly preserving the data for
2455 the other case of a UTF-8 character when checking ahead for a match
2456 while processing a minimizing repeat. If the check also involved
2457 matching a wide character, but failed, corruption could cause an
2458 erroneous result when trying to check for a repeat of the original
2459 character.
2460
2461 13. Some tidying changes to the testing mechanism:
2462
2463 (a) The RunTest script now detects the internal link size and whether there
2464 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2465 values substituted by "configure". (The RunGrepTest script already did
2466 this for UTF-8.) The configure.ac script no longer substitutes the
2467 relevant variables.
2468
2469 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2470 with length and offset values. This means that the output is different
2471 for different internal link sizes. Test 2 is skipped for link sizes
2472 other than 2 because of this, bypassing the problem. Unfortunately,
2473 there was also a test in test 3 (the locale tests) that used /B and
2474 failed for link sizes other than 2. Rather than cut the whole test out,
2475 I have added a new /Z option to pcretest that replaces the length and
2476 offset values with spaces. This is now used to make test 3 independent
2477 of link size. (Test 2 will be tidied up later.)
2478
2479 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2480 segmentation fault instead of returning the appropriate error message.
2481
2482 15. In multiline mode when the newline sequence was set to "any", the pattern
2483 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2484 This doesn't seem right; it now treats the CRLF combination as the line
2485 ending, and so does not match in that case. It's only a pattern such as ^$
2486 that would hit this one: something like ^ABC$ would have failed after \r
2487 and then tried again after \r\n.
2488
2489 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2490 in an attempt to make files that differ only in their line terminators
2491 compare equal. This works on Linux.
2492
2493 17. Under certain error circumstances pcregrep might try to free random memory
2494 as it exited. This is now fixed, thanks to valgrind.
2495
2496 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2497 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2498 was because its rules for how to advance for /g after matching an empty
2499 string at the end of a line did not allow for this case. They now check for
2500 it specially.
2501
2502 20. pcretest is supposed to handle patterns and data of any length, by
2503 extending its buffers when necessary. It was getting this wrong when the
2504 buffer for a data line had to be extended.
2505
2506 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2507 CRLF as a newline sequence.
2508
2509 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2510 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2511 I have nevertheless tidied it up.
2512
2513 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2514
2515 24. Added a man page for pcre-config.
2516
2517
2518 Version 7.0 19-Dec-06
2519 ---------------------
2520
2521 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2522 moving to gcc 4.1.1.
2523
2524 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2525 sys/time.h, which is documented as needed for this function. It doesn't
2526 seem to matter on Linux, but it showed up on some releases of OS X.
2527
2528 3. It seems that there are systems where bytes whose values are greater than
2529 127 match isprint() in the "C" locale. The "C" locale should be the
2530 default when a C program starts up. In most systems, only ASCII printing
2531 characters match isprint(). This difference caused the output from pcretest
2532 to vary, making some of the tests fail. I have changed pcretest so that:
2533
2534 (a) When it is outputting text in the compiled version of a pattern, bytes
2535 other than 32-126 are always shown as hex escapes.
2536
2537 (b) When it is outputting text that is a matched part of a subject string,
2538 it does the same, unless a different locale has been set for the match
2539 (using the /L modifier). In this case, it uses isprint() to decide.
2540
2541 4. Fixed a major bug that caused incorrect computation of the amount of memory
2542 required for a compiled pattern when options that changed within the
2543 pattern affected the logic of the preliminary scan that determines the
2544 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2545 that the computed length was too small. The symptoms of this bug were
2546 either the PCRE error "internal error: code overflow" from pcre_compile(),
2547 or a glibc crash with a message such as "pcretest: free(): invalid next
2548 size (fast)". Examples of patterns that provoked this bug (shown in
2549 pcretest format) are:
2550
2551 /(?-x: )/x
2552 /(?x)(?-x: \s*#\s*)/
2553 /((?i)[\x{c0}])/8
2554 /(?i:[\x{c0}])/8
2555
2556 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2557 is now done differently.
2558
2559 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2560 wrapper classes; (b) implement a new function in the C++ scanner that is
2561 more efficient than the old way of doing things because it avoids levels of
2562 recursion in the regex matching; (c) add a paragraph to the documentation
2563 for the FullMatch() function.
2564
2565 6. The escape sequence \n was being treated as whatever was defined as
2566 "newline". Not only was this contrary to the documentation, which states
2567 that \n is character 10 (hex 0A), but it also went horribly wrong when
2568 "newline" was defined as CRLF. This has been fixed.
2569
2570 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2571 was being set to -1 for the "end of line" case (supposedly a value that no
2572 character can have). Though this value is never used (the check for end of
2573 line is "zero bytes in current character"), it caused compiler complaints.
2574 I've changed it to 0xffffffff.
2575
2576 8. In pcre_version.c, the version string was being built by a sequence of
2577 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2578 string (as it is for production releases) called a macro with an empty
2579 argument. The C standard says the result of this is undefined. The gcc
2580 compiler treats it as an empty string (which was what was wanted) but it is
2581 reported that Visual C gives an error. The source has been hacked around to
2582 avoid this problem.
2583
2584 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2585 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2586 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2587 of them did).
2588
2589 10. Originally, pcretest opened its input and output without "b"; then I was
2590 told that "b" was needed in some environments, so it was added for release
2591 5.0 to both the input and output. (It makes no difference on Unix-like
2592 systems.) Later I was told that it is wrong for the input on Windows. I've
2593 now abstracted the modes into two macros, to make it easier to fiddle with
2594 them, and removed "b" from the input mode under Windows.
2595
2596 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2597
2598 12. Added -help and --help to pcretest as an official way of being reminded
2599 of the options.
2600
2601 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2602 and pcrecpp.cc because they annoy compilers at high warning levels.
2603
2604 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2605
2606 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2607 scripts are not C programs :-) and which was not noticed because it works
2608 on Linux.
2609
2610 16. pcretest is supposed to handle any length of pattern and data line (as one
2611 line or as a continued sequence of lines) by extending its input buffer if
2612 necessary. This feature was broken for very long pattern lines, leading to
2613 a string of junk being passed to pcre_compile() if the pattern was longer
2614 than about 50K.
2615
2616 17. I have done a major re-factoring of the way pcre_compile() computes the
2617 amount of memory needed for a compiled pattern. Previously, there was code
2618 that made a preliminary scan of the pattern in order to do this. That was
2619 OK when PCRE was new, but as the facilities have expanded, it has become
2620 harder and harder to keep it in step with the real compile phase, and there
2621 have been a number of bugs (see for example, 4 above). I have now found a
2622 cunning way of running the real compile function in a "fake" mode that
2623 enables it to compute how much memory it would need, while actually only
2624 ever using a few hundred bytes of working memory and without too many
2625 tests of the mode. This should make future maintenance and development
2626 easier. A side effect of this work is that the limit of 200 on the nesting
2627 depth of parentheses has been removed (though this was never a serious
2628 limitation, I suspect). However, there is a downside: pcre_compile() now
2629 runs more slowly than before (30% or more, depending on the pattern). I
2630 hope this isn't a big issue. There is no effect on runtime performance.
2631
2632 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2633 newline (only possible for the last line of a file) and it was a
2634 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2635
2636 19. Added additional timing features to pcretest. (1) The -tm option now times
2637 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2638 separate command line item, by a number that specifies the number of
2639 repeats to use when timing. The default is 50000; this gives better
2640 precision, but takes uncomfortably long for very large patterns.
2641
2642 20. Extended pcre_study() to be more clever in cases where a branch of a
2643 subpattern has no definite first character. For example, (a*|b*)[cd] would
2644 previously give no result from pcre_study(). Now it recognizes that the
2645 first character must be a, b, c, or d.
2646
2647 21. There was an incorrect error "recursive call could loop indefinitely" if
2648 a subpattern (or the entire pattern) that was being tested for matching an
2649 empty string contained only one non-empty item after a nested subpattern.
2650 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2651 incorrectly, because the \d was being skipped in the check.
2652
2653 22. The pcretest program now has a new pattern option /B and a command line
2654 option -b, which is equivalent to adding /B to every pattern. This causes
2655 it to show the compiled bytecode, without the additional information that
2656 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2657 is the same as /B/I).
2658
2659 23. A new optimization is now able automatically to treat some sequences such
2660 as a*b as a*+b. More specifically, if something simple (such as a character
2661 or a simple class like \d) has an unlimited quantifier, and is followed by
2662 something that cannot possibly match the quantified thing, the quantifier
2663 is automatically "possessified".
2664
2665 24. A recursive reference to a subpattern whose number was greater than 39
2666 went wrong under certain circumstances in UTF-8 mode. This bug could also
2667 have affected the operation of pcre_study().
2668
2669 25. Realized that a little bit of performance could be had by replacing
2670 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2671
2672 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2673
2674 27. Possessive quantifiers such as a++ were previously implemented by turning
2675 them into atomic groups such as ($>a+). Now they have their own opcodes,
2676 which improves performance. This includes the automatically created ones
2677 from 23 above.
2678
2679 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2680 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2681 the first matched character to be a colon. This applied both to named and
2682 numbered groups.
2683
2684 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2685
2686 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2687 building PCRE on Windows easier, so I have included it in the distribution.
2688
2689 31. There is now a check in pcretest against a ridiculously large number being
2690 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2691 loop, the loop is abandoned.
2692
2693 32. Forward references to subpatterns in conditions such as (?(2)...) where
2694 subpattern 2 is defined later cause pcre_compile() to search forwards in
2695 the pattern for the relevant set of parentheses. This search went wrong
2696 when there were unescaped parentheses in a character class, parentheses
2697 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2698
2699 33. "Subroutine" calls and backreferences were previously restricted to
2700 referencing subpatterns earlier in the regex. This restriction has now
2701 been removed.
2702
2703 34. Added a number of extra features that are going to be in Perl 5.10. On the
2704 whole, these are just syntactic alternatives for features that PCRE had
2705 previously implemented using the Python syntax or my own invention. The
2706 other formats are all retained for compatibility.
2707
2708 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2709 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2710 also .NET compatible.
2711
2712 (b) A recursion or subroutine call to a named group can now be defined as
2713 (?&name) as well as (?P>name).
2714
2715 (c) A backreference to a named group can now be defined as \k<name> or
2716 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2717 5.10, are also .NET compatible.
2718
2719 (d) A conditional reference to a named group can now use the syntax
2720 (?(<name>) or (?('name') as well as (?(name).
2721
2722 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2723 groups (named and numbered) that are never evaluated inline, but can be
2724 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2725 is always false. There may be only one alternative in such a group.
2726
2727 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2728 as the simple (?(R). The condition is true only if the most recent
2729 recursion is that of the given number or name. It does not search out
2730 through the entire recursion stack.
2731
2732 (g) The escape \gN or \g{N} has been added, where N is a positive or
2733 negative number, specifying an absolute or relative reference.
2734
2735 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2736 some "unreachable code" warnings.
2737
2738 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2739 things, this adds five new scripts.
2740
2741 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2742 There were also incompatibilities regarding the handling of \Q..\E inside
2743 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2744 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2745
2746 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2747 matches an empty string, and forcibly breaks the loop. There were bugs in
2748 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2749 against aaaa the result was just "a" rather than "aaaa", for example. Two
2750 separate and independent bugs (that affected different cases) have been
2751 fixed.
2752
2753 39. Refactored the code to abolish the use of different opcodes for small
2754 capturing bracket numbers. This is a tidy that I avoided doing when I
2755 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2756 The new approach is not only tidier, it makes it possible to reduce the
2757 memory needed to fix the previous bug (38).
2758
2759 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2760 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2761 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2762 mode.
2763
2764 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2765 report.
2766
2767 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2768 copy construction and assignment in the C++ wrapper.
2769
2770 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2771 couple of bugs that could have given wrong results in the "--newline=crlf"
2772 case.
2773
2774 44. Added a number of casts and did some reorganization of signed/unsigned int
2775 variables following suggestions from Dair Grant. Also renamed the variable
2776 "this" as "item" because it is a C++ keyword.
2777
2778 45. Arranged for dftables to add
2779
2780 #include "pcre_internal.h"
2781
2782 to pcre_chartables.c because without it, gcc 4.x may remove the array
2783 definition from the final binary if PCRE is built into a static library and
2784 dead code stripping is activated.
2785
2786 46. For an unanchored pattern, if a match attempt fails at the start of a
2787 newline sequence, and the newline setting is CRLF or ANY, and the next two
2788 characters are CRLF, advance by two characters instead of one.
2789
2790
2791 Version 6.7 04-Jul-06
2792 ---------------------
2793
2794 1. In order to handle tests when input lines are enormously long, pcretest has
2795 been re-factored so that it automatically extends its buffers when
2796 necessary. The code is crude, but this _is_ just a test program. The
2797 default size has been increased from 32K to 50K.
2798
2799 2. The code in pcre_study() was using the value of the re argument before
2800 testing it for NULL. (Of course, in any sensible call of the function, it
2801 won't be NULL.)
2802
2803 3. The memmove() emulation function in pcre_internal.h, which is used on
2804 systems that lack both memmove() and bcopy() - that is, hardly ever -
2805 was missing a "static" storage class specifier.
2806
2807 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2808 containing an extended class (one that cannot be represented by a bitmap
2809 because it contains high-valued characters or Unicode property items, e.g.
2810 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2811 pattern, but PCRE should not loop if you do not (it no longer does).
2812 [Detail: two cases were found: (a) a repeated subpattern containing an
2813 extended class; (b) a recursive reference to a subpattern that followed a
2814 previous extended class. It wasn't skipping over the extended class
2815 correctly when UTF-8 mode was not set.]
2816
2817 5. A negated single-character class was not being recognized as fixed-length
2818 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2819 compile error "lookbehind assertion is not fixed length".
2820
2821 6. The RunPerlTest auxiliary script was showing an unexpected difference
2822 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2823 write a Perl script that can interpret lines of an input file either as
2824 byte characters or as UTF-8, which is what "perltest" was being required to
2825 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2826 can't do is switch easily at run time between having the "use utf8;" pragma
2827 or not. In the end, I fudged it by using the RunPerlTest script to insert
2828 "use utf8;" explicitly for the UTF-8 tests.
2829
2830 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2831 the end of the subject string, contrary to the documentation and to what
2832 Perl does. This was true of both matching functions. Now it matches only at
2833 the start of the subject and immediately after *internal* newlines.
2834
2835 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2836 a pointer to an int instead of a pointer to an unsigned long int. This
2837 caused problems on 64-bit systems.
2838
2839 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2840 instance of the 'standard' template library not being so standard".
2841
2842 10. There was no check on the number of named subpatterns nor the maximum
2843 length of a subpattern name. The product of these values is used to compute
2844 the size of the memory block for a compiled pattern. By supplying a very
2845 long subpattern name and a large number of named subpatterns, the size
2846 computation could be caused to overflow. This is now prevented by limiting
2847 the length of names to 32 characters, and the number of named subpatterns
2848 to 10,000.
2849
2850 11. Subpatterns that are repeated with specific counts have to be replicated in
2851 the compiled pattern. The size of memory for this was computed from the
2852 length of the subpattern and the repeat count. The latter is limited to
2853 65535, but there was no limit on the former, meaning that integer overflow
2854 could in principle occur. The compiled length of a repeated subpattern is
2855 now limited to 30,000 bytes in order to prevent this.
2856
2857 12. Added the optional facility to have named substrings with the same name.
2858
2859 13. Added the ability to use a named substring as a condition, using the
2860 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2861 are numbers (not recommended). Forward references are permitted.
2862
2863 14. Added forward references in named backreferences (if you see what I mean).
2864
2865 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2866 pattern could run off the end of the subject. For example, the pattern
2867 "(?s)(.{1,5})"8 did this with the subject "ab".
2868
2869 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2870 PCRE_CASELESS was set when matching characters that were quantified with ?
2871 or *.
2872
2873 17. A character class other than a single negated character that had a minimum
2874 but no maximum quantifier - for example [ab]{6,} - was not handled
2875 correctly by pce_dfa_exec(). It would match only one character.
2876
2877 18. A valid (though odd) pattern that looked like a POSIX character
2878 class but used an invalid character after [ (for example [[,abc,]]) caused
2879 pcre_compile() to give the error "Failed: internal error: code overflow" or
2880 in some cases to crash with a glibc free() error. This could even happen if
2881 the pattern terminated after [[ but there just happened to be a sequence of
2882 letters, a binary zero, and a closing ] in the memory that followed.
2883
2884 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2885 over the years. Originally (before any Unicode support), just the bottom 8
2886 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2887 output from "man perlunicode" includes this:
2888
2889 The regular expression compiler produces polymorphic opcodes. That
2890 is, the pattern adapts to the data and automatically switches to
2891 the Unicode character scheme when presented with Unicode data--or
2892 instead uses a traditional byte scheme when presented with byte
2893 data.
2894
2895 Sadly, a wide octal escape does not cause a switch, and in a string with
2896 no other multibyte characters, these octal escapes are treated as before.
2897 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2898 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2899 Unicode string.
2900
2901 I have not perpetrated such confusion in PCRE. Up till now, it took just
2902 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2903 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2904 translate to the appropriate multibyte character.
2905
2906 29. Applied some refactoring to reduce the number of warnings from Microsoft
2907 and Borland compilers. This has included removing the fudge introduced
2908 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2909 a warning about an unused variable.
2910
2911 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2912 characters since release 4.0, because Perl (from release 5.004) does not.
2913 [Or at least, is documented not to: some releases seem to be in conflict
2914 with the documentation.] However, when a pattern was studied with
2915 pcre_study() and all its branches started with \s, PCRE still included VT
2916 as a possible starting character. Of course, this did no harm; it just
2917 caused an unnecessary match attempt.
2918
2919 22. Removed a now-redundant internal flag bit that recorded the fact that case
2920 dependency changed within the pattern. This was once needed for "required
2921 byte" processing, but is no longer used. This recovers a now-scarce options
2922 bit. Also moved the least significant internal flag bit to the most-
2923 significant bit of the word, which was not previously used (hangover from
2924 the days when it was an int rather than a uint) to free up another bit for
2925 the future.
2926
2927 23. Added support for CRLF line endings as well as CR and LF. As well as the
2928 default being selectable at build time, it can now be changed at runtime
2929 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2930 specify that it is scanning data with non-default line endings.
2931
2932 24. Changed the definition of CXXLINK to make it agree with the definition of
2933 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2934
2935 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2936 recursions. This makes a big different to stack usage for some patterns.
2937
2938 26. If a subpattern containing a named recursion or subroutine reference such
2939 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2940 the space required for the compiled pattern went wrong and gave too small a
2941 value. Depending on the environment, this could lead to "Failed: internal
2942 error: code overflow at offset 49" or "glibc detected double free or
2943 corruption" errors.
2944
2945 27. Applied patches from Google (a) to support the new newline modes and (b) to
2946 advance over multibyte UTF-8 characters in GlobalReplace.
2947
2948 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2949 difference for some implementation of PCRE in some Windows version.
2950
2951 29. Added some extra testing facilities to pcretest:
2952
2953 \q<number> in a data line sets the "match limit" value
2954 \Q<number> in a data line sets the "match recursion limt" value
2955 -S <number> sets the stack size, where <number> is in megabytes
2956
2957 The -S option isn't available for Windows.
2958
2959
2960 Version 6.6 06-Feb-06
2961 ---------------------
2962
2963 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2964 in pcreposix.h. I have copied the definition from pcre.h.
2965
2966 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2967 because pcre.h is no longer a built file.
2968
2969 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2970 not normally included in the compiled code.
2971
2972
2973 Version 6.5 01-Feb-06
2974 ---------------------
2975
2976 1. When using the partial match feature with pcre_dfa_exec(), it was not
2977 anchoring the second and subsequent partial matches at the new starting
2978 point. This could lead to incorrect results. For example, with the pattern
2979 /1234/, partially matching against "123" and then "a4" gave a match.
2980
2981 2. Changes to pcregrep:
2982
2983 (a) All non-match returns from pcre_exec() were being treated as failures
2984 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2985 error message is output. Some extra information is given for the
2986 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2987 probably the only errors that are likely to be caused by users (by
2988 specifying a regex that has nested indefinite repeats, for instance).
2989 If there are more than 20 of these errors, pcregrep is abandoned.
2990
2991 (b) A binary zero was treated as data while matching, but terminated the
2992 output line if it was written out. This has been fixed: binary zeroes
2993 are now no different to any other data bytes.
2994
2995 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2996 used to set a locale for matching. The --locale=xxxx long option has
2997 been added (no short equivalent) to specify a locale explicitly on the
2998 pcregrep command, overriding the environment variables.
2999
3000 (d) When -B was used with -n, some line numbers in the output were one less
3001 than they should have been.
3002
3003 (e) Added the -o (--only-matching) option.
3004
3005 (f) If -A or -C was used with -c (count only), some lines of context were
3006 accidentally printed for the final match.
3007
3008 (g) Added the -H (--with-filename) option.
3009
3010 (h) The combination of options -rh failed to suppress file names for files
3011 that were found from directory arguments.
3012
3013 (i) Added the -D (--devices) and -d (--directories) options.
3014
3015 (j) Added the -F (--fixed-strings) option.
3016
3017 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3018
3019 (l) Added the --colo(u)r option.
3020
3021 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3022 is not present by default.
3023
3024 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3025 items such as (?R) or (?1), when the recursion could match a number of
3026 alternatives. If it matched one of the alternatives, but subsequently,
3027 outside the recursion, there was a failure, the code tried to back up into
3028 the recursion. However, because of the way PCRE is implemented, this is not
3029 possible, and the result was an incorrect result from the match.
3030
3031 In order to prevent this happening, the specification of recursion has
3032 been changed so that all such subpatterns are automatically treated as
3033 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3034
3035 4. I had overlooked the fact that, in some locales, there are characters for
3036 which isalpha() is true but neither isupper() nor islower() are true. In
3037 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3038 and ordfeminine) are like this. This affected the treatment of \w and \W
3039 when they appeared in character classes, but not when they appeared outside
3040 a character class. The bit map for "word" characters is now created
3041 separately from the results of isalnum() instead of just taking it from the
3042 upper, lower, and digit maps. (Plus the underscore character, of course.)
3043
3044 5. The above bug also affected the handling of POSIX character classes such as
3045 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3046 permanent tables. Instead, the bit maps for such a class were previously
3047 created as the appropriate unions of the upper, lower, and digit bitmaps.
3048 Now they are created by subtraction from the [[:word:]] class, which has
3049 its own bitmap.
3050
3051 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3052 It is created by subtracting the vertical space characters (\x09, \x0a,
3053 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3054 subtraction was done in the overall bitmap for a character class, meaning
3055 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3056 be recognized. This bug has been fixed.
3057
3058 7. Patches from the folks at Google:
3059
3060 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3061 real life, but is still worth protecting against".
3062
3063 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3064 regular expressions".
3065
3066 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3067 have it.
3068
3069 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3070 "configure" and the latter not, in order to fix a problem somebody had
3071 with compiling the Arg class on HP-UX.
3072
3073 (e) Improve the error-handling of the C++ wrapper a little bit.
3074
3075 (f) New tests for checking recursion limiting.
3076
3077 8. The pcre_memmove() function, which is used only if the environment does not
3078 have a standard memmove() function (and is therefore rarely compiled),
3079 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3080 returning a result (though PCRE never actually uses the result).
3081
3082 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3083 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3084 returned instead of calling malloc() with an overflowing number that would
3085 most likely cause subsequent chaos.
3086
3087 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3088
3089 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3090 with this option is matched, the nmatch and pmatch options of regexec() are
3091 ignored.
3092
3093 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3094 provided in case anyone wants to the the POSIX interface with UTF-8
3095 strings.
3096
3097 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3098 C++ linking (needed for some HP-UX environments).
3099
3100 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3101 (unused parameter) and in the pcre_printint() function (omitted "default"
3102 switch label when the default is to do nothing).
3103
3104 15. Added some code to make it possible, when PCRE is compiled as a C++
3105 library, to replace subject pointers for pcre_exec() with a smart pointer
3106 class, thus making it possible to process discontinuous strings.
3107
3108 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3109 much the same function. They were added by different people who were trying
3110 to make PCRE easy to compile on non-Unix systems. It has been suggested
3111 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3112 for compiling on Windows systems. I have therefore replaced it with
3113 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3114 defaults to "extern" for C or "extern C" for C++, which works fine on
3115 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3116 SCOPE with something explicit in config.h. In addition:
3117
3118 (a) pcreposix.h still had just "extern" instead of either of these macros;
3119 I have replaced it with PCRE_DATA_SCOPE.
3120
3121 (b) Functions such as _pcre_xclass(), which are internal to the library,
3122 but external in the C sense, all had PCRE_EXPORT in their definitions.
3123 This is apparently wrong for the Windows case, so I have removed it.
3124 (It makes no difference on Unix-like systems.)
3125
3126 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3127 of recursive calls to match(). This is different to MATCH_LIMIT because
3128 that limits the total number of calls to match(), not all of which increase
3129 the depth of recursion. Limiting the recursion depth limits the amount of
3130 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3131 when PCRE is compiled, and changed at run time. A patch from Google adds
3132 this functionality to the C++ interface.
3133
3134 18. Changes to the handling of Unicode character properties:
3135
3136 (a) Updated the table to Unicode 4.1.0.
3137
3138 (b) Recognize characters that are not in the table as "Cn" (undefined).
3139
3140 (c) I revised the way the table is implemented to a much improved format
3141 which includes recognition of ranges. It now supports the ranges that
3142 are defined in UnicodeData.txt, and it also amalgamates other
3143 characters into ranges. This has reduced the number of entries in the
3144 table from around 16,000 to around 3,000, thus reducing its size
3145 considerably. I realized I did not need to use a tree structure after
3146 all - a binary chop search is just as efficient. Having reduced the
3147 number of entries, I extended their size from 6 bytes to 8 bytes to
3148 allow for more data.
3149
3150 (d) Added support for Unicode script names via properties such as \p{Han}.
3151
3152 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3153 matching that character.
3154
3155 20. When matching a repeated Unicode property with a minimum greater than zero,
3156 (for example \pL{2,}), PCRE could look past the end of the subject if it
3157 reached it while seeking the minimum number of characters. This could
3158 happen only if some of the characters were more than one byte long, because
3159 there is a check for at least the minimum number of bytes.
3160
3161 21. Refactored the implementation of \p and \P so as to be more general, to
3162 allow for more different types of property in future. This has changed the
3163 compiled form incompatibly. Anybody with saved compiled patterns that use
3164 \p or \P will have to recompile them.
3165
3166 22. Added "Any" and "L&" to the supported property types.
3167
3168 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3169 but give a compile time error if the value is greater than 0xff.
3170
3171 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3172 accidentally not being installed or uninstalled.
3173
3174 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3175 made were to insert the current release number. This seemed silly, because
3176 it made things harder for people building PCRE on systems that don't run
3177 "configure". I have turned pcre.h into a distributed file, no longer built
3178 by "configure", with the version identification directly included. There is
3179 no longer a pcre.h.in file.
3180
3181 However, this change necessitated a change to the pcre-config script as
3182 well. It is built from pcre-config.in, and one of the substitutions was the
3183 release number. I have updated configure.ac so that ./configure now finds
3184 the release number by grepping pcre.h.
3185
3186 26. Added the ability to run the tests under valgrind.
3187
3188
3189 Version 6.4 05-Sep-05
3190 ---------------------
3191
3192 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3193 "--" to be printed when multiple files were scanned, even when none of the
3194 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3195 consider it to be a bug, and have restored the previous behaviour.
3196
3197 2. A couple of code tidies to get rid of compiler warnings.
3198
3199 3. The pcretest program used to cheat by referring to symbols in the library
3200 whose names begin with _pcre_. These are internal symbols that are not
3201 really supposed to be visible externally, and in some environments it is
3202 possible to suppress them. The cheating is now confined to including
3203 certain files from the library's source, which is a bit cleaner.
3204
3205 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3206 file's purpose clearer.
3207
3208 5. Reorganized pcre_ucp_findchar().
3209
3210
3211 Version 6.3 15-Aug-05
3212 ---------------------
3213
3214 1. The file libpcre.pc.in did not have general read permission in the tarball.
3215
3216 2. There were some problems when building without C++ support:
3217
3218 (a) If C++ support was not built, "make install" and "make test" still
3219 tried to test it.
3220
3221 (b) There were problems when the value of CXX was explicitly set. Some
3222 changes have been made to try to fix these, and ...
3223
3224 (c) --disable-cpp can now be used to explicitly disable C++ support.
3225
3226 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3227 backslash in a target when C++ was disabled. This confuses some
3228 versions of "make", apparently. Using an intermediate variable solves
3229 this. (Same for CPP_LOBJ.)
3230
3231 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3232 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3233 necessary on certain architectures.
3234
3235 4. Added a setting of -export-symbols-regex to the link command to remove
3236 those symbols that are exported in the C sense, but actually are local
3237 within the library, and not documented. Their names all begin with
3238 "_pcre_". This is not a perfect job, because (a) we have to except some
3239 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3240 available (and never for static libraries). I have made a note to try to
3241 find a way round (a) in the future.
3242
3243
3244 Version 6.2 01-Aug-05
3245 ---------------------
3246
3247 1. There was no test for integer overflow of quantifier values. A construction
3248 such as {1111111111111111} would give undefined results. What is worse, if
3249 a minimum quantifier for a parenthesized subpattern overflowed and became
3250 negative, the calculation of the memory size went wrong. This could have
3251 led to memory overwriting.
3252
3253 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3254
3255 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3256 operating environments where this matters.
3257
3258 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3259 PCRE options from within the C++ wrapper.
3260
3261 5. Named capturing subpatterns were not being correctly counted when a pattern
3262 was compiled. This caused two problems: (a) If there were more than 100
3263 such subpatterns, the calculation of the memory needed for the whole
3264 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3265 back references of the form \12, where the number was greater than 9, were
3266 not recognized as back references, even though there were sufficient
3267 previous subpatterns.
3268
3269 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3270 versions of gcc, e.g. 2.95.4.
3271
3272
3273 Version 6.1 21-Jun-05
3274 ---------------------
3275
3276 1. There was one reference to the variable "posix" in pcretest.c that was not
3277 surrounded by "#if !defined NOPOSIX".
3278
3279 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3280 the cross-check on the old pcre_info() function, for the benefit of the
3281 cut-down version of PCRE that is currently imported into Exim.
3282
3283 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3284 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3285 patterns that start (?i) but I don't think that matters. The use of (?i) is
3286 just an example; this all applies to the other options as well.
3287
3288 4. Since libtool seems to echo the compile commands it is issuing, the output
3289 from "make" can be reduced a bit by putting "@" in front of each libtool
3290 compile command.
3291
3292 5. Patch from the folks at Google for configure.in to be a bit more thorough
3293 in checking for a suitable C++ installation before trying to compile the
3294 C++ stuff. This should fix a reported problem when a compiler was present,
3295 but no suitable headers.
3296
3297 6. The man pages all had just "PCRE" as their title. I have changed them to
3298 be the relevant file name. I have also arranged that these names are
3299 retained in the file doc/pcre.txt, which is a concatenation in text format
3300 of all the man pages except the little individual ones for each function.
3301
3302 7. The NON-UNIX-USE file had not been updated for the different set of source
3303 files that come with release 6. I also added a few comments about the C++
3304 wrapper.
3305
3306
3307 Version 6.0 07-Jun-05
3308 ---------------------
3309
3310 1. Some minor internal re-organization to help with my DFA experiments.
3311
3312 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3313 didn't matter for the library itself when fully configured, but did matter
3314 when compiling without UCP support, or within Exim, where the ucp files are
3315 not imported.
3316
3317 3. Refactoring of the library code to split up the various functions into
3318 different source modules. The addition of the new DFA matching code (see
3319 below) to a single monolithic source would have made it really too
3320 unwieldy, quite apart from causing all the code to be include in a
3321 statically linked application, when only some functions are used. This is
3322 relevant even without the DFA addition now that patterns can be compiled in
3323 one application and matched in another.
3324
3325 The downside of splitting up is that there have to be some external
3326 functions and data tables that are used internally in different modules of
3327 the library but which are not part of the API. These have all had their
3328 names changed to start with "_pcre_" so that they are unlikely to clash
3329 with other external names.
3330
3331 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3332 a different (DFA) algorithm. Although it is slower than the original
3333 function, it does have some advantages for certain types of matching
3334 problem.
3335
3336 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3337 including restarting after a partial match.
3338
3339 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3340 defined when compiling for Windows was sent to me. I have put it into the
3341 code, though I have no means of testing or verifying it.
3342
3343 7. Added the pcre_refcount() auxiliary function.
3344
3345 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3346 match before or at the first newline in the subject string. In pcretest,
3347 the /f option on a pattern can be used to set this.
3348
3349 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3350 would behave wrongly. This has been present in PCRE since release 4.0.
3351
3352 10. A number of changes to the pcregrep command:
3353
3354 (a) Refactored how -x works; insert ^(...)$ instead of setting
3355 PCRE_ANCHORED and checking the length, in preparation for adding
3356 something similar for -w.
3357
3358 (b) Added the -w (match as a word) option.
3359
3360 (c) Refactored the way lines are read and buffered so as to have more
3361 than one at a time available.
3362
3363 (d) Implemented a pcregrep test script.
3364
3365 (e) Added the -M (multiline match) option. This allows patterns to match
3366 over several lines of the subject. The buffering ensures that at least
3367 8K, or the rest of the document (whichever is the shorter) is available
3368 for matching (and similarly the previous 8K for lookbehind assertions).
3369
3370 (f) Changed the --help output so that it now says
3371
3372 -w, --word-regex(p)
3373
3374 instead of two lines, one with "regex" and the other with "regexp"
3375 because that confused at least one person since the short forms are the
3376 same. (This required a bit of code, as the output is generated
3377 automatically from a table. It wasn't just a text change.)
3378
3379 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3380 option but starts with a hyphen. Could be a pattern or a path name
3381 starting with a hyphen, for instance.
3382
3383 (h) "-" can be given as a file name to represent stdin.
3384
3385 (i) When file names are being printed, "(standard input)" is used for
3386 the standard input, for compatibility with GNU grep. Previously
3387 "<stdin>" was used.
3388
3389 (j) The option --label=xxx can be used to supply a name to be used for
3390 stdin when file names are being printed. There is no short form.
3391
3392 (k) Re-factored the options decoding logic because we are going to add
3393 two more options that take data. Such options can now be given in four
3394 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3395
3396 (l) Added the -A, -B, and -C options for requesting that lines of context
3397 around matches be printed.
3398
3399 (m) Added the -L option to print the names of files that do not contain
3400 any matching lines, that is, the complement of -l.
3401
3402 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3403 continue to scan other files.
3404
3405 (o) The -s option was incorrectly implemented. For compatibility with other
3406 greps, it now suppresses the error message for a non-existent or non-
3407 accessible file (but not the return code). There is a new option called
3408 -q that suppresses the output of matching lines, which was what -s was
3409 previously doing.
3410
3411 (p) Added --include and --exclude options to specify files for inclusion
3412 and exclusion when recursing.
3413
3414 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3415 Hopefully, it now does.
3416
3417 12. Missing cast in pcre_study().
3418
3419 13. Added an "uninstall" target to the makefile.
3420
3421 14. Replaced "extern" in the function prototypes in Makefile.in with
3422 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3423 world, but is set differently for Windows.
3424
3425 15. Added a second compiling function called pcre_compile2(). The only
3426 difference is that it has an extra argument, which is a pointer to an
3427 integer error code. When there is a compile-time failure, this is set
3428 non-zero, in addition to the error test pointer being set to point to an
3429 error message. The new argument may be NULL if no error number is required
3430 (but then you may as well call pcre_compile(), which is now just a
3431 wrapper). This facility is provided because some applications need a
3432 numeric error indication, but it has also enabled me to tidy up the way
3433 compile-time errors are handled in the POSIX wrapper.
3434
3435 16. Added VPATH=.libs to the makefile; this should help when building with one
3436 prefix path and installing with another. (Or so I'm told by someone who
3437 knows more about this stuff than I do.)
3438
3439 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3440 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3441 match everything, including newlines. This is not POSIX-compatible, but
3442 somebody wanted the feature. From pcretest it can be activated by using
3443 both the P and the s flags.
3444
3445 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3446
3447 19. libpcre.pc was being incorrectly installed as executable.
3448
3449 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3450 it now also looks for '\r' so that it will work unmodified on Windows.
3451
3452 21. Added Google's contributed C++ wrapper to the distribution.
3453
3454 22. Added some untidy missing memory free() calls in pcretest, to keep
3455 Electric Fence happy when testing.
3456
3457
3458
3459 Version 5.0 13-Sep-04
3460 ---------------------
3461
3462 1. Internal change: literal characters are no longer packed up into items
3463 containing multiple characters in a single byte-string. Each character
3464 is now matched using a separate opcode. However, there may be more than one
3465 byte in the character in UTF-8 mode.
3466
3467 2. The pcre_callout_block structure has two new fields: pattern_position and
3468 next_item_length. These contain the offset in the pattern to the next match
3469 item, and its length, respectively.
3470
3471 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3472 insertion of callouts before each pattern item. Added the /C option to
3473 pcretest to make use of this.
3474
3475 4. On the advice of a Windows user, the lines
3476
3477 #if defined(_WIN32) || defined(WIN32)
3478 _setmode( _fileno( stdout ), 0x8000 );
3479 #endif /* defined(_WIN32) || defined(WIN32) */
3480
3481 have been added to the source of pcretest. This apparently does useful
3482 magic in relation to line terminators.
3483
3484 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3485 for the benefit of those environments where the "b" makes a difference.
3486
3487 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3488 to know about it. I have put a hack into configure.in that adds in code
3489 to set GCC=yes if CC=icc. This seems to end up at a point in the
3490 generated configure script that is early enough to affect the setting of
3491 compiler options, which is what is needed, but I have no means of testing
3492 whether it really works. (The user who reported this had patched the
3493 generated configure script, which of course I cannot do.)
3494
3495 LATER: After change 22 below (new libtool files), the configure script
3496 seems to know about icc (and also ecc). Therefore, I have commented out
3497 this hack in configure.in.
3498
3499 7. Added support for pkg-config (2 patches were sent in).
3500
3501 8. Negated POSIX character classes that used a combination of internal tables
3502 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3503 [[:^ascii]]. Typically, they would match almost any characters. The other
3504 POSIX classes were not broken in this way.
3505
3506 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3507 to find the match, as PCRE was deluded into thinking that the match had to
3508 start at the start point or following a newline. The same bug applied to
3509 patterns with negative forward assertions or any backward assertions
3510 preceding ".*" at the start, unless the pattern required a fixed first
3511 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3512
3513 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3514 starting at the last subject character, bytes beyond the end of the subject
3515 string were read.
3516
3517 11. Renamed the variable "class" as "classbits" to make life easier for C++
3518 users. (Previously there was a macro definition, but it apparently wasn't
3519 enough.)
3520
3521 12. Added the new field "tables" to the extra data so that tables can be passed
3522 in at exec time, or the internal tables can be re-selected. This allows
3523 a compiled regex to be saved and re-used at a later time by a different
3524 program that might have everything at different addresses.
3525
3526 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3527 -R library as well as a -L library.
3528
3529 14. The debugging options of pcretest (-d on the command line or D on a
3530 pattern) showed incorrect output for anything following an extended class
3531 that contained multibyte characters and which was followed by a quantifier.
3532
3533 15. Added optional support for general category Unicode character properties
3534 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3535 support. It adds about 90K to the size of the library. The meanings of the
3536 inbuilt class escapes such as \d and \s have NOT been changed.
3537
3538 16. Updated pcredemo.c to include calls to free() to release the memory for the
3539 compiled pattern.
3540
3541 17. The generated file chartables.c was being created in the source directory
3542 instead of in the building directory. This caused the build to fail if the
3543 source directory was different from the building directory, and was
3544 read-only.
3545
3546 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3547 file. No doubt somebody will tell me if they don't make sense... Also added
3548 Dan Mooney's comments about building on OpenVMS.
3549
3550 19. Added support for partial matching via the PCRE_PARTIAL option for
3551 pcre_exec() and the \P data escape in pcretest.
3552
3553 20. Extended pcretest with 3 new pattern features:
3554
3555 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3556 write the compiled pattern to the file whose name is "rest-of-line".
3557 This is a straight binary dump of the data, with the saved pointer to
3558 the character tables forced to be NULL. The study data, if any, is
3559 written too. After writing, pcretest reads a new pattern.
3560
3561 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3562 compiled pattern from the given file. There must not be any
3563 occurrences of "<" in the file name (pretty unlikely); if there are,
3564 pcretest will instead treat the initial "<" as a pattern delimiter.
3565 After reading in the pattern, pcretest goes on to read data lines as
3566 usual.
3567
3568 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3569 and 16-bit fields in a compiled pattern, to simulate a pattern that
3570 was compiled on a host of opposite endianness.
3571
3572 21. The pcre-exec() function can now cope with patterns that were compiled on
3573 hosts of opposite endianness, with this restriction:
3574
3575 As for any compiled expression that is saved and used later, the tables
3576 pointer field cannot be preserved; the extra_data field in the arguments
3577 to pcre_exec() should be used to pass in a tables address if a value
3578 other than the default internal tables were used at compile time.
3579
3580 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3581 now diagnosed as an error. Previously, most of the time, a negative number
3582 would have been treated as zero, but if in addition "ovector" was passed as
3583 NULL, a crash could occur.
3584
3585 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3586 new versions from the libtool 1.5 distribution (the last one is a copy of
3587 a file called libtool.m4). This seems to have fixed the need to patch
3588 "configure" to support Darwin 1.3 (which I used to do). However, I still
3589 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3590 workstation).
3591
3592 24. Changed the PCRE licence to be the more standard "BSD" licence.
3593
3594
3595 Version 4.5 01-Dec-03
3596 ---------------------
3597
3598 1. There has been some re-arrangement of the code for the match() function so
3599 that it can be compiled in a version that does not call itself recursively.
3600 Instead, it keeps those local variables that need separate instances for
3601 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3602 needs to "recurse". Keeping track of where control must go is done by means
3603 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3604 hide most of the details from the main code, and operates only if
3605 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3606 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3607 operating.
3608
3609 To make it easier for callers to provide specially tailored get/free
3610 functions for this usage, two new functions, pcre_stack_malloc, and
3611 pcre_stack_free, are used. They are always called in strict stacking order,
3612 and the size of block requested is always the same.
3613
3614 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3615 PCRE has been compiled to use the stack or the heap for recursion. The
3616 -C option of pcretest uses this to show which version is compiled.
3617
3618 A new data escape \S, is added to pcretest; it causes the amounts of store
3619 obtained and freed by both kinds of malloc/free at match time to be added
3620 to the output.
3621
3622 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3623 what's available on my current Linux desktop machine.
3624
3625 3. When matching a UTF-8 string, the test for a valid string at the start has
3626 been extended. If start_offset is not zero, PCRE now checks that it points
3627 to a byte that is the start of a UTF-8 character. If not, it returns
3628 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3629 this is necessary because there may be backward assertions in the pattern.
3630 When matching the same subject several times, it may save resources to use
3631 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3632
3633 4. The code for checking the validity of UTF-8 strings has been tightened so
3634 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3635 containing "overlong sequences".
3636
3637 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3638 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3639 should have been "&", but it just so happened that all the cases this let
3640 through by mistake were picked up later in the function.
3641
3642 6. I had used a variable called "isblank" - this is a C99 function, causing
3643 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3644
3645 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3646 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3647 so the version is shown; (c) stop "make test" echoing "./RunTest".
3648
3649 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3650
3651 9. The prototype for memmove() for systems that don't have it was using
3652 size_t, but the inclusion of the header that defines size_t was later. I've
3653 moved the #includes for the C headers earlier to avoid this.
3654
3655 10. Added some adjustments to the code to make it easier to compiler on certain
3656 special systems:
3657
3658 (a) Some "const" qualifiers were missing.
3659 (b) Added the macro EXPORT before all exported functions; by default this
3660 is defined to be empty.
3661 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3662 that it reads its output file name as an argument instead of writing
3663 to the standard output and assuming this can be redirected.
3664
3665 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3666 class containing characters with values greater than 255, PCRE compilation
3667 went into a loop.
3668
3669 12. A recursive reference to a subpattern that was within another subpattern
3670 that had a minimum quantifier of zero caused PCRE to crash. For example,
3671 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3672 recursion. If the recursively-called subpattern itself had a zero repeat,
3673 that was OK.
3674
3675 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3676 buffer into which it was copied (for escape processing) was still set at
3677 1024, so long lines caused crashes.
3678
3679 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3680 "internal error: code overflow...". This applied to any character class
3681 that was followed by a possessive quantifier.
3682
3683 15. Modified the Makefile to add libpcre.la as a prerequisite for
3684 libpcreposix.la because I was told this is needed for a parallel build to
3685 work.
3686
3687 16. If a pattern that contained .* following optional items at the start was
3688 studied, the wrong optimizing data was generated, leading to matching
3689 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3690 matching string must start with a or b or c. The correct conclusion for
3691 this pattern is that a match can start with any character.
3692
3693
3694 Version 4.4 13-Aug-03
3695 ---------------------
3696
3697 1. In UTF-8 mode, a character class containing characters with values between
3698 127 and 255 was not handled correctly if the compiled pattern was studied.
3699 In fixing this, I have also improved the studying algorithm for such
3700 classes (slightly).
3701
3702 2. Three internal functions had redundant arguments passed to them. Removal
3703 might give a very teeny performance improvement.
3704
3705 3. Documentation bug: the value of the capture_top field in a callout is *one
3706 more than* the number of the hightest numbered captured substring.
3707
3708 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3709 in incorrectly linking with a previously installed version. They now link
3710 explicitly with libpcre.la.
3711
3712 5. configure.in no longer needs to recognize Cygwin specially.
3713
3714 6. A problem in pcre.in for Windows platforms is fixed.
3715
3716 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3717 pcretest, it used to include the size of the study block as part of its
3718 output. Unfortunately, the structure contains a field that has a different
3719 size on different hardware architectures. This meant that the tests that
3720 showed this size failed. As the block is currently always of a fixed size,
3721 this information isn't actually particularly useful in pcretest output, so
3722 I have just removed it.
3723
3724 8. Three pre-processor statements accidentally did not start in column 1.
3725 Sadly, there are *still* compilers around that complain, even though
3726 standard C has not required this for well over a decade. Sigh.
3727
3728 9. In pcretest, the code for checking callouts passed small integers in the
3729 callout_data field, which is a void * field. However, some picky compilers
3730 complained about the casts involved for this on 64-bit systems. Now
3731 pcretest passes the address of the small integer instead, which should get
3732 rid of the warnings.
3733
3734 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3735 both compile and run time, and gives an error if an invalid UTF-8 sequence
3736 is found. There is a option for disabling this check in cases where the
3737 string is known to be correct and/or the maximum performance is wanted.
3738
3739 11. In response to a bug report, I changed one line in Makefile.in from
3740
3741 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3742 to
3743 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3744
3745 to look similar to other lines, but I have no way of telling whether this
3746 is the right thing to do, as I do not use Windows. No doubt I'll get told
3747 if it's wrong...
3748
3749
3750 Version 4.3 21-May-03
3751 ---------------------
3752
3753 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3754 Makefile.
3755
3756 2. Some refactoring to improve the quality of the code:
3757
3758 (i) The utf8_table... variables are now declared "const".
3759
3760 (ii) The code for \cx, which used the "case flipping" table to upper case
3761 lower case letters, now just substracts 32. This is ASCII-specific,
3762 but the whole concept of \cx is ASCII-specific, so it seems
3763 reasonable.
3764
3765 (iii) PCRE was using its character types table to recognize decimal and
3766 hexadecimal digits in the pattern. This is silly, because it handles
3767 only 0-9, a-f, and A-F, but the character types table is locale-
3768 specific, which means strange things might happen. A private
3769 table is now used for this - though it costs 256 bytes, a table is
3770 much faster than multiple explicit tests. Of course, the standard
3771 character types table is still used for matching digits in subject
3772 strings against \d.
3773
3774 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3775 ending in _t are). So I've renamed it as ESC_tee.
3776
3777 3. The first argument for regexec() in the POSIX wrapper should have been
3778 defined as "const".
3779
3780 4. Changed pcretest to use malloc() for its buffers so that they can be
3781 Electric Fenced for debugging.
3782
3783 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3784 to read one or more bytes before the start of the subject string. Often this
3785 had no effect on PCRE's behaviour, but in some circumstances it could
3786 provoke a segmentation fault.
3787
3788 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3789 to try to read one or more bytes before the start of the subject string.
3790
3791 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3792 UTF-8 support could misbehave in various ways if the subject string
3793 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3794 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3795 back over UTF-8 characters.)
3796
3797
3798 Version 4.2 14-Apr-03
3799 ---------------------
3800
3801 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3802
3803 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3804 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3805 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3806 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3807 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3808 and BUILD_EXEEXT
3809 Note: automatic setting of the BUILD variables is not yet working
3810 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3811 compile-time but not at link-time
3812 [LINK]: use for linking executables only
3813 make different versions for Windows and non-Windows
3814 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3815 libraries
3816 [LINK_FOR_BUILD]: new variable
3817 [OBJEXT]: use throughout
3818 [EXEEXT]: use throughout
3819 <winshared>: new target
3820 <wininstall>: new target
3821 <dftables.o>: use native compiler
3822 <dftables>: use native linker
3823 <install>: handle Windows platform correctly
3824 <clean>: ditto
3825 <check>: ditto
3826 copy DLL to top builddir before testing
3827
3828 As part of these changes, -no-undefined was removed again. This was reported
3829 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3830 in any case.
3831
3832 3. Some tidies to get rid of compiler warnings:
3833
3834 . In the match_data structure, match_limit was an unsigned long int, whereas
3835 match_call_count was an int. I've made them both unsigned long ints.
3836
3837 . In pcretest the fact that a const uschar * doesn't automatically cast to
3838 a void * provoked a warning.
3839
3840 . Turning on some more compiler warnings threw up some "shadow" variables
3841 and a few more missing casts.
3842
3843 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3844 option, a class that contained a single character with a value between 128
3845 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3846
3847 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3848 option, a class that contained several characters, but with at least one
3849 whose value was between 128 and 255 caused PCRE to crash.
3850
3851
3852 Version 4.1 12-Mar-03
3853 ---------------------
3854
3855 1. Compiling with gcc -pedantic found a couple of places where casts were
3856 needed, and a string in dftables.c that was longer than standard compilers are
3857 required to support.
3858
3859 2. Compiling with Sun's compiler found a few more places where the code could
3860 be tidied up in order to avoid warnings.
3861
3862 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3863 first of these names is deprecated in the latest Autoconf in favour of the name
3864 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3865 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3866 analogy I have changed it to CFLAGS_FOR_BUILD.
3867
3868 4. Added -no-undefined to the linking command in the Makefile, because this is
3869 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3870 linking step for the pcreposix library.
3871
3872 5. PCRE was failing to diagnose the case of two named groups with the same
3873 name.
3874
3875 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3876 literal character that is needed in the subject for a match, and scans along to
3877 ensure that it is present before embarking on the full matching process. This
3878 saves time in cases of nested unlimited repeats that are never going to match.
3879 Problem: the scan can take a lot of time if the subject is very long (e.g.
3880 megabytes), thus penalizing straightforward matches. It is now done only if the
3881 amount of subject to be scanned is less than 1000 bytes.
3882
3883 7. A lesser problem with the same optimization is that it was recording the
3884 first character of an anchored pattern as "needed", thus provoking a search
3885 right along the subject, even when the first match of the pattern was going to
3886 fail. The "needed" character is now not set for anchored patterns, unless it
3887 follows something in the pattern that is of non-fixed length. Thus, it still
3888 fulfils its original purpose of finding quick non-matches in cases of nested
3889 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3890
3891
3892 Version 4.0 17-Feb-03
3893 ---------------------
3894
3895 1. If a comment in an extended regex that started immediately after a meta-item
3896 extended to the end of string, PCRE compiled incorrect data. This could lead to
3897 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3898
3899 2. Moved to autoconf 2.53 and libtool 1.4.2.
3900
3901 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3902 the special perltest8 script is no longer needed - all the tests can be run
3903 from a single perltest script.
3904
3905 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3906 by \s. It has now been removed in PCRE. This means it isn't recognized as
3907 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3908 class [:space:] *does* include VT, thereby creating a mess.
3909
3910 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3911 space and tab.
3912
3913 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3914 its new features into the main test script, reducing the number of scripts.
3915
3916 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3917 were backward compatible, and made the (?i) apply to the whole pattern, as if
3918 /i were given. Now it behaves more logically, and applies the option setting
3919 only to what follows. PCRE has been changed to follow suit. However, if it
3920 finds options settings right at the start of the pattern, it extracts them into
3921 the global options, as before. Thus, they show up in the info data.
3922
3923 8. Added support for the \Q...\E escape sequence. Characters in between are
3924 treated as literals. This is slightly different from Perl in that $ and @ are
3925 also handled as literals inside the quotes. In Perl, they will cause variable
3926 interpolation. Note the following examples:
3927
3928 Pattern PCRE matches Perl matches
3929
3930 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3931 \Qabc\$xyz\E abc\$xyz abc\$xyz
3932 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3933
3934 For compatibility with Perl, \Q...\E sequences are recognized inside character
3935 classes as well as outside them.
3936
3937 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3938 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3939 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3940 signed/unsigned warnings.
3941
3942 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3943 option for pcretest, so I've replaced it by a simple function that does just
3944 that job.
3945
3946 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3947 "pcregrep -".
3948
3949 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3950 Java package. This provides some syntactic sugar for simple cases of what my
3951 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3952 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3953 item, you can use this simplified notation. Note that only makes sense with
3954 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3955 greediness, whatever the setting of the PCRE_UNGREEDY option.
3956
3957 13. A change of greediness default within a pattern was not taking effect at
3958 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3959 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3960 was abstracted outside.
3961
3962 14. PCRE now supports the \G assertion. It is true when the current matching
3963 position is at the start point of the match. This differs from \A when the
3964 starting offset is non-zero. Used with the /g option of pcretest (or similar
3965 code), it works in the same way as it does for Perl's /g option. If all
3966 alternatives of a regex begin with \G, the expression is anchored to the start
3967 match position, and the "anchored" flag is set in the compiled expression.
3968
3969 15. Some bugs concerning the handling of certain option changes within patterns
3970 have been fixed. These applied to options other than (?ims). For example,
3971 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3972 been the other way round. Some of this was related to change 7 above.
3973
3974 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3975 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3976 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3977 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3978
3979 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3980 mode. Unlike ".", it always matches newline, whatever the setting of
3981 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3982 assertions. Perl allows it, but it doesn't (in general) work because it can't
3983 calculate the length of the lookbehind. At least, that's the case for Perl
3984 5.8.0 - I've been told they are going to document that it doesn't work in
3985 future.
3986
3987 18. Added an error diagnosis for escapes that PCRE does not support: these are
3988 \L, \l, \N, \P, \p, \U, \u, and \X.
3989
3990 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3991 reading past the end of the pattern in cases such as /[abcd/.
3992
3993 20. PCRE was getting more memory than necessary for patterns with classes that
3994 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3995
3996 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3997 compiling PCRE for use with Virtual Pascal.
3998
3999 22. Small fix to the Makefile to make it work properly if the build is done
4000 outside the source tree.
4001
4002 23. Added a new extension: a condition to go with recursion. If a conditional
4003 subpattern starts with (?(R) the "true" branch is used if recursion has
4004 happened, whereas the "false" branch is used only at the top level.
4005
4006 24. When there was a very long string of literal characters (over 255 bytes
4007 without UTF support, over 250 bytes with UTF support), the computation of how
4008 much memory was required could be incorrect, leading to segfaults or other
4009 strange effects.
4010
4011 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4012 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4013 there was a subsequent back reference to those brackets. This meant that, for
4014 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4015 possible to check for precisely this case. All we can do is abandon the
4016 optimization if .* occurs inside capturing brackets when there are any back
4017 references whatsoever. (See below for a better fix that came later.)
4018
4019 26. The handling of the optimization for finding the first character of a
4020 non-anchored pattern, and for finding a character that is required later in the
4021 match were failing in some cases. This didn't break the matching; it just
4022 failed to optimize when it could. The way this is done has been re-implemented.
4023
4024 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4025
4026 28. Added a new feature that provides some of the functionality that Perl
4027 provides with (?{...}). The facility is termed a "callout". The way it is done
4028 in PCRE is for the caller to provide an optional function, by setting
4029 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4030 global variable. By default it is unset, which disables all calling out. To get
4031 the function called, the regex must include (?C) at appropriate points. This
4032 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4033 This provides a means of identifying different callout points. When PCRE
4034 reaches such a point in the regex, if pcre_callout has been set, the external
4035 function is called. It is provided with data in a structure called
4036 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4037 matching continues; if it returns a non-zero value, the match at the current
4038 point fails. However, backtracking will occur if possible. [This was changed
4039 later and other features added - see item 49 below.]
4040
4041 29. pcretest is upgraded to test the callout functionality. It provides a
4042 callout function that displays information. By default, it shows the start of
4043 the match and the current position in the text. There are some new data escapes
4044 to vary what happens:
4045
4046 \C+ in addition, show current contents of captured substrings
4047 \C- do not supply a callout function
4048 \C!n return 1 when callout number n is reached
4049 \C!n!m return 1 when callout number n is reached for the mth time
4050
4051 30. If pcregrep was called with the -l option and just a single file name, it
4052 output "<stdin>" if a match was found, instead of the file name.
4053
4054 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4055 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4056 pcre_exec(). This saves a malloc/free per call. The default value of
4057 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4058 when configuring.
4059
4060 32. The default maximum size of a compiled pattern is 64K. There have been a
4061 few cases of people hitting this limit. The code now uses macros to handle the
4062 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4063 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4064 configuring. Tests 2 and 5 work only with 2-byte links because they output
4065 debugging information about compiled patterns.
4066
4067 33. Internal code re-arrangements:
4068
4069 (a) Moved the debugging function for printing out a compiled regex into
4070 its own source file (printint.c) and used #include to pull it into
4071 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4072 separate copies.
4073
4074 (b) Defined the list of op-code names for debugging as a macro in
4075 internal.h so that it is next to the definition of the opcodes.
4076
4077 (c) Defined a table of op-code lengths for simpler skipping along compiled
4078 code. This is again a macro in internal.h so that it is next to the
4079 definition of the opcodes.
4080
4081 34. Added support for recursive calls to individual subpatterns, along the
4082 lines of Robin Houston's patch (but implemented somewhat differently).
4083
4084 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4085 allow it to read and process whole directories in Win32. This code was
4086 contributed by Lionel Fourquaux; it has not been tested by me.
4087
4088 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4089 used to name a group. Names consist of alphanumerics and underscores, and must
4090 be unique. Back references use the syntax (?P=name) and recursive calls use
4091 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4092 numbers. The function pcre_fullinfo() can be used after compilation to extract
4093 a name/number map. There are three relevant calls:
4094
4095 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4096 PCRE_INFO_NAMECOUNT yields the number of entries
4097 PCRE_INFO_NAMETABLE yields a pointer to the map.
4098
4099 The map is a vector of fixed-size entries. The size of each entry depends on
4100 the length of the longest name used. The first two bytes of each entry are the
4101 group number, most significant byte first. There follows the corresponding
4102 name, zero terminated. The names are in alphabetical order.
4103
4104 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4105 case instead of 255. Making it the same both with and without UTF-8 support
4106 means that the same test output works with both.
4107
4108 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4109 calling malloc() with a zero argument.
4110
4111 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4112 optimization. I've improved things by keeping a bitmap of backreferences with
4113 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4114 fact referenced, the optimization can be applied. It is unlikely that a
4115 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4116 the match to follow \n) will appear inside brackets with a number greater than
4117 31, but if it does, any back reference > 31 suppresses the optimization.
4118
4119 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4120 of disabling numbered capturing parentheses. Any opening parenthesis that is
4121 not followed by ? behaves as if it were followed by ?: but named parentheses
4122 can still be used for capturing (and they will acquire numbers in the usual
4123 way).
4124
4125 41. Redesigned the return codes from the match() function into yes/no/error so
4126 that errors can be passed back from deep inside the nested calls. A malloc
4127 failure while inside a recursive subpattern call now causes the
4128 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4129
4130 42. It is now possible to set a limit on the number of times the match()
4131 function is called in a call to pcre_exec(). This facility makes it possible to
4132 limit the amount of recursion and backtracking, though not in a directly
4133 obvious way, because the match() function is used in a number of different
4134 circumstances. The count starts from zero for each position in the subject
4135 string (for non-anchored patterns). The default limit is, for compatibility, a
4136 large number, namely 10 000 000. You can change this in two ways:
4137
4138 (a) When configuring PCRE before making, you can use --with-match-limit=n
4139 to set a default value for the compiled library.
4140
4141 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4142 a different value is set. See 45 below.
4143
4144 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4145
4146 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4147 of things that can be changed at compile time. The first argument specifies
4148 what is wanted and the second points to where the information is to be placed.
4149 The current list of available information is:
4150
4151 PCRE_CONFIG_UTF8
4152
4153 The output is an integer that is set to one if UTF-8 support is available;
4154 otherwise it is set to zero.
4155
4156 PCRE_CONFIG_NEWLINE
4157
4158 The output is an integer that it set to the value of the code that is used for
4159 newline. It is either LF (10) or CR (13).
4160
4161 PCRE_CONFIG_LINK_SIZE
4162
4163 The output is an integer that contains the number of bytes used for internal
4164 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4165
4166 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4167
4168 The output is an integer that contains the threshold above which the POSIX
4169 interface uses malloc() for output vectors. See item 31 above.
4170
4171 PCRE_CONFIG_MATCH_LIMIT
4172
4173 The output is an unsigned integer that contains the default limit of the number
4174 of match() calls in a pcre_exec() execution. See 42 above.
4175
4176 44. pcretest has been upgraded by the addition of the -C option. This causes it
4177 to extract all the available output from the new pcre_config() function, and to
4178 output it. The program then exits immediately.
4179
4180 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4181 order to support additional features. One way would have been to define
4182 pcre_exec2() (for example) with extra arguments, but this would not have been
4183 extensible, and would also have required all calls to the original function to
4184 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4185 is used for passing in "extra" data from pcre_study().
4186
4187 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4188 contains the following fields:
4189
4190 flags a bitmap indicating which of the following fields are set
4191 study_data opaque data from pcre_study()
4192 match_limit a way of specifying a limit on match() calls for a specific
4193 call to pcre_exec()
4194 callout_data data for callouts (see 49 below)
4195
4196 The flag bits are also defined in pcre.h, and are
4197
4198 PCRE_EXTRA_STUDY_DATA
4199 PCRE_EXTRA_MATCH_LIMIT
4200 PCRE_EXTRA_CALLOUT_DATA
4201
4202 The pcre_study() function now returns one of these new pcre_extra blocks, with
4203 the actual study data pointed to by the study_data field, and the
4204 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4205 before. That is, this change is entirely upwards-compatible and requires no
4206 change to existing code.
4207
4208 If you want to pass in additional data to pcre_exec(), you can either place it
4209 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4210 block.
4211
4212 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4213 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4214 times with different match limits, until it finds the minimum value needed for
4215 pcre_exec() to complete. The value is then output. This can be instructive; for
4216 most simple matches the number is quite small, but for pathological cases it
4217 gets very large very quickly.
4218
4219 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4220 returns the size of the data block pointed to by the study_data field in a
4221 pcre_extra block, that is, the value that was passed as the argument to
4222 pcre_malloc() when PCRE was getting memory in which to place the information
4223 created by pcre_study(). The fourth argument should point to a size_t variable.
4224 pcretest has been extended so that this information is shown after a successful
4225 pcre_study() call when information about the compiled regex is being displayed.
4226
4227 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4228 because what follows is always an absolute path. (Later: it turns out that this
4229 is more than cosmetic for MinGW, because it doesn't like empty path
4230 components.)
4231
4232 49. Some changes have been made to the callout feature (see 28 above):
4233
4234 (i) A callout function now has three choices for what it returns:
4235
4236 0 => success, carry on matching
4237 > 0 => failure at this point, but backtrack if possible
4238 < 0 => serious error, return this value from pcre_exec()
4239
4240 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4241 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4242 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4243 use by callout functions. It will never be used by PCRE itself.
4244
4245 (ii) The pcre_extra structure (see 45 above) has a void * field called
4246 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4247 pcre_callout_block structure has a field of the same name. The contents of
4248 the field passed in the pcre_extra structure are passed to the callout
4249 function in the corresponding field in the callout block. This makes it
4250 easier to use the same callout-containing regex from multiple threads. For
4251 testing, the pcretest program has a new data escape
4252
4253 \C*n pass the number n (may be negative) as callout_data
4254
4255 If the callout function in pcretest receives a non-zero value as
4256 callout_data, it returns that value.
4257
4258 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4259 there were some redundant $(CFLAGS) in commands that are now specified as
4260 $(LINK), which already includes $(CFLAGS).
4261
4262 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4263 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4264 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4265 one-byte characters throughout. Note that case-insensitive matching applies
4266 only to characters whose values are less than 256. PCRE doesn't support the
4267 notion of cases for higher-valued characters.
4268
4269 (i) A character class whose characters are all within 0-255 is handled as
4270 a bit map, and the map is inverted for negative classes. Previously, a
4271 character > 255 always failed to match such a class; however it should
4272 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4273
4274 (ii) A negated character class with a single character < 255 is coded as
4275 "not this character" (OP_NOT). This wasn't working properly when the test
4276 character was multibyte, either singly or repeated.
4277
4278 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4279 mode, for example: \x{100}{2,3}.
4280
4281 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4282 singly or repeated) now correctly test multibyte characters. However,
4283 PCRE doesn't recognize any characters with values greater than 255 as
4284 digits, spaces, or word characters. Such characters always match \D, \S,
4285 and \W, and never match \d, \s, or \w.
4286
4287 (v) Classes may now contain characters and character ranges with values
4288 greater than 255. For example: [ab\x{100}-\x{400}].
4289
4290 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4291 PCRE in UTF-8 mode.
4292
4293 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4294 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4295 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4296 value.)
4297
4298 53. The single man page has become too large. I have therefore split it up into
4299 a number of separate man pages. These also give rise to individual HTML pages;
4300 these are now put in a separate directory, and there is an index.html page that
4301 lists them all. Some hyperlinking between the pages has been installed.
4302
4303 54. Added convenience functions for handling named capturing parentheses.
4304
4305 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4306 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4307 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4308 are faulted.
4309
4310 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4311 calling configure. These values are used when compiling the dftables.c program
4312 which is run to generate the source of the default character tables. They
4313 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4314 you will need to set these values.
4315
4316 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4317
4318
4319 Version 3.9 02-Jan-02
4320 ---------------------
4321
4322 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4323
4324 2. If --disable-static was given, the building process failed when trying to
4325 build pcretest and pcregrep. (For some reason it was using libtool to compile
4326 them, which is not right, as they aren't part of the library.)
4327
4328
4329 Version 3.8 18-Dec-01
4330 ---------------------
4331
4332 1. The experimental UTF-8 code was completely screwed up. It was packing the
4333 bytes in the wrong order. How dumb can you get?
4334
4335
4336 Version 3.7 29-Oct-01
4337 ---------------------
4338
4339 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4340 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4341 this didn't happen under Solaris 8, where I normally test things.
4342
4343 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4344 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4345 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4346 this fix an hour or so after the initial 3.7 release.)
4347
4348
4349 Version 3.6 23-Oct-01
4350 ---------------------
4351
4352 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4353 offsets passed as NULL with zero offset count.
4354
4355 2. The config.guess and config.sub files had not been updated when I moved to
4356 the latest autoconf.
4357
4358
4359 Version 3.5 15-Aug-01
4360 ---------------------
4361
4362 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4363 had been forgotten.
4364
4365 2. By using declared but undefined structures, we can avoid using "void"
4366 definitions in pcre.h while keeping the internal definitions of the structures
4367 private.
4368
4369 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4370 user point of view, this means that both static and shared libraries are built
4371 by default, but this can be individually controlled. More of the work of
4372 handling this static/shared cases is now inside libtool instead of PCRE's make
4373 file.
4374
4375 4. The pcretest utility is now installed along with pcregrep because it is
4376 useful for users (to test regexs) and by doing this, it automatically gets
4377 relinked by libtool. The documentation has been turned into a man page, so
4378 there are now .1, .txt, and .html versions in /doc.
4379
4380 5. Upgrades to pcregrep:
4381 (i) Added long-form option names like gnu grep.
4382 (ii) Added --help to list all options with an explanatory phrase.
4383 (iii) Added -r, --recursive to recurse into sub-directories.
4384 (iv) Added -f, --file to read patterns from a file.
4385
4386 6. pcre_exec() was referring to its "code" argument before testing that
4387 argument for NULL (and giving an error if it was NULL).
4388
4389 7. Upgraded Makefile.in to allow for compiling in a different directory from
4390 the source directory.
4391
4392 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4393 options bits, the pointer it was passed was to an int instead of to an unsigned
4394 long int. This mattered only on 64-bit systems.
4395
4396 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4397 generated) instead of pcre.in, which it its source. Also made the same change
4398 in several of the .c files.
4399
4400 10. A new release of gcc defines printf() as a macro, which broke pcretest
4401 because it had an ifdef in the middle of a string argument for printf(). Fixed
4402 by using separate calls to printf().
4403
4404 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4405 script, to force use of CR or LF instead of \n in the source. On non-Unix
4406 systems, the value can be set in config.h.
4407
4408 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4409 absolute limit. Changed the text of the error message to make this clear, and
4410 likewise updated the man page.
4411
4412 13. The limit of 99 on the number of capturing subpatterns has been removed.
4413 The new limit is 65535, which I hope will not be a "real" limit.
4414
4415
4416 Version 3.4 22-Aug-00
4417 ---------------------
4418
4419 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4420
4421 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4422
4423
4424 Version 3.3 01-Aug-00
4425 ---------------------
4426
4427 1. If an octal character was given, but the value was greater than \377, it
4428 was not getting masked to the least significant bits, as documented. This could
4429 lead to crashes in some systems.
4430
4431 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4432 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4433
4434 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4435 These just pass their arguments on to (pcre_free)(), but they are provided
4436 because some uses of PCRE bind it to non-C systems that can call its functions,
4437 but cannot call free() or pcre_free() directly.
4438
4439 4. Add "make test" as a synonym for "make check". Corrected some comments in
4440 the Makefile.
4441
4442 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4443 Makefile.
4444
4445 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4446 command called pgrep for grepping around the active processes.
4447
4448 7. Added the beginnings of support for UTF-8 character strings.
4449
4450 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4451 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4452 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4453 out for the ar command.)
4454
4455
4456 Version 3.2 12-May-00
4457 ---------------------
4458
4459 This is purely a bug fixing release.
4460
4461 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4462 of ZA. This was just one example of several cases that could provoke this bug,
4463 which was introduced by change 9 of version 2.00. The code for breaking
4464 infinite loops after an iteration that matches an empty string was't working
4465 correctly.
4466
4467 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4468 when matched against abbab (for example). After matching an empty string, it
4469 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4470 caused it to match further down the string than it should.
4471
4472 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4473 was there because it doesn't seem to be needed, and it causes trouble on some
4474 systems, as it is not a Standard C header. It has been removed.
4475
4476 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4477 were reported on the Macintosh. The changes were from
4478
4479 while ((c = *(++ptr)) != 0 && c != '\n');
4480 to
4481 while ((c = *(++ptr)) != 0 && c != '\n') ;
4482
4483 Totally extraordinary, but if that's what it takes...
4484
4485 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4486 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4487 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4488 assumes the way PCRE uses memmove() (always moving upwards).
4489
4490 6. PCRE is being used in one environment where strchr() is not available. There
4491 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4492 faster code anyway.
4493
4494
4495 Version 3.1 09-Feb-00
4496 ---------------------
4497
4498 The only change in this release is the fixing of some bugs in Makefile.in for
4499 the "install" target:
4500
4501 (1) It was failing to install pcreposix.h.
4502
4503 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4504
4505
4506 Version 3.0 01-Feb-00
4507 ---------------------
4508
4509 1. Add support for the /+ modifier to perltest (to output $` like it does in
4510 pcretest).
4511
4512 2. Add support for the /g modifier to perltest.
4513
4514 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4515 matches null strings.
4516
4517 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4518 pattern. Perl treats empty patterns specially - it reuses the most recent
4519 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4520 effect.
4521
4522 5. The POSIX interface was broken in that it was just handing over the POSIX
4523 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4524 required a bigger vector, with some working space on the end. This means that
4525 the POSIX wrapper now has to get and free some memory, and copy the results.
4526
4527 6. Added some simple autoconf support, placing the test data and the
4528 documentation in separate directories, re-organizing some of the
4529 information files, and making it build pcre-config (a GNU standard). Also added
4530 libtool support for building PCRE as a shared library, which is now the
4531 default.
4532
4533 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4534 09 are not valid octal constants. Single digits will be used for minor values
4535 less than 10.
4536
4537 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4538 existing programs that set these in the POSIX interface can use PCRE without
4539 modification.
4540
4541 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4542 return all that pcre_info() returns, plus additional data. The pcre_info()
4543 function is retained for compatibility, but is considered to be obsolete.
4544
4545 10. Added experimental recursion feature (?R) to handle one common case that
4546 Perl 5.6 will be able to do with (?p{...}).
4547
4548 11. Added support for POSIX character classes like [:alpha:], which Perl is
4549 adopting.
4550
4551
4552 Version 2.08 31-Aug-99
4553 ----------------------
4554
4555 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4556 trying to match at the startoffset position, but instead was moving forward to
4557 the next newline as if a previous match had failed.
4558
4559 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4560 and could get into a loop if a null string was matched other than at the start
4561 of the subject.
4562
4563 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4564 be distinguished at compile time, and for completeness also added PCRE_DATE.
4565
4566 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4567 in GnuWin32 environments.
4568
4569
4570 Version 2.07 29-Jul-99
4571 ----------------------
4572
4573 1. The documentation is now supplied in plain text form and HTML as well as in
4574 the form of man page sources.
4575
4576 2. C++ compilers don't like assigning (void *) values to other pointer types.
4577 In particular this affects malloc(). Although there is no problem in Standard
4578 C, I've put in casts to keep C++ compilers happy.
4579
4580 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4581 should be (const char *).
4582
4583 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4584 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4585 However, I haven't made this a standard facility. The documentation doesn't
4586 mention it, and the Makefile doesn't support it.
4587
4588 5. The Makefile now contains an "install" target, with editable destinations at
4589 the top of the file. The pcretest program is not installed.
4590
4591 6. pgrep -V now gives the PCRE version number and date.
4592
4593 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4594 causing the entire string to be ignored, instead of just the last character.
4595
4596 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4597 non-matching string, it can take a very, very long time, even for strings of
4598 quite modest length, because of the nested recursion. PCRE now does better in
4599 some of these cases. It does this by remembering the last required literal
4600 character in the pattern, and pre-searching the subject to ensure it is present
4601 before running the real match. In other words, it applies a heuristic to detect
4602 some types of certain failure quickly, and in the above example, if presented
4603 with a string that has no trailing " it gives "no match" very quickly.
4604
4605 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4606 other alternatives are tried instead.
4607
4608
4609 Version 2.06 09-Jun-99
4610 ----------------------
4611
4612 1. Change pcretest's output for amount of store used to show just the code
4613 space, because the remainder (the data block) varies in size between 32-bit and
4614 64-bit systems.
4615
4616 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4617 start matching at. This allows lookbehinds to work when searching for multiple
4618 occurrences in a string.
4619
4620 3. Added additional options to pcretest for testing multiple occurrences:
4621
4622 /+ outputs the rest of the string that follows a match
4623 /g loops for multiple occurrences, using the new startoffset argument
4624 /G loops for multiple occurrences by passing an incremented pointer
4625
4626 4. PCRE wasn't doing the "first character" optimization for patterns starting
4627 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4628 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4629 the letter 'x'. On long subject strings, this gives a significant speed-up.
4630
4631
4632 Version 2.05 21-Apr-99
4633 ----------------------
4634
4635 1. Changed the type of magic_number from int to long int so that it works
4636 properly on 16-bit systems.
4637
4638 2. Fixed a bug which caused patterns starting with .* not to work correctly
4639 when the subject string contained newline characters. PCRE was assuming
4640 anchoring for such patterns in all cases, which is not correct because .* will
4641 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4642 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4643 must be retried after every newline in the subject.
4644
4645
4646 Version 2.04 18-Feb-99
4647 ----------------------
4648
4649 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4650 computation of the store needed to hold the pattern was incorrect (too large).
4651 If such patterns were nested a few deep, this could multiply and become a real
4652 problem.
4653
4654 2. Added /M option to pcretest to show the memory requirement of a specific
4655 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4656
4657 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4658 compiled in such a way that the backtracking after subsequent failure was
4659 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4660 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4661
4662
4663 Version 2.03 02-Feb-99
4664 ----------------------
4665
4666 1. Fixed typo and small mistake in man page.
4667
4668 2. Added 4th condition (GPL supersedes if conflict) and created separate
4669 LICENCE file containing the conditions.
4670
4671 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4672 Perl, that is the internal \ allows the delimiter to be included in the
4673 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4674 the final delimiter, add \ to the end of the pattern (to test the error).
4675
4676 4. Added the convenience functions for extracting substrings after a successful
4677 match. Updated pcretest to make it able to test these functions.
4678
4679
4680 Version 2.02 14-Jan-99
4681 ----------------------
4682
4683 1. Initialized the working variables associated with each extraction so that
4684 their saving and restoring doesn't refer to uninitialized store.
4685
4686 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4687 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4688 fix the problem.
4689
4690 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4691 calls, and wasn't printing the correct value for compiling calls. Increased the
4692 default value of LOOPREPEAT, and the number of significant figures in the
4693 times.
4694
4695 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4696
4697 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4698 a building problem on Windows NT with a FAT file system.
4699
4700
4701 Version 2.01 21-Oct-98
4702 ----------------------
4703
4704 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4705 to character tables built by pcre_maketables() in the current locale. If NULL
4706 is passed, the default tables are used.
4707
4708
4709 Version 2.00 24-Sep-98
4710 ----------------------
4711
4712 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4713 it any more.
4714
4715 2. Allow quantification of (?>) groups, and make it work correctly.
4716
4717 3. The first character computation wasn't working for (?>) groups.
4718
4719 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4720 end of the subject) and add 5.005's \z, which really does match only at the
4721 very end of the subject.
4722
4723 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4724
4725 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4726 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4727 localized options. All options to pcre_study() were also removed.
4728
4729 7. Add other new features from 5.005:
4730
4731 $(?<= positive lookbehind
4732 $(?<! negative lookbehind
4733 (?imsx-imsx) added the unsetting capability
4734 such a setting is global if at outer level; local otherwise
4735 (?imsx-imsx:) non-capturing groups with option setting
4736 (?(cond)re|re) conditional pattern matching
4737
4738 A backreference to itself in a repeated group matches the previous
4739 captured string.
4740
4741 8. General tidying up of studying (both automatic and via "study")
4742 consequential on the addition of new assertions.
4743
4744 9. As in 5.005, unlimited repeated groups that could match an empty substring
4745 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4746 runtime if any iteration does actually match an empty substring.
4747
4748 10. Include the RunTest script in the distribution.
4749
4750 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4751 discrepancies, some of which were old and were also with respect to 5.004. They
4752 have now been fixed.
4753
4754
4755 Version 1.09 28-Apr-98
4756 ----------------------
4757
4758 1. A negated single character class followed by a quantifier with a minimum
4759 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4760 program crashes, or just wrong answers. This did not apply to negated classes
4761 containing more than one character, or to minima other than one.
4762
4763
4764 Version 1.08 27-Mar-98
4765 ----------------------
4766
4767 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4768
4769 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4770 latter must appear before anything that relies on it in the pattern.
4771
4772
4773 Version 1.07 16-Feb-98
4774 ----------------------
4775
4776 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4777 repeat of a potentially empty string).
4778
4779
4780 Version 1.06 23-Jan-98
4781 ----------------------
4782
4783 1. Added Markus Oberhumer's little patches for C++.
4784
4785 2. Literal strings longer than 255 characters were broken.
4786
4787
4788 Version 1.05 23-Dec-97
4789 ----------------------
4790
4791 1. Negated character classes containing more than one character were failing if
4792 PCRE_CASELESS was set at run time.
4793
4794
4795 Version 1.04 19-Dec-97
4796 ----------------------
4797
4798 1. Corrected the man page, where some "const" qualifiers had been omitted.
4799
4800 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4801 input syntax.
4802
4803 3. Fixed memory leak which occurred when a regex with back references was
4804 matched with an offsets vector that wasn't big enough. The temporary memory
4805 that is used in this case wasn't being freed if the match failed.
4806
4807 4. Tidied pcretest to ensure it frees memory that it gets.
4808
4809 5. Temporary memory was being obtained in the case where the passed offsets
4810 vector was exactly big enough.
4811
4812 6. Corrected definition of offsetof() from change 5 below.
4813
4814 7. I had screwed up change 6 below and broken the rules for the use of
4815 setjmp(). Now fixed.
4816
4817
4818 Version 1.03 18-Dec-97
4819 ----------------------
4820
4821 1. A erroneous regex with a missing opening parenthesis was correctly
4822 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4823 on some systems.
4824
4825 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4826 it was reported that one broken compiler failed on the former because "code" is
4827 also an independent variable.
4828
4829 3. The erroneous regex a[]b caused an array overrun reference.
4830
4831 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4832 fail on data ending with that character. (It was going on too far, and checking
4833 the next character, typically a binary zero.) This was specific to the
4834 optimized code for single-character negative classes.
4835
4836 5. Added a contributed patch from the TIN world which does the following:
4837
4838 + Add an undef for memmove, in case the the system defines a macro for it.
4839
4840 + Add a definition of offsetof(), in case there isn't one. (I don't know
4841 the reason behind this - offsetof() is part of the ANSI standard - but
4842 it does no harm).
4843
4844 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4845 most of the places where whitespace preceded '#'. I have given up and
4846 allowed the remaining 2 cases to be at the margin.
4847
4848 + Rename some variables in pcre to eliminate shadowing. This seems very
4849 pedantic, but does no harm, of course.
4850
4851 6. Moved the call to setjmp() into its own function, to get rid of warnings
4852 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4853
4854 7. Constructs such as \d{8,} were compiling into the equivalent of
4855 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4856 outcome, but in this particular case used more store than had been allocated,
4857 which caused the bug to be discovered because it threw up an internal error.
4858
4859 8. The debugging code in both pcre and pcretest for outputting the compiled
4860 form of a regex was going wrong in the case of back references followed by
4861 curly-bracketed repeats.
4862
4863
4864 Version 1.02 12-Dec-97
4865 ----------------------
4866
4867 1. Typos in pcre.3 and comments in the source fixed.
4868
4869 2. Applied a contributed patch to get rid of places where it used to remove
4870 'const' from variables, and fixed some signed/unsigned and uninitialized
4871 variable warnings.
4872
4873 3. Added the "runtest" target to Makefile.
4874
4875 4. Set default compiler flag to -O2 rather than just -O.
4876
4877
4878 Version 1.01 19-Nov-97
4879 ----------------------
4880
4881 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4882 like /([ab]*)*/, that is, for classes with more than one character in them.
4883
4884 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4885 as /((?>a*))*/ (a PCRE_EXTRA facility).
4886
4887
4888 Version 1.00 18-Nov-97
4889 ----------------------
4890
4891 1. Added compile-time macros to support systems such as SunOS4 which don't have
4892 memmove() or strerror() but have other things that can be used instead.
4893
4894 2. Arranged that "make clean" removes the executables.
4895
4896
4897 Version 0.99 27-Oct-97
4898 ----------------------
4899
4900 1. Fixed bug in code for optimizing classes with only one character. It was
4901 initializing a 32-byte map regardless, which could cause it to run off the end
4902 of the memory it had got.
4903
4904 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4905
4906
4907 Version 0.98 22-Oct-97
4908 ----------------------
4909
4910 1. Fixed bug in code for handling temporary memory usage when there are more
4911 back references than supplied space in the ovector. This could cause segfaults.
4912
4913
4914 Version 0.97 21-Oct-97
4915 ----------------------
4916
4917 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4918
4919 2. Optimized negated single characters not to use a bit map.
4920
4921 3. Brought error texts together as macro definitions; clarified some of them;
4922 fixed one that was wrong - it said "range out of order" when it meant "invalid
4923 escape sequence".
4924
4925 4. Changed some char * arguments to const char *.
4926
4927 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4928
4929 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4930 pcretest.
4931
4932
4933 Version 0.96 16-Oct-97
4934 ----------------------
4935
4936 1. Added a simple "pgrep" utility to the distribution.
4937
4938 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4939 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4940 where "ddd" means "one or more decimal digits".
4941
4942 3. Fixed serious bug. If a pattern had a back reference, but the call to
4943 pcre_exec() didn't supply a large enough ovector to record the related
4944 identifying subpattern, the match always failed. PCRE now remembers the number
4945 of the largest back reference, and gets some temporary memory in which to save
4946 the offsets during matching if necessary, in order to ensure that
4947 backreferences always work.
4948
4949 4. Increased the compatibility with Perl in a number of ways:
4950
4951 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4952 to request this handling. The option can be set at compile or exec time.
4953
4954 (b) $ matches before a terminating newline by default; an option
4955 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4956 mode). The option can be set at compile or exec time.
4957
4958 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4959 the same as Perl's. If the decimal number it represents is less than 10
4960 or there aren't that many previous left capturing parentheses, an octal
4961 escape is read. Inside a character class, it's always an octal escape,
4962 even if it is a single digit.
4963
4964 (d) An escaped but undefined alphabetic character is taken as a literal,
4965 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4966 escapes.
4967
4968 (e) {0} is now permitted. (The previous item is removed from the compiled
4969 pattern).
4970
4971 5. Changed all the names of code files so that the basic parts are no longer
4972 than 10 characters, and abolished the teeny "globals.c" file.
4973
4974 6. Changed the handling of character classes; they are now done with a 32-byte
4975 bit map always.
4976
4977 7. Added the -d and /D options to pcretest to make it possible to look at the
4978 internals of compilation without having to recompile pcre.
4979
4980
4981 Version 0.95 23-Sep-97
4982 ----------------------
4983
4984 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4985 \x20 at the start of a run of normal characters. These were being treated as
4986 real characters, instead of the source characters being re-checked.
4987
4988
4989 Version 0.94 18-Sep-97
4990 ----------------------
4991
4992 1. The functions are now thread-safe, with the caveat that the global variables
4993 containing pointers to malloc() and free() or alternative functions are the
4994 same for all threads.
4995
4996 2. Get pcre_study() to generate a bitmap of initial characters for non-
4997 anchored patterns when this is possible, and use it if passed to pcre_exec().
4998
4999
5000 Version 0.93 15-Sep-97
5001 ----------------------
5002
5003 1. /(b)|(:+)/ was computing an incorrect first character.
5004
5005 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5006 but not actually doing anything yet.
5007
5008 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5009 as Perl does (e.g. [-az] or [az-]).
5010
5011 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5012 all possible positions.
5013
5014 5. Split up into different modules to avoid including unneeded functions in a
5015 compiled binary. However, compile and exec are still in one module. The "study"
5016 function is split off.
5017
5018 6. The character tables are now in a separate module whose source is generated
5019 by an auxiliary program - but can then be edited by hand if required. There are
5020 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5021 toupper() in the code.
5022
5023 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5024 make them global. Abolish the function for setting them, as the caller can now
5025 set them directly.
5026
5027
5028 Version 0.92 11-Sep-97
5029 ----------------------
5030
5031 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5032 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5033
5034 2. Caseless matching was not working in character classes if the characters in
5035 the pattern were in upper case.
5036
5037 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5038
5039 4. Make PCRE_ANCHORED public and accept as a compile option.
5040
5041 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5042 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5043 pass them.
5044
5045 6. Give an error if bad option bits passed at compile or run time.
5046
5047 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5048 pcretest to cause it to pass that flag.
5049
5050 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5051 options, and the first character, if set.
5052
5053 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5054
5055
5056 Version 0.91 10-Sep-97
5057 ----------------------
5058
5059 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5060 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5061
5062 2. PCRE was looping on encountering an indefinitely repeated back reference to
5063 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5064 Perl does - treats the match as successful.
5065
5066 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5