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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1271 - (show annotations)
Wed Mar 6 16:50:38 2013 UTC (6 years, 8 months ago) by ph10
File size: 244375 byte(s)
Error occurred while calculating annotation data.
Confine PRUNE and SKIP to recursive subpattern calls, like COMMIT.
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 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89 there was a conditional group that depended on an assertion, if the
90 assertion was false, the callout that immediately followed the alternation
91 in the condition was skipped when pcre_exec() was used for matching.
92
93 23. Allow an explicit callout to be inserted before an assertion that is the
94 condition for a conditional group, for compatibility with automatic
95 callouts, which always insert a callout at this point.
96
97 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99
100
101 Version 8.32 30-November-2012
102 -----------------------------
103
104 1. Improved JIT compiler optimizations for first character search and single
105 character iterators.
106
107 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
108 Patch by Daniel Richard G.
109
110 3. Single character iterator optimizations in the JIT compiler.
111
112 4. Improved JIT compiler optimizations for character ranges.
113
114 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
115 Reported by Giuseppe D'Angelo.
116
117 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
118 of a line, was being set incorrectly in cases where .* appeared inside
119 atomic brackets at the start of a pattern, or where there was a subsequent
120 *PRUNE or *SKIP.
121
122 7. Improved instruction cache flush for POWER/PowerPC.
123 Patch by Daniel Richard G.
124
125 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
126 grep:
127
128 (a) There is now no limit to the number of patterns to be matched.
129
130 (b) An error is given if a pattern is too long.
131
132 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
133 are now supported.
134
135 (d) --exclude-from and --include-from (multiple use) have been added.
136
137 (e) Exclusions and inclusions now apply to all files and directories, not
138 just to those obtained from scanning a directory recursively.
139
140 (f) Multiple uses of -f and --file-list are now supported.
141
142 (g) In a Windows environment, the default for -d has been changed from
143 "read" (the GNU grep default) to "skip", because otherwise the presence
144 of a directory in the file list provokes an error.
145
146 (h) The documentation has been revised and clarified in places.
147
148 9. Improve the matching speed of capturing brackets.
149
150 10. Changed the meaning of \X so that it now matches a Unicode extended
151 grapheme cluster.
152
153 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
154 out POSIX threads when JIT support is configured.
155
156 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
157
158 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
159 structure could go wrong in environments where size_t is not the same size
160 as int.
161
162 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
163 set.
164
165 15. The EBCDIC support had decayed; later updates to the code had included
166 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
167 general tidy up of EBCDIC-related issues, and the documentation was also
168 not quite right. There is now a test that can be run on ASCII systems to
169 check some of the EBCDIC-related things (but is it not a full test).
170
171 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
172 in a small tidy to the code.
173
174 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
175
176 18. If the --only-matching (-o) option in pcregrep is specified multiple
177 times, each one causes appropriate output. For example, -o1 -o2 outputs the
178 substrings matched by the 1st and 2nd capturing parentheses. A separating
179 string can be specified by --om-separator (default empty).
180
181 19. Improving the first n character searches.
182
183 20. Turn case lists for horizontal and vertical white space into macros so that
184 they are defined only once.
185
186 21. This set of changes together give more compatible Unicode case-folding
187 behaviour for characters that have more than one other case when UCP
188 support is available.
189
190 (a) The Unicode property table now has offsets into a new table of sets of
191 three or more characters that are case-equivalent. The MultiStage2.py
192 script that generates these tables (the pcre_ucd.c file) now scans
193 CaseFolding.txt instead of UnicodeData.txt for character case
194 information.
195
196 (b) The code for adding characters or ranges of characters to a character
197 class has been abstracted into a generalized function that also handles
198 case-independence. In UTF-mode with UCP support, this uses the new data
199 to handle characters with more than one other case.
200
201 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
202 whose other case is greater than 256 are now correctly matched
203 caselessly. Previously, the high codepoint matched the low one, but not
204 vice versa.
205
206 (d) The processing of \h, \H, \v, and \ in character classes now makes use
207 of the new class addition function, using character lists defined as
208 macros alongside the case definitions of 20 above.
209
210 (e) Caseless back references now work with characters that have more than
211 one other case.
212
213 (f) General caseless matching of characters with more than one other case
214 is supported.
215
216 22. Unicode character properties were updated from Unicode 6.2.0
217
218 23. Improved CMake support under Windows. Patch by Daniel Richard G.
219
220 24. Add support for 32-bit character strings, and UTF-32
221
222 25. Major JIT compiler update (code refactoring and bugfixing).
223 Experimental Sparc 32 support is added.
224
225 26. Applied a modified version of Daniel Richard G's patch to create
226 pcre.h.generic and config.h.generic by "make" instead of in the
227 PrepareRelease script.
228
229 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
230 pcre_compile.c when checking for a zero character.
231
232 28. Introducing a native interface for JIT. Through this interface, the compiled
233 machine code can be directly executed. The purpose of this interface is to
234 provide fast pattern matching, so several sanity checks are not performed.
235 However, feature tests are still performed. The new interface provides
236 1.4x speedup compared to the old one.
237
238 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
239 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
240 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
241
242 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
243 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
244 the "old" RFC 2279). Instead, it ended up passing a negative length to
245 pcre_exec().
246
247 31. Add support for GCC's visibility feature to hide internal functions.
248
249 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
250 "unknown -C option" after outputting 0 or 1.
251
252 33. There is now support for generating a code coverage report for the test
253 suite in environments where gcc is the compiler and lcov is installed. This
254 is mainly for the benefit of the developers.
255
256 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
257 unaddressable using valgrind annotations, allowing valgrind to detect
258 invalid memory accesses. This is mainly for the benefit of the developers.
259
260 25. (*UTF) can now be used to start a pattern in any of the three libraries.
261
262 26. Give configure error if --enable-cpp but no C++ compiler found.
263
264
265 Version 8.31 06-July-2012
266 -------------------------
267
268 1. Fixing a wrong JIT test case and some compiler warnings.
269
270 2. Removed a bashism from the RunTest script.
271
272 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
273 to unsigned type, result still unsigned" that was given by an MS compiler
274 on encountering the code "-sizeof(xxx)".
275
276 4. Partial matching support is added to the JIT compiler.
277
278 5. Fixed several bugs concerned with partial matching of items that consist
279 of more than one character:
280
281 (a) /^(..)\1/ did not partially match "aba" because checking references was
282 done on an "all or nothing" basis. This also applied to repeated
283 references.
284
285 (b) \R did not give a hard partial match if \r was found at the end of the
286 subject.
287
288 (c) \X did not give a hard partial match after matching one or more
289 characters at the end of the subject.
290
291 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
292 a partial match for the string "\r".
293
294 (e) When newline was set to CRLF, the metacharacter "." did not recognize
295 a partial match for a CR character at the end of the subject string.
296
297 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
298 running pcretest, the text "(JIT)" added to the output whenever JIT is
299 actually used to run the match.
300
301 7. Individual JIT compile options can be set in pcretest by following -s+[+]
302 or /S+[+] with a digit between 1 and 7.
303
304 8. OP_NOT now supports any UTF character not just single-byte ones.
305
306 9. (*MARK) control verb is now supported by the JIT compiler.
307
308 10. The command "./RunTest list" lists the available tests without actually
309 running any of them. (Because I keep forgetting what they all are.)
310
311 11. Add PCRE_INFO_MAXLOOKBEHIND.
312
313 12. Applied a (slightly modified) user-supplied patch that improves performance
314 when the heap is used for recursion (compiled with --disable-stack-for-
315 recursion). Instead of malloc and free for each heap frame each time a
316 logical recursion happens, frames are retained on a chain and re-used where
317 possible. This sometimes gives as much as 30% improvement.
318
319 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
320 call.
321
322 14. As documented, (*COMMIT) is now confined to within a positive assertion.
323
324 15. It is now possible to link pcretest with libedit as an alternative to
325 libreadline.
326
327 16. (*COMMIT) control verb is now supported by the JIT compiler.
328
329 17. The Unicode data tables have been updated to Unicode 6.1.0.
330
331 18. Added --file-list option to pcregrep.
332
333 19. Added binary file support to pcregrep, including the -a, --binary-files,
334 -I, and --text options.
335
336 20. The madvise function is renamed for posix_madvise for QNX compatibility
337 reasons. Fixed by Giuseppe D'Angelo.
338
339 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
340
341 22. Changed the option for creating symbolic links for 16-bit man pages from
342 -s to -sf so that re-installing does not cause issues.
343
344 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
345
346 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
347 matches in certain environments (the workspace was not being correctly
348 retained). Also added to pcre_dfa_exec() a simple plausibility check on
349 some of the workspace data at the beginning of a restart.
350
351 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
352 was not doing so when it should - probably a typo introduced by SVN 528
353 (change 8.10/14).
354
355 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
356 \w+ when the character tables indicated that \x{c4} was a word character.
357 There were several related cases, all because the tests for doing a table
358 lookup were testing for characters less than 127 instead of 255.
359
360 27. If a pattern contains capturing parentheses that are not used in a match,
361 their slots in the ovector are set to -1. For those that are higher than
362 any matched groups, this happens at the end of processing. In the case when
363 there were back references that the ovector was too small to contain
364 (causing temporary malloc'd memory to be used during matching), and the
365 highest capturing number was not used, memory off the end of the ovector
366 was incorrectly being set to -1. (It was using the size of the temporary
367 memory instead of the true size.)
368
369 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
370 ovector size, it uses memory at the end of the block that it has got.
371
372 29. Check for an overlong MARK name and give an error at compile time. The
373 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
374
375 30. JIT compiler update.
376
377 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
378 Rill for the patch.
379
380 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
381
382 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
383
384 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
385 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
386 was enabled.
387
388 35. Improve JIT code generation for greedy plus quantifier.
389
390 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
391 1 to "aa" instead of to an empty string. The bug affected repeated groups
392 that could potentially match an empty string.
393
394 37. Optimizing single character iterators in JIT.
395
396 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
397 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
398 codepoints that are too big for the mode are faulted, and in a UTF mode,
399 disallowed codepoints are also faulted.
400
401 39. If PCRE was compiled with UTF support, in three places in the DFA
402 matcher there was code that should only have been obeyed in UTF mode, but
403 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
404 processing when bytes with values greater than 127 were present. In 16-bit
405 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
406 both cases the values are those that cannot be the first data item in a UTF
407 character. The three items that might have provoked this were recursions,
408 possessively repeated groups, and atomic groups.
409
410 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
411 and pcregrep, because some OS require shared objects to be explicitly
412 passed to ld, causing the link step to fail if they are not.
413
414 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
415 mode, patterns that started with \h* or \R* might be incorrectly matched.
416
417
418 Version 8.30 04-February-2012
419 -----------------------------
420
421 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
422 name is defined in ctype.h.
423
424 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
425 only in quite long subpatterns.
426
427 3. Removed the function pcre_info(), which has been obsolete and deprecated
428 since it was replaced by pcre_fullinfo() in February 2000.
429
430 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
431 match a (*MARK), and the match failed at the start of the subject, a
432 reference to memory before the start of the subject could occur. This bug
433 was introduced by fix 17 of release 8.21.
434
435 5. A reference to an unset group with zero minimum repetition was giving
436 totally wrong answers (in non-JavaScript-compatibility mode). For example,
437 /(another)?(\1?)test/ matched against "hello world test". This bug was
438 introduced in release 8.13.
439
440 6. Add support for 16-bit character strings (a large amount of work involving
441 many changes and refactorings).
442
443 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
444 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
445 from a file.
446
447 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
448 rounding is not applied in this particular case).
449
450 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
451 if they appear, or are escaped, in patterns.
452
453 10. Get rid of a number of -Wunused-but-set-variable warnings.
454
455 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
456 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
457 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
458 also returns the mark "x". This bug applied to capturing parentheses,
459 non-capturing parentheses, and atomic parentheses. It also applied to some
460 assertions.
461
462 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
463 information out of configure.ac instead of relying on pcre.h.generic, which
464 is not stored in the repository.
465
466 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
467 -lreadline.
468
469 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
470
471 15. Applied Graycode's patch to put the top-level frame on the stack rather
472 than the heap when not using the stack for recursion. This gives a
473 performance improvement in many cases when recursion is not deep.
474
475 16. Experimental code added to "pcretest -C" to output the stack frame size.
476
477
478 Version 8.21 12-Dec-2011
479 ------------------------
480
481 1. Updating the JIT compiler.
482
483 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
484 are added as well.
485
486 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
487 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
488 calling _pcre_jit_exec. Some extra comments are added.
489
490 4. (*MARK) settings inside atomic groups that do not contain any capturing
491 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
492 was introduced by change 18 for 8.20.
493
494 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
495 ECMA-262 standard.
496
497 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
498 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
499 This bug was probably introduced by change 9 of 8.13.
500
501 7. While fixing 6 above, I noticed that a number of other items were being
502 incorrectly rejected as "not fixed length". This arose partly because newer
503 opcodes had not been added to the fixed-length checking code. I have (a)
504 corrected the bug and added tests for these items, and (b) arranged for an
505 error to occur if an unknown opcode is encountered while checking for fixed
506 length instead of just assuming "not fixed length". The items that were
507 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
508 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
509 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
510
511 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
512 being incorrectly compiled and would have given unpredicatble results.
513
514 9. A possessively repeated subpattern with minimum repeat count greater than
515 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
516 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
517 the first (A) could occur when it should not.
518
519 10. Add a cast and remove a redundant test from the code.
520
521 11. JIT should use pcre_malloc/pcre_free for allocation.
522
523 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
524 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
525 is anything other than /usr, -L is still shown).
526
527 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
528
529 14. Perl does not support \N without a following name in a [] class; PCRE now
530 also gives an error.
531
532 15. If a forward reference was repeated with an upper limit of around 2000,
533 it caused the error "internal error: overran compiling workspace". The
534 maximum number of forward references (including repeats) was limited by the
535 internal workspace, and dependent on the LINK_SIZE. The code has been
536 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
537 the default depends on LINK_SIZE. There is a new upper limit (for safety)
538 of around 200,000 forward references. While doing this, I also speeded up
539 the filling in of repeated forward references.
540
541 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
542 incorrectly expecting the subject to contain another "a" after the start.
543
544 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
545 in the match, the SKIP should be ignored. This was not happening; instead
546 the SKIP was being treated as NOMATCH. For patterns such as
547 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
548 tested.
549
550 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
551 now much more compatible with Perl, in particular in cases where the result
552 is a non-match for a non-anchored pattern. For example, if
553 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
554 "m", where previously it did not return a name. A side effect of this
555 change is that for partial matches, the last encountered mark name is
556 returned, as for non matches. A number of tests that were previously not
557 Perl-compatible have been moved into the Perl-compatible test files. The
558 refactoring has had the pleasing side effect of removing one argument from
559 the match() function, thus reducing its stack requirements.
560
561 19. If the /S+ option was used in pcretest to study a pattern using JIT,
562 subsequent uses of /S (without +) incorrectly behaved like /S+.
563
564 21. Retrieve executable code size support for the JIT compiler and fixing
565 some warnings.
566
567 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
568 not work when the shorter character appeared right at the end of the
569 subject string.
570
571 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
572 systems.
573
574 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
575 output it when the /M option is used in pcretest.
576
577 25. The CheckMan script was not being included in the distribution. Also, added
578 an explicit "perl" to run Perl scripts from the PrepareRelease script
579 because this is reportedly needed in Windows.
580
581 26. If study data was being save in a file and studying had not found a set of
582 "starts with" bytes for the pattern, the data written to the file (though
583 never used) was taken from uninitialized memory and so caused valgrind to
584 complain.
585
586 27. Updated RunTest.bat as provided by Sheri Pierce.
587
588 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
589
590 29. Computation of memory usage for the table of capturing group names was
591 giving an unnecessarily large value.
592
593
594 Version 8.20 21-Oct-2011
595 ------------------------
596
597 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
598 a POSIX class. After further experiments with Perl, which convinced me that
599 Perl has bugs and confusions, a closing square bracket is no longer allowed
600 in a POSIX name. This bug also affected patterns with classes that started
601 with full stops.
602
603 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
604 captured substring, but while checking the failing first alternative,
605 substring 1 is temporarily captured. If the output vector supplied to
606 pcre_exec() was not big enough for this capture, the yield of the function
607 was still zero ("insufficient space for captured substrings"). This cannot
608 be totally fixed without adding another stack variable, which seems a lot
609 of expense for a edge case. However, I have improved the situation in cases
610 such as /(a)(b)x|abc/ matched against "abc", where the return code
611 indicates that fewer than the maximum number of slots in the ovector have
612 been set.
613
614 3. Related to (2) above: when there are more back references in a pattern than
615 slots in the output vector, pcre_exec() uses temporary memory during
616 matching, and copies in the captures as far as possible afterwards. It was
617 using the entire output vector, but this conflicts with the specification
618 that only 2/3 is used for passing back captured substrings. Now it uses
619 only the first 2/3, for compatibility. This is, of course, another edge
620 case.
621
622 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
623 main code base, and can be used by building with --enable-jit. When this is
624 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
625 runtime --no-jit option is given.
626
627 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
628 ovector, the return from the function was zero, implying that there were
629 other matches that did not fit. The correct "exactly full" value is now
630 returned.
631
632 6. If a subpattern that was called recursively or as a subroutine contained
633 (*PRUNE) or any other control that caused it to give a non-standard return,
634 invalid errors such as "Error -26 (nested recursion at the same subject
635 position)" or even infinite loops could occur.
636
637 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
638 computing the minimum length on reaching *ACCEPT, and so ended up with the
639 wrong value of 1 rather than 0. Further investigation indicates that
640 computing a minimum subject length in the presence of *ACCEPT is difficult
641 (think back references, subroutine calls), and so I have changed the code
642 so that no minimum is registered for a pattern that contains *ACCEPT.
643
644 8. If (*THEN) was present in the first (true) branch of a conditional group,
645 it was not handled as intended. [But see 16 below.]
646
647 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
648 Sheri Pierce.
649
650 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
651 the first byte in a match must be "a".
652
653 11. Change 17 for 8.13 increased the recursion depth for patterns like
654 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
655 pattern contains any instances of (*THEN). If it does not, the old
656 optimizations are restored. It would be nice to do this on a per-group
657 basis, but at the moment that is not feasible.
658
659 12. In some environments, the output of pcretest -C is CRLF terminated. This
660 broke RunTest's code that checks for the link size. A single white space
661 character after the value is now allowed for.
662
663 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
664 For "fr", it uses the Windows-specific input and output files.
665
666 14. If (*THEN) appeared in a group that was called recursively or as a
667 subroutine, it did not work as intended. [But see next item.]
668
669 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
670 pattern fragments (but not containing any | characters). If A and B are
671 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
672 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
673 D. In other words, Perl considers parentheses that do not contain any |
674 characters to be part of a surrounding alternative, whereas PCRE was
675 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
676 differently. PCRE now behaves in the same way as Perl, except in the case
677 of subroutine/recursion calls such as (?1) which have in any case always
678 been different (but PCRE had them first :-).
679
680 16. Related to 15 above: Perl does not treat the | in a conditional group as
681 creating alternatives. Such a group is treated in the same way as an
682 ordinary group without any | characters when processing (*THEN). PCRE has
683 been changed to match Perl's behaviour.
684
685 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
686 RunGrepTest script failed.
687
688 18. Change 22 for version 13 caused atomic groups to use more stack. This is
689 inevitable for groups that contain captures, but it can lead to a lot of
690 stack use in large patterns. The old behaviour has been restored for atomic
691 groups that do not contain any capturing parentheses.
692
693 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
694 suppress the check for a minimum subject length at run time. (If it was
695 given to pcre_exec() or pcre_dfa_exec() it did work.)
696
697 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
698 fail to work when decoding hex characters in data strings in EBCDIC
699 environments.
700
701 21. It appears that in at least one Mac OS environment, the isxdigit() function
702 is implemented as a macro that evaluates to its argument more than once,
703 contravening the C 90 Standard (I haven't checked a later standard). There
704 was an instance in pcretest which caused it to go wrong when processing
705 \x{...} escapes in subject strings. The has been rewritten to avoid using
706 things like p++ in the argument of isxdigit().
707
708
709 Version 8.13 16-Aug-2011
710 ------------------------
711
712 1. The Unicode data tables have been updated to Unicode 6.0.0.
713
714 2. Two minor typos in pcre_internal.h have been fixed.
715
716 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
717 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
718 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
719
720 4. There were a number of related bugs in the code for matching backrefences
721 caselessly in UTF-8 mode when codes for the characters concerned were
722 different numbers of bytes. For example, U+023A and U+2C65 are an upper
723 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
724 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
725 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
726 2-byte code at the end of the subject (it thought there wasn't enough data
727 left).
728
729 5. Comprehensive information about what went wrong is now returned by
730 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
731 as the output vector has at least 2 elements. The offset of the start of
732 the failing character and a reason code are placed in the vector.
733
734 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
735 now returned is for the first byte of the failing character, instead of the
736 last byte inspected. This is an incompatible change, but I hope it is small
737 enough not to be a problem. It makes the returned offset consistent with
738 pcre_exec() and pcre_dfa_exec().
739
740 7. pcretest now gives a text phrase as well as the error number when
741 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
742 failure, the offset and reason code are output.
743
744 8. When \R was used with a maximizing quantifier it failed to skip backwards
745 over a \r\n pair if the subsequent match failed. Instead, it just skipped
746 back over a single character (\n). This seems wrong (because it treated the
747 two characters as a single entity when going forwards), conflicts with the
748 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
749 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
750 has been changed.
751
752 9. Some internal refactoring has changed the processing so that the handling
753 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
754 time (the PCRE_DOTALL option was changed this way some time ago: version
755 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
756 which was always a bit of a fudge. It also means that there is one less
757 argument for the match() function, which reduces its stack requirements
758 slightly. This change also fixes an incompatibility with Perl: the pattern
759 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
760
761 10. More internal refactoring has drastically reduced the number of recursive
762 calls to match() for possessively repeated groups such as (abc)++ when
763 using pcre_exec().
764
765 11. While implementing 10, a number of bugs in the handling of groups were
766 discovered and fixed:
767
768 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
769 (a|)*(?1) gave a compile-time internal error.
770 ((a|)+)+ did not notice that the outer group could match an empty string.
771 (^a|^)+ was not marked as anchored.
772 (.*a|.*)+ was not marked as matching at start or after a newline.
773
774 12. Yet more internal refactoring has removed another argument from the match()
775 function. Special calls to this function are now indicated by setting a
776 value in a variable in the "match data" data block.
777
778 13. Be more explicit in pcre_study() instead of relying on "default" for
779 opcodes that mean there is no starting character; this means that when new
780 ones are added and accidentally left out of pcre_study(), testing should
781 pick them up.
782
783 14. The -s option of pcretest has been documented for ages as being an old
784 synonym of -m (show memory usage). I have changed it to mean "force study
785 for every regex", that is, assume /S for every regex. This is similar to -i
786 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
787 using it. It makes it easier to run collections of tests with and without
788 study enabled, and thereby test pcre_study() more easily. All the standard
789 tests are now run with and without -s (but some patterns can be marked as
790 "never study" - see 20 below).
791
792 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
793 restoration of the capturing data to the outer values was not happening
794 correctly.
795
796 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
797 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
798 pattern had matched an empty string, and so incorrectly returned a no
799 match.
800
801 17. There was optimizing code for the last branch of non-capturing parentheses,
802 and also for the obeyed branch of a conditional subexpression, which used
803 tail recursion to cut down on stack usage. Unfortunately, now that there is
804 the possibility of (*THEN) occurring in these branches, tail recursion is
805 no longer possible because the return has to be checked for (*THEN). These
806 two optimizations have therefore been removed. [But see 8.20/11 above.]
807
808 18. If a pattern containing \R was studied, it was assumed that \R always
809 matched two bytes, thus causing the minimum subject length to be
810 incorrectly computed because \R can also match just one byte.
811
812 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
813 was incorrectly computed.
814
815 20. If /S is present twice on a test pattern in pcretest input, it now
816 *disables* studying, thereby overriding the use of -s on the command line
817 (see 14 above). This is necessary for one or two tests to keep the output
818 identical in both cases.
819
820 21. When (*ACCEPT) was used in an assertion that matched an empty string and
821 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
822
823 22. When an atomic group that contained a capturing parenthesis was
824 successfully matched, but the branch in which it appeared failed, the
825 capturing was not being forgotten if a higher numbered group was later
826 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
827 group 1 to "a", when in fact it should be unset. This applied to multi-
828 branched capturing and non-capturing groups, repeated or not, and also to
829 positive assertions (capturing in negative assertions does not happen
830 in PCRE) and also to nested atomic groups.
831
832 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
833 subject after a captured substring, to make it easier to tell which of a
834 number of identical substrings has been captured.
835
836 24. The way atomic groups are processed by pcre_exec() has been changed so that
837 if they are repeated, backtracking one repetition now resets captured
838 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
839 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
840 "aaa". Previously, it would have been "a". As part of this code
841 refactoring, the way recursive calls are handled has also been changed.
842
843 25. If an assertion condition captured any substrings, they were not passed
844 back unless some other capturing happened later. For example, if
845 (?(?=(a))a) was matched against "a", no capturing was returned.
846
847 26. When studying a pattern that contained subroutine calls or assertions,
848 the code for finding the minimum length of a possible match was handling
849 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
850 group 1 called group 2 while simultaneously a separate group 2 called group
851 1). A stack overflow occurred in this case. I have fixed this by limiting
852 the recursion depth to 10.
853
854 27. Updated RunTest.bat in the distribution to the version supplied by Tom
855 Fortmann. This supports explicit test numbers on the command line, and has
856 argument validation and error reporting.
857
858 28. An instance of \X with an unlimited repeat could fail if at any point the
859 first character it looked at was a mark character.
860
861 29. Some minor code refactoring concerning Unicode properties and scripts
862 should reduce the stack requirement of match() slightly.
863
864 30. Added the '=' option to pcretest to check the setting of unused capturing
865 slots at the end of the pattern, which are documented as being -1, but are
866 not included in the return count.
867
868 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
869 compiled something random. Now it gives a compile-time error (as does
870 Perl).
871
872 32. A *MARK encountered during the processing of a positive assertion is now
873 recorded and passed back (compatible with Perl).
874
875 33. If --only-matching or --colour was set on a pcregrep call whose pattern
876 had alternative anchored branches, the search for a second match in a line
877 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
878 matched the line "0102" twice. The same bug affected patterns that started
879 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
880 twice.
881
882 34. Previously, PCRE did not allow quantification of assertions. However, Perl
883 does, and because of capturing effects, quantifying parenthesized
884 assertions may at times be useful. Quantifiers are now allowed for
885 parenthesized assertions.
886
887 35. A minor code tidy in pcre_compile() when checking options for \R usage.
888
889 36. \g was being checked for fancy things in a character class, when it should
890 just be a literal "g".
891
892 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
893 appearance of a nested POSIX class supersedes an apparent external class.
894 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
895 unescaped square brackets may also appear as part of class names. For
896 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
897 more like Perl. (But see 8.20/1 above.)
898
899 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
900 was because it thought it was \N{name}, which is not supported).
901
902 39. Add minix to OS list not supporting the -S option in pcretest.
903
904 40. PCRE tries to detect cases of infinite recursion at compile time, but it
905 cannot analyze patterns in sufficient detail to catch mutual recursions
906 such as ((?1))((?2)). There is now a runtime test that gives an error if a
907 subgroup is called recursively as a subpattern for a second time at the
908 same position in the subject string. In previous releases this might have
909 been caught by the recursion limit, or it might have run out of stack.
910
911 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
912 happen only once. PCRE was, however incorrectly giving a compile time error
913 "recursive call could loop indefinitely" because it cannot analyze the
914 pattern in sufficient detail. The compile time test no longer happens when
915 PCRE is compiling a conditional subpattern, but actual runaway loops are
916 now caught at runtime (see 40 above).
917
918 42. It seems that Perl allows any characters other than a closing parenthesis
919 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
920 has been changed to be the same.
921
922 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
923 as not to get warnings when autogen.sh is called. Also changed
924 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
925
926 44. To help people who use pcregrep to scan files containing exceedingly long
927 lines, the following changes have been made:
928
929 (a) The default value of the buffer size parameter has been increased from
930 8K to 20K. (The actual buffer used is three times this size.)
931
932 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
933 PCRE is built.
934
935 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
936 to be set at run time.
937
938 (d) Numerical values in pcregrep options can be followed by K or M, for
939 example --buffer-size=50K.
940
941 (e) If a line being scanned overflows pcregrep's buffer, an error is now
942 given and the return code is set to 2.
943
944 45. Add a pointer to the latest mark to the callout data block.
945
946 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
947 partial match of an empty string instead of no match. This was specific to
948 the use of ".".
949
950 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
951 complete match instead of a partial match. This bug was dependent on both
952 the PCRE_UTF8 and PCRE_DOTALL options being set.
953
954 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
955 starting byte set, because \b was not being ignored.
956
957
958 Version 8.12 15-Jan-2011
959 ------------------------
960
961 1. Fixed some typos in the markup of the man pages, and wrote a script that
962 checks for such things as part of the documentation building process.
963
964 2. On a big-endian 64-bit system, pcregrep did not correctly process the
965 --match-limit and --recursion-limit options (added for 8.11). In
966 particular, this made one of the standard tests fail. (The integer value
967 went into the wrong half of a long int.)
968
969 3. If the --colour option was given to pcregrep with -v (invert match), it
970 did strange things, either producing crazy output, or crashing. It should,
971 of course, ignore a request for colour when reporting lines that do not
972 match.
973
974 4. Another pcregrep bug caused similar problems if --colour was specified with
975 -M (multiline) and the pattern match finished with a line ending.
976
977 5. In pcregrep, when a pattern that ended with a literal newline sequence was
978 matched in multiline mode, the following line was shown as part of the
979 match. This seems wrong, so I have changed it.
980
981 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
982 the check for further matches in the same line (so they could be coloured)
983 to overrun the end of the current line. If another match was found, it was
984 incorrectly shown (and then shown again when found in the next line).
985
986 7. If pcregrep was compiled under Windows, there was a reference to the
987 function pcregrep_exit() before it was defined. I am assuming this was
988 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
989 reported by a user. I've moved the definition above the reference.
990
991
992 Version 8.11 10-Dec-2010
993 ------------------------
994
995 1. (*THEN) was not working properly if there were untried alternatives prior
996 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
997 backtracked to try for "b" instead of moving to the next alternative branch
998 at the same level (in this case, to look for "c"). The Perl documentation
999 is clear that when (*THEN) is backtracked onto, it goes to the "next
1000 alternative in the innermost enclosing group".
1001
1002 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1003 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1004 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1005 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1006 (*THEN).
1007
1008 3. If \s appeared in a character class, it removed the VT character from
1009 the class, even if it had been included by some previous item, for example
1010 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1011 of \s, but is part of the POSIX "space" class.)
1012
1013 4. A partial match never returns an empty string (because you can always
1014 match an empty string at the end of the subject); however the checking for
1015 an empty string was starting at the "start of match" point. This has been
1016 changed to the "earliest inspected character" point, because the returned
1017 data for a partial match starts at this character. This means that, for
1018 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1019 (previously it gave "no match").
1020
1021 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1022 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1023 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1024 has an implication that the given string is incomplete (because a partial
1025 match is preferred over a full match). For this reason, these items now
1026 give a partial match in this situation. [Aside: previously, the one case
1027 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1028 match rather than a full match, which was wrong by the old rules, but is
1029 now correct.]
1030
1031 6. There was a bug in the handling of #-introduced comments, recognized when
1032 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1033 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1034 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1035 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1036 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1037 places in pcre_compile().
1038
1039 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1040 comments when looking ahead for named forward references to subpatterns,
1041 the only newline sequence it recognized was NL. It now handles newlines
1042 according to the set newline convention.
1043
1044 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1045 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1046 cater for a lack of strerror(). These oversights have been fixed.
1047
1048 9. Added --match-limit and --recursion-limit to pcregrep.
1049
1050 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1051
1052 11. When the -o option was used, pcregrep was setting a return code of 1, even
1053 when matches were found, and --line-buffered was not being honoured.
1054
1055 12. Added an optional parentheses number to the -o and --only-matching options
1056 of pcregrep.
1057
1058 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1059 can match an empty string. The code to do it in pcretest and pcredemo
1060 needed fixing:
1061
1062 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1063 only one byte after an empty string match just before CRLF (this case
1064 just got forgotten; "any" and "anycrlf" were OK).
1065
1066 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1067 mode when an empty string match preceded an ASCII character followed by
1068 a non-ASCII character. (The code for advancing by one character rather
1069 than one byte was nonsense.)
1070
1071 (c) The pcredemo.c sample program did not have any code at all to handle
1072 the cases when CRLF is a valid newline sequence.
1073
1074 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1075 as a starting offset was within the subject string. There is now a new
1076 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1077 negative or greater than the length of the string. In order to test this,
1078 pcretest is extended to allow the setting of negative starting offsets.
1079
1080 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1081 starting offset points to the beginning of a UTF-8 character was
1082 unnecessarily clumsy. I tidied it up.
1083
1084 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1085 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1086
1087 17. Nobody had reported that the --include_dir option, which was added in
1088 release 7.7 should have been called --include-dir (hyphen, not underscore)
1089 for compatibility with GNU grep. I have changed it to --include-dir, but
1090 left --include_dir as an undocumented synonym, and the same for
1091 --exclude-dir, though that is not available in GNU grep, at least as of
1092 release 2.5.4.
1093
1094 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1095 characters from a string of bytes) have been redefined so as not to use
1096 loops, in order to improve performance in some environments. At the same
1097 time, I abstracted some of the common code into auxiliary macros to save
1098 repetition (this should not affect the compiled code).
1099
1100 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1101 compile-time error is now given if \c is not followed by an ASCII
1102 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1103 different, and any byte value is allowed.)
1104
1105 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1106 START_OPTIMIZE option, which is now allowed at compile time - but just
1107 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1108 to pcregrep and other applications that have no direct access to PCRE
1109 options. The new /Y option in pcretest sets this option when calling
1110 pcre_compile().
1111
1112 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1113 back references. Groups containing recursive back references were forced to
1114 be atomic by that change, but in the case of named groups, the amount of
1115 memory required was incorrectly computed, leading to "Failed: internal
1116 error: code overflow". This has been fixed.
1117
1118 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1119 pcretest.c, to avoid build problems in some Borland environments.
1120
1121
1122 Version 8.10 25-Jun-2010
1123 ------------------------
1124
1125 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1126 THEN.
1127
1128 2. (*ACCEPT) was not working when inside an atomic group.
1129
1130 3. Inside a character class, \B is treated as a literal by default, but
1131 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1132 causes the error). The code is unchanged, but I tidied the documentation.
1133
1134 4. Inside a character class, PCRE always treated \R and \X as literals,
1135 whereas Perl faults them if its -w option is set. I have changed PCRE so
1136 that it faults them when PCRE_EXTRA is set.
1137
1138 5. Added support for \N, which always matches any character other than
1139 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1140
1141 6. When compiling pcregrep with newer versions of gcc which may have
1142 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1143 declared with attribute warn_unused_result" were given. Just casting the
1144 result to (void) does not stop the warnings; a more elaborate fudge is
1145 needed. I've used a macro to implement this.
1146
1147 7. Minor change to pcretest.c to avoid a compiler warning.
1148
1149 8. Added four artifical Unicode properties to help with an option to make
1150 \s etc use properties (see next item). The new properties are: Xan
1151 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1152
1153 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1154 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1155 this option. Modified pcretest to add /W to test this facility. Added
1156 REG_UCP to make it available via the POSIX interface.
1157
1158 10. Added --line-buffered to pcregrep.
1159
1160 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1161 studied, and the match started with a letter with a code point greater than
1162 127 whose first byte was different to the first byte of the other case of
1163 the letter, the other case of this starting letter was not recognized
1164 (#976).
1165
1166 12. If a pattern that was studied started with a repeated Unicode property
1167 test, for example, \p{Nd}+, there was the theoretical possibility of
1168 setting up an incorrect bitmap of starting bytes, but fortunately it could
1169 not have actually happened in practice until change 8 above was made (it
1170 added property types that matched character-matching opcodes).
1171
1172 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1173 possible starting bytes for non-anchored patterns.
1174
1175 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1176 \R, and also a number of cases that involve Unicode properties, both
1177 explicit and implicit when PCRE_UCP is set.
1178
1179 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1180 input, it could crash or give wrong results if characters with values
1181 greater than 0xc0 were present in the subject string. (Detail: it assumed
1182 UTF-8 input when processing these items.)
1183
1184 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1185 size_t is 64-bit (#991).
1186
1187 17. Added a check for running out of memory when PCRE is compiled with
1188 --disable-stack-for-recursion (#990).
1189
1190 18. If the last data line in a file for pcretest does not have a newline on
1191 the end, a newline was missing in the output.
1192
1193 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1194 less than 128) in its various bitmaps. However, there is a facility for
1195 generating tables according to the current locale when PCRE is compiled. It
1196 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1197 characters, are recognized by isspace() and therefore were getting set in
1198 these tables, and indeed these tables seem to approximate to ISO 8859. This
1199 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1200 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1201 which of course cannot start UTF-8 characters. I have changed the code so
1202 that only real ASCII characters (less than 128) and the correct starting
1203 bytes for UTF-8 encodings are set for characters greater than 127 when in
1204 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1205 altogether.)
1206
1207 20. Added the /T option to pcretest so as to be able to run tests with non-
1208 standard character tables, thus making it possible to include the tests
1209 used for 19 above in the standard set of tests.
1210
1211 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1212 reference to a subpattern the other side of a comment that contains an
1213 opening parenthesis caused either an internal compiling error, or a
1214 reference to the wrong subpattern.
1215
1216
1217 Version 8.02 19-Mar-2010
1218 ------------------------
1219
1220 1. The Unicode data tables have been updated to Unicode 5.2.0.
1221
1222 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1223 configured.
1224
1225 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1226 original author of that file, following a query about its status.
1227
1228 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1229 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1230
1231 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1232 quantifier applied to a forward-referencing subroutine call, could compile
1233 incorrect code or give the error "internal error: previously-checked
1234 referenced subpattern not found".
1235
1236 6. Both MS Visual Studio and Symbian OS have problems with initializing
1237 variables to point to external functions. For these systems, therefore,
1238 pcre_malloc etc. are now initialized to local functions that call the
1239 relevant global functions.
1240
1241 7. There were two entries missing in the vectors called coptable and poptable
1242 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1243 I've fixed the data, and added a kludgy way of testing at compile time that
1244 the lengths are correct (equal to the number of opcodes).
1245
1246 8. Following on from 7, I added a similar kludge to check the length of the
1247 eint vector in pcreposix.c.
1248
1249 9. Error texts for pcre_compile() are held as one long string to avoid too
1250 much relocation at load time. To find a text, the string is searched,
1251 counting zeros. There was no check for running off the end of the string,
1252 which could happen if a new error number was added without updating the
1253 string.
1254
1255 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1256
1257 11. \K was not working if it appeared in an atomic group or in a group that
1258 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1259 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1260 the assertion is positive, but not if it is negative.
1261
1262 12. Change 11 fortuitously reduced the size of the stack frame used in the
1263 "match()" function of pcre_exec.c by one pointer. Forthcoming
1264 implementation of support for (*MARK) will need an extra pointer on the
1265 stack; I have reserved it now, so that the stack frame size does not
1266 decrease.
1267
1268 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1269 item in branch that calls a recursion is a subroutine call - as in the
1270 second branch in the above example - was incorrectly given the compile-
1271 time error "recursive call could loop indefinitely" because pcre_compile()
1272 was not correctly checking the subroutine for matching a non-empty string.
1273
1274 14. The checks for overrunning compiling workspace could trigger after an
1275 overrun had occurred. This is a "should never occur" error, but it can be
1276 triggered by pathological patterns such as hundreds of nested parentheses.
1277 The checks now trigger 100 bytes before the end of the workspace.
1278
1279 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1280
1281
1282 Version 8.01 19-Jan-2010
1283 ------------------------
1284
1285 1. If a pattern contained a conditional subpattern with only one branch (in
1286 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1287 computed the wrong minimum data length (which is of course zero for such
1288 subpatterns). This could cause incorrect "no match" results.
1289
1290 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1291 the pattern is reset in the first branch, pcre_compile() failed with
1292 "internal error: code overflow at offset...". This happened only when
1293 the reset was to the original external option setting. (An optimization
1294 abstracts leading options settings into an external setting, which was the
1295 cause of this.)
1296
1297 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1298 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1299 assertion pattern did not match (meaning that the assertion was true), it
1300 was incorrectly treated as false if the SKIP had been reached during the
1301 matching. This also applied to assertions used as conditions.
1302
1303 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1304 assertion subpattern, including such a pattern used as a condition,
1305 unpredictable results occurred, instead of the error return
1306 PCRE_ERROR_DFA_UITEM.
1307
1308 5. The C++ GlobalReplace function was not working like Perl for the special
1309 situation when an empty string is matched. It now does the fancy magic
1310 stuff that is necessary.
1311
1312 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1313 removed. (These were left over from very, very early versions of PCRE.)
1314
1315 7. Some cosmetic changes to the code to make life easier when compiling it
1316 as part of something else:
1317
1318 (a) Change DEBUG to PCRE_DEBUG.
1319
1320 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1321 called "current" as "current_branch", to prevent a collision with the
1322 Linux macro when compiled as a kernel module.
1323
1324 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1325 prevent a collision with the Linux macro when compiled as a kernel
1326 module.
1327
1328 8. In pcre_compile() there are some checks for integer overflows that used to
1329 cast potentially large values to (double). This has been changed to that
1330 when building, a check for int64_t is made, and if it is found, it is used
1331 instead, thus avoiding the use of floating point arithmetic. (There is no
1332 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1333 double.
1334
1335 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1336 2005 (difference between two addresses compared to an unsigned value).
1337
1338 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1339 custom one, because of the following reported problem in Windows:
1340
1341 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1342 under Win32.
1343 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1344 therefore missing the function definition.
1345 - The compiler thus generates a "C" signature for the test function.
1346 - The linker fails to find the "C" function.
1347 - PCRE fails to configure if asked to do so against libbz2.
1348
1349 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1350 messages were output:
1351
1352 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1353 rerunning libtoolize, to keep the correct libtool macros in-tree.
1354 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1355
1356 I have done both of these things.
1357
1358 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1359 most of the time, it *can* run out if it is given a pattern that contains a
1360 runaway infinite recursion. I updated the discussion in the pcrestack man
1361 page.
1362
1363 13. Now that we have gone to the x.xx style of version numbers, the minor
1364 version may start with zero. Using 08 or 09 is a bad idea because users
1365 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1366 interpreted as invalid octal numbers. I've updated the previous comment in
1367 configure.ac, and also added a check that gives an error if 08 or 09 are
1368 used.
1369
1370 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1371 causing partial matching to fail when the end of the subject matched \W
1372 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1373
1374 15. There were some discrepancies between the declarations in pcre_internal.h
1375 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1376 their definitions. The declarations used "const uschar *" and the
1377 definitions used USPTR. Even though USPTR is normally defined as "const
1378 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1379 reported that: "This difference in casting confuses some C++ compilers, for
1380 example, SunCC recognizes above declarations as different functions and
1381 generates broken code for hbpcre." I have changed the declarations to use
1382 USPTR.
1383
1384 16. GNU libtool is named differently on some systems. The autogen.sh script now
1385 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1386 (FreeBSD).
1387
1388 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1389 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1390 comment: "Figure out how to create a longlong from a string: strtoll and
1391 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1392 instance, but it only takes 2 args instead of 3!"
1393
1394 18. A subtle bug concerned with back references has been fixed by a change of
1395 specification, with a corresponding code fix. A pattern such as
1396 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1397 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1398 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1399 same bug. Such groups have to be quantified to be useful, or contained
1400 inside another quantified group. (If there's no repetition, the reference
1401 can never match.) The problem arises because, having left the group and
1402 moved on to the rest of the pattern, a later failure that backtracks into
1403 the group uses the captured value from the final iteration of the group
1404 rather than the correct earlier one. I have fixed this in PCRE by forcing
1405 any group that contains a reference to itself to be an atomic group; that
1406 is, there cannot be any backtracking into it once it has completed. This is
1407 similar to recursive and subroutine calls.
1408
1409
1410 Version 8.00 19-Oct-09
1411 ----------------------
1412
1413 1. The table for translating pcre_compile() error codes into POSIX error codes
1414 was out-of-date, and there was no check on the pcre_compile() error code
1415 being within the table. This could lead to an OK return being given in
1416 error.
1417
1418 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1419 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1420 in a Windows environment.
1421
1422 3. The pcregrep --count option prints the count for each file even when it is
1423 zero, as does GNU grep. However, pcregrep was also printing all files when
1424 --files-with-matches was added. Now, when both options are given, it prints
1425 counts only for those files that have at least one match. (GNU grep just
1426 prints the file name in this circumstance, but including the count seems
1427 more useful - otherwise, why use --count?) Also ensured that the
1428 combination -clh just lists non-zero counts, with no names.
1429
1430 4. The long form of the pcregrep -F option was incorrectly implemented as
1431 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1432 but it seems right to fix it, and I didn't think it was worth preserving
1433 the old behaviour.
1434
1435 5. The command line items --regex=pattern and --regexp=pattern were not
1436 recognized by pcregrep, which required --regex pattern or --regexp pattern
1437 (with a space rather than an '='). The man page documented the '=' forms,
1438 which are compatible with GNU grep; these now work.
1439
1440 6. No libpcreposix.pc file was created for pkg-config; there was just
1441 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1442
1443 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1444 when UCP support is not needed, by modifying the Python script that
1445 generates it from Unicode data files. This should not matter if the module
1446 is correctly used as a library, but I received one complaint about 50K of
1447 unwanted data. My guess is that the person linked everything into his
1448 program rather than using a library. Anyway, it does no harm.
1449
1450 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1451 was a minimum greater than 1 for a wide character in a possessive
1452 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1453 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1454 character. Chaos in the form of incorrect output or a compiling loop could
1455 result.
1456
1457 9. The restrictions on what a pattern can contain when partial matching is
1458 requested for pcre_exec() have been removed. All patterns can now be
1459 partially matched by this function. In addition, if there are at least two
1460 slots in the offset vector, the offset of the earliest inspected character
1461 for the match and the offset of the end of the subject are set in them when
1462 PCRE_ERROR_PARTIAL is returned.
1463
1464 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1465 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1466 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1467 and may be more useful for multi-segment matching.
1468
1469 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1470 used to be given if ever the end of the subject was reached; now it is
1471 given only if matching could not proceed because another character was
1472 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1473 string "Z", which now yields "no match" instead of "partial match". In the
1474 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1475 final character ended with (*FAIL).
1476
1477 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1478 if the pattern had a "must contain" character that was already found in the
1479 earlier partial match, unless partial matching was again requested. For
1480 example, with the pattern /dog.(body)?/, the "must contain" character is
1481 "g". If the first part-match was for the string "dog", restarting with
1482 "sbody" failed. This bug has been fixed.
1483
1484 13. The string returned by pcre_dfa_exec() after a partial match has been
1485 changed so that it starts at the first inspected character rather than the
1486 first character of the match. This makes a difference only if the pattern
1487 starts with a lookbehind assertion or \b or \B (\K is not supported by
1488 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1489 matching functions compatible, and I think it's the right thing to do.
1490
1491 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1492 so that the demonstration program is easily available in environments where
1493 PCRE has not been installed from source.
1494
1495 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1496 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1497 library.
1498
1499 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1500 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1501 is not the first non-POSIX option to be added. Clearly some people find
1502 these options useful.
1503
1504 17. If a caller to the POSIX matching function regexec() passes a non-zero
1505 value for nmatch with a NULL value for pmatch, the value of
1506 nmatch is forced to zero.
1507
1508 18. RunGrepTest did not have a test for the availability of the -u option of
1509 the diff command, as RunTest does. It now checks in the same way as
1510 RunTest, and also checks for the -b option.
1511
1512 19. If an odd number of negated classes containing just a single character
1513 interposed, within parentheses, between a forward reference to a named
1514 subpattern and the definition of the subpattern, compilation crashed with
1515 an internal error, complaining that it could not find the referenced
1516 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1517 [The bug was that it was starting one character too far in when skipping
1518 over the character class, thus treating the ] as data rather than
1519 terminating the class. This meant it could skip too much.]
1520
1521 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1522 /g option in pcretest when the pattern contains \K, which makes it possible
1523 to have an empty string match not at the start, even when the pattern is
1524 anchored. Updated pcretest and pcredemo to use this option.
1525
1526 21. If the maximum number of capturing subpatterns in a recursion was greater
1527 than the maximum at the outer level, the higher number was returned, but
1528 with unset values at the outer level. The correct (outer level) value is
1529 now given.
1530
1531 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1532 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1533 make it do so. The string so far is captured, making this feature
1534 compatible with Perl.
1535
1536 23. The tests have been re-organized, adding tests 11 and 12, to make it
1537 possible to check the Perl 5.10 features against Perl 5.10.
1538
1539 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1540 pattern matches a fixed length string. PCRE did not allow this; now it
1541 does. Neither allows recursion.
1542
1543 25. I finally figured out how to implement a request to provide the minimum
1544 length of subject string that was needed in order to match a given pattern.
1545 (It was back references and recursion that I had previously got hung up
1546 on.) This code has now been added to pcre_study(); it finds a lower bound
1547 to the length of subject needed. It is not necessarily the greatest lower
1548 bound, but using it to avoid searching strings that are too short does give
1549 some useful speed-ups. The value is available to calling programs via
1550 pcre_fullinfo().
1551
1552 26. While implementing 25, I discovered to my embarrassment that pcretest had
1553 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1554 study optimizations had never been tested with that matching function.
1555 Oops. What is worse, even when it was passed study data, there was a bug in
1556 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1557 were also very few tests of studied patterns with pcre_dfa_exec().
1558
1559 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1560 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1561 on the other side of the coin, they are no longer allowed to have different
1562 names, because these cannot be distinguished in PCRE, and this has caused
1563 confusion. (This is a difference from Perl.)
1564
1565 28. When duplicate subpattern names are present (necessarily with different
1566 numbers, as required by 27 above), and a test is made by name in a
1567 conditional pattern, either for a subpattern having been matched, or for
1568 recursion in such a pattern, all the associated numbered subpatterns are
1569 tested, and the overall condition is true if the condition is true for any
1570 one of them. This is the way Perl works, and is also more like the way
1571 testing by number works.
1572
1573
1574 Version 7.9 11-Apr-09
1575 ---------------------
1576
1577 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1578 (pcretest), all targets were linked against these libraries. This included
1579 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1580 libraries. This caused unwanted dependencies to be created. This problem
1581 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1582 pcretest is linked with readline.
1583
1584 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1585 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1586 moved outside it again, because FALSE and TRUE are already defined in AIX,
1587 but BOOL is not.
1588
1589 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1590 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1591
1592 4. The pcregrep documentation said spaces were inserted as well as colons (or
1593 hyphens) following file names and line numbers when outputting matching
1594 lines. This is not true; no spaces are inserted. I have also clarified the
1595 wording for the --colour (or --color) option.
1596
1597 5. In pcregrep, when --colour was used with -o, the list of matching strings
1598 was not coloured; this is different to GNU grep, so I have changed it to be
1599 the same.
1600
1601 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1602 each matching line was coloured. Now it goes on to look for further matches
1603 of any of the test patterns, which is the same behaviour as GNU grep.
1604
1605 7. A pattern that could match an empty string could cause pcregrep to loop; it
1606 doesn't make sense to accept an empty string match in pcregrep, so I have
1607 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1608 seems to be how GNU grep behaves.
1609
1610 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1611 start or after a newline", because the conditional assertion was not being
1612 correctly handled. The rule now is that both the assertion and what follows
1613 in the first alternative must satisfy the test.
1614
1615 9. If auto-callout was enabled in a pattern with a conditional group whose
1616 condition was an assertion, PCRE could crash during matching, both with
1617 pcre_exec() and pcre_dfa_exec().
1618
1619 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1620 used for matching.
1621
1622 11. Unicode property support in character classes was not working for
1623 characters (bytes) greater than 127 when not in UTF-8 mode.
1624
1625 12. Added the -M command line option to pcretest.
1626
1627 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1628
1629 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1630
1631 16. Added comments and documentation about mis-use of no_arg in the C++
1632 wrapper.
1633
1634 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1635 from Martin Jerabek that uses macro names for all relevant character and
1636 string constants.
1637
1638 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1639 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1640 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1641 these, but not everybody uses configure.
1642
1643 19. A conditional group that had only one branch was not being correctly
1644 recognized as an item that could match an empty string. This meant that an
1645 enclosing group might also not be so recognized, causing infinite looping
1646 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1647 with the subject "ab", where knowledge that the repeated group can match
1648 nothing is needed in order to break the loop.
1649
1650 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1651 exec(), but without supplying a callout function, matching went wrong.
1652
1653 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1654 leak if the size of the offset vector was greater than 30. When the vector
1655 is smaller, the saved offsets during recursion go onto a local stack
1656 vector, but for larger vectors malloc() is used. It was failing to free
1657 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1658 error, in fact).
1659
1660 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1661 heapframe that is used only when UTF-8 support is enabled. This caused no
1662 problem, but was untidy.
1663
1664 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1665 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1666 included within another project.
1667
1668 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1669 slightly modified by me:
1670
1671 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1672 not building pcregrep.
1673
1674 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1675 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1676
1677 25. Forward references, both numeric and by name, in patterns that made use of
1678 duplicate group numbers, could behave incorrectly or give incorrect errors,
1679 because when scanning forward to find the reference group, PCRE was not
1680 taking into account the duplicate group numbers. A pattern such as
1681 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1682
1683 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1684 the feature of a custom pointer more persuasive (as requested by a user).
1685
1686 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1687 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1688 pre-defined.
1689
1690 28. Added support for (*UTF8) at the start of a pattern.
1691
1692 29. Arrange for flags added by the "release type" setting in CMake to be shown
1693 in the configuration summary.
1694
1695
1696 Version 7.8 05-Sep-08
1697 ---------------------
1698
1699 1. Replaced UCP searching code with optimized version as implemented for Ad
1700 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1701 stage table and inline lookup instead of a function, giving speed ups of 2
1702 to 5 times on some simple patterns that I tested. Permission was given to
1703 distribute the MultiStage2.py script that generates the tables (it's not in
1704 the tarball, but is in the Subversion repository).
1705
1706 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1707 scripts.
1708
1709 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1710 a group with a zero qualifier. The result of the study could be incorrect,
1711 or the function might crash, depending on the pattern.
1712
1713 4. Caseless matching was not working for non-ASCII characters in back
1714 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1715 It now works when Unicode Property Support is available.
1716
1717 5. In pcretest, an escape such as \x{de} in the data was always generating
1718 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1719 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1720 truncation.
1721
1722 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1723
1724 7. Added two (int) casts to pcregrep when printing the difference of two
1725 pointers, in case they are 64-bit values.
1726
1727 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1728 test 2 if it fails.
1729
1730 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1731 and a #define of that name to empty if it is not externally set. This is to
1732 allow users of MSVC to set it if necessary.
1733
1734 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1735 the convenience functions in the pcre_get.c source file.
1736
1737 11. An option change at the start of a pattern that had top-level alternatives
1738 could cause overwriting and/or a crash. This command provoked a crash in
1739 some environments:
1740
1741 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1742
1743 This potential security problem was recorded as CVE-2008-2371.
1744
1745 12. For a pattern where the match had to start at the beginning or immediately
1746 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1747 pcre_dfa_exec() could read past the end of the passed subject if there was
1748 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1749 pcretest so that it places the subject at the end of its malloc-ed buffer.
1750
1751 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1752 exec() might read past the end of the data buffer in UTF-8 mode.
1753
1754 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1755 the data contained the byte 0x85 as part of a UTF-8 character within its
1756 first line. This applied both to normal and DFA matching.
1757
1758 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1759 /^[^d]*?$/8 failed to match "abc".
1760
1761 16. Added a missing copyright notice to pcrecpp_internal.h.
1762
1763 17. Make it more clear in the documentation that values returned from
1764 pcre_exec() in ovector are byte offsets, not character counts.
1765
1766 18. Tidied a few places to stop certain compilers from issuing warnings.
1767
1768 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1769 supplied by Stefan Weber. I made a further small update for 7.8 because
1770 there is a change of source arrangements: the pcre_searchfuncs.c module is
1771 replaced by pcre_ucd.c.
1772
1773
1774 Version 7.7 07-May-08
1775 ---------------------
1776
1777 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1778 a string to a long long, pretend we don't even have a long long." This is
1779 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1780
1781 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1782 pre-7.6 versions, which defined a global no_arg variable instead of putting
1783 it in the RE class. (See also #8 below.)
1784
1785 3. Remove a line of dead code, identified by coverity and reported by Nuno
1786 Lopes.
1787
1788 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1789
1790 (1) The include/exclude patterns were being applied to the whole pathnames
1791 of files, instead of just to the final components.
1792
1793 (2) If there was more than one level of directory, the subdirectories were
1794 skipped unless they satisfied the include/exclude conditions. This is
1795 inconsistent with GNU grep (and could even be seen as contrary to the
1796 pcregrep specification - which I improved to make it absolutely clear).
1797 The action now is always to scan all levels of directory, and just
1798 apply the include/exclude patterns to regular files.
1799
1800 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1801 --exclude_dir in the tests to avoid scanning .svn directories.
1802
1803 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1804 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1805 doesn't support NULs in patterns.
1806
1807 7. Added some missing "const"s to declarations of static tables in
1808 pcre_compile.c and pcre_dfa_exec.c.
1809
1810 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1811 caused by fix #2 above. (Subsequently also a second patch to fix the
1812 first patch. And a third patch - this was a messy problem.)
1813
1814 9. Applied Craig's patch to remove the use of push_back().
1815
1816 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1817 matching function regexec().
1818
1819 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1820 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1821 references. PCRE supports relative numbers with this syntax (I don't think
1822 Oniguruma does).
1823
1824 12. Previously, a group with a zero repeat such as (...){0} was completely
1825 omitted from the compiled regex. However, this means that if the group
1826 was called as a subroutine from elsewhere in the pattern, things went wrong
1827 (an internal error was given). Such groups are now left in the compiled
1828 pattern, with a new opcode that causes them to be skipped at execution
1829 time.
1830
1831 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1832 to the way PCRE behaves:
1833
1834 (a) A lone ] character is dis-allowed (Perl treats it as data).
1835
1836 (b) A back reference to an unmatched subpattern matches an empty string
1837 (Perl fails the current match path).
1838
1839 (c) A data ] in a character class must be notated as \] because if the
1840 first data character in a class is ], it defines an empty class. (In
1841 Perl it is not possible to have an empty class.) The empty class []
1842 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1843 The negative empty class [^] matches any one character, independently
1844 of the DOTALL setting.
1845
1846 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1847 non-existent subpattern following a character class starting with ']' and
1848 containing () gave an internal compiling error instead of "reference to
1849 non-existent subpattern". Fortunately, when the pattern did exist, the
1850 compiled code was correct. (When scanning forwards to check for the
1851 existencd of the subpattern, it was treating the data ']' as terminating
1852 the class, so got the count wrong. When actually compiling, the reference
1853 was subsequently set up correctly.)
1854
1855 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1856 it was being rejected as not supported by pcre_dfa_exec(), even though
1857 other assertions are supported. I have made pcre_dfa_exec() support
1858 (*FAIL).
1859
1860 16. The implementation of 13c above involved the invention of a new opcode,
1861 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1862 cannot be changed at match time, I realized I could make a small
1863 improvement to matching performance by compiling OP_ALLANY instead of
1864 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1865 on the OP_ANY path.
1866
1867 17. Compiling pcretest on Windows with readline support failed without the
1868 following two fixes: (1) Make the unistd.h include conditional on
1869 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1870
1871 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1872 ncurses library to be included for pcretest when ReadLine support is
1873 requested, but also to allow for it to be overridden. This patch came from
1874 Daniel Bergström.
1875
1876 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1877 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1878 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1879 spotting this.
1880
1881
1882 Version 7.6 28-Jan-08
1883 ---------------------
1884
1885 1. A character class containing a very large number of characters with
1886 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1887 overflow.
1888
1889 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1890 HAVE_LONG_LONG is not defined.
1891
1892 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1893 bring them up to date and include new features. This patch includes:
1894
1895 - Fixed PH's badly added libz and libbz2 support.
1896 - Fixed a problem with static linking.
1897 - Added pcredemo. [But later removed - see 7 below.]
1898 - Fixed dftables problem and added an option.
1899 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1900 HAVE_LONG_LONG.
1901 - Added readline support for pcretest.
1902 - Added an listing of the option settings after cmake has run.
1903
1904 4. A user submitted a patch to Makefile that makes it easy to create
1905 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1906 Makefile.am that cause it to include this special target, without
1907 affecting anything else. Note that the same mingw target plus all
1908 the other distribution libraries and programs are now supported
1909 when configuring with CMake (see 6 below) instead of with
1910 Configure/Make.
1911
1912 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1913 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1914 exported in the Windows port". It has not yet been confirmed that the patch
1915 solves the problem, but it does no harm.
1916
1917 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1918 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1919 with CMake, and also correct the comment about stack recursion.
1920
1921 7. Remove the automatic building of pcredemo from the ./configure system and
1922 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1923 of a program that users should build themselves after PCRE is installed, so
1924 building it automatically is not really right. What is more, it gave
1925 trouble in some build environments.
1926
1927 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1928
1929
1930 Version 7.5 10-Jan-08
1931 ---------------------
1932
1933 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1934 values in parens when parsing an RE using the C++ wrapper."
1935
1936 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1937 Characters greater than 255 were excluded from the class instead of being
1938 included.
1939
1940 3. The same bug as (2) above applied to negated POSIX classes such as
1941 [:^space:].
1942
1943 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1944 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1945 I have changed it.
1946
1947 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1948 first named subpattern) and a construct such as (?&a) would reference the
1949 first named subpattern whose name started with "a" (in other words, the
1950 length check was missing). Both these problems are fixed. "Subpattern name
1951 expected" is now given for (?&) (a zero-length name), and this patch also
1952 makes it give the same error for \k'' (previously it complained that that
1953 was a reference to a non-existent subpattern).
1954
1955 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1956 this is right because (?- can be followed by option settings as well as by
1957 digits. I have, however, made the messages clearer.
1958
1959 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1960 than the number used in the conditional) now cause a compile-time error.
1961 This is actually not compatible with Perl, which accepts such patterns, but
1962 treats the conditional as always being FALSE (as PCRE used to), but it
1963 seems to me that giving a diagnostic is better.
1964
1965 8. Change "alphameric" to the more common word "alphanumeric" in comments
1966 and messages.
1967
1968 9. Fix two occurrences of "backslash" in comments that should have been
1969 "backspace".
1970
1971 10. Remove two redundant lines of code that can never be obeyed (their function
1972 was moved elsewhere).
1973
1974 11. The program that makes PCRE's Unicode character property table had a bug
1975 which caused it to generate incorrect table entries for sequences of
1976 characters that have the same character type, but are in different scripts.
1977 It amalgamated them into a single range, with the script of the first of
1978 them. In other words, some characters were in the wrong script. There were
1979 thirteen such cases, affecting characters in the following ranges:
1980
1981 U+002b0 - U+002c1
1982 U+0060c - U+0060d
1983 U+0061e - U+00612
1984 U+0064b - U+0065e
1985 U+0074d - U+0076d
1986 U+01800 - U+01805
1987 U+01d00 - U+01d77
1988 U+01d9b - U+01dbf
1989 U+0200b - U+0200f
1990 U+030fc - U+030fe
1991 U+03260 - U+0327f
1992 U+0fb46 - U+0fbb1
1993 U+10450 - U+1049d
1994
1995 12. The -o option (show only the matching part of a line) for pcregrep was not
1996 compatible with GNU grep in that, if there was more than one match in a
1997 line, it showed only the first of them. It now behaves in the same way as
1998 GNU grep.
1999
2000 13. If the -o and -v options were combined for pcregrep, it printed a blank
2001 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2002 does the same. The return code can be used to tell if there were any
2003 non-matching lines.
2004
2005 14. Added --file-offsets and --line-offsets to pcregrep.
2006
2007 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2008 infinitely looping recursion. The bug was that positive lookaheads were not
2009 being skipped when checking for a possible empty match (negative lookaheads
2010 and both kinds of lookbehind were skipped).
2011
2012 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2013 inclusion of <windows.h> to before rather than after the definition of
2014 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2015
2016 17. Specifying a possessive quantifier with a specific limit for a Unicode
2017 character property caused pcre_compile() to compile bad code, which led at
2018 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2019 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2020 caused the error; without that there was no problem.
2021
2022 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2023
2024 19. Added --enable-pcretest-libreadline.
2025
2026 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2027 RE::GlobalReplace(). As a result, the number of replacements returned was
2028 double what it should be. I removed one of the increments, but Craig sent a
2029 later patch that removed the other one (the right fix) and added unit tests
2030 that check the return values (which was not done before).
2031
2032 21. Several CMake things:
2033
2034 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2035 the names libpcre and libpcreposix, not just pcre and pcreposix.
2036
2037 (2) The above change means that pcretest and pcregrep are now correctly
2038 linked with the newly-built libraries, not previously installed ones.
2039
2040 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2041
2042 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2043 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2044 UTF-8 newline character). The key issue is that the pattern starts .*;
2045 this means that the match must be either at the beginning, or after a
2046 newline. The bug was in the code for advancing after a failed match and
2047 checking that the new position followed a newline. It was not taking
2048 account of UTF-8 characters correctly.
2049
2050 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2051 character classes. PCRE was not treating the sequence [:...:] as a
2052 character class unless the ... were all letters. Perl, however, seems to
2053 allow any characters between [: and :], though of course it rejects as
2054 unknown any "names" that contain non-letters, because all the known class
2055 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2056 for example, whereas PCRE did not - it did not recognize a POSIX character
2057 class. This seemed a bit dangerous, so the code has been changed to be
2058 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2059 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2060 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2061 Perl does, and where it didn't before.
2062
2063 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2064 Windows environments %n is disabled by default.
2065
2066
2067 Version 7.4 21-Sep-07
2068 ---------------------
2069
2070 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2071 means that a class such as [\s] counted as "explicit reference to CR or
2072 LF". That isn't really right - the whole point of the change was to try to
2073 help when there was an actual mention of one of the two characters. So now
2074 the change happens only if \r or \n (or a literal CR or LF) character is
2075 encountered.
2076
2077 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2078 of both had grown to the point where there were only 3 bits left.
2079 Fortunately, there was spare space in the data structure, and so I have
2080 moved the internal flags into a new 16-bit field to free up more option
2081 bits.
2082
2083 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2084 but did not set the internal JCHANGED flag - either of these is enough to
2085 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2086 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2087 start sets both bits.
2088
2089 4. Added options (at build time, compile time, exec time) to change \R from
2090 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2091
2092 5. doc/pcresyntax.html was missing from the distribution.
2093
2094 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2095 compatibility, even though it is no longer used.
2096
2097 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2098 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2099 windows.h file is present (where different names are used). [This was
2100 reversed later after testing - see 16 below.]
2101
2102 8. Changed all #include <config.h> to #include "config.h". There were also
2103 some further <pcre.h> cases that I changed to "pcre.h".
2104
2105 9. When pcregrep was used with the --colour option, it missed the line ending
2106 sequence off the lines that it output.
2107
2108 10. It was pointed out to me that arrays of string pointers cause lots of
2109 relocations when a shared library is dynamically loaded. A technique of
2110 using a single long string with a table of offsets can drastically reduce
2111 these. I have refactored PCRE in four places to do this. The result is
2112 dramatic:
2113
2114 Originally: 290
2115 After changing UCP table: 187
2116 After changing error message table: 43
2117 After changing table of "verbs" 36
2118 After changing table of Posix names 22
2119
2120 Thanks to the folks working on Gregex for glib for this insight.
2121
2122 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2123 unicode-properties was also set.
2124
2125 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2126
2127 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2128 checked only for CRLF.
2129
2130 14. Added casts to pcretest.c to avoid compiler warnings.
2131
2132 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2133
2134 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2135 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2136 entirely. This removes changes made in 7 above.
2137
2138 17. The CMake files have been updated, and there is now more information about
2139 building with CMake in the NON-UNIX-USE document.
2140
2141
2142 Version 7.3 28-Aug-07
2143 ---------------------
2144
2145 1. In the rejigging of the build system that eventually resulted in 7.1, the
2146 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2147 brackets there is not right, since it causes compilers to look for an
2148 installed pcre.h, not the version that is in the source that is being
2149 compiled (which of course may be different). I have changed it back to:
2150
2151 #include "pcre.h"
2152
2153 I have a vague recollection that the change was concerned with compiling in
2154 different directories, but in the new build system, that is taken care of
2155 by the VPATH setting the Makefile.
2156
2157 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2158 when the subject happened to end in the byte 0x85 (e.g. if the last
2159 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2160 characters but of course it shouldn't be taken as a newline when it is part
2161 of another character. The bug was that, for an unlimited repeat of . in
2162 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2163 characters when looking for a newline.
2164
2165 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2166
2167 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2168 in debug output.
2169
2170 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2171 long printing in the pcrecpp unittest when running under MinGW.
2172
2173 6. ESC_K was left out of the EBCDIC table.
2174
2175 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2176 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2177 limit also applies to "virtual nesting" when a pattern is recursive, and in
2178 this case 1000 isn't so big. I have been able to remove this limit at the
2179 expense of backing off one optimization in certain circumstances. Normally,
2180 when pcre_exec() would call its internal match() function recursively and
2181 immediately return the result unconditionally, it uses a "tail recursion"
2182 feature to save stack. However, when a subpattern that can match an empty
2183 string has an unlimited repetition quantifier, it no longer makes this
2184 optimization. That gives it a stack frame in which to save the data for
2185 checking that an empty string has been matched. Previously this was taken
2186 from the 1000-entry workspace that had been reserved. So now there is no
2187 explicit limit, but more stack is used.
2188
2189 8. Applied Daniel's patches to solve problems with the import/export magic
2190 syntax that is required for Windows, and which was going wrong for the
2191 pcreposix and pcrecpp parts of the library. These were overlooked when this
2192 problem was solved for the main library.
2193
2194 9. There were some crude static tests to avoid integer overflow when computing
2195 the size of patterns that contain repeated groups with explicit upper
2196 limits. As the maximum quantifier is 65535, the maximum group length was
2197 set at 30,000 so that the product of these two numbers did not overflow a
2198 32-bit integer. However, it turns out that people want to use groups that
2199 are longer than 30,000 bytes (though not repeat them that many times).
2200 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2201 made it possible to implement the integer overflow checks in a much more
2202 dynamic way, which I have now done. The artificial limitation on group
2203 length has been removed - we now have only the limit on the total length of
2204 the compiled pattern, which depends on the LINK_SIZE setting.
2205
2206 10. Fixed a bug in the documentation for get/copy named substring when
2207 duplicate names are permitted. If none of the named substrings are set, the
2208 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2209 empty string.
2210
2211 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2212 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2213 because the ] is interpreted as the first data character and the
2214 terminating ] is not found. PCRE has been made compatible with Perl in this
2215 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2216 cause memory overwriting.
2217
2218 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2219 string has been matched (to stop an infinite loop). It was not recognizing
2220 a conditional subpattern that could match an empty string if that
2221 subpattern was within another subpattern. For example, it looped when
2222 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2223 condition was not nested. This bug has been fixed.
2224
2225 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2226 past the start of the subject in the presence of bytes with the top bit
2227 set, for example "\x8aBCD".
2228
2229 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2230 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2231
2232 14. Optimized (?!) to (*FAIL).
2233
2234 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2235 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2236 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2237 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2238 does: it's just the validity check that is more restrictive.
2239
2240 16. Inserted checks for integer overflows during escape sequence (backslash)
2241 processing, and also fixed erroneous offset values for syntax errors during
2242 backslash processing.
2243
2244 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2245 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2246
2247 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2248 caused an overrun.
2249
2250 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2251 something other than just ASCII characters) inside a group that had an
2252 unlimited repeat caused a loop at compile time (while checking to see
2253 whether the group could match an empty string).
2254
2255 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2256 [\P{Any}] did so. (Error in the code for printing property names.)
2257
2258 21. An orphan \E inside a character class could cause a crash.
2259
2260 22. A repeated capturing bracket such as (A)? could cause a wild memory
2261 reference during compilation.
2262
2263 23. There are several functions in pcre_compile() that scan along a compiled
2264 expression for various reasons (e.g. to see if it's fixed length for look
2265 behind). There were bugs in these functions when a repeated \p or \P was
2266 present in the pattern. These operators have additional parameters compared
2267 with \d, etc, and these were not being taken into account when moving along
2268 the compiled data. Specifically:
2269
2270 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2271 length.
2272
2273 (b) An item such as \pL+ within a repeated group could cause crashes or
2274 loops.
2275
2276 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2277 "reference to non-existent subpattern" error.
2278
2279 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2280
2281 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2282 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2283
2284 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2285
2286 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2287 character were causing crashes (broken optimization).
2288
2289 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2290 \p or \P) caused a compile-time loop.
2291
2292 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2293 break. For example, the unstudied pattern [\r\n]A does not match the string
2294 "\r\nA" because change 7.0/46 below moves the current point on by two
2295 characters after failing to match at the start. However, the pattern \nA
2296 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2297 the same is true. There doesn't seem any very clean way out of this, but
2298 what I have chosen to do makes the common cases work: PCRE now takes note
2299 of whether there can be an explicit match for \r or \n anywhere in the
2300 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2301 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2302 pattern has explicit CR or LF references.
2303
2304 29. Added (*CR) etc for changing newline setting at start of pattern.
2305
2306
2307 Version 7.2 19-Jun-07
2308 ---------------------
2309
2310 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2311 which is apparently normally available under Windows.
2312
2313 2. Re-jig the pcregrep tests with different newline settings in an attempt
2314 to make them independent of the local environment's newline setting.
2315
2316 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2317
2318 4. Some of the "internals" tests were previously cut out when the link size
2319 was not 2, because the output contained actual offsets. The recent new
2320 "Z" feature of pcretest means that these can be cut out, making the tests
2321 usable with all link sizes.
2322
2323 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2324 stack recursion. This gives a massive performance boost under BSD, but just
2325 a small improvement under Linux. However, it saves one field in the frame
2326 in all cases.
2327
2328 6. Added more features from the forthcoming Perl 5.10:
2329
2330 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2331 recursion call. It refers to the nth most recently opened parentheses.
2332
2333 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2334 to be opened parentheses.
2335
2336 (c) Conditions that refer to capturing parentheses can be specified
2337 relatively, for example, (?(-2)... or (?(+3)...
2338
2339 (d) \K resets the start of the current match so that everything before
2340 is not part of it.
2341
2342 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2343
2344 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2345 reference syntax.
2346
2347 (g) (?| introduces a group in which the numbering of parentheses in each
2348 alternative starts with the same number.
2349
2350 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2351
2352 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2353 PCRE_INFO_JCHANGED.
2354
2355 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2356 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2357 for detecting groups that can match an empty string.
2358
2359 9. A pattern with a very large number of alternatives (more than several
2360 hundred) was running out of internal workspace during the pre-compile
2361 phase, where pcre_compile() figures out how much memory will be needed. A
2362 bit of new cunning has reduced the workspace needed for groups with
2363 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2364 workspace instead of running out of the 4096 that are available.
2365
2366 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2367
2368 11. Applied patch from Google to remove an optimization that didn't quite work.
2369 The report of the bug said:
2370
2371 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2372 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2373 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2374
2375 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2376 it matched the wrong number of bytes.
2377
2378
2379 Version 7.1 24-Apr-07
2380 ---------------------
2381
2382 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2383 that is more "standard", making use of automake and other Autotools. There
2384 is some re-arrangement of the files and adjustment of comments consequent
2385 on this.
2386
2387 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2388 for recursive directory scanning broke on some systems because the files
2389 are not scanned in any specific order and on different systems the order
2390 was different. A call to "sort" has been inserted into RunGrepTest for the
2391 approprate test as a short-term fix. In the longer term there may be an
2392 alternative.
2393
2394 3. I had an email from Eric Raymond about problems translating some of PCRE's
2395 man pages to HTML (despite the fact that I distribute HTML pages, some
2396 people do their own conversions for various reasons). The problems
2397 concerned the use of low-level troff macros .br and .in. I have therefore
2398 removed all such uses from the man pages (some were redundant, some could
2399 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2400 HTML has been updated to handle .nf/.fi and to complain if it encounters
2401 .br or .in.
2402
2403 4. Updated comments in configure.ac that get placed in config.h.in and also
2404 arranged for config.h to be included in the distribution, with the name
2405 config.h.generic, for the benefit of those who have to compile without
2406 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2407
2408 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2409 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2410 makevp.bat for the current PCRE, using the additional files
2411 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2412
2413 6. A Windows user reported a minor discrepancy with test 2, which turned out
2414 to be caused by a trailing space on an input line that had got lost in his
2415 copy. The trailing space was an accident, so I've just removed it.
2416
2417 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2418 that is needed.
2419
2420 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2421 as "const" (a) because they are and (b) because it helps the PHP
2422 maintainers who have recently made a script to detect big data structures
2423 in the php code that should be moved to the .rodata section. I remembered
2424 to update Builducptable as well, so it won't revert if ucptable.h is ever
2425 re-created.
2426
2427 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2428 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2429 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2430 support is not required. This saves 1.5-2K of code, which is important in
2431 some applications.
2432
2433 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2434 so as not to refer to the tables, even though these functions will never be
2435 called when UTF-8 support is disabled. Otherwise there are problems with a
2436 shared library.
2437
2438 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2439
2440 (a) It was defining its arguments as char * instead of void *.
2441
2442 (b) It was assuming that all moves were upwards in memory; this was true
2443 a long time ago when I wrote it, but is no longer the case.
2444
2445 The emulated memove() is provided for those environments that have neither
2446 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2447 is clearly not the case, as these two bugs were recently reported.
2448
2449 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2450 and Detrail to create the HTML documentation, the .txt form of the man
2451 pages, and it removes trailing spaces from listed files. It also creates
2452 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2453 case, it wraps all the #defines with #ifndefs. This script should be run
2454 before "make dist".
2455
2456 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2457 with Unicode property support.
2458
2459 (a) For a maximizing quantifier, if the two different cases of the
2460 character were of different lengths in their UTF-8 codings (there are
2461 some cases like this - I found 11), and the matching function had to
2462 back up over a mixture of the two cases, it incorrectly assumed they
2463 were both the same length.
2464
2465 (b) When PCRE was configured to use the heap rather than the stack for
2466 recursion during matching, it was not correctly preserving the data for
2467 the other case of a UTF-8 character when checking ahead for a match
2468 while processing a minimizing repeat. If the check also involved
2469 matching a wide character, but failed, corruption could cause an
2470 erroneous result when trying to check for a repeat of the original
2471 character.
2472
2473 13. Some tidying changes to the testing mechanism:
2474
2475 (a) The RunTest script now detects the internal link size and whether there
2476 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2477 values substituted by "configure". (The RunGrepTest script already did
2478 this for UTF-8.) The configure.ac script no longer substitutes the
2479 relevant variables.
2480
2481 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2482 with length and offset values. This means that the output is different
2483 for different internal link sizes. Test 2 is skipped for link sizes
2484 other than 2 because of this, bypassing the problem. Unfortunately,
2485 there was also a test in test 3 (the locale tests) that used /B and
2486 failed for link sizes other than 2. Rather than cut the whole test out,
2487 I have added a new /Z option to pcretest that replaces the length and
2488 offset values with spaces. This is now used to make test 3 independent
2489 of link size. (Test 2 will be tidied up later.)
2490
2491 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2492 segmentation fault instead of returning the appropriate error message.
2493
2494 15. In multiline mode when the newline sequence was set to "any", the pattern
2495 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2496 This doesn't seem right; it now treats the CRLF combination as the line
2497 ending, and so does not match in that case. It's only a pattern such as ^$
2498 that would hit this one: something like ^ABC$ would have failed after \r
2499 and then tried again after \r\n.
2500
2501 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2502 in an attempt to make files that differ only in their line terminators
2503 compare equal. This works on Linux.
2504
2505 17. Under certain error circumstances pcregrep might try to free random memory
2506 as it exited. This is now fixed, thanks to valgrind.
2507
2508 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2509 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2510 was because its rules for how to advance for /g after matching an empty
2511 string at the end of a line did not allow for this case. They now check for
2512 it specially.
2513
2514 20. pcretest is supposed to handle patterns and data of any length, by
2515 extending its buffers when necessary. It was getting this wrong when the
2516 buffer for a data line had to be extended.
2517
2518 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2519 CRLF as a newline sequence.
2520
2521 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2522 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2523 I have nevertheless tidied it up.
2524
2525 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2526
2527 24. Added a man page for pcre-config.
2528
2529
2530 Version 7.0 19-Dec-06
2531 ---------------------
2532
2533 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2534 moving to gcc 4.1.1.
2535
2536 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2537 sys/time.h, which is documented as needed for this function. It doesn't
2538 seem to matter on Linux, but it showed up on some releases of OS X.
2539
2540 3. It seems that there are systems where bytes whose values are greater than
2541 127 match isprint() in the "C" locale. The "C" locale should be the
2542 default when a C program starts up. In most systems, only ASCII printing
2543 characters match isprint(). This difference caused the output from pcretest
2544 to vary, making some of the tests fail. I have changed pcretest so that:
2545
2546 (a) When it is outputting text in the compiled version of a pattern, bytes
2547 other than 32-126 are always shown as hex escapes.
2548
2549 (b) When it is outputting text that is a matched part of a subject string,
2550 it does the same, unless a different locale has been set for the match
2551 (using the /L modifier). In this case, it uses isprint() to decide.
2552
2553 4. Fixed a major bug that caused incorrect computation of the amount of memory
2554 required for a compiled pattern when options that changed within the
2555 pattern affected the logic of the preliminary scan that determines the
2556 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2557 that the computed length was too small. The symptoms of this bug were
2558 either the PCRE error "internal error: code overflow" from pcre_compile(),
2559 or a glibc crash with a message such as "pcretest: free(): invalid next
2560 size (fast)". Examples of patterns that provoked this bug (shown in
2561 pcretest format) are:
2562
2563 /(?-x: )/x
2564 /(?x)(?-x: \s*#\s*)/
2565 /((?i)[\x{c0}])/8
2566 /(?i:[\x{c0}])/8
2567
2568 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2569 is now done differently.
2570
2571 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2572 wrapper classes; (b) implement a new function in the C++ scanner that is
2573 more efficient than the old way of doing things because it avoids levels of
2574 recursion in the regex matching; (c) add a paragraph to the documentation
2575 for the FullMatch() function.
2576
2577 6. The escape sequence \n was being treated as whatever was defined as
2578 "newline". Not only was this contrary to the documentation, which states
2579 that \n is character 10 (hex 0A), but it also went horribly wrong when
2580 "newline" was defined as CRLF. This has been fixed.
2581
2582 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2583 was being set to -1 for the "end of line" case (supposedly a value that no
2584 character can have). Though this value is never used (the check for end of
2585 line is "zero bytes in current character"), it caused compiler complaints.
2586 I've changed it to 0xffffffff.
2587
2588 8. In pcre_version.c, the version string was being built by a sequence of
2589 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2590 string (as it is for production releases) called a macro with an empty
2591 argument. The C standard says the result of this is undefined. The gcc
2592 compiler treats it as an empty string (which was what was wanted) but it is
2593 reported that Visual C gives an error. The source has been hacked around to
2594 avoid this problem.
2595
2596 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2597 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2598 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2599 of them did).
2600
2601 10. Originally, pcretest opened its input and output without "b"; then I was
2602 told that "b" was needed in some environments, so it was added for release
2603 5.0 to both the input and output. (It makes no difference on Unix-like
2604 systems.) Later I was told that it is wrong for the input on Windows. I've
2605 now abstracted the modes into two macros, to make it easier to fiddle with
2606 them, and removed "b" from the input mode under Windows.
2607
2608 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2609
2610 12. Added -help and --help to pcretest as an official way of being reminded
2611 of the options.
2612
2613 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2614 and pcrecpp.cc because they annoy compilers at high warning levels.
2615
2616 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2617
2618 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2619 scripts are not C programs :-) and which was not noticed because it works
2620 on Linux.
2621
2622 16. pcretest is supposed to handle any length of pattern and data line (as one
2623 line or as a continued sequence of lines) by extending its input buffer if
2624 necessary. This feature was broken for very long pattern lines, leading to
2625 a string of junk being passed to pcre_compile() if the pattern was longer
2626 than about 50K.
2627
2628 17. I have done a major re-factoring of the way pcre_compile() computes the
2629 amount of memory needed for a compiled pattern. Previously, there was code
2630 that made a preliminary scan of the pattern in order to do this. That was
2631 OK when PCRE was new, but as the facilities have expanded, it has become
2632 harder and harder to keep it in step with the real compile phase, and there
2633 have been a number of bugs (see for example, 4 above). I have now found a
2634 cunning way of running the real compile function in a "fake" mode that
2635 enables it to compute how much memory it would need, while actually only
2636 ever using a few hundred bytes of working memory and without too many
2637 tests of the mode. This should make future maintenance and development
2638 easier. A side effect of this work is that the limit of 200 on the nesting
2639 depth of parentheses has been removed (though this was never a serious
2640 limitation, I suspect). However, there is a downside: pcre_compile() now
2641 runs more slowly than before (30% or more, depending on the pattern). I
2642 hope this isn't a big issue. There is no effect on runtime performance.
2643
2644 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2645 newline (only possible for the last line of a file) and it was a
2646 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2647
2648 19. Added additional timing features to pcretest. (1) The -tm option now times
2649 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2650 separate command line item, by a number that specifies the number of
2651 repeats to use when timing. The default is 50000; this gives better
2652 precision, but takes uncomfortably long for very large patterns.
2653
2654 20. Extended pcre_study() to be more clever in cases where a branch of a
2655 subpattern has no definite first character. For example, (a*|b*)[cd] would
2656 previously give no result from pcre_study(). Now it recognizes that the
2657 first character must be a, b, c, or d.
2658
2659 21. There was an incorrect error "recursive call could loop indefinitely" if
2660 a subpattern (or the entire pattern) that was being tested for matching an
2661 empty string contained only one non-empty item after a nested subpattern.
2662 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2663 incorrectly, because the \d was being skipped in the check.
2664
2665 22. The pcretest program now has a new pattern option /B and a command line
2666 option -b, which is equivalent to adding /B to every pattern. This causes
2667 it to show the compiled bytecode, without the additional information that
2668 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2669 is the same as /B/I).
2670
2671 23. A new optimization is now able automatically to treat some sequences such
2672 as a*b as a*+b. More specifically, if something simple (such as a character
2673 or a simple class like \d) has an unlimited quantifier, and is followed by
2674 something that cannot possibly match the quantified thing, the quantifier
2675 is automatically "possessified".
2676
2677 24. A recursive reference to a subpattern whose number was greater than 39
2678 went wrong under certain circumstances in UTF-8 mode. This bug could also
2679 have affected the operation of pcre_study().
2680
2681 25. Realized that a little bit of performance could be had by replacing
2682 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2683
2684 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2685
2686 27. Possessive quantifiers such as a++ were previously implemented by turning
2687 them into atomic groups such as ($>a+). Now they have their own opcodes,
2688 which improves performance. This includes the automatically created ones
2689 from 23 above.
2690
2691 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2692 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2693 the first matched character to be a colon. This applied both to named and
2694 numbered groups.
2695
2696 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2697
2698 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2699 building PCRE on Windows easier, so I have included it in the distribution.
2700
2701 31. There is now a check in pcretest against a ridiculously large number being
2702 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2703 loop, the loop is abandoned.
2704
2705 32. Forward references to subpatterns in conditions such as (?(2)...) where
2706 subpattern 2 is defined later cause pcre_compile() to search forwards in
2707 the pattern for the relevant set of parentheses. This search went wrong
2708 when there were unescaped parentheses in a character class, parentheses
2709 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2710
2711 33. "Subroutine" calls and backreferences were previously restricted to
2712 referencing subpatterns earlier in the regex. This restriction has now
2713 been removed.
2714
2715 34. Added a number of extra features that are going to be in Perl 5.10. On the
2716 whole, these are just syntactic alternatives for features that PCRE had
2717 previously implemented using the Python syntax or my own invention. The
2718 other formats are all retained for compatibility.
2719
2720 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2721 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2722 also .NET compatible.
2723
2724 (b) A recursion or subroutine call to a named group can now be defined as
2725 (?&name) as well as (?P>name).
2726
2727 (c) A backreference to a named group can now be defined as \k<name> or
2728 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2729 5.10, are also .NET compatible.
2730
2731 (d) A conditional reference to a named group can now use the syntax
2732 (?(<name>) or (?('name') as well as (?(name).
2733
2734 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2735 groups (named and numbered) that are never evaluated inline, but can be
2736 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2737 is always false. There may be only one alternative in such a group.
2738
2739 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2740 as the simple (?(R). The condition is true only if the most recent
2741 recursion is that of the given number or name. It does not search out
2742 through the entire recursion stack.
2743
2744 (g) The escape \gN or \g{N} has been added, where N is a positive or
2745 negative number, specifying an absolute or relative reference.
2746
2747 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2748 some "unreachable code" warnings.
2749
2750 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2751 things, this adds five new scripts.
2752
2753 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2754 There were also incompatibilities regarding the handling of \Q..\E inside
2755 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2756 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2757
2758 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2759 matches an empty string, and forcibly breaks the loop. There were bugs in
2760 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2761 against aaaa the result was just "a" rather than "aaaa", for example. Two
2762 separate and independent bugs (that affected different cases) have been
2763 fixed.
2764
2765 39. Refactored the code to abolish the use of different opcodes for small
2766 capturing bracket numbers. This is a tidy that I avoided doing when I
2767 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2768 The new approach is not only tidier, it makes it possible to reduce the
2769 memory needed to fix the previous bug (38).
2770
2771 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2772 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2773 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2774 mode.
2775
2776 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2777 report.
2778
2779 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2780 copy construction and assignment in the C++ wrapper.
2781
2782 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2783 couple of bugs that could have given wrong results in the "--newline=crlf"
2784 case.
2785
2786 44. Added a number of casts and did some reorganization of signed/unsigned int
2787 variables following suggestions from Dair Grant. Also renamed the variable
2788 "this" as "item" because it is a C++ keyword.
2789
2790 45. Arranged for dftables to add
2791
2792 #include "pcre_internal.h"
2793
2794 to pcre_chartables.c because without it, gcc 4.x may remove the array
2795 definition from the final binary if PCRE is built into a static library and
2796 dead code stripping is activated.
2797
2798 46. For an unanchored pattern, if a match attempt fails at the start of a
2799 newline sequence, and the newline setting is CRLF or ANY, and the next two
2800 characters are CRLF, advance by two characters instead of one.
2801
2802
2803 Version 6.7 04-Jul-06
2804 ---------------------
2805
2806 1. In order to handle tests when input lines are enormously long, pcretest has
2807 been re-factored so that it automatically extends its buffers when
2808 necessary. The code is crude, but this _is_ just a test program. The
2809 default size has been increased from 32K to 50K.
2810
2811 2. The code in pcre_study() was using the value of the re argument before
2812 testing it for NULL. (Of course, in any sensible call of the function, it
2813 won't be NULL.)
2814
2815 3. The memmove() emulation function in pcre_internal.h, which is used on
2816 systems that lack both memmove() and bcopy() - that is, hardly ever -
2817 was missing a "static" storage class specifier.
2818
2819 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2820 containing an extended class (one that cannot be represented by a bitmap
2821 because it contains high-valued characters or Unicode property items, e.g.
2822 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2823 pattern, but PCRE should not loop if you do not (it no longer does).
2824 [Detail: two cases were found: (a) a repeated subpattern containing an
2825 extended class; (b) a recursive reference to a subpattern that followed a
2826 previous extended class. It wasn't skipping over the extended class
2827 correctly when UTF-8 mode was not set.]
2828
2829 5. A negated single-character class was not being recognized as fixed-length
2830 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2831 compile error "lookbehind assertion is not fixed length".
2832
2833 6. The RunPerlTest auxiliary script was showing an unexpected difference
2834 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2835 write a Perl script that can interpret lines of an input file either as
2836 byte characters or as UTF-8, which is what "perltest" was being required to
2837 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2838 can't do is switch easily at run time between having the "use utf8;" pragma
2839 or not. In the end, I fudged it by using the RunPerlTest script to insert
2840 "use utf8;" explicitly for the UTF-8 tests.
2841
2842 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2843 the end of the subject string, contrary to the documentation and to what
2844 Perl does. This was true of both matching functions. Now it matches only at
2845 the start of the subject and immediately after *internal* newlines.
2846
2847 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2848 a pointer to an int instead of a pointer to an unsigned long int. This
2849 caused problems on 64-bit systems.
2850
2851 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2852 instance of the 'standard' template library not being so standard".
2853
2854 10. There was no check on the number of named subpatterns nor the maximum
2855 length of a subpattern name. The product of these values is used to compute
2856 the size of the memory block for a compiled pattern. By supplying a very
2857 long subpattern name and a large number of named subpatterns, the size
2858 computation could be caused to overflow. This is now prevented by limiting
2859 the length of names to 32 characters, and the number of named subpatterns
2860 to 10,000.
2861
2862 11. Subpatterns that are repeated with specific counts have to be replicated in
2863 the compiled pattern. The size of memory for this was computed from the
2864 length of the subpattern and the repeat count. The latter is limited to
2865 65535, but there was no limit on the former, meaning that integer overflow
2866 could in principle occur. The compiled length of a repeated subpattern is
2867 now limited to 30,000 bytes in order to prevent this.
2868
2869 12. Added the optional facility to have named substrings with the same name.
2870
2871 13. Added the ability to use a named substring as a condition, using the
2872 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2873 are numbers (not recommended). Forward references are permitted.
2874
2875 14. Added forward references in named backreferences (if you see what I mean).
2876
2877 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2878 pattern could run off the end of the subject. For example, the pattern
2879 "(?s)(.{1,5})"8 did this with the subject "ab".
2880
2881 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2882 PCRE_CASELESS was set when matching characters that were quantified with ?
2883 or *.
2884
2885 17. A character class other than a single negated character that had a minimum
2886 but no maximum quantifier - for example [ab]{6,} - was not handled
2887 correctly by pce_dfa_exec(). It would match only one character.
2888
2889 18. A valid (though odd) pattern that looked like a POSIX character
2890 class but used an invalid character after [ (for example [[,abc,]]) caused
2891 pcre_compile() to give the error "Failed: internal error: code overflow" or
2892 in some cases to crash with a glibc free() error. This could even happen if
2893 the pattern terminated after [[ but there just happened to be a sequence of
2894 letters, a binary zero, and a closing ] in the memory that followed.
2895
2896 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2897 over the years. Originally (before any Unicode support), just the bottom 8
2898 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2899 output from "man perlunicode" includes this:
2900
2901 The regular expression compiler produces polymorphic opcodes. That
2902 is, the pattern adapts to the data and automatically switches to
2903 the Unicode character scheme when presented with Unicode data--or
2904 instead uses a traditional byte scheme when presented with byte
2905 data.
2906
2907 Sadly, a wide octal escape does not cause a switch, and in a string with
2908 no other multibyte characters, these octal escapes are treated as before.
2909 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2910 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2911 Unicode string.
2912
2913 I have not perpetrated such confusion in PCRE. Up till now, it took just
2914 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2915 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2916 translate to the appropriate multibyte character.
2917
2918 29. Applied some refactoring to reduce the number of warnings from Microsoft
2919 and Borland compilers. This has included removing the fudge introduced
2920 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2921 a warning about an unused variable.
2922
2923 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2924 characters since release 4.0, because Perl (from release 5.004) does not.
2925 [Or at least, is documented not to: some releases seem to be in conflict
2926 with the documentation.] However, when a pattern was studied with
2927 pcre_study() and all its branches started with \s, PCRE still included VT
2928 as a possible starting character. Of course, this did no harm; it just
2929 caused an unnecessary match attempt.
2930
2931 22. Removed a now-redundant internal flag bit that recorded the fact that case
2932 dependency changed within the pattern. This was once needed for "required
2933 byte" processing, but is no longer used. This recovers a now-scarce options
2934 bit. Also moved the least significant internal flag bit to the most-
2935 significant bit of the word, which was not previously used (hangover from
2936 the days when it was an int rather than a uint) to free up another bit for
2937 the future.
2938
2939 23. Added support for CRLF line endings as well as CR and LF. As well as the
2940 default being selectable at build time, it can now be changed at runtime
2941 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2942 specify that it is scanning data with non-default line endings.
2943
2944 24. Changed the definition of CXXLINK to make it agree with the definition of
2945 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2946
2947 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2948 recursions. This makes a big different to stack usage for some patterns.
2949
2950 26. If a subpattern containing a named recursion or subroutine reference such
2951 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2952 the space required for the compiled pattern went wrong and gave too small a
2953 value. Depending on the environment, this could lead to "Failed: internal
2954 error: code overflow at offset 49" or "glibc detected double free or
2955 corruption" errors.
2956
2957 27. Applied patches from Google (a) to support the new newline modes and (b) to
2958 advance over multibyte UTF-8 characters in GlobalReplace.
2959
2960 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2961 difference for some implementation of PCRE in some Windows version.
2962
2963 29. Added some extra testing facilities to pcretest:
2964
2965 \q<number> in a data line sets the "match limit" value
2966 \Q<number> in a data line sets the "match recursion limt" value
2967 -S <number> sets the stack size, where <number> is in megabytes
2968
2969 The -S option isn't available for Windows.
2970
2971
2972 Version 6.6 06-Feb-06
2973 ---------------------
2974
2975 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2976 in pcreposix.h. I have copied the definition from pcre.h.
2977
2978 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2979 because pcre.h is no longer a built file.
2980
2981 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2982 not normally included in the compiled code.
2983
2984
2985 Version 6.5 01-Feb-06
2986 ---------------------
2987
2988 1. When using the partial match feature with pcre_dfa_exec(), it was not
2989 anchoring the second and subsequent partial matches at the new starting
2990 point. This could lead to incorrect results. For example, with the pattern
2991 /1234/, partially matching against "123" and then "a4" gave a match.
2992
2993 2. Changes to pcregrep:
2994
2995 (a) All non-match returns from pcre_exec() were being treated as failures
2996 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2997 error message is output. Some extra information is given for the
2998 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2999 probably the only errors that are likely to be caused by users (by
3000 specifying a regex that has nested indefinite repeats, for instance).
3001 If there are more than 20 of these errors, pcregrep is abandoned.
3002
3003 (b) A binary zero was treated as data while matching, but terminated the
3004 output line if it was written out. This has been fixed: binary zeroes
3005 are now no different to any other data bytes.
3006
3007 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3008 used to set a locale for matching. The --locale=xxxx long option has
3009 been added (no short equivalent) to specify a locale explicitly on the
3010 pcregrep command, overriding the environment variables.
3011
3012 (d) When -B was used with -n, some line numbers in the output were one less
3013 than they should have been.
3014
3015 (e) Added the -o (--only-matching) option.
3016
3017 (f) If -A or -C was used with -c (count only), some lines of context were
3018 accidentally printed for the final match.
3019
3020 (g) Added the -H (--with-filename) option.
3021
3022 (h) The combination of options -rh failed to suppress file names for files
3023 that were found from directory arguments.
3024
3025 (i) Added the -D (--devices) and -d (--directories) options.
3026
3027 (j) Added the -F (--fixed-strings) option.
3028
3029 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3030
3031 (l) Added the --colo(u)r option.
3032
3033 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3034 is not present by default.
3035
3036 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3037 items such as (?R) or (?1), when the recursion could match a number of
3038 alternatives. If it matched one of the alternatives, but subsequently,
3039 outside the recursion, there was a failure, the code tried to back up into
3040 the recursion. However, because of the way PCRE is implemented, this is not
3041 possible, and the result was an incorrect result from the match.
3042
3043 In order to prevent this happening, the specification of recursion has
3044 been changed so that all such subpatterns are automatically treated as
3045 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3046
3047 4. I had overlooked the fact that, in some locales, there are characters for
3048 which isalpha() is true but neither isupper() nor islower() are true. In
3049 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3050 and ordfeminine) are like this. This affected the treatment of \w and \W
3051 when they appeared in character classes, but not when they appeared outside
3052 a character class. The bit map for "word" characters is now created
3053 separately from the results of isalnum() instead of just taking it from the
3054 upper, lower, and digit maps. (Plus the underscore character, of course.)
3055
3056 5. The above bug also affected the handling of POSIX character classes such as
3057 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3058 permanent tables. Instead, the bit maps for such a class were previously
3059 created as the appropriate unions of the upper, lower, and digit bitmaps.
3060 Now they are created by subtraction from the [[:word:]] class, which has
3061 its own bitmap.
3062
3063 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3064 It is created by subtracting the vertical space characters (\x09, \x0a,
3065 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3066 subtraction was done in the overall bitmap for a character class, meaning
3067 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3068 be recognized. This bug has been fixed.
3069
3070 7. Patches from the folks at Google:
3071
3072 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3073 real life, but is still worth protecting against".
3074
3075 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3076 regular expressions".
3077
3078 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3079 have it.
3080
3081 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3082 "configure" and the latter not, in order to fix a problem somebody had
3083 with compiling the Arg class on HP-UX.
3084
3085 (e) Improve the error-handling of the C++ wrapper a little bit.
3086
3087 (f) New tests for checking recursion limiting.
3088
3089 8. The pcre_memmove() function, which is used only if the environment does not
3090 have a standard memmove() function (and is therefore rarely compiled),
3091 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3092 returning a result (though PCRE never actually uses the result).
3093
3094 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3095 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3096 returned instead of calling malloc() with an overflowing number that would
3097 most likely cause subsequent chaos.
3098
3099 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3100
3101 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3102 with this option is matched, the nmatch and pmatch options of regexec() are
3103 ignored.
3104
3105 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3106 provided in case anyone wants to the the POSIX interface with UTF-8
3107 strings.
3108
3109 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3110 C++ linking (needed for some HP-UX environments).
3111
3112 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3113 (unused parameter) and in the pcre_printint() function (omitted "default"
3114 switch label when the default is to do nothing).
3115
3116 15. Added some code to make it possible, when PCRE is compiled as a C++
3117 library, to replace subject pointers for pcre_exec() with a smart pointer
3118 class, thus making it possible to process discontinuous strings.
3119
3120 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3121 much the same function. They were added by different people who were trying
3122 to make PCRE easy to compile on non-Unix systems. It has been suggested
3123 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3124 for compiling on Windows systems. I have therefore replaced it with
3125 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3126 defaults to "extern" for C or "extern C" for C++, which works fine on
3127 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3128 SCOPE with something explicit in config.h. In addition:
3129
3130 (a) pcreposix.h still had just "extern" instead of either of these macros;
3131 I have replaced it with PCRE_DATA_SCOPE.
3132
3133 (b) Functions such as _pcre_xclass(), which are internal to the library,
3134 but external in the C sense, all had PCRE_EXPORT in their definitions.
3135 This is apparently wrong for the Windows case, so I have removed it.
3136 (It makes no difference on Unix-like systems.)
3137
3138 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3139 of recursive calls to match(). This is different to MATCH_LIMIT because
3140 that limits the total number of calls to match(), not all of which increase
3141 the depth of recursion. Limiting the recursion depth limits the amount of
3142 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3143 when PCRE is compiled, and changed at run time. A patch from Google adds
3144 this functionality to the C++ interface.
3145
3146 18. Changes to the handling of Unicode character properties:
3147
3148 (a) Updated the table to Unicode 4.1.0.
3149
3150 (b) Recognize characters that are not in the table as "Cn" (undefined).
3151
3152 (c) I revised the way the table is implemented to a much improved format
3153 which includes recognition of ranges. It now supports the ranges that
3154 are defined in UnicodeData.txt, and it also amalgamates other
3155 characters into ranges. This has reduced the number of entries in the
3156 table from around 16,000 to around 3,000, thus reducing its size
3157 considerably. I realized I did not need to use a tree structure after
3158 all - a binary chop search is just as efficient. Having reduced the
3159 number of entries, I extended their size from 6 bytes to 8 bytes to
3160 allow for more data.
3161
3162 (d) Added support for Unicode script names via properties such as \p{Han}.
3163
3164 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3165 matching that character.
3166
3167 20. When matching a repeated Unicode property with a minimum greater than zero,
3168 (for example \pL{2,}), PCRE could look past the end of the subject if it
3169 reached it while seeking the minimum number of characters. This could
3170 happen only if some of the characters were more than one byte long, because
3171 there is a check for at least the minimum number of bytes.
3172
3173 21. Refactored the implementation of \p and \P so as to be more general, to
3174 allow for more different types of property in future. This has changed the
3175 compiled form incompatibly. Anybody with saved compiled patterns that use
3176 \p or \P will have to recompile them.
3177
3178 22. Added "Any" and "L&" to the supported property types.
3179
3180 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3181 but give a compile time error if the value is greater than 0xff.
3182
3183 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3184 accidentally not being installed or uninstalled.
3185
3186 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3187 made were to insert the current release number. This seemed silly, because
3188 it made things harder for people building PCRE on systems that don't run
3189 "configure". I have turned pcre.h into a distributed file, no longer built
3190 by "configure", with the version identification directly included. There is
3191 no longer a pcre.h.in file.
3192
3193 However, this change necessitated a change to the pcre-config script as
3194 well. It is built from pcre-config.in, and one of the substitutions was the
3195 release number. I have updated configure.ac so that ./configure now finds
3196 the release number by grepping pcre.h.
3197
3198 26. Added the ability to run the tests under valgrind.
3199
3200
3201 Version 6.4 05-Sep-05
3202 ---------------------
3203
3204 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3205 "--" to be printed when multiple files were scanned, even when none of the
3206 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3207 consider it to be a bug, and have restored the previous behaviour.
3208
3209 2. A couple of code tidies to get rid of compiler warnings.
3210
3211 3. The pcretest program used to cheat by referring to symbols in the library
3212 whose names begin with _pcre_. These are internal symbols that are not
3213 really supposed to be visible externally, and in some environments it is
3214 possible to suppress them. The cheating is now confined to including
3215 certain files from the library's source, which is a bit cleaner.
3216
3217 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3218 file's purpose clearer.
3219
3220 5. Reorganized pcre_ucp_findchar().
3221
3222
3223 Version 6.3 15-Aug-05
3224 ---------------------
3225
3226 1. The file libpcre.pc.in did not have general read permission in the tarball.
3227
3228 2. There were some problems when building without C++ support:
3229
3230 (a) If C++ support was not built, "make install" and "make test" still
3231 tried to test it.
3232
3233 (b) There were problems when the value of CXX was explicitly set. Some
3234 changes have been made to try to fix these, and ...
3235
3236 (c) --disable-cpp can now be used to explicitly disable C++ support.
3237
3238 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3239 backslash in a target when C++ was disabled. This confuses some
3240 versions of "make", apparently. Using an intermediate variable solves
3241 this. (Same for CPP_LOBJ.)
3242
3243 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3244 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3245 necessary on certain architectures.
3246
3247 4. Added a setting of -export-symbols-regex to the link command to remove
3248 those symbols that are exported in the C sense, but actually are local
3249 within the library, and not documented. Their names all begin with
3250 "_pcre_". This is not a perfect job, because (a) we have to except some
3251 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3252 available (and never for static libraries). I have made a note to try to
3253 find a way round (a) in the future.
3254
3255
3256 Version 6.2 01-Aug-05
3257 ---------------------
3258
3259 1. There was no test for integer overflow of quantifier values. A construction
3260 such as {1111111111111111} would give undefined results. What is worse, if
3261 a minimum quantifier for a parenthesized subpattern overflowed and became
3262 negative, the calculation of the memory size went wrong. This could have
3263 led to memory overwriting.
3264
3265 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3266
3267 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3268 operating environments where this matters.
3269
3270 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3271 PCRE options from within the C++ wrapper.
3272
3273 5. Named capturing subpatterns were not being correctly counted when a pattern
3274 was compiled. This caused two problems: (a) If there were more than 100
3275 such subpatterns, the calculation of the memory needed for the whole
3276 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3277 back references of the form \12, where the number was greater than 9, were
3278 not recognized as back references, even though there were sufficient
3279 previous subpatterns.
3280
3281 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3282 versions of gcc, e.g. 2.95.4.
3283
3284
3285 Version 6.1 21-Jun-05
3286 ---------------------
3287
3288 1. There was one reference to the variable "posix" in pcretest.c that was not
3289 surrounded by "#if !defined NOPOSIX".
3290
3291 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3292 the cross-check on the old pcre_info() function, for the benefit of the
3293 cut-down version of PCRE that is currently imported into Exim.
3294
3295 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3296 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3297 patterns that start (?i) but I don't think that matters. The use of (?i) is
3298 just an example; this all applies to the other options as well.
3299
3300 4. Since libtool seems to echo the compile commands it is issuing, the output
3301 from "make" can be reduced a bit by putting "@" in front of each libtool
3302 compile command.
3303
3304 5. Patch from the folks at Google for configure.in to be a bit more thorough
3305 in checking for a suitable C++ installation before trying to compile the
3306 C++ stuff. This should fix a reported problem when a compiler was present,
3307 but no suitable headers.
3308
3309 6. The man pages all had just "PCRE" as their title. I have changed them to
3310 be the relevant file name. I have also arranged that these names are
3311 retained in the file doc/pcre.txt, which is a concatenation in text format
3312 of all the man pages except the little individual ones for each function.
3313
3314 7. The NON-UNIX-USE file had not been updated for the different set of source
3315 files that come with release 6. I also added a few comments about the C++
3316 wrapper.
3317
3318
3319 Version 6.0 07-Jun-05
3320 ---------------------
3321
3322 1. Some minor internal re-organization to help with my DFA experiments.
3323
3324 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3325 didn't matter for the library itself when fully configured, but did matter
3326 when compiling without UCP support, or within Exim, where the ucp files are
3327 not imported.
3328
3329 3. Refactoring of the library code to split up the various functions into
3330 different source modules. The addition of the new DFA matching code (see
3331 below) to a single monolithic source would have made it really too
3332 unwieldy, quite apart from causing all the code to be include in a
3333 statically linked application, when only some functions are used. This is
3334 relevant even without the DFA addition now that patterns can be compiled in
3335 one application and matched in another.
3336
3337 The downside of splitting up is that there have to be some external
3338 functions and data tables that are used internally in different modules of
3339 the library but which are not part of the API. These have all had their
3340 names changed to start with "_pcre_" so that they are unlikely to clash
3341 with other external names.
3342
3343 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3344 a different (DFA) algorithm. Although it is slower than the original
3345 function, it does have some advantages for certain types of matching
3346 problem.
3347
3348 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3349 including restarting after a partial match.
3350
3351 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3352 defined when compiling for Windows was sent to me. I have put it into the
3353 code, though I have no means of testing or verifying it.
3354
3355 7. Added the pcre_refcount() auxiliary function.
3356
3357 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3358 match before or at the first newline in the subject string. In pcretest,
3359 the /f option on a pattern can be used to set this.
3360
3361 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3362 would behave wrongly. This has been present in PCRE since release 4.0.
3363
3364 10. A number of changes to the pcregrep command:
3365
3366 (a) Refactored how -x works; insert ^(...)$ instead of setting
3367 PCRE_ANCHORED and checking the length, in preparation for adding
3368 something similar for -w.
3369
3370 (b) Added the -w (match as a word) option.
3371
3372 (c) Refactored the way lines are read and buffered so as to have more
3373 than one at a time available.
3374
3375 (d) Implemented a pcregrep test script.
3376
3377 (e) Added the -M (multiline match) option. This allows patterns to match
3378 over several lines of the subject. The buffering ensures that at least
3379 8K, or the rest of the document (whichever is the shorter) is available
3380 for matching (and similarly the previous 8K for lookbehind assertions).
3381
3382 (f) Changed the --help output so that it now says
3383
3384 -w, --word-regex(p)
3385
3386 instead of two lines, one with "regex" and the other with "regexp"
3387 because that confused at least one person since the short forms are the
3388 same. (This required a bit of code, as the output is generated
3389 automatically from a table. It wasn't just a text change.)
3390
3391 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3392 option but starts with a hyphen. Could be a pattern or a path name
3393 starting with a hyphen, for instance.
3394
3395 (h) "-" can be given as a file name to represent stdin.
3396
3397 (i) When file names are being printed, "(standard input)" is used for
3398 the standard input, for compatibility with GNU grep. Previously
3399 "<stdin>" was used.
3400
3401 (j) The option --label=xxx can be used to supply a name to be used for
3402 stdin when file names are being printed. There is no short form.
3403
3404 (k) Re-factored the options decoding logic because we are going to add
3405 two more options that take data. Such options can now be given in four
3406 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3407
3408 (l) Added the -A, -B, and -C options for requesting that lines of context
3409 around matches be printed.
3410
3411 (m) Added the -L option to print the names of files that do not contain
3412 any matching lines, that is, the complement of -l.
3413
3414 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3415 continue to scan other files.
3416
3417 (o) The -s option was incorrectly implemented. For compatibility with other
3418 greps, it now suppresses the error message for a non-existent or non-
3419 accessible file (but not the return code). There is a new option called
3420 -q that suppresses the output of matching lines, which was what -s was
3421 previously doing.
3422
3423 (p) Added --include and --exclude options to specify files for inclusion
3424 and exclusion when recursing.
3425
3426 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3427 Hopefully, it now does.
3428
3429 12. Missing cast in pcre_study().
3430
3431 13. Added an "uninstall" target to the makefile.
3432
3433 14. Replaced "extern" in the function prototypes in Makefile.in with
3434 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3435 world, but is set differently for Windows.
3436
3437 15. Added a second compiling function called pcre_compile2(). The only
3438 difference is that it has an extra argument, which is a pointer to an
3439 integer error code. When there is a compile-time failure, this is set
3440 non-zero, in addition to the error test pointer being set to point to an
3441 error message. The new argument may be NULL if no error number is required
3442 (but then you may as well call pcre_compile(), which is now just a
3443 wrapper). This facility is provided because some applications need a
3444 numeric error indication, but it has also enabled me to tidy up the way
3445 compile-time errors are handled in the POSIX wrapper.
3446
3447 16. Added VPATH=.libs to the makefile; this should help when building with one
3448 prefix path and installing with another. (Or so I'm told by someone who
3449 knows more about this stuff than I do.)
3450
3451 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3452 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3453 match everything, including newlines. This is not POSIX-compatible, but
3454 somebody wanted the feature. From pcretest it can be activated by using
3455 both the P and the s flags.
3456
3457 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3458
3459 19. libpcre.pc was being incorrectly installed as executable.
3460
3461 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3462 it now also looks for '\r' so that it will work unmodified on Windows.
3463
3464 21. Added Google's contributed C++ wrapper to the distribution.
3465
3466 22. Added some untidy missing memory free() calls in pcretest, to keep
3467 Electric Fence happy when testing.
3468
3469
3470
3471 Version 5.0 13-Sep-04
3472 ---------------------
3473
3474 1. Internal change: literal characters are no longer packed up into items
3475 containing multiple characters in a single byte-string. Each character
3476 is now matched using a separate opcode. However, there may be more than one
3477 byte in the character in UTF-8 mode.
3478
3479 2. The pcre_callout_block structure has two new fields: pattern_position and
3480 next_item_length. These contain the offset in the pattern to the next match
3481 item, and its length, respectively.
3482
3483 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3484 insertion of callouts before each pattern item. Added the /C option to
3485 pcretest to make use of this.
3486
3487 4. On the advice of a Windows user, the lines
3488
3489 #if defined(_WIN32) || defined(WIN32)
3490 _setmode( _fileno( stdout ), 0x8000 );
3491 #endif /* defined(_WIN32) || defined(WIN32) */
3492
3493 have been added to the source of pcretest. This apparently does useful
3494 magic in relation to line terminators.
3495
3496 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3497 for the benefit of those environments where the "b" makes a difference.
3498
3499 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3500 to know about it. I have put a hack into configure.in that adds in code
3501 to set GCC=yes if CC=icc. This seems to end up at a point in the
3502 generated configure script that is early enough to affect the setting of
3503 compiler options, which is what is needed, but I have no means of testing
3504 whether it really works. (The user who reported this had patched the
3505 generated configure script, which of course I cannot do.)
3506
3507 LATER: After change 22 below (new libtool files), the configure script
3508 seems to know about icc (and also ecc). Therefore, I have commented out
3509 this hack in configure.in.
3510
3511 7. Added support for pkg-config (2 patches were sent in).
3512
3513 8. Negated POSIX character classes that used a combination of internal tables
3514 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3515 [[:^ascii]]. Typically, they would match almost any characters. The other
3516 POSIX classes were not broken in this way.
3517
3518 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3519 to find the match, as PCRE was deluded into thinking that the match had to
3520 start at the start point or following a newline. The same bug applied to
3521 patterns with negative forward assertions or any backward assertions
3522 preceding ".*" at the start, unless the pattern required a fixed first
3523 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3524
3525 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3526 starting at the last subject character, bytes beyond the end of the subject
3527 string were read.
3528
3529 11. Renamed the variable "class" as "classbits" to make life easier for C++
3530 users. (Previously there was a macro definition, but it apparently wasn't
3531 enough.)
3532
3533 12. Added the new field "tables" to the extra data so that tables can be passed
3534 in at exec time, or the internal tables can be re-selected. This allows
3535 a compiled regex to be saved and re-used at a later time by a different
3536 program that might have everything at different addresses.
3537
3538 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3539 -R library as well as a -L library.
3540
3541 14. The debugging options of pcretest (-d on the command line or D on a
3542 pattern) showed incorrect output for anything following an extended class
3543 that contained multibyte characters and which was followed by a quantifier.
3544
3545 15. Added optional support for general category Unicode character properties
3546 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3547 support. It adds about 90K to the size of the library. The meanings of the
3548 inbuilt class escapes such as \d and \s have NOT been changed.
3549
3550 16. Updated pcredemo.c to include calls to free() to release the memory for the
3551 compiled pattern.
3552
3553 17. The generated file chartables.c was being created in the source directory
3554 instead of in the building directory. This caused the build to fail if the
3555 source directory was different from the building directory, and was
3556 read-only.
3557
3558 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3559 file. No doubt somebody will tell me if they don't make sense... Also added
3560 Dan Mooney's comments about building on OpenVMS.
3561
3562 19. Added support for partial matching via the PCRE_PARTIAL option for
3563 pcre_exec() and the \P data escape in pcretest.
3564
3565 20. Extended pcretest with 3 new pattern features:
3566
3567 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3568 write the compiled pattern to the file whose name is "rest-of-line".
3569 This is a straight binary dump of the data, with the saved pointer to
3570 the character tables forced to be NULL. The study data, if any, is
3571 written too. After writing, pcretest reads a new pattern.
3572
3573 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3574 compiled pattern from the given file. There must not be any
3575 occurrences of "<" in the file name (pretty unlikely); if there are,
3576 pcretest will instead treat the initial "<" as a pattern delimiter.
3577 After reading in the pattern, pcretest goes on to read data lines as
3578 usual.
3579
3580 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3581 and 16-bit fields in a compiled pattern, to simulate a pattern that
3582 was compiled on a host of opposite endianness.
3583
3584 21. The pcre-exec() function can now cope with patterns that were compiled on
3585 hosts of opposite endianness, with this restriction:
3586
3587 As for any compiled expression that is saved and used later, the tables
3588 pointer field cannot be preserved; the extra_data field in the arguments
3589 to pcre_exec() should be used to pass in a tables address if a value
3590 other than the default internal tables were used at compile time.
3591
3592 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3593 now diagnosed as an error. Previously, most of the time, a negative number
3594 would have been treated as zero, but if in addition "ovector" was passed as
3595 NULL, a crash could occur.
3596
3597 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3598 new versions from the libtool 1.5 distribution (the last one is a copy of
3599 a file called libtool.m4). This seems to have fixed the need to patch
3600 "configure" to support Darwin 1.3 (which I used to do). However, I still
3601 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3602 workstation).
3603
3604 24. Changed the PCRE licence to be the more standard "BSD" licence.
3605
3606
3607 Version 4.5 01-Dec-03
3608 ---------------------
3609
3610 1. There has been some re-arrangement of the code for the match() function so
3611 that it can be compiled in a version that does not call itself recursively.
3612 Instead, it keeps those local variables that need separate instances for
3613 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3614 needs to "recurse". Keeping track of where control must go is done by means
3615 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3616 hide most of the details from the main code, and operates only if
3617 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3618 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3619 operating.
3620
3621 To make it easier for callers to provide specially tailored get/free
3622 functions for this usage, two new functions, pcre_stack_malloc, and
3623 pcre_stack_free, are used. They are always called in strict stacking order,
3624 and the size of block requested is always the same.
3625
3626 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3627 PCRE has been compiled to use the stack or the heap for recursion. The
3628 -C option of pcretest uses this to show which version is compiled.
3629
3630 A new data escape \S, is added to pcretest; it causes the amounts of store
3631 obtained and freed by both kinds of malloc/free at match time to be added
3632 to the output.
3633
3634 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3635 what's available on my current Linux desktop machine.
3636
3637 3. When matching a UTF-8 string, the test for a valid string at the start has
3638 been extended. If start_offset is not zero, PCRE now checks that it points
3639 to a byte that is the start of a UTF-8 character. If not, it returns
3640 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3641 this is necessary because there may be backward assertions in the pattern.
3642 When matching the same subject several times, it may save resources to use
3643 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3644
3645 4. The code for checking the validity of UTF-8 strings has been tightened so
3646 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3647 containing "overlong sequences".
3648
3649 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3650 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3651 should have been "&", but it just so happened that all the cases this let
3652 through by mistake were picked up later in the function.
3653
3654 6. I had used a variable called "isblank" - this is a C99 function, causing
3655 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3656
3657 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3658 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3659 so the version is shown; (c) stop "make test" echoing "./RunTest".
3660
3661 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3662
3663 9. The prototype for memmove() for systems that don't have it was using
3664 size_t, but the inclusion of the header that defines size_t was later. I've
3665 moved the #includes for the C headers earlier to avoid this.
3666
3667 10. Added some adjustments to the code to make it easier to compiler on certain
3668 special systems:
3669
3670 (a) Some "const" qualifiers were missing.
3671 (b) Added the macro EXPORT before all exported functions; by default this
3672 is defined to be empty.
3673 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3674 that it reads its output file name as an argument instead of writing
3675 to the standard output and assuming this can be redirected.
3676
3677 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3678 class containing characters with values greater than 255, PCRE compilation
3679 went into a loop.
3680
3681 12. A recursive reference to a subpattern that was within another subpattern
3682 that had a minimum quantifier of zero caused PCRE to crash. For example,
3683 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3684 recursion. If the recursively-called subpattern itself had a zero repeat,
3685 that was OK.
3686
3687 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3688 buffer into which it was copied (for escape processing) was still set at
3689 1024, so long lines caused crashes.
3690
3691 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3692 "internal error: code overflow...". This applied to any character class
3693 that was followed by a possessive quantifier.
3694
3695 15. Modified the Makefile to add libpcre.la as a prerequisite for
3696 libpcreposix.la because I was told this is needed for a parallel build to
3697 work.
3698
3699 16. If a pattern that contained .* following optional items at the start was
3700 studied, the wrong optimizing data was generated, leading to matching
3701 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3702 matching string must start with a or b or c. The correct conclusion for
3703 this pattern is that a match can start with any character.
3704
3705
3706 Version 4.4 13-Aug-03
3707 ---------------------
3708
3709 1. In UTF-8 mode, a character class containing characters with values between
3710 127 and 255 was not handled correctly if the compiled pattern was studied.
3711 In fixing this, I have also improved the studying algorithm for such
3712 classes (slightly).
3713
3714 2. Three internal functions had redundant arguments passed to them. Removal
3715 might give a very teeny performance improvement.
3716
3717 3. Documentation bug: the value of the capture_top field in a callout is *one
3718 more than* the number of the hightest numbered captured substring.
3719
3720 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3721 in incorrectly linking with a previously installed version. They now link
3722 explicitly with libpcre.la.
3723
3724 5. configure.in no longer needs to recognize Cygwin specially.
3725
3726 6. A problem in pcre.in for Windows platforms is fixed.
3727
3728 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3729 pcretest, it used to include the size of the study block as part of its
3730 output. Unfortunately, the structure contains a field that has a different
3731 size on different hardware architectures. This meant that the tests that
3732 showed this size failed. As the block is currently always of a fixed size,
3733 this information isn't actually particularly useful in pcretest output, so
3734 I have just removed it.
3735
3736 8. Three pre-processor statements accidentally did not start in column 1.
3737 Sadly, there are *still* compilers around that complain, even though
3738 standard C has not required this for well over a decade. Sigh.
3739
3740 9. In pcretest, the code for checking callouts passed small integers in the
3741 callout_data field, which is a void * field. However, some picky compilers
3742 complained about the casts involved for this on 64-bit systems. Now
3743 pcretest passes the address of the small integer instead, which should get
3744 rid of the warnings.
3745
3746 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3747 both compile and run time, and gives an error if an invalid UTF-8 sequence
3748 is found. There is a option for disabling this check in cases where the
3749 string is known to be correct and/or the maximum performance is wanted.
3750
3751 11. In response to a bug report, I changed one line in Makefile.in from
3752
3753 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3754 to
3755 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3756
3757 to look similar to other lines, but I have no way of telling whether this
3758 is the right thing to do, as I do not use Windows. No doubt I'll get told
3759 if it's wrong...
3760
3761
3762 Version 4.3 21-May-03
3763 ---------------------
3764
3765 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3766 Makefile.
3767
3768 2. Some refactoring to improve the quality of the code:
3769
3770 (i) The utf8_table... variables are now declared "const".
3771
3772 (ii) The code for \cx, which used the "case flipping" table to upper case
3773 lower case letters, now just substracts 32. This is ASCII-specific,
3774 but the whole concept of \cx is ASCII-specific, so it seems
3775 reasonable.
3776
3777 (iii) PCRE was using its character types table to recognize decimal and
3778 hexadecimal digits in the pattern. This is silly, because it handles
3779 only 0-9, a-f, and A-F, but the character types table is locale-
3780 specific, which means strange things might happen. A private
3781 table is now used for this - though it costs 256 bytes, a table is
3782 much faster than multiple explicit tests. Of course, the standard
3783 character types table is still used for matching digits in subject
3784 strings against \d.
3785
3786 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3787 ending in _t are). So I've renamed it as ESC_tee.
3788
3789 3. The first argument for regexec() in the POSIX wrapper should have been
3790 defined as "const".
3791
3792 4. Changed pcretest to use malloc() for its buffers so that they can be
3793 Electric Fenced for debugging.
3794
3795 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3796 to read one or more bytes before the start of the subject string. Often this
3797 had no effect on PCRE's behaviour, but in some circumstances it could
3798 provoke a segmentation fault.
3799
3800 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3801 to try to read one or more bytes before the start of the subject string.
3802
3803 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3804 UTF-8 support could misbehave in various ways if the subject string
3805 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3806 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3807 back over UTF-8 characters.)
3808
3809
3810 Version 4.2 14-Apr-03
3811 ---------------------
3812
3813 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3814
3815 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3816 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3817 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3818 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3819 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3820 and BUILD_EXEEXT
3821 Note: automatic setting of the BUILD variables is not yet working
3822 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3823 compile-time but not at link-time
3824 [LINK]: use for linking executables only
3825 make different versions for Windows and non-Windows
3826 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3827 libraries
3828 [LINK_FOR_BUILD]: new variable
3829 [OBJEXT]: use throughout
3830 [EXEEXT]: use throughout
3831 <winshared>: new target
3832 <wininstall>: new target
3833 <dftables.o>: use native compiler
3834 <dftables>: use native linker
3835 <install>: handle Windows platform correctly
3836 <clean>: ditto
3837 <check>: ditto
3838 copy DLL to top builddir before testing
3839
3840 As part of these changes, -no-undefined was removed again. This was reported
3841 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3842 in any case.
3843
3844 3. Some tidies to get rid of compiler warnings:
3845
3846 . In the match_data structure, match_limit was an unsigned long int, whereas
3847 match_call_count was an int. I've made them both unsigned long ints.
3848
3849 . In pcretest the fact that a const uschar * doesn't automatically cast to
3850 a void * provoked a warning.
3851
3852 . Turning on some more compiler warnings threw up some "shadow" variables
3853 and a few more missing casts.
3854
3855 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3856 option, a class that contained a single character with a value between 128
3857 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3858
3859 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3860 option, a class that contained several characters, but with at least one
3861 whose value was between 128 and 255 caused PCRE to crash.
3862
3863
3864 Version 4.1 12-Mar-03
3865 ---------------------
3866
3867 1. Compiling with gcc -pedantic found a couple of places where casts were
3868 needed, and a string in dftables.c that was longer than standard compilers are
3869 required to support.
3870
3871 2. Compiling with Sun's compiler found a few more places where the code could
3872 be tidied up in order to avoid warnings.
3873
3874 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3875 first of these names is deprecated in the latest Autoconf in favour of the name
3876 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3877 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3878 analogy I have changed it to CFLAGS_FOR_BUILD.
3879
3880 4. Added -no-undefined to the linking command in the Makefile, because this is
3881 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3882 linking step for the pcreposix library.
3883
3884 5. PCRE was failing to diagnose the case of two named groups with the same
3885 name.
3886
3887 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3888 literal character that is needed in the subject for a match, and scans along to
3889 ensure that it is present before embarking on the full matching process. This
3890 saves time in cases of nested unlimited repeats that are never going to match.
3891 Problem: the scan can take a lot of time if the subject is very long (e.g.
3892 megabytes), thus penalizing straightforward matches. It is now done only if the
3893 amount of subject to be scanned is less than 1000 bytes.
3894
3895 7. A lesser problem with the same optimization is that it was recording the
3896 first character of an anchored pattern as "needed", thus provoking a search
3897 right along the subject, even when the first match of the pattern was going to
3898 fail. The "needed" character is now not set for anchored patterns, unless it
3899 follows something in the pattern that is of non-fixed length. Thus, it still
3900 fulfils its original purpose of finding quick non-matches in cases of nested
3901 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3902
3903
3904 Version 4.0 17-Feb-03
3905 ---------------------
3906
3907 1. If a comment in an extended regex that started immediately after a meta-item
3908 extended to the end of string, PCRE compiled incorrect data. This could lead to
3909 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3910
3911 2. Moved to autoconf 2.53 and libtool 1.4.2.
3912
3913 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3914 the special perltest8 script is no longer needed - all the tests can be run
3915 from a single perltest script.
3916
3917 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3918 by \s. It has now been removed in PCRE. This means it isn't recognized as
3919 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3920 class [:space:] *does* include VT, thereby creating a mess.
3921
3922 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3923 space and tab.
3924
3925 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3926 its new features into the main test script, reducing the number of scripts.
3927
3928 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3929 were backward compatible, and made the (?i) apply to the whole pattern, as if
3930 /i were given. Now it behaves more logically, and applies the option setting
3931 only to what follows. PCRE has been changed to follow suit. However, if it
3932 finds options settings right at the start of the pattern, it extracts them into
3933 the global options, as before. Thus, they show up in the info data.
3934
3935 8. Added support for the \Q...\E escape sequence. Characters in between are
3936 treated as literals. This is slightly different from Perl in that $ and @ are
3937 also handled as literals inside the quotes. In Perl, they will cause variable
3938 interpolation. Note the following examples:
3939
3940 Pattern PCRE matches Perl matches
3941
3942 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3943 \Qabc\$xyz\E abc\$xyz abc\$xyz
3944 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3945
3946 For compatibility with Perl, \Q...\E sequences are recognized inside character
3947 classes as well as outside them.
3948
3949 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3950 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3951 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3952 signed/unsigned warnings.
3953
3954 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3955 option for pcretest, so I've replaced it by a simple function that does just
3956 that job.
3957
3958 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3959 "pcregrep -".
3960
3961 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3962 Java package. This provides some syntactic sugar for simple cases of what my
3963 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3964 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3965 item, you can use this simplified notation. Note that only makes sense with
3966 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3967 greediness, whatever the setting of the PCRE_UNGREEDY option.
3968
3969 13. A change of greediness default within a pattern was not taking effect at
3970 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3971 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3972 was abstracted outside.
3973
3974 14. PCRE now supports the \G assertion. It is true when the current matching
3975 position is at the start point of the match. This differs from \A when the
3976 starting offset is non-zero. Used with the /g option of pcretest (or similar
3977 code), it works in the same way as it does for Perl's /g option. If all
3978 alternatives of a regex begin with \G, the expression is anchored to the start
3979 match position, and the "anchored" flag is set in the compiled expression.
3980
3981 15. Some bugs concerning the handling of certain option changes within patterns
3982 have been fixed. These applied to options other than (?ims). For example,
3983 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3984 been the other way round. Some of this was related to change 7 above.
3985
3986 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3987 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3988 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3989 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3990
3991 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3992 mode. Unlike ".", it always matches newline, whatever the setting of
3993 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3994 assertions. Perl allows it, but it doesn't (in general) work because it can't
3995 calculate the length of the lookbehind. At least, that's the case for Perl
3996 5.8.0 - I've been told they are going to document that it doesn't work in
3997 future.
3998
3999 18. Added an error diagnosis for escapes that PCRE does not support: these are
4000 \L, \l, \N, \P, \p, \U, \u, and \X.
4001
4002 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4003 reading past the end of the pattern in cases such as /[abcd/.
4004
4005 20. PCRE was getting more memory than necessary for patterns with classes that
4006 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4007
4008 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4009 compiling PCRE for use with Virtual Pascal.
4010
4011 22. Small fix to the Makefile to make it work properly if the build is done
4012 outside the source tree.
4013
4014 23. Added a new extension: a condition to go with recursion. If a conditional
4015 subpattern starts with (?(R) the "true" branch is used if recursion has
4016 happened, whereas the "false" branch is used only at the top level.
4017
4018 24. When there was a very long string of literal characters (over 255 bytes
4019 without UTF support, over 250 bytes with UTF support), the computation of how
4020 much memory was required could be incorrect, leading to segfaults or other
4021 strange effects.
4022
4023 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4024 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4025 there was a subsequent back reference to those brackets. This meant that, for
4026 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4027 possible to check for precisely this case. All we can do is abandon the
4028 optimization if .* occurs inside capturing brackets when there are any back
4029 references whatsoever. (See below for a better fix that came later.)
4030
4031 26. The handling of the optimization for finding the first character of a
4032 non-anchored pattern, and for finding a character that is required later in the
4033 match were failing in some cases. This didn't break the matching; it just
4034 failed to optimize when it could. The way this is done has been re-implemented.
4035
4036 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4037
4038 28. Added a new feature that provides some of the functionality that Perl
4039 provides with (?{...}). The facility is termed a "callout". The way it is done
4040 in PCRE is for the caller to provide an optional function, by setting
4041 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4042 global variable. By default it is unset, which disables all calling out. To get
4043 the function called, the regex must include (?C) at appropriate points. This
4044 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4045 This provides a means of identifying different callout points. When PCRE
4046 reaches such a point in the regex, if pcre_callout has been set, the external
4047 function is called. It is provided with data in a structure called
4048 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4049 matching continues; if it returns a non-zero value, the match at the current
4050 point fails. However, backtracking will occur if possible. [This was changed
4051 later and other features added - see item 49 below.]
4052
4053 29. pcretest is upgraded to test the callout functionality. It provides a
4054 callout function that displays information. By default, it shows the start of
4055 the match and the current position in the text. There are some new data escapes
4056 to vary what happens:
4057
4058 \C+ in addition, show current contents of captured substrings
4059 \C- do not supply a callout function
4060 \C!n return 1 when callout number n is reached
4061 \C!n!m return 1 when callout number n is reached for the mth time
4062
4063 30. If pcregrep was called with the -l option and just a single file name, it
4064 output "<stdin>" if a match was found, instead of the file name.
4065
4066 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4067 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4068 pcre_exec(). This saves a malloc/free per call. The default value of
4069 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4070 when configuring.
4071
4072 32. The default maximum size of a compiled pattern is 64K. There have been a
4073 few cases of people hitting this limit. The code now uses macros to handle the
4074 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4075 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4076 configuring. Tests 2 and 5 work only with 2-byte links because they output
4077 debugging information about compiled patterns.
4078
4079 33. Internal code re-arrangements:
4080
4081 (a) Moved the debugging function for printing out a compiled regex into
4082 its own source file (printint.c) and used #include to pull it into
4083 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4084 separate copies.
4085
4086 (b) Defined the list of op-code names for debugging as a macro in
4087 internal.h so that it is next to the definition of the opcodes.
4088
4089 (c) Defined a table of op-code lengths for simpler skipping along compiled
4090 code. This is again a macro in internal.h so that it is next to the
4091 definition of the opcodes.
4092
4093 34. Added support for recursive calls to individual subpatterns, along the
4094 lines of Robin Houston's patch (but implemented somewhat differently).
4095
4096 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4097 allow it to read and process whole directories in Win32. This code was
4098 contributed by Lionel Fourquaux; it has not been tested by me.
4099
4100 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4101 used to name a group. Names consist of alphanumerics and underscores, and must
4102 be unique. Back references use the syntax (?P=name) and recursive calls use
4103 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4104 numbers. The function pcre_fullinfo() can be used after compilation to extract
4105 a name/number map. There are three relevant calls:
4106
4107 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4108 PCRE_INFO_NAMECOUNT yields the number of entries
4109 PCRE_INFO_NAMETABLE yields a pointer to the map.
4110
4111 The map is a vector of fixed-size entries. The size of each entry depends on
4112 the length of the longest name used. The first two bytes of each entry are the
4113 group number, most significant byte first. There follows the corresponding
4114 name, zero terminated. The names are in alphabetical order.
4115
4116 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4117 case instead of 255. Making it the same both with and without UTF-8 support
4118 means that the same test output works with both.
4119
4120 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4121 calling malloc() with a zero argument.
4122
4123 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4124 optimization. I've improved things by keeping a bitmap of backreferences with
4125 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4126 fact referenced, the optimization can be applied. It is unlikely that a
4127 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4128 the match to follow \n) will appear inside brackets with a number greater than
4129 31, but if it does, any back reference > 31 suppresses the optimization.
4130
4131 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4132 of disabling numbered capturing parentheses. Any opening parenthesis that is
4133 not followed by ? behaves as if it were followed by ?: but named parentheses
4134 can still be used for capturing (and they will acquire numbers in the usual
4135 way).
4136
4137 41. Redesigned the return codes from the match() function into yes/no/error so
4138 that errors can be passed back from deep inside the nested calls. A malloc
4139 failure while inside a recursive subpattern call now causes the
4140 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4141
4142 42. It is now possible to set a limit on the number of times the match()
4143 function is called in a call to pcre_exec(). This facility makes it possible to
4144 limit the amount of recursion and backtracking, though not in a directly
4145 obvious way, because the match() function is used in a number of different
4146 circumstances. The count starts from zero for each position in the subject
4147 string (for non-anchored patterns). The default limit is, for compatibility, a
4148 large number, namely 10 000 000. You can change this in two ways:
4149
4150 (a) When configuring PCRE before making, you can use --with-match-limit=n
4151 to set a default value for the compiled library.
4152
4153 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4154 a different value is set. See 45 below.
4155
4156 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4157
4158 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4159 of things that can be changed at compile time. The first argument specifies
4160 what is wanted and the second points to where the information is to be placed.
4161 The current list of available information is:
4162
4163 PCRE_CONFIG_UTF8
4164
4165 The output is an integer that is set to one if UTF-8 support is available;
4166 otherwise it is set to zero.
4167
4168 PCRE_CONFIG_NEWLINE
4169
4170 The output is an integer that it set to the value of the code that is used for
4171 newline. It is either LF (10) or CR (13).
4172
4173 PCRE_CONFIG_LINK_SIZE
4174
4175 The output is an integer that contains the number of bytes used for internal
4176 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4177
4178 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4179
4180 The output is an integer that contains the threshold above which the POSIX
4181 interface uses malloc() for output vectors. See item 31 above.
4182
4183 PCRE_CONFIG_MATCH_LIMIT
4184
4185 The output is an unsigned integer that contains the default limit of the number
4186 of match() calls in a pcre_exec() execution. See 42 above.
4187
4188 44. pcretest has been upgraded by the addition of the -C option. This causes it
4189 to extract all the available output from the new pcre_config() function, and to
4190 output it. The program then exits immediately.
4191
4192 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4193 order to support additional features. One way would have been to define
4194 pcre_exec2() (for example) with extra arguments, but this would not have been
4195 extensible, and would also have required all calls to the original function to
4196 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4197 is used for passing in "extra" data from pcre_study().
4198
4199 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4200 contains the following fields:
4201
4202 flags a bitmap indicating which of the following fields are set
4203 study_data opaque data from pcre_study()
4204 match_limit a way of specifying a limit on match() calls for a specific
4205 call to pcre_exec()
4206 callout_data data for callouts (see 49 below)
4207
4208 The flag bits are also defined in pcre.h, and are
4209
4210 PCRE_EXTRA_STUDY_DATA
4211 PCRE_EXTRA_MATCH_LIMIT
4212 PCRE_EXTRA_CALLOUT_DATA
4213
4214 The pcre_study() function now returns one of these new pcre_extra blocks, with
4215 the actual study data pointed to by the study_data field, and the
4216 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4217 before. That is, this change is entirely upwards-compatible and requires no
4218 change to existing code.
4219
4220 If you want to pass in additional data to pcre_exec(), you can either place it
4221 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4222 block.
4223
4224 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4225 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4226 times with different match limits, until it finds the minimum value needed for
4227 pcre_exec() to complete. The value is then output. This can be instructive; for
4228 most simple matches the number is quite small, but for pathological cases it
4229 gets very large very quickly.
4230
4231 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4232 returns the size of the data block pointed to by the study_data field in a
4233 pcre_extra block, that is, the value that was passed as the argument to
4234 pcre_malloc() when PCRE was getting memory in which to place the information
4235 created by pcre_study(). The fourth argument should point to a size_t variable.
4236 pcretest has been extended so that this information is shown after a successful
4237 pcre_study() call when information about the compiled regex is being displayed.
4238
4239 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4240 because what follows is always an absolute path. (Later: it turns out that this
4241 is more than cosmetic for MinGW, because it doesn't like empty path
4242 components.)
4243
4244 49. Some changes have been made to the callout feature (see 28 above):
4245
4246 (i) A callout function now has three choices for what it returns:
4247
4248 0 => success, carry on matching
4249 > 0 => failure at this point, but backtrack if possible
4250 < 0 => serious error, return this value from pcre_exec()
4251
4252 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4253 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4254 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4255 use by callout functions. It will never be used by PCRE itself.
4256
4257 (ii) The pcre_extra structure (see 45 above) has a void * field called
4258 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4259 pcre_callout_block structure has a field of the same name. The contents of
4260 the field passed in the pcre_extra structure are passed to the callout
4261 function in the corresponding field in the callout block. This makes it
4262 easier to use the same callout-containing regex from multiple threads. For
4263 testing, the pcretest program has a new data escape
4264
4265 \C*n pass the number n (may be negative) as callout_data
4266
4267 If the callout function in pcretest receives a non-zero value as
4268 callout_data, it returns that value.
4269
4270 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4271 there were some redundant $(CFLAGS) in commands that are now specified as
4272 $(LINK), which already includes $(CFLAGS).
4273
4274 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4275 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4276 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4277 one-byte characters throughout. Note that case-insensitive matching applies
4278 only to characters whose values are less than 256. PCRE doesn't support the
4279 notion of cases for higher-valued characters.
4280
4281 (i) A character class whose characters are all within 0-255 is handled as
4282 a bit map, and the map is inverted for negative classes. Previously, a
4283 character > 255 always failed to match such a class; however it should
4284 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4285
4286 (ii) A negated character class with a single character < 255 is coded as
4287 "not this character" (OP_NOT). This wasn't working properly when the test
4288 character was multibyte, either singly or repeated.
4289
4290 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4291 mode, for example: \x{100}{2,3}.
4292
4293 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4294 singly or repeated) now correctly test multibyte characters. However,
4295 PCRE doesn't recognize any characters with values greater than 255 as
4296 digits, spaces, or word characters. Such characters always match \D, \S,
4297 and \W, and never match \d, \s, or \w.
4298
4299 (v) Classes may now contain characters and character ranges with values
4300 greater than 255. For example: [ab\x{100}-\x{400}].
4301
4302 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4303 PCRE in UTF-8 mode.
4304
4305 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4306 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4307 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4308 value.)
4309
4310 53. The single man page has become too large. I have therefore split it up into
4311 a number of separate man pages. These also give rise to individual HTML pages;
4312 these are now put in a separate directory, and there is an index.html page that
4313 lists them all. Some hyperlinking between the pages has been installed.
4314
4315 54. Added convenience functions for handling named capturing parentheses.
4316
4317 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4318 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4319 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4320 are faulted.
4321
4322 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4323 calling configure. These values are used when compiling the dftables.c program
4324 which is run to generate the source of the default character tables. They
4325 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4326 you will need to set these values.
4327
4328 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4329
4330
4331 Version 3.9 02-Jan-02
4332 ---------------------
4333
4334 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4335
4336 2. If --disable-static was given, the building process failed when trying to
4337 build pcretest and pcregrep. (For some reason it was using libtool to compile
4338 them, which is not right, as they aren't part of the library.)
4339
4340
4341 Version 3.8 18-Dec-01
4342 ---------------------
4343
4344 1. The experimental UTF-8 code was completely screwed up. It was packing the
4345 bytes in the wrong order. How dumb can you get?
4346
4347
4348 Version 3.7 29-Oct-01
4349 ---------------------
4350
4351 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4352 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4353 this didn't happen under Solaris 8, where I normally test things.
4354
4355 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4356 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4357 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4358 this fix an hour or so after the initial 3.7 release.)
4359
4360
4361 Version 3.6 23-Oct-01
4362 ---------------------
4363
4364 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4365 offsets passed as NULL with zero offset count.
4366
4367 2. The config.guess and config.sub files had not been updated when I moved to
4368 the latest autoconf.
4369
4370
4371 Version 3.5 15-Aug-01
4372 ---------------------
4373
4374 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4375 had been forgotten.
4376
4377 2. By using declared but undefined structures, we can avoid using "void"
4378 definitions in pcre.h while keeping the internal definitions of the structures
4379 private.
4380
4381 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4382 user point of view, this means that both static and shared libraries are built
4383 by default, but this can be individually controlled. More of the work of
4384 handling this static/shared cases is now inside libtool instead of PCRE's make
4385 file.
4386
4387 4. The pcretest utility is now installed along with pcregrep because it is
4388 useful for users (to test regexs) and by doing this, it automatically gets
4389 relinked by libtool. The documentation has been turned into a man page, so
4390 there are now .1, .txt, and .html versions in /doc.
4391
4392 5. Upgrades to pcregrep:
4393 (i) Added long-form option names like gnu grep.
4394 (ii) Added --help to list all options with an explanatory phrase.
4395 (iii) Added -r, --recursive to recurse into sub-directories.
4396 (iv) Added -f, --file to read patterns from a file.
4397
4398 6. pcre_exec() was referring to its "code" argument before testing that
4399 argument for NULL (and giving an error if it was NULL).
4400
4401 7. Upgraded Makefile.in to allow for compiling in a different directory from
4402 the source directory.
4403
4404 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4405 options bits, the pointer it was passed was to an int instead of to an unsigned
4406 long int. This mattered only on 64-bit systems.
4407
4408 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4409 generated) instead of pcre.in, which it its source. Also made the same change
4410 in several of the .c files.
4411
4412 10. A new release of gcc defines printf() as a macro, which broke pcretest
4413 because it had an ifdef in the middle of a string argument for printf(). Fixed
4414 by using separate calls to printf().
4415
4416 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4417 script, to force use of CR or LF instead of \n in the source. On non-Unix
4418 systems, the value can be set in config.h.
4419
4420 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4421 absolute limit. Changed the text of the error message to make this clear, and
4422 likewise updated the man page.
4423
4424 13. The limit of 99 on the number of capturing subpatterns has been removed.
4425 The new limit is 65535, which I hope will not be a "real" limit.
4426
4427
4428 Version 3.4 22-Aug-00
4429 ---------------------
4430
4431 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4432
4433 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4434
4435
4436 Version 3.3 01-Aug-00
4437 ---------------------
4438
4439 1. If an octal character was given, but the value was greater than \377, it
4440 was not getting masked to the least significant bits, as documented. This could
4441 lead to crashes in some systems.
4442
4443 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4444 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4445
4446 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4447 These just pass their arguments on to (pcre_free)(), but they are provided
4448 because some uses of PCRE bind it to non-C systems that can call its functions,
4449 but cannot call free() or pcre_free() directly.
4450
4451 4. Add "make test" as a synonym for "make check". Corrected some comments in
4452 the Makefile.
4453
4454 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4455 Makefile.
4456
4457 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4458 command called pgrep for grepping around the active processes.
4459
4460 7. Added the beginnings of support for UTF-8 character strings.
4461
4462 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4463 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4464 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4465 out for the ar command.)
4466
4467
4468 Version 3.2 12-May-00
4469 ---------------------
4470
4471 This is purely a bug fixing release.
4472
4473 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4474 of ZA. This was just one example of several cases that could provoke this bug,
4475 which was introduced by change 9 of version 2.00. The code for breaking
4476 infinite loops after an iteration that matches an empty string was't working
4477 correctly.
4478
4479 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4480 when matched against abbab (for example). After matching an empty string, it
4481 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4482 caused it to match further down the string than it should.
4483
4484 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4485 was there because it doesn't seem to be needed, and it causes trouble on some
4486 systems, as it is not a Standard C header. It has been removed.
4487
4488 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4489 were reported on the Macintosh. The changes were from
4490
4491 while ((c = *(++ptr)) != 0 && c != '\n');
4492 to
4493 while ((c = *(++ptr)) != 0 && c != '\n') ;
4494
4495 Totally extraordinary, but if that's what it takes...
4496
4497 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4498 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4499 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4500 assumes the way PCRE uses memmove() (always moving upwards).
4501
4502 6. PCRE is being used in one environment where strchr() is not available. There
4503 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4504 faster code anyway.
4505
4506
4507 Version 3.1 09-Feb-00
4508 ---------------------
4509
4510 The only change in this release is the fixing of some bugs in Makefile.in for
4511 the "install" target:
4512
4513 (1) It was failing to install pcreposix.h.
4514
4515 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4516
4517
4518 Version 3.0 01-Feb-00
4519 ---------------------
4520
4521 1. Add support for the /+ modifier to perltest (to output $` like it does in
4522 pcretest).
4523
4524 2. Add support for the /g modifier to perltest.
4525
4526 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4527 matches null strings.
4528
4529 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4530 pattern. Perl treats empty patterns specially - it reuses the most recent
4531 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4532 effect.
4533
4534 5. The POSIX interface was broken in that it was just handing over the POSIX
4535 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4536 required a bigger vector, with some working space on the end. This means that
4537 the POSIX wrapper now has to get and free some memory, and copy the results.
4538
4539 6. Added some simple autoconf support, placing the test data and the
4540 documentation in separate directories, re-organizing some of the
4541 information files, and making it build pcre-config (a GNU standard). Also added
4542 libtool support for building PCRE as a shared library, which is now the
4543 default.
4544
4545 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4546 09 are not valid octal constants. Single digits will be used for minor values
4547 less than 10.
4548
4549 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4550 existing programs that set these in the POSIX interface can use PCRE without
4551 modification.
4552
4553 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4554 return all that pcre_info() returns, plus additional data. The pcre_info()
4555 function is retained for compatibility, but is considered to be obsolete.
4556
4557 10. Added experimental recursion feature (?R) to handle one common case that
4558 Perl 5.6 will be able to do with (?p{...}).
4559
4560 11. Added support for POSIX character classes like [:alpha:], which Perl is
4561 adopting.
4562
4563
4564 Version 2.08 31-Aug-99
4565 ----------------------
4566
4567 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4568 trying to match at the startoffset position, but instead was moving forward to
4569 the next newline as if a previous match had failed.
4570
4571 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4572 and could get into a loop if a null string was matched other than at the start
4573 of the subject.
4574
4575 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4576 be distinguished at compile time, and for completeness also added PCRE_DATE.
4577
4578 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4579 in GnuWin32 environments.
4580
4581
4582 Version 2.07 29-Jul-99
4583 ----------------------
4584
4585 1. The documentation is now supplied in plain text form and HTML as well as in
4586 the form of man page sources.
4587
4588 2. C++ compilers don't like assigning (void *) values to other pointer types.
4589 In particular this affects malloc(). Although there is no problem in Standard
4590 C, I've put in casts to keep C++ compilers happy.
4591
4592 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4593 should be (const char *).
4594
4595 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4596 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4597 However, I haven't made this a standard facility. The documentation doesn't
4598 mention it, and the Makefile doesn't support it.
4599
4600 5. The Makefile now contains an "install" target, with editable destinations at
4601 the top of the file. The pcretest program is not installed.
4602
4603 6. pgrep -V now gives the PCRE version number and date.
4604
4605 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4606 causing the entire string to be ignored, instead of just the last character.
4607
4608 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4609 non-matching string, it can take a very, very long time, even for strings of
4610 quite modest length, because of the nested recursion. PCRE now does better in
4611 some of these cases. It does this by remembering the last required literal
4612 character in the pattern, and pre-searching the subject to ensure it is present
4613 before running the real match. In other words, it applies a heuristic to detect
4614 some types of certain failure quickly, and in the above example, if presented
4615 with a string that has no trailing " it gives "no match" very quickly.
4616
4617 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4618 other alternatives are tried instead.
4619
4620
4621 Version 2.06 09-Jun-99
4622 ----------------------
4623
4624 1. Change pcretest's output for amount of store used to show just the code
4625 space, because the remainder (the data block) varies in size between 32-bit and
4626 64-bit systems.
4627
4628 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4629 start matching at. This allows lookbehinds to work when searching for multiple
4630 occurrences in a string.
4631
4632 3. Added additional options to pcretest for testing multiple occurrences:
4633
4634 /+ outputs the rest of the string that follows a match
4635 /g loops for multiple occurrences, using the new startoffset argument
4636 /G loops for multiple occurrences by passing an incremented pointer
4637
4638 4. PCRE wasn't doing the "first character" optimization for patterns starting
4639 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4640 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4641 the letter 'x'. On long subject strings, this gives a significant speed-up.
4642
4643
4644 Version 2.05 21-Apr-99
4645 ----------------------
4646
4647 1. Changed the type of magic_number from int to long int so that it works
4648 properly on 16-bit systems.
4649
4650 2. Fixed a bug which caused patterns starting with .* not to work correctly
4651 when the subject string contained newline characters. PCRE was assuming
4652 anchoring for such patterns in all cases, which is not correct because .* will
4653 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4654 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4655 must be retried after every newline in the subject.
4656
4657
4658 Version 2.04 18-Feb-99
4659 ----------------------
4660
4661 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4662 computation of the store needed to hold the pattern was incorrect (too large).
4663 If such patterns were nested a few deep, this could multiply and become a real
4664 problem.
4665
4666 2. Added /M option to pcretest to show the memory requirement of a specific
4667 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4668
4669 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4670 compiled in such a way that the backtracking after subsequent failure was
4671 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4672 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4673
4674
4675 Version 2.03 02-Feb-99
4676 ----------------------
4677
4678 1. Fixed typo and small mistake in man page.
4679
4680 2. Added 4th condition (GPL supersedes if conflict) and created separate
4681 LICENCE file containing the conditions.
4682
4683 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4684 Perl, that is the internal \ allows the delimiter to be included in the
4685 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4686 the final delimiter, add \ to the end of the pattern (to test the error).
4687
4688 4. Added the convenience functions for extracting substrings after a successful
4689 match. Updated pcretest to make it able to test these functions.
4690
4691
4692 Version 2.02 14-Jan-99
4693 ----------------------
4694
4695 1. Initialized the working variables associated with each extraction so that
4696 their saving and restoring doesn't refer to uninitialized store.
4697
4698 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4699 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4700 fix the problem.
4701
4702 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4703 calls, and wasn't printing the correct value for compiling calls. Increased the
4704 default value of LOOPREPEAT, and the number of significant figures in the
4705 times.
4706
4707 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4708
4709 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4710 a building problem on Windows NT with a FAT file system.
4711
4712
4713 Version 2.01 21-Oct-98
4714 ----------------------
4715
4716 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4717 to character tables built by pcre_maketables() in the current locale. If NULL
4718 is passed, the default tables are used.
4719
4720
4721 Version 2.00 24-Sep-98
4722 ----------------------
4723
4724 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4725 it any more.
4726
4727 2. Allow quantification of (?>) groups, and make it work correctly.
4728
4729 3. The first character computation wasn't working for (?>) groups.
4730
4731 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4732 end of the subject) and add 5.005's \z, which really does match only at the
4733 very end of the subject.
4734
4735 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4736
4737 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4738 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4739 localized options. All options to pcre_study() were also removed.
4740
4741 7. Add other new features from 5.005:
4742
4743 $(?<= positive lookbehind
4744 $(?<! negative lookbehind
4745 (?imsx-imsx) added the unsetting capability
4746 such a setting is global if at outer level; local otherwise
4747 (?imsx-imsx:) non-capturing groups with option setting
4748 (?(cond)re|re) conditional pattern matching
4749
4750 A backreference to itself in a repeated group matches the previous
4751 captured string.
4752
4753 8. General tidying up of studying (both automatic and via "study")
4754 consequential on the addition of new assertions.
4755
4756 9. As in 5.005, unlimited repeated groups that could match an empty substring
4757 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4758 runtime if any iteration does actually match an empty substring.
4759
4760 10. Include the RunTest script in the distribution.
4761
4762 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4763 discrepancies, some of which were old and were also with respect to 5.004. They
4764 have now been fixed.
4765
4766
4767 Version 1.09 28-Apr-98
4768 ----------------------
4769
4770 1. A negated single character class followed by a quantifier with a minimum
4771 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4772 program crashes, or just wrong answers. This did not apply to negated classes
4773 containing more than one character, or to minima other than one.
4774
4775
4776 Version 1.08 27-Mar-98
4777 ----------------------
4778
4779 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4780
4781 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4782 latter must appear before anything that relies on it in the pattern.
4783
4784
4785 Version 1.07 16-Feb-98
4786 ----------------------
4787
4788 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4789 repeat of a potentially empty string).
4790
4791
4792 Version 1.06 23-Jan-98
4793 ----------------------
4794
4795 1. Added Markus Oberhumer's little patches for C++.
4796
4797 2. Literal strings longer than 255 characters were broken.
4798
4799
4800 Version 1.05 23-Dec-97
4801 ----------------------
4802
4803 1. Negated character classes containing more than one character were failing if
4804 PCRE_CASELESS was set at run time.
4805
4806
4807 Version 1.04 19-Dec-97
4808 ----------------------
4809
4810 1. Corrected the man page, where some "const" qualifiers had been omitted.
4811
4812 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4813 input syntax.
4814
4815 3. Fixed memory leak which occurred when a regex with back references was
4816 matched with an offsets vector that wasn't big enough. The temporary memory
4817 that is used in this case wasn't being freed if the match failed.
4818
4819 4. Tidied pcretest to ensure it frees memory that it gets.
4820
4821 5. Temporary memory was being obtained in the case where the passed offsets
4822 vector was exactly big enough.
4823
4824 6. Corrected definition of offsetof() from change 5 below.
4825
4826 7. I had screwed up change 6 below and broken the rules for the use of
4827 setjmp(). Now fixed.
4828
4829
4830 Version 1.03 18-Dec-97
4831 ----------------------
4832
4833 1. A erroneous regex with a missing opening parenthesis was correctly
4834 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4835 on some systems.
4836
4837 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4838 it was reported that one broken compiler failed on the former because "code" is
4839 also an independent variable.
4840
4841 3. The erroneous regex a[]b caused an array overrun reference.
4842
4843 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4844 fail on data ending with that character. (It was going on too far, and checking
4845 the next character, typically a binary zero.) This was specific to the
4846 optimized code for single-character negative classes.
4847
4848 5. Added a contributed patch from the TIN world which does the following:
4849
4850 + Add an undef for memmove, in case the the system defines a macro for it.
4851
4852 + Add a definition of offsetof(), in case there isn't one. (I don't know
4853 the reason behind this - offsetof() is part of the ANSI standard - but
4854 it does no harm).
4855
4856 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4857 most of the places where whitespace preceded '#'. I have given up and
4858 allowed the remaining 2 cases to be at the margin.
4859
4860 + Rename some variables in pcre to eliminate shadowing. This seems very
4861 pedantic, but does no harm, of course.
4862
4863 6. Moved the call to setjmp() into its own function, to get rid of warnings
4864 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4865
4866 7. Constructs such as \d{8,} were compiling into the equivalent of
4867 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4868 outcome, but in this particular case used more store than had been allocated,
4869 which caused the bug to be discovered because it threw up an internal error.
4870
4871 8. The debugging code in both pcre and pcretest for outputting the compiled
4872 form of a regex was going wrong in the case of back references followed by
4873 curly-bracketed repeats.
4874
4875
4876 Version 1.02 12-Dec-97
4877 ----------------------
4878
4879 1. Typos in pcre.3 and comments in the source fixed.
4880
4881 2. Applied a contributed patch to get rid of places where it used to remove
4882 'const' from variables, and fixed some signed/unsigned and uninitialized
4883 variable warnings.
4884
4885 3. Added the "runtest" target to Makefile.
4886
4887 4. Set default compiler flag to -O2 rather than just -O.
4888
4889
4890 Version 1.01 19-Nov-97
4891 ----------------------
4892
4893 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4894 like /([ab]*)*/, that is, for classes with more than one character in them.
4895
4896 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4897 as /((?>a*))*/ (a PCRE_EXTRA facility).
4898
4899
4900 Version 1.00 18-Nov-97
4901 ----------------------
4902
4903 1. Added compile-time macros to support systems such as SunOS4 which don't have
4904 memmove() or strerror() but have other things that can be used instead.
4905
4906 2. Arranged that "make clean" removes the executables.
4907
4908
4909 Version 0.99 27-Oct-97
4910 ----------------------
4911
4912 1. Fixed bug in code for optimizing classes with only one character. It was
4913 initializing a 32-byte map regardless, which could cause it to run off the end
4914 of the memory it had got.
4915
4916 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4917
4918
4919 Version 0.98 22-Oct-97
4920 ----------------------
4921
4922 1. Fixed bug in code for handling temporary memory usage when there are more
4923 back references than supplied space in the ovector. This could cause segfaults.
4924
4925
4926 Version 0.97 21-Oct-97
4927 ----------------------
4928
4929 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4930
4931 2. Optimized negated single characters not to use a bit map.
4932
4933 3. Brought error texts together as macro definitions; clarified some of them;
4934 fixed one that was wrong - it said "range out of order" when it meant "invalid
4935 escape sequence".
4936
4937 4. Changed some char * arguments to const char *.
4938
4939 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4940
4941 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4942 pcretest.
4943
4944
4945 Version 0.96 16-Oct-97
4946 ----------------------
4947
4948 1. Added a simple "pgrep" utility to the distribution.
4949
4950 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4951 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4952 where "ddd" means "one or more decimal digits".
4953
4954 3. Fixed serious bug. If a pattern had a back reference, but the call to
4955 pcre_exec() didn't supply a large enough ovector to record the related
4956 identifying subpattern, the match always failed. PCRE now remembers the number
4957 of the largest back reference, and gets some temporary memory in which to save
4958 the offsets during matching if necessary, in order to ensure that
4959 backreferences always work.
4960
4961 4. Increased the compatibility with Perl in a number of ways:
4962
4963 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4964 to request this handling. The option can be set at compile or exec time.
4965
4966 (b) $ matches before a terminating newline by default; an option
4967 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4968 mode). The option can be set at compile or exec time.
4969
4970 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4971 the same as Perl's. If the decimal number it represents is less than 10
4972 or there aren't that many previous left capturing parentheses, an octal
4973 escape is read. Inside a character class, it's always an octal escape,
4974 even if it is a single digit.
4975
4976 (d) An escaped but undefined alphabetic character is taken as a literal,
4977 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4978 escapes.
4979
4980 (e) {0} is now permitted. (The previous item is removed from the compiled
4981 pattern).
4982
4983 5. Changed all the names of code files so that the basic parts are no longer
4984 than 10 characters, and abolished the teeny "globals.c" file.
4985
4986 6. Changed the handling of character classes; they are now done with a 32-byte
4987 bit map always.
4988
4989 7. Added the -d and /D options to pcretest to make it possible to look at the
4990 internals of compilation without having to recompile pcre.
4991
4992
4993 Version 0.95 23-Sep-97
4994 ----------------------
4995
4996 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4997 \x20 at the start of a run of normal characters. These were being treated as
4998 real characters, instead of the source characters being re-checked.
4999
5000
5001 Version 0.94 18-Sep-97
5002 ----------------------
5003
5004 1. The functions are now thread-safe, with the caveat that the global variables
5005 containing pointers to malloc() and free() or alternative functions are the
5006 same for all threads.
5007
5008 2. Get pcre_study() to generate a bitmap of initial characters for non-
5009 anchored patterns when this is possible, and use it if passed to pcre_exec().
5010
5011
5012 Version 0.93 15-Sep-97
5013 ----------------------
5014
5015 1. /(b)|(:+)/ was computing an incorrect first character.
5016
5017 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5018 but not actually doing anything yet.
5019
5020 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5021 as Perl does (e.g. [-az] or [az-]).
5022
5023 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5024 all possible positions.
5025
5026 5. Split up into different modules to avoid including unneeded functions in a
5027 compiled binary. However, compile and exec are still in one module. The "study"
5028 function is split off.
5029
5030 6. The character tables are now in a separate module whose source is generated
5031 by an auxiliary program - but can then be edited by hand if required. There are
5032 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5033 toupper() in the code.
5034
5035 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5036 make them global. Abolish the function for setting them, as the caller can now
5037 set them directly.
5038
5039
5040 Version 0.92 11-Sep-97
5041 ----------------------
5042
5043 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5044 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5045
5046 2. Caseless matching was not working in character classes if the characters in
5047 the pattern were in upper case.
5048
5049 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5050
5051 4. Make PCRE_ANCHORED public and accept as a compile option.
5052
5053 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5054 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5055 pass them.
5056
5057 6. Give an error if bad option bits passed at compile or run time.
5058
5059 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5060 pcretest to cause it to pass that flag.
5061
5062 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5063 options, and the first character, if set.
5064
5065 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5066
5067
5068 Version 0.91 10-Sep-97
5069 ----------------------
5070
5071 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5072 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5073
5074 2. PCRE was looping on encountering an indefinitely repeated back reference to
5075 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5076 Perl does - treats the match as successful.
5077
5078 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5