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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1239 - (show annotations)
Fri Jan 18 08:20:44 2013 UTC (6 years, 8 months ago) by zherczeg
File size: 241077 byte(s)
Error occurred while calculating annotation data.
Fix forward search in JIT when link size is 3 or greater.
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
33 Version 8.32 30-November-2012
34 -----------------------------
35
36 1. Improved JIT compiler optimizations for first character search and single
37 character iterators.
38
39 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
40 Patch by Daniel Richard G.
41
42 3. Single character iterator optimizations in the JIT compiler.
43
44 4. Improved JIT compiler optimizations for character ranges.
45
46 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
47 Reported by Giuseppe D'Angelo.
48
49 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
50 of a line, was being set incorrectly in cases where .* appeared inside
51 atomic brackets at the start of a pattern, or where there was a subsequent
52 *PRUNE or *SKIP.
53
54 7. Improved instruction cache flush for POWER/PowerPC.
55 Patch by Daniel Richard G.
56
57 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
58 grep:
59
60 (a) There is now no limit to the number of patterns to be matched.
61
62 (b) An error is given if a pattern is too long.
63
64 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
65 are now supported.
66
67 (d) --exclude-from and --include-from (multiple use) have been added.
68
69 (e) Exclusions and inclusions now apply to all files and directories, not
70 just to those obtained from scanning a directory recursively.
71
72 (f) Multiple uses of -f and --file-list are now supported.
73
74 (g) In a Windows environment, the default for -d has been changed from
75 "read" (the GNU grep default) to "skip", because otherwise the presence
76 of a directory in the file list provokes an error.
77
78 (h) The documentation has been revised and clarified in places.
79
80 9. Improve the matching speed of capturing brackets.
81
82 10. Changed the meaning of \X so that it now matches a Unicode extended
83 grapheme cluster.
84
85 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
86 out POSIX threads when JIT support is configured.
87
88 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
89
90 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
91 structure could go wrong in environments where size_t is not the same size
92 as int.
93
94 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
95 set.
96
97 15. The EBCDIC support had decayed; later updates to the code had included
98 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
99 general tidy up of EBCDIC-related issues, and the documentation was also
100 not quite right. There is now a test that can be run on ASCII systems to
101 check some of the EBCDIC-related things (but is it not a full test).
102
103 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
104 in a small tidy to the code.
105
106 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
107
108 18. If the --only-matching (-o) option in pcregrep is specified multiple
109 times, each one causes appropriate output. For example, -o1 -o2 outputs the
110 substrings matched by the 1st and 2nd capturing parentheses. A separating
111 string can be specified by --om-separator (default empty).
112
113 19. Improving the first n character searches.
114
115 20. Turn case lists for horizontal and vertical white space into macros so that
116 they are defined only once.
117
118 21. This set of changes together give more compatible Unicode case-folding
119 behaviour for characters that have more than one other case when UCP
120 support is available.
121
122 (a) The Unicode property table now has offsets into a new table of sets of
123 three or more characters that are case-equivalent. The MultiStage2.py
124 script that generates these tables (the pcre_ucd.c file) now scans
125 CaseFolding.txt instead of UnicodeData.txt for character case
126 information.
127
128 (b) The code for adding characters or ranges of characters to a character
129 class has been abstracted into a generalized function that also handles
130 case-independence. In UTF-mode with UCP support, this uses the new data
131 to handle characters with more than one other case.
132
133 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
134 whose other case is greater than 256 are now correctly matched
135 caselessly. Previously, the high codepoint matched the low one, but not
136 vice versa.
137
138 (d) The processing of \h, \H, \v, and \ in character classes now makes use
139 of the new class addition function, using character lists defined as
140 macros alongside the case definitions of 20 above.
141
142 (e) Caseless back references now work with characters that have more than
143 one other case.
144
145 (f) General caseless matching of characters with more than one other case
146 is supported.
147
148 22. Unicode character properties were updated from Unicode 6.2.0
149
150 23. Improved CMake support under Windows. Patch by Daniel Richard G.
151
152 24. Add support for 32-bit character strings, and UTF-32
153
154 25. Major JIT compiler update (code refactoring and bugfixing).
155 Experimental Sparc 32 support is added.
156
157 26. Applied a modified version of Daniel Richard G's patch to create
158 pcre.h.generic and config.h.generic by "make" instead of in the
159 PrepareRelease script.
160
161 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
162 pcre_compile.c when checking for a zero character.
163
164 28. Introducing a native interface for JIT. Through this interface, the compiled
165 machine code can be directly executed. The purpose of this interface is to
166 provide fast pattern matching, so several sanity checks are not performed.
167 However, feature tests are still performed. The new interface provides
168 1.4x speedup compared to the old one.
169
170 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
171 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
172 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
173
174 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
175 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
176 the "old" RFC 2279). Instead, it ended up passing a negative length to
177 pcre_exec().
178
179 31. Add support for GCC's visibility feature to hide internal functions.
180
181 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
182 "unknown -C option" after outputting 0 or 1.
183
184 33. There is now support for generating a code coverage report for the test
185 suite in environments where gcc is the compiler and lcov is installed. This
186 is mainly for the benefit of the developers.
187
188 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
189 unaddressable using valgrind annotations, allowing valgrind to detect
190 invalid memory accesses. This is mainly for the benefit of the developers.
191
192 25. (*UTF) can now be used to start a pattern in any of the three libraries.
193
194 26. Give configure error if --enable-cpp but no C++ compiler found.
195
196
197 Version 8.31 06-July-2012
198 -------------------------
199
200 1. Fixing a wrong JIT test case and some compiler warnings.
201
202 2. Removed a bashism from the RunTest script.
203
204 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
205 to unsigned type, result still unsigned" that was given by an MS compiler
206 on encountering the code "-sizeof(xxx)".
207
208 4. Partial matching support is added to the JIT compiler.
209
210 5. Fixed several bugs concerned with partial matching of items that consist
211 of more than one character:
212
213 (a) /^(..)\1/ did not partially match "aba" because checking references was
214 done on an "all or nothing" basis. This also applied to repeated
215 references.
216
217 (b) \R did not give a hard partial match if \r was found at the end of the
218 subject.
219
220 (c) \X did not give a hard partial match after matching one or more
221 characters at the end of the subject.
222
223 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
224 a partial match for the string "\r".
225
226 (e) When newline was set to CRLF, the metacharacter "." did not recognize
227 a partial match for a CR character at the end of the subject string.
228
229 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
230 running pcretest, the text "(JIT)" added to the output whenever JIT is
231 actually used to run the match.
232
233 7. Individual JIT compile options can be set in pcretest by following -s+[+]
234 or /S+[+] with a digit between 1 and 7.
235
236 8. OP_NOT now supports any UTF character not just single-byte ones.
237
238 9. (*MARK) control verb is now supported by the JIT compiler.
239
240 10. The command "./RunTest list" lists the available tests without actually
241 running any of them. (Because I keep forgetting what they all are.)
242
243 11. Add PCRE_INFO_MAXLOOKBEHIND.
244
245 12. Applied a (slightly modified) user-supplied patch that improves performance
246 when the heap is used for recursion (compiled with --disable-stack-for-
247 recursion). Instead of malloc and free for each heap frame each time a
248 logical recursion happens, frames are retained on a chain and re-used where
249 possible. This sometimes gives as much as 30% improvement.
250
251 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
252 call.
253
254 14. As documented, (*COMMIT) is now confined to within a positive assertion.
255
256 15. It is now possible to link pcretest with libedit as an alternative to
257 libreadline.
258
259 16. (*COMMIT) control verb is now supported by the JIT compiler.
260
261 17. The Unicode data tables have been updated to Unicode 6.1.0.
262
263 18. Added --file-list option to pcregrep.
264
265 19. Added binary file support to pcregrep, including the -a, --binary-files,
266 -I, and --text options.
267
268 20. The madvise function is renamed for posix_madvise for QNX compatibility
269 reasons. Fixed by Giuseppe D'Angelo.
270
271 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
272
273 22. Changed the option for creating symbolic links for 16-bit man pages from
274 -s to -sf so that re-installing does not cause issues.
275
276 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
277
278 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
279 matches in certain environments (the workspace was not being correctly
280 retained). Also added to pcre_dfa_exec() a simple plausibility check on
281 some of the workspace data at the beginning of a restart.
282
283 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
284 was not doing so when it should - probably a typo introduced by SVN 528
285 (change 8.10/14).
286
287 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
288 \w+ when the character tables indicated that \x{c4} was a word character.
289 There were several related cases, all because the tests for doing a table
290 lookup were testing for characters less than 127 instead of 255.
291
292 27. If a pattern contains capturing parentheses that are not used in a match,
293 their slots in the ovector are set to -1. For those that are higher than
294 any matched groups, this happens at the end of processing. In the case when
295 there were back references that the ovector was too small to contain
296 (causing temporary malloc'd memory to be used during matching), and the
297 highest capturing number was not used, memory off the end of the ovector
298 was incorrectly being set to -1. (It was using the size of the temporary
299 memory instead of the true size.)
300
301 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
302 ovector size, it uses memory at the end of the block that it has got.
303
304 29. Check for an overlong MARK name and give an error at compile time. The
305 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
306
307 30. JIT compiler update.
308
309 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
310 Rill for the patch.
311
312 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
313
314 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
315
316 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
317 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
318 was enabled.
319
320 35. Improve JIT code generation for greedy plus quantifier.
321
322 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
323 1 to "aa" instead of to an empty string. The bug affected repeated groups
324 that could potentially match an empty string.
325
326 37. Optimizing single character iterators in JIT.
327
328 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
329 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
330 codepoints that are too big for the mode are faulted, and in a UTF mode,
331 disallowed codepoints are also faulted.
332
333 39. If PCRE was compiled with UTF support, in three places in the DFA
334 matcher there was code that should only have been obeyed in UTF mode, but
335 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
336 processing when bytes with values greater than 127 were present. In 16-bit
337 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
338 both cases the values are those that cannot be the first data item in a UTF
339 character. The three items that might have provoked this were recursions,
340 possessively repeated groups, and atomic groups.
341
342 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
343 and pcregrep, because some OS require shared objects to be explicitly
344 passed to ld, causing the link step to fail if they are not.
345
346 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
347 mode, patterns that started with \h* or \R* might be incorrectly matched.
348
349
350 Version 8.30 04-February-2012
351 -----------------------------
352
353 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
354 name is defined in ctype.h.
355
356 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
357 only in quite long subpatterns.
358
359 3. Removed the function pcre_info(), which has been obsolete and deprecated
360 since it was replaced by pcre_fullinfo() in February 2000.
361
362 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
363 match a (*MARK), and the match failed at the start of the subject, a
364 reference to memory before the start of the subject could occur. This bug
365 was introduced by fix 17 of release 8.21.
366
367 5. A reference to an unset group with zero minimum repetition was giving
368 totally wrong answers (in non-JavaScript-compatibility mode). For example,
369 /(another)?(\1?)test/ matched against "hello world test". This bug was
370 introduced in release 8.13.
371
372 6. Add support for 16-bit character strings (a large amount of work involving
373 many changes and refactorings).
374
375 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
376 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
377 from a file.
378
379 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
380 rounding is not applied in this particular case).
381
382 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
383 if they appear, or are escaped, in patterns.
384
385 10. Get rid of a number of -Wunused-but-set-variable warnings.
386
387 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
388 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
389 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
390 also returns the mark "x". This bug applied to capturing parentheses,
391 non-capturing parentheses, and atomic parentheses. It also applied to some
392 assertions.
393
394 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
395 information out of configure.ac instead of relying on pcre.h.generic, which
396 is not stored in the repository.
397
398 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
399 -lreadline.
400
401 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
402
403 15. Applied Graycode's patch to put the top-level frame on the stack rather
404 than the heap when not using the stack for recursion. This gives a
405 performance improvement in many cases when recursion is not deep.
406
407 16. Experimental code added to "pcretest -C" to output the stack frame size.
408
409
410 Version 8.21 12-Dec-2011
411 ------------------------
412
413 1. Updating the JIT compiler.
414
415 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
416 are added as well.
417
418 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
419 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
420 calling _pcre_jit_exec. Some extra comments are added.
421
422 4. (*MARK) settings inside atomic groups that do not contain any capturing
423 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
424 was introduced by change 18 for 8.20.
425
426 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
427 ECMA-262 standard.
428
429 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
430 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
431 This bug was probably introduced by change 9 of 8.13.
432
433 7. While fixing 6 above, I noticed that a number of other items were being
434 incorrectly rejected as "not fixed length". This arose partly because newer
435 opcodes had not been added to the fixed-length checking code. I have (a)
436 corrected the bug and added tests for these items, and (b) arranged for an
437 error to occur if an unknown opcode is encountered while checking for fixed
438 length instead of just assuming "not fixed length". The items that were
439 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
440 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
441 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
442
443 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
444 being incorrectly compiled and would have given unpredicatble results.
445
446 9. A possessively repeated subpattern with minimum repeat count greater than
447 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
448 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
449 the first (A) could occur when it should not.
450
451 10. Add a cast and remove a redundant test from the code.
452
453 11. JIT should use pcre_malloc/pcre_free for allocation.
454
455 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
456 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
457 is anything other than /usr, -L is still shown).
458
459 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
460
461 14. Perl does not support \N without a following name in a [] class; PCRE now
462 also gives an error.
463
464 15. If a forward reference was repeated with an upper limit of around 2000,
465 it caused the error "internal error: overran compiling workspace". The
466 maximum number of forward references (including repeats) was limited by the
467 internal workspace, and dependent on the LINK_SIZE. The code has been
468 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
469 the default depends on LINK_SIZE. There is a new upper limit (for safety)
470 of around 200,000 forward references. While doing this, I also speeded up
471 the filling in of repeated forward references.
472
473 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
474 incorrectly expecting the subject to contain another "a" after the start.
475
476 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
477 in the match, the SKIP should be ignored. This was not happening; instead
478 the SKIP was being treated as NOMATCH. For patterns such as
479 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
480 tested.
481
482 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
483 now much more compatible with Perl, in particular in cases where the result
484 is a non-match for a non-anchored pattern. For example, if
485 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
486 "m", where previously it did not return a name. A side effect of this
487 change is that for partial matches, the last encountered mark name is
488 returned, as for non matches. A number of tests that were previously not
489 Perl-compatible have been moved into the Perl-compatible test files. The
490 refactoring has had the pleasing side effect of removing one argument from
491 the match() function, thus reducing its stack requirements.
492
493 19. If the /S+ option was used in pcretest to study a pattern using JIT,
494 subsequent uses of /S (without +) incorrectly behaved like /S+.
495
496 21. Retrieve executable code size support for the JIT compiler and fixing
497 some warnings.
498
499 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
500 not work when the shorter character appeared right at the end of the
501 subject string.
502
503 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
504 systems.
505
506 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
507 output it when the /M option is used in pcretest.
508
509 25. The CheckMan script was not being included in the distribution. Also, added
510 an explicit "perl" to run Perl scripts from the PrepareRelease script
511 because this is reportedly needed in Windows.
512
513 26. If study data was being save in a file and studying had not found a set of
514 "starts with" bytes for the pattern, the data written to the file (though
515 never used) was taken from uninitialized memory and so caused valgrind to
516 complain.
517
518 27. Updated RunTest.bat as provided by Sheri Pierce.
519
520 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
521
522 29. Computation of memory usage for the table of capturing group names was
523 giving an unnecessarily large value.
524
525
526 Version 8.20 21-Oct-2011
527 ------------------------
528
529 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
530 a POSIX class. After further experiments with Perl, which convinced me that
531 Perl has bugs and confusions, a closing square bracket is no longer allowed
532 in a POSIX name. This bug also affected patterns with classes that started
533 with full stops.
534
535 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
536 captured substring, but while checking the failing first alternative,
537 substring 1 is temporarily captured. If the output vector supplied to
538 pcre_exec() was not big enough for this capture, the yield of the function
539 was still zero ("insufficient space for captured substrings"). This cannot
540 be totally fixed without adding another stack variable, which seems a lot
541 of expense for a edge case. However, I have improved the situation in cases
542 such as /(a)(b)x|abc/ matched against "abc", where the return code
543 indicates that fewer than the maximum number of slots in the ovector have
544 been set.
545
546 3. Related to (2) above: when there are more back references in a pattern than
547 slots in the output vector, pcre_exec() uses temporary memory during
548 matching, and copies in the captures as far as possible afterwards. It was
549 using the entire output vector, but this conflicts with the specification
550 that only 2/3 is used for passing back captured substrings. Now it uses
551 only the first 2/3, for compatibility. This is, of course, another edge
552 case.
553
554 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
555 main code base, and can be used by building with --enable-jit. When this is
556 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
557 runtime --no-jit option is given.
558
559 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
560 ovector, the return from the function was zero, implying that there were
561 other matches that did not fit. The correct "exactly full" value is now
562 returned.
563
564 6. If a subpattern that was called recursively or as a subroutine contained
565 (*PRUNE) or any other control that caused it to give a non-standard return,
566 invalid errors such as "Error -26 (nested recursion at the same subject
567 position)" or even infinite loops could occur.
568
569 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
570 computing the minimum length on reaching *ACCEPT, and so ended up with the
571 wrong value of 1 rather than 0. Further investigation indicates that
572 computing a minimum subject length in the presence of *ACCEPT is difficult
573 (think back references, subroutine calls), and so I have changed the code
574 so that no minimum is registered for a pattern that contains *ACCEPT.
575
576 8. If (*THEN) was present in the first (true) branch of a conditional group,
577 it was not handled as intended. [But see 16 below.]
578
579 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
580 Sheri Pierce.
581
582 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
583 the first byte in a match must be "a".
584
585 11. Change 17 for 8.13 increased the recursion depth for patterns like
586 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
587 pattern contains any instances of (*THEN). If it does not, the old
588 optimizations are restored. It would be nice to do this on a per-group
589 basis, but at the moment that is not feasible.
590
591 12. In some environments, the output of pcretest -C is CRLF terminated. This
592 broke RunTest's code that checks for the link size. A single white space
593 character after the value is now allowed for.
594
595 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
596 For "fr", it uses the Windows-specific input and output files.
597
598 14. If (*THEN) appeared in a group that was called recursively or as a
599 subroutine, it did not work as intended. [But see next item.]
600
601 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
602 pattern fragments (but not containing any | characters). If A and B are
603 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
604 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
605 D. In other words, Perl considers parentheses that do not contain any |
606 characters to be part of a surrounding alternative, whereas PCRE was
607 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
608 differently. PCRE now behaves in the same way as Perl, except in the case
609 of subroutine/recursion calls such as (?1) which have in any case always
610 been different (but PCRE had them first :-).
611
612 16. Related to 15 above: Perl does not treat the | in a conditional group as
613 creating alternatives. Such a group is treated in the same way as an
614 ordinary group without any | characters when processing (*THEN). PCRE has
615 been changed to match Perl's behaviour.
616
617 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
618 RunGrepTest script failed.
619
620 18. Change 22 for version 13 caused atomic groups to use more stack. This is
621 inevitable for groups that contain captures, but it can lead to a lot of
622 stack use in large patterns. The old behaviour has been restored for atomic
623 groups that do not contain any capturing parentheses.
624
625 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
626 suppress the check for a minimum subject length at run time. (If it was
627 given to pcre_exec() or pcre_dfa_exec() it did work.)
628
629 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
630 fail to work when decoding hex characters in data strings in EBCDIC
631 environments.
632
633 21. It appears that in at least one Mac OS environment, the isxdigit() function
634 is implemented as a macro that evaluates to its argument more than once,
635 contravening the C 90 Standard (I haven't checked a later standard). There
636 was an instance in pcretest which caused it to go wrong when processing
637 \x{...} escapes in subject strings. The has been rewritten to avoid using
638 things like p++ in the argument of isxdigit().
639
640
641 Version 8.13 16-Aug-2011
642 ------------------------
643
644 1. The Unicode data tables have been updated to Unicode 6.0.0.
645
646 2. Two minor typos in pcre_internal.h have been fixed.
647
648 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
649 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
650 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
651
652 4. There were a number of related bugs in the code for matching backrefences
653 caselessly in UTF-8 mode when codes for the characters concerned were
654 different numbers of bytes. For example, U+023A and U+2C65 are an upper
655 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
656 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
657 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
658 2-byte code at the end of the subject (it thought there wasn't enough data
659 left).
660
661 5. Comprehensive information about what went wrong is now returned by
662 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
663 as the output vector has at least 2 elements. The offset of the start of
664 the failing character and a reason code are placed in the vector.
665
666 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
667 now returned is for the first byte of the failing character, instead of the
668 last byte inspected. This is an incompatible change, but I hope it is small
669 enough not to be a problem. It makes the returned offset consistent with
670 pcre_exec() and pcre_dfa_exec().
671
672 7. pcretest now gives a text phrase as well as the error number when
673 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
674 failure, the offset and reason code are output.
675
676 8. When \R was used with a maximizing quantifier it failed to skip backwards
677 over a \r\n pair if the subsequent match failed. Instead, it just skipped
678 back over a single character (\n). This seems wrong (because it treated the
679 two characters as a single entity when going forwards), conflicts with the
680 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
681 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
682 has been changed.
683
684 9. Some internal refactoring has changed the processing so that the handling
685 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
686 time (the PCRE_DOTALL option was changed this way some time ago: version
687 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
688 which was always a bit of a fudge. It also means that there is one less
689 argument for the match() function, which reduces its stack requirements
690 slightly. This change also fixes an incompatibility with Perl: the pattern
691 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
692
693 10. More internal refactoring has drastically reduced the number of recursive
694 calls to match() for possessively repeated groups such as (abc)++ when
695 using pcre_exec().
696
697 11. While implementing 10, a number of bugs in the handling of groups were
698 discovered and fixed:
699
700 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
701 (a|)*(?1) gave a compile-time internal error.
702 ((a|)+)+ did not notice that the outer group could match an empty string.
703 (^a|^)+ was not marked as anchored.
704 (.*a|.*)+ was not marked as matching at start or after a newline.
705
706 12. Yet more internal refactoring has removed another argument from the match()
707 function. Special calls to this function are now indicated by setting a
708 value in a variable in the "match data" data block.
709
710 13. Be more explicit in pcre_study() instead of relying on "default" for
711 opcodes that mean there is no starting character; this means that when new
712 ones are added and accidentally left out of pcre_study(), testing should
713 pick them up.
714
715 14. The -s option of pcretest has been documented for ages as being an old
716 synonym of -m (show memory usage). I have changed it to mean "force study
717 for every regex", that is, assume /S for every regex. This is similar to -i
718 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
719 using it. It makes it easier to run collections of tests with and without
720 study enabled, and thereby test pcre_study() more easily. All the standard
721 tests are now run with and without -s (but some patterns can be marked as
722 "never study" - see 20 below).
723
724 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
725 restoration of the capturing data to the outer values was not happening
726 correctly.
727
728 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
729 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
730 pattern had matched an empty string, and so incorrectly returned a no
731 match.
732
733 17. There was optimizing code for the last branch of non-capturing parentheses,
734 and also for the obeyed branch of a conditional subexpression, which used
735 tail recursion to cut down on stack usage. Unfortunately, now that there is
736 the possibility of (*THEN) occurring in these branches, tail recursion is
737 no longer possible because the return has to be checked for (*THEN). These
738 two optimizations have therefore been removed. [But see 8.20/11 above.]
739
740 18. If a pattern containing \R was studied, it was assumed that \R always
741 matched two bytes, thus causing the minimum subject length to be
742 incorrectly computed because \R can also match just one byte.
743
744 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
745 was incorrectly computed.
746
747 20. If /S is present twice on a test pattern in pcretest input, it now
748 *disables* studying, thereby overriding the use of -s on the command line
749 (see 14 above). This is necessary for one or two tests to keep the output
750 identical in both cases.
751
752 21. When (*ACCEPT) was used in an assertion that matched an empty string and
753 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
754
755 22. When an atomic group that contained a capturing parenthesis was
756 successfully matched, but the branch in which it appeared failed, the
757 capturing was not being forgotten if a higher numbered group was later
758 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
759 group 1 to "a", when in fact it should be unset. This applied to multi-
760 branched capturing and non-capturing groups, repeated or not, and also to
761 positive assertions (capturing in negative assertions does not happen
762 in PCRE) and also to nested atomic groups.
763
764 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
765 subject after a captured substring, to make it easier to tell which of a
766 number of identical substrings has been captured.
767
768 24. The way atomic groups are processed by pcre_exec() has been changed so that
769 if they are repeated, backtracking one repetition now resets captured
770 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
771 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
772 "aaa". Previously, it would have been "a". As part of this code
773 refactoring, the way recursive calls are handled has also been changed.
774
775 25. If an assertion condition captured any substrings, they were not passed
776 back unless some other capturing happened later. For example, if
777 (?(?=(a))a) was matched against "a", no capturing was returned.
778
779 26. When studying a pattern that contained subroutine calls or assertions,
780 the code for finding the minimum length of a possible match was handling
781 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
782 group 1 called group 2 while simultaneously a separate group 2 called group
783 1). A stack overflow occurred in this case. I have fixed this by limiting
784 the recursion depth to 10.
785
786 27. Updated RunTest.bat in the distribution to the version supplied by Tom
787 Fortmann. This supports explicit test numbers on the command line, and has
788 argument validation and error reporting.
789
790 28. An instance of \X with an unlimited repeat could fail if at any point the
791 first character it looked at was a mark character.
792
793 29. Some minor code refactoring concerning Unicode properties and scripts
794 should reduce the stack requirement of match() slightly.
795
796 30. Added the '=' option to pcretest to check the setting of unused capturing
797 slots at the end of the pattern, which are documented as being -1, but are
798 not included in the return count.
799
800 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
801 compiled something random. Now it gives a compile-time error (as does
802 Perl).
803
804 32. A *MARK encountered during the processing of a positive assertion is now
805 recorded and passed back (compatible with Perl).
806
807 33. If --only-matching or --colour was set on a pcregrep call whose pattern
808 had alternative anchored branches, the search for a second match in a line
809 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
810 matched the line "0102" twice. The same bug affected patterns that started
811 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
812 twice.
813
814 34. Previously, PCRE did not allow quantification of assertions. However, Perl
815 does, and because of capturing effects, quantifying parenthesized
816 assertions may at times be useful. Quantifiers are now allowed for
817 parenthesized assertions.
818
819 35. A minor code tidy in pcre_compile() when checking options for \R usage.
820
821 36. \g was being checked for fancy things in a character class, when it should
822 just be a literal "g".
823
824 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
825 appearance of a nested POSIX class supersedes an apparent external class.
826 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
827 unescaped square brackets may also appear as part of class names. For
828 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
829 more like Perl. (But see 8.20/1 above.)
830
831 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
832 was because it thought it was \N{name}, which is not supported).
833
834 39. Add minix to OS list not supporting the -S option in pcretest.
835
836 40. PCRE tries to detect cases of infinite recursion at compile time, but it
837 cannot analyze patterns in sufficient detail to catch mutual recursions
838 such as ((?1))((?2)). There is now a runtime test that gives an error if a
839 subgroup is called recursively as a subpattern for a second time at the
840 same position in the subject string. In previous releases this might have
841 been caught by the recursion limit, or it might have run out of stack.
842
843 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
844 happen only once. PCRE was, however incorrectly giving a compile time error
845 "recursive call could loop indefinitely" because it cannot analyze the
846 pattern in sufficient detail. The compile time test no longer happens when
847 PCRE is compiling a conditional subpattern, but actual runaway loops are
848 now caught at runtime (see 40 above).
849
850 42. It seems that Perl allows any characters other than a closing parenthesis
851 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
852 has been changed to be the same.
853
854 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
855 as not to get warnings when autogen.sh is called. Also changed
856 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
857
858 44. To help people who use pcregrep to scan files containing exceedingly long
859 lines, the following changes have been made:
860
861 (a) The default value of the buffer size parameter has been increased from
862 8K to 20K. (The actual buffer used is three times this size.)
863
864 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
865 PCRE is built.
866
867 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
868 to be set at run time.
869
870 (d) Numerical values in pcregrep options can be followed by K or M, for
871 example --buffer-size=50K.
872
873 (e) If a line being scanned overflows pcregrep's buffer, an error is now
874 given and the return code is set to 2.
875
876 45. Add a pointer to the latest mark to the callout data block.
877
878 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
879 partial match of an empty string instead of no match. This was specific to
880 the use of ".".
881
882 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
883 complete match instead of a partial match. This bug was dependent on both
884 the PCRE_UTF8 and PCRE_DOTALL options being set.
885
886 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
887 starting byte set, because \b was not being ignored.
888
889
890 Version 8.12 15-Jan-2011
891 ------------------------
892
893 1. Fixed some typos in the markup of the man pages, and wrote a script that
894 checks for such things as part of the documentation building process.
895
896 2. On a big-endian 64-bit system, pcregrep did not correctly process the
897 --match-limit and --recursion-limit options (added for 8.11). In
898 particular, this made one of the standard tests fail. (The integer value
899 went into the wrong half of a long int.)
900
901 3. If the --colour option was given to pcregrep with -v (invert match), it
902 did strange things, either producing crazy output, or crashing. It should,
903 of course, ignore a request for colour when reporting lines that do not
904 match.
905
906 4. Another pcregrep bug caused similar problems if --colour was specified with
907 -M (multiline) and the pattern match finished with a line ending.
908
909 5. In pcregrep, when a pattern that ended with a literal newline sequence was
910 matched in multiline mode, the following line was shown as part of the
911 match. This seems wrong, so I have changed it.
912
913 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
914 the check for further matches in the same line (so they could be coloured)
915 to overrun the end of the current line. If another match was found, it was
916 incorrectly shown (and then shown again when found in the next line).
917
918 7. If pcregrep was compiled under Windows, there was a reference to the
919 function pcregrep_exit() before it was defined. I am assuming this was
920 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
921 reported by a user. I've moved the definition above the reference.
922
923
924 Version 8.11 10-Dec-2010
925 ------------------------
926
927 1. (*THEN) was not working properly if there were untried alternatives prior
928 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
929 backtracked to try for "b" instead of moving to the next alternative branch
930 at the same level (in this case, to look for "c"). The Perl documentation
931 is clear that when (*THEN) is backtracked onto, it goes to the "next
932 alternative in the innermost enclosing group".
933
934 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
935 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
936 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
937 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
938 (*THEN).
939
940 3. If \s appeared in a character class, it removed the VT character from
941 the class, even if it had been included by some previous item, for example
942 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
943 of \s, but is part of the POSIX "space" class.)
944
945 4. A partial match never returns an empty string (because you can always
946 match an empty string at the end of the subject); however the checking for
947 an empty string was starting at the "start of match" point. This has been
948 changed to the "earliest inspected character" point, because the returned
949 data for a partial match starts at this character. This means that, for
950 example, /(?<=abc)def/ gives a partial match for the subject "abc"
951 (previously it gave "no match").
952
953 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
954 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
955 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
956 has an implication that the given string is incomplete (because a partial
957 match is preferred over a full match). For this reason, these items now
958 give a partial match in this situation. [Aside: previously, the one case
959 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
960 match rather than a full match, which was wrong by the old rules, but is
961 now correct.]
962
963 6. There was a bug in the handling of #-introduced comments, recognized when
964 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
965 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
966 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
967 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
968 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
969 places in pcre_compile().
970
971 7. Related to (6) above, when pcre_compile() was skipping #-introduced
972 comments when looking ahead for named forward references to subpatterns,
973 the only newline sequence it recognized was NL. It now handles newlines
974 according to the set newline convention.
975
976 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
977 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
978 cater for a lack of strerror(). These oversights have been fixed.
979
980 9. Added --match-limit and --recursion-limit to pcregrep.
981
982 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
983
984 11. When the -o option was used, pcregrep was setting a return code of 1, even
985 when matches were found, and --line-buffered was not being honoured.
986
987 12. Added an optional parentheses number to the -o and --only-matching options
988 of pcregrep.
989
990 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
991 can match an empty string. The code to do it in pcretest and pcredemo
992 needed fixing:
993
994 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
995 only one byte after an empty string match just before CRLF (this case
996 just got forgotten; "any" and "anycrlf" were OK).
997
998 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
999 mode when an empty string match preceded an ASCII character followed by
1000 a non-ASCII character. (The code for advancing by one character rather
1001 than one byte was nonsense.)
1002
1003 (c) The pcredemo.c sample program did not have any code at all to handle
1004 the cases when CRLF is a valid newline sequence.
1005
1006 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1007 as a starting offset was within the subject string. There is now a new
1008 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1009 negative or greater than the length of the string. In order to test this,
1010 pcretest is extended to allow the setting of negative starting offsets.
1011
1012 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1013 starting offset points to the beginning of a UTF-8 character was
1014 unnecessarily clumsy. I tidied it up.
1015
1016 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1017 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1018
1019 17. Nobody had reported that the --include_dir option, which was added in
1020 release 7.7 should have been called --include-dir (hyphen, not underscore)
1021 for compatibility with GNU grep. I have changed it to --include-dir, but
1022 left --include_dir as an undocumented synonym, and the same for
1023 --exclude-dir, though that is not available in GNU grep, at least as of
1024 release 2.5.4.
1025
1026 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1027 characters from a string of bytes) have been redefined so as not to use
1028 loops, in order to improve performance in some environments. At the same
1029 time, I abstracted some of the common code into auxiliary macros to save
1030 repetition (this should not affect the compiled code).
1031
1032 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1033 compile-time error is now given if \c is not followed by an ASCII
1034 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1035 different, and any byte value is allowed.)
1036
1037 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1038 START_OPTIMIZE option, which is now allowed at compile time - but just
1039 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1040 to pcregrep and other applications that have no direct access to PCRE
1041 options. The new /Y option in pcretest sets this option when calling
1042 pcre_compile().
1043
1044 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1045 back references. Groups containing recursive back references were forced to
1046 be atomic by that change, but in the case of named groups, the amount of
1047 memory required was incorrectly computed, leading to "Failed: internal
1048 error: code overflow". This has been fixed.
1049
1050 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1051 pcretest.c, to avoid build problems in some Borland environments.
1052
1053
1054 Version 8.10 25-Jun-2010
1055 ------------------------
1056
1057 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1058 THEN.
1059
1060 2. (*ACCEPT) was not working when inside an atomic group.
1061
1062 3. Inside a character class, \B is treated as a literal by default, but
1063 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1064 causes the error). The code is unchanged, but I tidied the documentation.
1065
1066 4. Inside a character class, PCRE always treated \R and \X as literals,
1067 whereas Perl faults them if its -w option is set. I have changed PCRE so
1068 that it faults them when PCRE_EXTRA is set.
1069
1070 5. Added support for \N, which always matches any character other than
1071 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1072
1073 6. When compiling pcregrep with newer versions of gcc which may have
1074 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1075 declared with attribute warn_unused_result" were given. Just casting the
1076 result to (void) does not stop the warnings; a more elaborate fudge is
1077 needed. I've used a macro to implement this.
1078
1079 7. Minor change to pcretest.c to avoid a compiler warning.
1080
1081 8. Added four artifical Unicode properties to help with an option to make
1082 \s etc use properties (see next item). The new properties are: Xan
1083 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1084
1085 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1086 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1087 this option. Modified pcretest to add /W to test this facility. Added
1088 REG_UCP to make it available via the POSIX interface.
1089
1090 10. Added --line-buffered to pcregrep.
1091
1092 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1093 studied, and the match started with a letter with a code point greater than
1094 127 whose first byte was different to the first byte of the other case of
1095 the letter, the other case of this starting letter was not recognized
1096 (#976).
1097
1098 12. If a pattern that was studied started with a repeated Unicode property
1099 test, for example, \p{Nd}+, there was the theoretical possibility of
1100 setting up an incorrect bitmap of starting bytes, but fortunately it could
1101 not have actually happened in practice until change 8 above was made (it
1102 added property types that matched character-matching opcodes).
1103
1104 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1105 possible starting bytes for non-anchored patterns.
1106
1107 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1108 \R, and also a number of cases that involve Unicode properties, both
1109 explicit and implicit when PCRE_UCP is set.
1110
1111 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1112 input, it could crash or give wrong results if characters with values
1113 greater than 0xc0 were present in the subject string. (Detail: it assumed
1114 UTF-8 input when processing these items.)
1115
1116 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1117 size_t is 64-bit (#991).
1118
1119 17. Added a check for running out of memory when PCRE is compiled with
1120 --disable-stack-for-recursion (#990).
1121
1122 18. If the last data line in a file for pcretest does not have a newline on
1123 the end, a newline was missing in the output.
1124
1125 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1126 less than 128) in its various bitmaps. However, there is a facility for
1127 generating tables according to the current locale when PCRE is compiled. It
1128 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1129 characters, are recognized by isspace() and therefore were getting set in
1130 these tables, and indeed these tables seem to approximate to ISO 8859. This
1131 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1132 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1133 which of course cannot start UTF-8 characters. I have changed the code so
1134 that only real ASCII characters (less than 128) and the correct starting
1135 bytes for UTF-8 encodings are set for characters greater than 127 when in
1136 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1137 altogether.)
1138
1139 20. Added the /T option to pcretest so as to be able to run tests with non-
1140 standard character tables, thus making it possible to include the tests
1141 used for 19 above in the standard set of tests.
1142
1143 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1144 reference to a subpattern the other side of a comment that contains an
1145 opening parenthesis caused either an internal compiling error, or a
1146 reference to the wrong subpattern.
1147
1148
1149 Version 8.02 19-Mar-2010
1150 ------------------------
1151
1152 1. The Unicode data tables have been updated to Unicode 5.2.0.
1153
1154 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1155 configured.
1156
1157 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1158 original author of that file, following a query about its status.
1159
1160 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1161 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1162
1163 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1164 quantifier applied to a forward-referencing subroutine call, could compile
1165 incorrect code or give the error "internal error: previously-checked
1166 referenced subpattern not found".
1167
1168 6. Both MS Visual Studio and Symbian OS have problems with initializing
1169 variables to point to external functions. For these systems, therefore,
1170 pcre_malloc etc. are now initialized to local functions that call the
1171 relevant global functions.
1172
1173 7. There were two entries missing in the vectors called coptable and poptable
1174 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1175 I've fixed the data, and added a kludgy way of testing at compile time that
1176 the lengths are correct (equal to the number of opcodes).
1177
1178 8. Following on from 7, I added a similar kludge to check the length of the
1179 eint vector in pcreposix.c.
1180
1181 9. Error texts for pcre_compile() are held as one long string to avoid too
1182 much relocation at load time. To find a text, the string is searched,
1183 counting zeros. There was no check for running off the end of the string,
1184 which could happen if a new error number was added without updating the
1185 string.
1186
1187 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1188
1189 11. \K was not working if it appeared in an atomic group or in a group that
1190 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1191 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1192 the assertion is positive, but not if it is negative.
1193
1194 12. Change 11 fortuitously reduced the size of the stack frame used in the
1195 "match()" function of pcre_exec.c by one pointer. Forthcoming
1196 implementation of support for (*MARK) will need an extra pointer on the
1197 stack; I have reserved it now, so that the stack frame size does not
1198 decrease.
1199
1200 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1201 item in branch that calls a recursion is a subroutine call - as in the
1202 second branch in the above example - was incorrectly given the compile-
1203 time error "recursive call could loop indefinitely" because pcre_compile()
1204 was not correctly checking the subroutine for matching a non-empty string.
1205
1206 14. The checks for overrunning compiling workspace could trigger after an
1207 overrun had occurred. This is a "should never occur" error, but it can be
1208 triggered by pathological patterns such as hundreds of nested parentheses.
1209 The checks now trigger 100 bytes before the end of the workspace.
1210
1211 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1212
1213
1214 Version 8.01 19-Jan-2010
1215 ------------------------
1216
1217 1. If a pattern contained a conditional subpattern with only one branch (in
1218 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1219 computed the wrong minimum data length (which is of course zero for such
1220 subpatterns). This could cause incorrect "no match" results.
1221
1222 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1223 the pattern is reset in the first branch, pcre_compile() failed with
1224 "internal error: code overflow at offset...". This happened only when
1225 the reset was to the original external option setting. (An optimization
1226 abstracts leading options settings into an external setting, which was the
1227 cause of this.)
1228
1229 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1230 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1231 assertion pattern did not match (meaning that the assertion was true), it
1232 was incorrectly treated as false if the SKIP had been reached during the
1233 matching. This also applied to assertions used as conditions.
1234
1235 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1236 assertion subpattern, including such a pattern used as a condition,
1237 unpredictable results occurred, instead of the error return
1238 PCRE_ERROR_DFA_UITEM.
1239
1240 5. The C++ GlobalReplace function was not working like Perl for the special
1241 situation when an empty string is matched. It now does the fancy magic
1242 stuff that is necessary.
1243
1244 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1245 removed. (These were left over from very, very early versions of PCRE.)
1246
1247 7. Some cosmetic changes to the code to make life easier when compiling it
1248 as part of something else:
1249
1250 (a) Change DEBUG to PCRE_DEBUG.
1251
1252 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1253 called "current" as "current_branch", to prevent a collision with the
1254 Linux macro when compiled as a kernel module.
1255
1256 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1257 prevent a collision with the Linux macro when compiled as a kernel
1258 module.
1259
1260 8. In pcre_compile() there are some checks for integer overflows that used to
1261 cast potentially large values to (double). This has been changed to that
1262 when building, a check for int64_t is made, and if it is found, it is used
1263 instead, thus avoiding the use of floating point arithmetic. (There is no
1264 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1265 double.
1266
1267 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1268 2005 (difference between two addresses compared to an unsigned value).
1269
1270 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1271 custom one, because of the following reported problem in Windows:
1272
1273 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1274 under Win32.
1275 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1276 therefore missing the function definition.
1277 - The compiler thus generates a "C" signature for the test function.
1278 - The linker fails to find the "C" function.
1279 - PCRE fails to configure if asked to do so against libbz2.
1280
1281 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1282 messages were output:
1283
1284 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1285 rerunning libtoolize, to keep the correct libtool macros in-tree.
1286 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1287
1288 I have done both of these things.
1289
1290 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1291 most of the time, it *can* run out if it is given a pattern that contains a
1292 runaway infinite recursion. I updated the discussion in the pcrestack man
1293 page.
1294
1295 13. Now that we have gone to the x.xx style of version numbers, the minor
1296 version may start with zero. Using 08 or 09 is a bad idea because users
1297 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1298 interpreted as invalid octal numbers. I've updated the previous comment in
1299 configure.ac, and also added a check that gives an error if 08 or 09 are
1300 used.
1301
1302 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1303 causing partial matching to fail when the end of the subject matched \W
1304 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1305
1306 15. There were some discrepancies between the declarations in pcre_internal.h
1307 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1308 their definitions. The declarations used "const uschar *" and the
1309 definitions used USPTR. Even though USPTR is normally defined as "const
1310 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1311 reported that: "This difference in casting confuses some C++ compilers, for
1312 example, SunCC recognizes above declarations as different functions and
1313 generates broken code for hbpcre." I have changed the declarations to use
1314 USPTR.
1315
1316 16. GNU libtool is named differently on some systems. The autogen.sh script now
1317 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1318 (FreeBSD).
1319
1320 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1321 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1322 comment: "Figure out how to create a longlong from a string: strtoll and
1323 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1324 instance, but it only takes 2 args instead of 3!"
1325
1326 18. A subtle bug concerned with back references has been fixed by a change of
1327 specification, with a corresponding code fix. A pattern such as
1328 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1329 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1330 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1331 same bug. Such groups have to be quantified to be useful, or contained
1332 inside another quantified group. (If there's no repetition, the reference
1333 can never match.) The problem arises because, having left the group and
1334 moved on to the rest of the pattern, a later failure that backtracks into
1335 the group uses the captured value from the final iteration of the group
1336 rather than the correct earlier one. I have fixed this in PCRE by forcing
1337 any group that contains a reference to itself to be an atomic group; that
1338 is, there cannot be any backtracking into it once it has completed. This is
1339 similar to recursive and subroutine calls.
1340
1341
1342 Version 8.00 19-Oct-09
1343 ----------------------
1344
1345 1. The table for translating pcre_compile() error codes into POSIX error codes
1346 was out-of-date, and there was no check on the pcre_compile() error code
1347 being within the table. This could lead to an OK return being given in
1348 error.
1349
1350 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1351 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1352 in a Windows environment.
1353
1354 3. The pcregrep --count option prints the count for each file even when it is
1355 zero, as does GNU grep. However, pcregrep was also printing all files when
1356 --files-with-matches was added. Now, when both options are given, it prints
1357 counts only for those files that have at least one match. (GNU grep just
1358 prints the file name in this circumstance, but including the count seems
1359 more useful - otherwise, why use --count?) Also ensured that the
1360 combination -clh just lists non-zero counts, with no names.
1361
1362 4. The long form of the pcregrep -F option was incorrectly implemented as
1363 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1364 but it seems right to fix it, and I didn't think it was worth preserving
1365 the old behaviour.
1366
1367 5. The command line items --regex=pattern and --regexp=pattern were not
1368 recognized by pcregrep, which required --regex pattern or --regexp pattern
1369 (with a space rather than an '='). The man page documented the '=' forms,
1370 which are compatible with GNU grep; these now work.
1371
1372 6. No libpcreposix.pc file was created for pkg-config; there was just
1373 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1374
1375 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1376 when UCP support is not needed, by modifying the Python script that
1377 generates it from Unicode data files. This should not matter if the module
1378 is correctly used as a library, but I received one complaint about 50K of
1379 unwanted data. My guess is that the person linked everything into his
1380 program rather than using a library. Anyway, it does no harm.
1381
1382 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1383 was a minimum greater than 1 for a wide character in a possessive
1384 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1385 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1386 character. Chaos in the form of incorrect output or a compiling loop could
1387 result.
1388
1389 9. The restrictions on what a pattern can contain when partial matching is
1390 requested for pcre_exec() have been removed. All patterns can now be
1391 partially matched by this function. In addition, if there are at least two
1392 slots in the offset vector, the offset of the earliest inspected character
1393 for the match and the offset of the end of the subject are set in them when
1394 PCRE_ERROR_PARTIAL is returned.
1395
1396 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1397 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1398 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1399 and may be more useful for multi-segment matching.
1400
1401 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1402 used to be given if ever the end of the subject was reached; now it is
1403 given only if matching could not proceed because another character was
1404 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1405 string "Z", which now yields "no match" instead of "partial match". In the
1406 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1407 final character ended with (*FAIL).
1408
1409 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1410 if the pattern had a "must contain" character that was already found in the
1411 earlier partial match, unless partial matching was again requested. For
1412 example, with the pattern /dog.(body)?/, the "must contain" character is
1413 "g". If the first part-match was for the string "dog", restarting with
1414 "sbody" failed. This bug has been fixed.
1415
1416 13. The string returned by pcre_dfa_exec() after a partial match has been
1417 changed so that it starts at the first inspected character rather than the
1418 first character of the match. This makes a difference only if the pattern
1419 starts with a lookbehind assertion or \b or \B (\K is not supported by
1420 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1421 matching functions compatible, and I think it's the right thing to do.
1422
1423 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1424 so that the demonstration program is easily available in environments where
1425 PCRE has not been installed from source.
1426
1427 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1428 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1429 library.
1430
1431 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1432 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1433 is not the first non-POSIX option to be added. Clearly some people find
1434 these options useful.
1435
1436 17. If a caller to the POSIX matching function regexec() passes a non-zero
1437 value for nmatch with a NULL value for pmatch, the value of
1438 nmatch is forced to zero.
1439
1440 18. RunGrepTest did not have a test for the availability of the -u option of
1441 the diff command, as RunTest does. It now checks in the same way as
1442 RunTest, and also checks for the -b option.
1443
1444 19. If an odd number of negated classes containing just a single character
1445 interposed, within parentheses, between a forward reference to a named
1446 subpattern and the definition of the subpattern, compilation crashed with
1447 an internal error, complaining that it could not find the referenced
1448 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1449 [The bug was that it was starting one character too far in when skipping
1450 over the character class, thus treating the ] as data rather than
1451 terminating the class. This meant it could skip too much.]
1452
1453 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1454 /g option in pcretest when the pattern contains \K, which makes it possible
1455 to have an empty string match not at the start, even when the pattern is
1456 anchored. Updated pcretest and pcredemo to use this option.
1457
1458 21. If the maximum number of capturing subpatterns in a recursion was greater
1459 than the maximum at the outer level, the higher number was returned, but
1460 with unset values at the outer level. The correct (outer level) value is
1461 now given.
1462
1463 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1464 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1465 make it do so. The string so far is captured, making this feature
1466 compatible with Perl.
1467
1468 23. The tests have been re-organized, adding tests 11 and 12, to make it
1469 possible to check the Perl 5.10 features against Perl 5.10.
1470
1471 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1472 pattern matches a fixed length string. PCRE did not allow this; now it
1473 does. Neither allows recursion.
1474
1475 25. I finally figured out how to implement a request to provide the minimum
1476 length of subject string that was needed in order to match a given pattern.
1477 (It was back references and recursion that I had previously got hung up
1478 on.) This code has now been added to pcre_study(); it finds a lower bound
1479 to the length of subject needed. It is not necessarily the greatest lower
1480 bound, but using it to avoid searching strings that are too short does give
1481 some useful speed-ups. The value is available to calling programs via
1482 pcre_fullinfo().
1483
1484 26. While implementing 25, I discovered to my embarrassment that pcretest had
1485 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1486 study optimizations had never been tested with that matching function.
1487 Oops. What is worse, even when it was passed study data, there was a bug in
1488 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1489 were also very few tests of studied patterns with pcre_dfa_exec().
1490
1491 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1492 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1493 on the other side of the coin, they are no longer allowed to have different
1494 names, because these cannot be distinguished in PCRE, and this has caused
1495 confusion. (This is a difference from Perl.)
1496
1497 28. When duplicate subpattern names are present (necessarily with different
1498 numbers, as required by 27 above), and a test is made by name in a
1499 conditional pattern, either for a subpattern having been matched, or for
1500 recursion in such a pattern, all the associated numbered subpatterns are
1501 tested, and the overall condition is true if the condition is true for any
1502 one of them. This is the way Perl works, and is also more like the way
1503 testing by number works.
1504
1505
1506 Version 7.9 11-Apr-09
1507 ---------------------
1508
1509 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1510 (pcretest), all targets were linked against these libraries. This included
1511 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1512 libraries. This caused unwanted dependencies to be created. This problem
1513 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1514 pcretest is linked with readline.
1515
1516 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1517 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1518 moved outside it again, because FALSE and TRUE are already defined in AIX,
1519 but BOOL is not.
1520
1521 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1522 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1523
1524 4. The pcregrep documentation said spaces were inserted as well as colons (or
1525 hyphens) following file names and line numbers when outputting matching
1526 lines. This is not true; no spaces are inserted. I have also clarified the
1527 wording for the --colour (or --color) option.
1528
1529 5. In pcregrep, when --colour was used with -o, the list of matching strings
1530 was not coloured; this is different to GNU grep, so I have changed it to be
1531 the same.
1532
1533 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1534 each matching line was coloured. Now it goes on to look for further matches
1535 of any of the test patterns, which is the same behaviour as GNU grep.
1536
1537 7. A pattern that could match an empty string could cause pcregrep to loop; it
1538 doesn't make sense to accept an empty string match in pcregrep, so I have
1539 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1540 seems to be how GNU grep behaves.
1541
1542 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1543 start or after a newline", because the conditional assertion was not being
1544 correctly handled. The rule now is that both the assertion and what follows
1545 in the first alternative must satisfy the test.
1546
1547 9. If auto-callout was enabled in a pattern with a conditional group whose
1548 condition was an assertion, PCRE could crash during matching, both with
1549 pcre_exec() and pcre_dfa_exec().
1550
1551 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1552 used for matching.
1553
1554 11. Unicode property support in character classes was not working for
1555 characters (bytes) greater than 127 when not in UTF-8 mode.
1556
1557 12. Added the -M command line option to pcretest.
1558
1559 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1560
1561 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1562
1563 16. Added comments and documentation about mis-use of no_arg in the C++
1564 wrapper.
1565
1566 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1567 from Martin Jerabek that uses macro names for all relevant character and
1568 string constants.
1569
1570 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1571 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1572 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1573 these, but not everybody uses configure.
1574
1575 19. A conditional group that had only one branch was not being correctly
1576 recognized as an item that could match an empty string. This meant that an
1577 enclosing group might also not be so recognized, causing infinite looping
1578 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1579 with the subject "ab", where knowledge that the repeated group can match
1580 nothing is needed in order to break the loop.
1581
1582 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1583 exec(), but without supplying a callout function, matching went wrong.
1584
1585 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1586 leak if the size of the offset vector was greater than 30. When the vector
1587 is smaller, the saved offsets during recursion go onto a local stack
1588 vector, but for larger vectors malloc() is used. It was failing to free
1589 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1590 error, in fact).
1591
1592 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1593 heapframe that is used only when UTF-8 support is enabled. This caused no
1594 problem, but was untidy.
1595
1596 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1597 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1598 included within another project.
1599
1600 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1601 slightly modified by me:
1602
1603 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1604 not building pcregrep.
1605
1606 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1607 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1608
1609 25. Forward references, both numeric and by name, in patterns that made use of
1610 duplicate group numbers, could behave incorrectly or give incorrect errors,
1611 because when scanning forward to find the reference group, PCRE was not
1612 taking into account the duplicate group numbers. A pattern such as
1613 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1614
1615 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1616 the feature of a custom pointer more persuasive (as requested by a user).
1617
1618 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1619 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1620 pre-defined.
1621
1622 28. Added support for (*UTF8) at the start of a pattern.
1623
1624 29. Arrange for flags added by the "release type" setting in CMake to be shown
1625 in the configuration summary.
1626
1627
1628 Version 7.8 05-Sep-08
1629 ---------------------
1630
1631 1. Replaced UCP searching code with optimized version as implemented for Ad
1632 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1633 stage table and inline lookup instead of a function, giving speed ups of 2
1634 to 5 times on some simple patterns that I tested. Permission was given to
1635 distribute the MultiStage2.py script that generates the tables (it's not in
1636 the tarball, but is in the Subversion repository).
1637
1638 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1639 scripts.
1640
1641 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1642 a group with a zero qualifier. The result of the study could be incorrect,
1643 or the function might crash, depending on the pattern.
1644
1645 4. Caseless matching was not working for non-ASCII characters in back
1646 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1647 It now works when Unicode Property Support is available.
1648
1649 5. In pcretest, an escape such as \x{de} in the data was always generating
1650 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1651 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1652 truncation.
1653
1654 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1655
1656 7. Added two (int) casts to pcregrep when printing the difference of two
1657 pointers, in case they are 64-bit values.
1658
1659 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1660 test 2 if it fails.
1661
1662 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1663 and a #define of that name to empty if it is not externally set. This is to
1664 allow users of MSVC to set it if necessary.
1665
1666 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1667 the convenience functions in the pcre_get.c source file.
1668
1669 11. An option change at the start of a pattern that had top-level alternatives
1670 could cause overwriting and/or a crash. This command provoked a crash in
1671 some environments:
1672
1673 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1674
1675 This potential security problem was recorded as CVE-2008-2371.
1676
1677 12. For a pattern where the match had to start at the beginning or immediately
1678 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1679 pcre_dfa_exec() could read past the end of the passed subject if there was
1680 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1681 pcretest so that it places the subject at the end of its malloc-ed buffer.
1682
1683 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1684 exec() might read past the end of the data buffer in UTF-8 mode.
1685
1686 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1687 the data contained the byte 0x85 as part of a UTF-8 character within its
1688 first line. This applied both to normal and DFA matching.
1689
1690 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1691 /^[^d]*?$/8 failed to match "abc".
1692
1693 16. Added a missing copyright notice to pcrecpp_internal.h.
1694
1695 17. Make it more clear in the documentation that values returned from
1696 pcre_exec() in ovector are byte offsets, not character counts.
1697
1698 18. Tidied a few places to stop certain compilers from issuing warnings.
1699
1700 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1701 supplied by Stefan Weber. I made a further small update for 7.8 because
1702 there is a change of source arrangements: the pcre_searchfuncs.c module is
1703 replaced by pcre_ucd.c.
1704
1705
1706 Version 7.7 07-May-08
1707 ---------------------
1708
1709 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1710 a string to a long long, pretend we don't even have a long long." This is
1711 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1712
1713 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1714 pre-7.6 versions, which defined a global no_arg variable instead of putting
1715 it in the RE class. (See also #8 below.)
1716
1717 3. Remove a line of dead code, identified by coverity and reported by Nuno
1718 Lopes.
1719
1720 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1721
1722 (1) The include/exclude patterns were being applied to the whole pathnames
1723 of files, instead of just to the final components.
1724
1725 (2) If there was more than one level of directory, the subdirectories were
1726 skipped unless they satisfied the include/exclude conditions. This is
1727 inconsistent with GNU grep (and could even be seen as contrary to the
1728 pcregrep specification - which I improved to make it absolutely clear).
1729 The action now is always to scan all levels of directory, and just
1730 apply the include/exclude patterns to regular files.
1731
1732 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1733 --exclude_dir in the tests to avoid scanning .svn directories.
1734
1735 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1736 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1737 doesn't support NULs in patterns.
1738
1739 7. Added some missing "const"s to declarations of static tables in
1740 pcre_compile.c and pcre_dfa_exec.c.
1741
1742 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1743 caused by fix #2 above. (Subsequently also a second patch to fix the
1744 first patch. And a third patch - this was a messy problem.)
1745
1746 9. Applied Craig's patch to remove the use of push_back().
1747
1748 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1749 matching function regexec().
1750
1751 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1752 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1753 references. PCRE supports relative numbers with this syntax (I don't think
1754 Oniguruma does).
1755
1756 12. Previously, a group with a zero repeat such as (...){0} was completely
1757 omitted from the compiled regex. However, this means that if the group
1758 was called as a subroutine from elsewhere in the pattern, things went wrong
1759 (an internal error was given). Such groups are now left in the compiled
1760 pattern, with a new opcode that causes them to be skipped at execution
1761 time.
1762
1763 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1764 to the way PCRE behaves:
1765
1766 (a) A lone ] character is dis-allowed (Perl treats it as data).
1767
1768 (b) A back reference to an unmatched subpattern matches an empty string
1769 (Perl fails the current match path).
1770
1771 (c) A data ] in a character class must be notated as \] because if the
1772 first data character in a class is ], it defines an empty class. (In
1773 Perl it is not possible to have an empty class.) The empty class []
1774 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1775 The negative empty class [^] matches any one character, independently
1776 of the DOTALL setting.
1777
1778 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1779 non-existent subpattern following a character class starting with ']' and
1780 containing () gave an internal compiling error instead of "reference to
1781 non-existent subpattern". Fortunately, when the pattern did exist, the
1782 compiled code was correct. (When scanning forwards to check for the
1783 existencd of the subpattern, it was treating the data ']' as terminating
1784 the class, so got the count wrong. When actually compiling, the reference
1785 was subsequently set up correctly.)
1786
1787 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1788 it was being rejected as not supported by pcre_dfa_exec(), even though
1789 other assertions are supported. I have made pcre_dfa_exec() support
1790 (*FAIL).
1791
1792 16. The implementation of 13c above involved the invention of a new opcode,
1793 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1794 cannot be changed at match time, I realized I could make a small
1795 improvement to matching performance by compiling OP_ALLANY instead of
1796 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1797 on the OP_ANY path.
1798
1799 17. Compiling pcretest on Windows with readline support failed without the
1800 following two fixes: (1) Make the unistd.h include conditional on
1801 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1802
1803 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1804 ncurses library to be included for pcretest when ReadLine support is
1805 requested, but also to allow for it to be overridden. This patch came from
1806 Daniel Bergström.
1807
1808 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1809 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1810 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1811 spotting this.
1812
1813
1814 Version 7.6 28-Jan-08
1815 ---------------------
1816
1817 1. A character class containing a very large number of characters with
1818 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1819 overflow.
1820
1821 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1822 HAVE_LONG_LONG is not defined.
1823
1824 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1825 bring them up to date and include new features. This patch includes:
1826
1827 - Fixed PH's badly added libz and libbz2 support.
1828 - Fixed a problem with static linking.
1829 - Added pcredemo. [But later removed - see 7 below.]
1830 - Fixed dftables problem and added an option.
1831 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1832 HAVE_LONG_LONG.
1833 - Added readline support for pcretest.
1834 - Added an listing of the option settings after cmake has run.
1835
1836 4. A user submitted a patch to Makefile that makes it easy to create
1837 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1838 Makefile.am that cause it to include this special target, without
1839 affecting anything else. Note that the same mingw target plus all
1840 the other distribution libraries and programs are now supported
1841 when configuring with CMake (see 6 below) instead of with
1842 Configure/Make.
1843
1844 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1845 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1846 exported in the Windows port". It has not yet been confirmed that the patch
1847 solves the problem, but it does no harm.
1848
1849 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1850 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1851 with CMake, and also correct the comment about stack recursion.
1852
1853 7. Remove the automatic building of pcredemo from the ./configure system and
1854 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1855 of a program that users should build themselves after PCRE is installed, so
1856 building it automatically is not really right. What is more, it gave
1857 trouble in some build environments.
1858
1859 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1860
1861
1862 Version 7.5 10-Jan-08
1863 ---------------------
1864
1865 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1866 values in parens when parsing an RE using the C++ wrapper."
1867
1868 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1869 Characters greater than 255 were excluded from the class instead of being
1870 included.
1871
1872 3. The same bug as (2) above applied to negated POSIX classes such as
1873 [:^space:].
1874
1875 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1876 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1877 I have changed it.
1878
1879 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1880 first named subpattern) and a construct such as (?&a) would reference the
1881 first named subpattern whose name started with "a" (in other words, the
1882 length check was missing). Both these problems are fixed. "Subpattern name
1883 expected" is now given for (?&) (a zero-length name), and this patch also
1884 makes it give the same error for \k'' (previously it complained that that
1885 was a reference to a non-existent subpattern).
1886
1887 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1888 this is right because (?- can be followed by option settings as well as by
1889 digits. I have, however, made the messages clearer.
1890
1891 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1892 than the number used in the conditional) now cause a compile-time error.
1893 This is actually not compatible with Perl, which accepts such patterns, but
1894 treats the conditional as always being FALSE (as PCRE used to), but it
1895 seems to me that giving a diagnostic is better.
1896
1897 8. Change "alphameric" to the more common word "alphanumeric" in comments
1898 and messages.
1899
1900 9. Fix two occurrences of "backslash" in comments that should have been
1901 "backspace".
1902
1903 10. Remove two redundant lines of code that can never be obeyed (their function
1904 was moved elsewhere).
1905
1906 11. The program that makes PCRE's Unicode character property table had a bug
1907 which caused it to generate incorrect table entries for sequences of
1908 characters that have the same character type, but are in different scripts.
1909 It amalgamated them into a single range, with the script of the first of
1910 them. In other words, some characters were in the wrong script. There were
1911 thirteen such cases, affecting characters in the following ranges:
1912
1913 U+002b0 - U+002c1
1914 U+0060c - U+0060d
1915 U+0061e - U+00612
1916 U+0064b - U+0065e
1917 U+0074d - U+0076d
1918 U+01800 - U+01805
1919 U+01d00 - U+01d77
1920 U+01d9b - U+01dbf
1921 U+0200b - U+0200f
1922 U+030fc - U+030fe
1923 U+03260 - U+0327f
1924 U+0fb46 - U+0fbb1
1925 U+10450 - U+1049d
1926
1927 12. The -o option (show only the matching part of a line) for pcregrep was not
1928 compatible with GNU grep in that, if there was more than one match in a
1929 line, it showed only the first of them. It now behaves in the same way as
1930 GNU grep.
1931
1932 13. If the -o and -v options were combined for pcregrep, it printed a blank
1933 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1934 does the same. The return code can be used to tell if there were any
1935 non-matching lines.
1936
1937 14. Added --file-offsets and --line-offsets to pcregrep.
1938
1939 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1940 infinitely looping recursion. The bug was that positive lookaheads were not
1941 being skipped when checking for a possible empty match (negative lookaheads
1942 and both kinds of lookbehind were skipped).
1943
1944 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1945 inclusion of <windows.h> to before rather than after the definition of
1946 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1947
1948 17. Specifying a possessive quantifier with a specific limit for a Unicode
1949 character property caused pcre_compile() to compile bad code, which led at
1950 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1951 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1952 caused the error; without that there was no problem.
1953
1954 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1955
1956 19. Added --enable-pcretest-libreadline.
1957
1958 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1959 RE::GlobalReplace(). As a result, the number of replacements returned was
1960 double what it should be. I removed one of the increments, but Craig sent a
1961 later patch that removed the other one (the right fix) and added unit tests
1962 that check the return values (which was not done before).
1963
1964 21. Several CMake things:
1965
1966 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1967 the names libpcre and libpcreposix, not just pcre and pcreposix.
1968
1969 (2) The above change means that pcretest and pcregrep are now correctly
1970 linked with the newly-built libraries, not previously installed ones.
1971
1972 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1973
1974 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1975 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1976 UTF-8 newline character). The key issue is that the pattern starts .*;
1977 this means that the match must be either at the beginning, or after a
1978 newline. The bug was in the code for advancing after a failed match and
1979 checking that the new position followed a newline. It was not taking
1980 account of UTF-8 characters correctly.
1981
1982 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1983 character classes. PCRE was not treating the sequence [:...:] as a
1984 character class unless the ... were all letters. Perl, however, seems to
1985 allow any characters between [: and :], though of course it rejects as
1986 unknown any "names" that contain non-letters, because all the known class
1987 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1988 for example, whereas PCRE did not - it did not recognize a POSIX character
1989 class. This seemed a bit dangerous, so the code has been changed to be
1990 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1991 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1992 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1993 Perl does, and where it didn't before.
1994
1995 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1996 Windows environments %n is disabled by default.
1997
1998
1999 Version 7.4 21-Sep-07
2000 ---------------------
2001
2002 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2003 means that a class such as [\s] counted as "explicit reference to CR or
2004 LF". That isn't really right - the whole point of the change was to try to
2005 help when there was an actual mention of one of the two characters. So now
2006 the change happens only if \r or \n (or a literal CR or LF) character is
2007 encountered.
2008
2009 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2010 of both had grown to the point where there were only 3 bits left.
2011 Fortunately, there was spare space in the data structure, and so I have
2012 moved the internal flags into a new 16-bit field to free up more option
2013 bits.
2014
2015 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2016 but did not set the internal JCHANGED flag - either of these is enough to
2017 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2018 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2019 start sets both bits.
2020
2021 4. Added options (at build time, compile time, exec time) to change \R from
2022 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2023
2024 5. doc/pcresyntax.html was missing from the distribution.
2025
2026 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2027 compatibility, even though it is no longer used.
2028
2029 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2030 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2031 windows.h file is present (where different names are used). [This was
2032 reversed later after testing - see 16 below.]
2033
2034 8. Changed all #include <config.h> to #include "config.h". There were also
2035 some further <pcre.h> cases that I changed to "pcre.h".
2036
2037 9. When pcregrep was used with the --colour option, it missed the line ending
2038 sequence off the lines that it output.
2039
2040 10. It was pointed out to me that arrays of string pointers cause lots of
2041 relocations when a shared library is dynamically loaded. A technique of
2042 using a single long string with a table of offsets can drastically reduce
2043 these. I have refactored PCRE in four places to do this. The result is
2044 dramatic:
2045
2046 Originally: 290
2047 After changing UCP table: 187
2048 After changing error message table: 43
2049 After changing table of "verbs" 36
2050 After changing table of Posix names 22
2051
2052 Thanks to the folks working on Gregex for glib for this insight.
2053
2054 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2055 unicode-properties was also set.
2056
2057 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2058
2059 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2060 checked only for CRLF.
2061
2062 14. Added casts to pcretest.c to avoid compiler warnings.
2063
2064 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2065
2066 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2067 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2068 entirely. This removes changes made in 7 above.
2069
2070 17. The CMake files have been updated, and there is now more information about
2071 building with CMake in the NON-UNIX-USE document.
2072
2073
2074 Version 7.3 28-Aug-07
2075 ---------------------
2076
2077 1. In the rejigging of the build system that eventually resulted in 7.1, the
2078 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2079 brackets there is not right, since it causes compilers to look for an
2080 installed pcre.h, not the version that is in the source that is being
2081 compiled (which of course may be different). I have changed it back to:
2082
2083 #include "pcre.h"
2084
2085 I have a vague recollection that the change was concerned with compiling in
2086 different directories, but in the new build system, that is taken care of
2087 by the VPATH setting the Makefile.
2088
2089 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2090 when the subject happened to end in the byte 0x85 (e.g. if the last
2091 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2092 characters but of course it shouldn't be taken as a newline when it is part
2093 of another character. The bug was that, for an unlimited repeat of . in
2094 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2095 characters when looking for a newline.
2096
2097 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2098
2099 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2100 in debug output.
2101
2102 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2103 long printing in the pcrecpp unittest when running under MinGW.
2104
2105 6. ESC_K was left out of the EBCDIC table.
2106
2107 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2108 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2109 limit also applies to "virtual nesting" when a pattern is recursive, and in
2110 this case 1000 isn't so big. I have been able to remove this limit at the
2111 expense of backing off one optimization in certain circumstances. Normally,
2112 when pcre_exec() would call its internal match() function recursively and
2113 immediately return the result unconditionally, it uses a "tail recursion"
2114 feature to save stack. However, when a subpattern that can match an empty
2115 string has an unlimited repetition quantifier, it no longer makes this
2116 optimization. That gives it a stack frame in which to save the data for
2117 checking that an empty string has been matched. Previously this was taken
2118 from the 1000-entry workspace that had been reserved. So now there is no
2119 explicit limit, but more stack is used.
2120
2121 8. Applied Daniel's patches to solve problems with the import/export magic
2122 syntax that is required for Windows, and which was going wrong for the
2123 pcreposix and pcrecpp parts of the library. These were overlooked when this
2124 problem was solved for the main library.
2125
2126 9. There were some crude static tests to avoid integer overflow when computing
2127 the size of patterns that contain repeated groups with explicit upper
2128 limits. As the maximum quantifier is 65535, the maximum group length was
2129 set at 30,000 so that the product of these two numbers did not overflow a
2130 32-bit integer. However, it turns out that people want to use groups that
2131 are longer than 30,000 bytes (though not repeat them that many times).
2132 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2133 made it possible to implement the integer overflow checks in a much more
2134 dynamic way, which I have now done. The artificial limitation on group
2135 length has been removed - we now have only the limit on the total length of
2136 the compiled pattern, which depends on the LINK_SIZE setting.
2137
2138 10. Fixed a bug in the documentation for get/copy named substring when
2139 duplicate names are permitted. If none of the named substrings are set, the
2140 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2141 empty string.
2142
2143 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2144 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2145 because the ] is interpreted as the first data character and the
2146 terminating ] is not found. PCRE has been made compatible with Perl in this
2147 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2148 cause memory overwriting.
2149
2150 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2151 string has been matched (to stop an infinite loop). It was not recognizing
2152 a conditional subpattern that could match an empty string if that
2153 subpattern was within another subpattern. For example, it looped when
2154 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2155 condition was not nested. This bug has been fixed.
2156
2157 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2158 past the start of the subject in the presence of bytes with the top bit
2159 set, for example "\x8aBCD".
2160
2161 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2162 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2163
2164 14. Optimized (?!) to (*FAIL).
2165
2166 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2167 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2168 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2169 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2170 does: it's just the validity check that is more restrictive.
2171
2172 16. Inserted checks for integer overflows during escape sequence (backslash)
2173 processing, and also fixed erroneous offset values for syntax errors during
2174 backslash processing.
2175
2176 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2177 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2178
2179 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2180 caused an overrun.
2181
2182 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2183 something other than just ASCII characters) inside a group that had an
2184 unlimited repeat caused a loop at compile time (while checking to see
2185 whether the group could match an empty string).
2186
2187 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2188 [\P{Any}] did so. (Error in the code for printing property names.)
2189
2190 21. An orphan \E inside a character class could cause a crash.
2191
2192 22. A repeated capturing bracket such as (A)? could cause a wild memory
2193 reference during compilation.
2194
2195 23. There are several functions in pcre_compile() that scan along a compiled
2196 expression for various reasons (e.g. to see if it's fixed length for look
2197 behind). There were bugs in these functions when a repeated \p or \P was
2198 present in the pattern. These operators have additional parameters compared
2199 with \d, etc, and these were not being taken into account when moving along
2200 the compiled data. Specifically:
2201
2202 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2203 length.
2204
2205 (b) An item such as \pL+ within a repeated group could cause crashes or
2206 loops.
2207
2208 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2209 "reference to non-existent subpattern" error.
2210
2211 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2212
2213 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2214 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2215
2216 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2217
2218 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2219 character were causing crashes (broken optimization).
2220
2221 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2222 \p or \P) caused a compile-time loop.
2223
2224 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2225 break. For example, the unstudied pattern [\r\n]A does not match the string
2226 "\r\nA" because change 7.0/46 below moves the current point on by two
2227 characters after failing to match at the start. However, the pattern \nA
2228 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2229 the same is true. There doesn't seem any very clean way out of this, but
2230 what I have chosen to do makes the common cases work: PCRE now takes note
2231 of whether there can be an explicit match for \r or \n anywhere in the
2232 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2233 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2234 pattern has explicit CR or LF references.
2235
2236 29. Added (*CR) etc for changing newline setting at start of pattern.
2237
2238
2239 Version 7.2 19-Jun-07
2240 ---------------------
2241
2242 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2243 which is apparently normally available under Windows.
2244
2245 2. Re-jig the pcregrep tests with different newline settings in an attempt
2246 to make them independent of the local environment's newline setting.
2247
2248 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2249
2250 4. Some of the "internals" tests were previously cut out when the link size
2251 was not 2, because the output contained actual offsets. The recent new
2252 "Z" feature of pcretest means that these can be cut out, making the tests
2253 usable with all link sizes.
2254
2255 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2256 stack recursion. This gives a massive performance boost under BSD, but just
2257 a small improvement under Linux. However, it saves one field in the frame
2258 in all cases.
2259
2260 6. Added more features from the forthcoming Perl 5.10:
2261
2262 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2263 recursion call. It refers to the nth most recently opened parentheses.
2264
2265 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2266 to be opened parentheses.
2267
2268 (c) Conditions that refer to capturing parentheses can be specified
2269 relatively, for example, (?(-2)... or (?(+3)...
2270
2271 (d) \K resets the start of the current match so that everything before
2272 is not part of it.
2273
2274 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2275
2276 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2277 reference syntax.
2278
2279 (g) (?| introduces a group in which the numbering of parentheses in each
2280 alternative starts with the same number.
2281
2282 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2283
2284 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2285 PCRE_INFO_JCHANGED.
2286
2287 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2288 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2289 for detecting groups that can match an empty string.
2290
2291 9. A pattern with a very large number of alternatives (more than several
2292 hundred) was running out of internal workspace during the pre-compile
2293 phase, where pcre_compile() figures out how much memory will be needed. A
2294 bit of new cunning has reduced the workspace needed for groups with
2295 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2296 workspace instead of running out of the 4096 that are available.
2297
2298 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2299
2300 11. Applied patch from Google to remove an optimization that didn't quite work.
2301 The report of the bug said:
2302
2303 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2304 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2305 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2306
2307 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2308 it matched the wrong number of bytes.
2309
2310
2311 Version 7.1 24-Apr-07
2312 ---------------------
2313
2314 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2315 that is more "standard", making use of automake and other Autotools. There
2316 is some re-arrangement of the files and adjustment of comments consequent
2317 on this.
2318
2319 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2320 for recursive directory scanning broke on some systems because the files
2321 are not scanned in any specific order and on different systems the order
2322 was different. A call to "sort" has been inserted into RunGrepTest for the
2323 approprate test as a short-term fix. In the longer term there may be an
2324 alternative.
2325
2326 3. I had an email from Eric Raymond about problems translating some of PCRE's
2327 man pages to HTML (despite the fact that I distribute HTML pages, some
2328 people do their own conversions for various reasons). The problems
2329 concerned the use of low-level troff macros .br and .in. I have therefore
2330 removed all such uses from the man pages (some were redundant, some could
2331 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2332 HTML has been updated to handle .nf/.fi and to complain if it encounters
2333 .br or .in.
2334
2335 4. Updated comments in configure.ac that get placed in config.h.in and also
2336 arranged for config.h to be included in the distribution, with the name
2337 config.h.generic, for the benefit of those who have to compile without
2338 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2339
2340 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2341 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2342 makevp.bat for the current PCRE, using the additional files
2343 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2344
2345 6. A Windows user reported a minor discrepancy with test 2, which turned out
2346 to be caused by a trailing space on an input line that had got lost in his
2347 copy. The trailing space was an accident, so I've just removed it.
2348
2349 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2350 that is needed.
2351
2352 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2353 as "const" (a) because they are and (b) because it helps the PHP
2354 maintainers who have recently made a script to detect big data structures
2355 in the php code that should be moved to the .rodata section. I remembered
2356 to update Builducptable as well, so it won't revert if ucptable.h is ever
2357 re-created.
2358
2359 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2360 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2361 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2362 support is not required. This saves 1.5-2K of code, which is important in
2363 some applications.
2364
2365 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2366 so as not to refer to the tables, even though these functions will never be
2367 called when UTF-8 support is disabled. Otherwise there are problems with a
2368 shared library.
2369
2370 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2371
2372 (a) It was defining its arguments as char * instead of void *.
2373
2374 (b) It was assuming that all moves were upwards in memory; this was true
2375 a long time ago when I wrote it, but is no longer the case.
2376
2377 The emulated memove() is provided for those environments that have neither
2378 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2379 is clearly not the case, as these two bugs were recently reported.
2380
2381 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2382 and Detrail to create the HTML documentation, the .txt form of the man
2383 pages, and it removes trailing spaces from listed files. It also creates
2384 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2385 case, it wraps all the #defines with #ifndefs. This script should be run
2386 before "make dist".
2387
2388 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2389 with Unicode property support.
2390
2391 (a) For a maximizing quantifier, if the two different cases of the
2392 character were of different lengths in their UTF-8 codings (there are
2393 some cases like this - I found 11), and the matching function had to
2394 back up over a mixture of the two cases, it incorrectly assumed they
2395 were both the same length.
2396
2397 (b) When PCRE was configured to use the heap rather than the stack for
2398 recursion during matching, it was not correctly preserving the data for
2399 the other case of a UTF-8 character when checking ahead for a match
2400 while processing a minimizing repeat. If the check also involved
2401 matching a wide character, but failed, corruption could cause an
2402 erroneous result when trying to check for a repeat of the original
2403 character.
2404
2405 13. Some tidying changes to the testing mechanism:
2406
2407 (a) The RunTest script now detects the internal link size and whether there
2408 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2409 values substituted by "configure". (The RunGrepTest script already did
2410 this for UTF-8.) The configure.ac script no longer substitutes the
2411 relevant variables.
2412
2413 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2414 with length and offset values. This means that the output is different
2415 for different internal link sizes. Test 2 is skipped for link sizes
2416 other than 2 because of this, bypassing the problem. Unfortunately,
2417 there was also a test in test 3 (the locale tests) that used /B and
2418 failed for link sizes other than 2. Rather than cut the whole test out,
2419 I have added a new /Z option to pcretest that replaces the length and
2420 offset values with spaces. This is now used to make test 3 independent
2421 of link size. (Test 2 will be tidied up later.)
2422
2423 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2424 segmentation fault instead of returning the appropriate error message.
2425
2426 15. In multiline mode when the newline sequence was set to "any", the pattern
2427 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2428 This doesn't seem right; it now treats the CRLF combination as the line
2429 ending, and so does not match in that case. It's only a pattern such as ^$
2430 that would hit this one: something like ^ABC$ would have failed after \r
2431 and then tried again after \r\n.
2432
2433 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2434 in an attempt to make files that differ only in their line terminators
2435 compare equal. This works on Linux.
2436
2437 17. Under certain error circumstances pcregrep might try to free random memory
2438 as it exited. This is now fixed, thanks to valgrind.
2439
2440 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2441 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2442 was because its rules for how to advance for /g after matching an empty
2443 string at the end of a line did not allow for this case. They now check for
2444 it specially.
2445
2446 20. pcretest is supposed to handle patterns and data of any length, by
2447 extending its buffers when necessary. It was getting this wrong when the
2448 buffer for a data line had to be extended.
2449
2450 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2451 CRLF as a newline sequence.
2452
2453 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2454 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2455 I have nevertheless tidied it up.
2456
2457 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2458
2459 24. Added a man page for pcre-config.
2460
2461
2462 Version 7.0 19-Dec-06
2463 ---------------------
2464
2465 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2466 moving to gcc 4.1.1.
2467
2468 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2469 sys/time.h, which is documented as needed for this function. It doesn't
2470 seem to matter on Linux, but it showed up on some releases of OS X.
2471
2472 3. It seems that there are systems where bytes whose values are greater than
2473 127 match isprint() in the "C" locale. The "C" locale should be the
2474 default when a C program starts up. In most systems, only ASCII printing
2475 characters match isprint(). This difference caused the output from pcretest
2476 to vary, making some of the tests fail. I have changed pcretest so that:
2477
2478 (a) When it is outputting text in the compiled version of a pattern, bytes
2479 other than 32-126 are always shown as hex escapes.
2480
2481 (b) When it is outputting text that is a matched part of a subject string,
2482 it does the same, unless a different locale has been set for the match
2483 (using the /L modifier). In this case, it uses isprint() to decide.
2484
2485 4. Fixed a major bug that caused incorrect computation of the amount of memory
2486 required for a compiled pattern when options that changed within the
2487 pattern affected the logic of the preliminary scan that determines the
2488 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2489 that the computed length was too small. The symptoms of this bug were
2490 either the PCRE error "internal error: code overflow" from pcre_compile(),
2491 or a glibc crash with a message such as "pcretest: free(): invalid next
2492 size (fast)". Examples of patterns that provoked this bug (shown in
2493 pcretest format) are:
2494
2495 /(?-x: )/x
2496 /(?x)(?-x: \s*#\s*)/
2497 /((?i)[\x{c0}])/8
2498 /(?i:[\x{c0}])/8
2499
2500 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2501 is now done differently.
2502
2503 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2504 wrapper classes; (b) implement a new function in the C++ scanner that is
2505 more efficient than the old way of doing things because it avoids levels of
2506 recursion in the regex matching; (c) add a paragraph to the documentation
2507 for the FullMatch() function.
2508
2509 6. The escape sequence \n was being treated as whatever was defined as
2510 "newline". Not only was this contrary to the documentation, which states
2511 that \n is character 10 (hex 0A), but it also went horribly wrong when
2512 "newline" was defined as CRLF. This has been fixed.
2513
2514 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2515 was being set to -1 for the "end of line" case (supposedly a value that no
2516 character can have). Though this value is never used (the check for end of
2517 line is "zero bytes in current character"), it caused compiler complaints.
2518 I've changed it to 0xffffffff.
2519
2520 8. In pcre_version.c, the version string was being built by a sequence of
2521 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2522 string (as it is for production releases) called a macro with an empty
2523 argument. The C standard says the result of this is undefined. The gcc
2524 compiler treats it as an empty string (which was what was wanted) but it is
2525 reported that Visual C gives an error. The source has been hacked around to
2526 avoid this problem.
2527
2528 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2529 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2530 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2531 of them did).
2532
2533 10. Originally, pcretest opened its input and output without "b"; then I was
2534 told that "b" was needed in some environments, so it was added for release
2535 5.0 to both the input and output. (It makes no difference on Unix-like
2536 systems.) Later I was told that it is wrong for the input on Windows. I've
2537 now abstracted the modes into two macros, to make it easier to fiddle with
2538 them, and removed "b" from the input mode under Windows.
2539
2540 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2541
2542 12. Added -help and --help to pcretest as an official way of being reminded
2543 of the options.
2544
2545 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2546 and pcrecpp.cc because they annoy compilers at high warning levels.
2547
2548 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2549
2550 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2551 scripts are not C programs :-) and which was not noticed because it works
2552 on Linux.
2553
2554 16. pcretest is supposed to handle any length of pattern and data line (as one
2555 line or as a continued sequence of lines) by extending its input buffer if
2556 necessary. This feature was broken for very long pattern lines, leading to
2557 a string of junk being passed to pcre_compile() if the pattern was longer
2558 than about 50K.
2559
2560 17. I have done a major re-factoring of the way pcre_compile() computes the
2561 amount of memory needed for a compiled pattern. Previously, there was code
2562 that made a preliminary scan of the pattern in order to do this. That was
2563 OK when PCRE was new, but as the facilities have expanded, it has become
2564 harder and harder to keep it in step with the real compile phase, and there
2565 have been a number of bugs (see for example, 4 above). I have now found a
2566 cunning way of running the real compile function in a "fake" mode that
2567 enables it to compute how much memory it would need, while actually only
2568 ever using a few hundred bytes of working memory and without too many
2569 tests of the mode. This should make future maintenance and development
2570 easier. A side effect of this work is that the limit of 200 on the nesting
2571 depth of parentheses has been removed (though this was never a serious
2572 limitation, I suspect). However, there is a downside: pcre_compile() now
2573 runs more slowly than before (30% or more, depending on the pattern). I
2574 hope this isn't a big issue. There is no effect on runtime performance.
2575
2576 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2577 newline (only possible for the last line of a file) and it was a
2578 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2579
2580 19. Added additional timing features to pcretest. (1) The -tm option now times
2581 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2582 separate command line item, by a number that specifies the number of
2583 repeats to use when timing. The default is 50000; this gives better
2584 precision, but takes uncomfortably long for very large patterns.
2585
2586 20. Extended pcre_study() to be more clever in cases where a branch of a
2587 subpattern has no definite first character. For example, (a*|b*)[cd] would
2588 previously give no result from pcre_study(). Now it recognizes that the
2589 first character must be a, b, c, or d.
2590
2591 21. There was an incorrect error "recursive call could loop indefinitely" if
2592 a subpattern (or the entire pattern) that was being tested for matching an
2593 empty string contained only one non-empty item after a nested subpattern.
2594 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2595 incorrectly, because the \d was being skipped in the check.
2596
2597 22. The pcretest program now has a new pattern option /B and a command line
2598 option -b, which is equivalent to adding /B to every pattern. This causes
2599 it to show the compiled bytecode, without the additional information that
2600 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2601 is the same as /B/I).
2602
2603 23. A new optimization is now able automatically to treat some sequences such
2604 as a*b as a*+b. More specifically, if something simple (such as a character
2605 or a simple class like \d) has an unlimited quantifier, and is followed by
2606 something that cannot possibly match the quantified thing, the quantifier
2607 is automatically "possessified".
2608
2609 24. A recursive reference to a subpattern whose number was greater than 39
2610 went wrong under certain circumstances in UTF-8 mode. This bug could also
2611 have affected the operation of pcre_study().
2612
2613 25. Realized that a little bit of performance could be had by replacing
2614 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2615
2616 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2617
2618 27. Possessive quantifiers such as a++ were previously implemented by turning
2619 them into atomic groups such as ($>a+). Now they have their own opcodes,
2620 which improves performance. This includes the automatically created ones
2621 from 23 above.
2622
2623 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2624 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2625 the first matched character to be a colon. This applied both to named and
2626 numbered groups.
2627
2628 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2629
2630 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2631 building PCRE on Windows easier, so I have included it in the distribution.
2632
2633 31. There is now a check in pcretest against a ridiculously large number being
2634 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2635 loop, the loop is abandoned.
2636
2637 32. Forward references to subpatterns in conditions such as (?(2)...) where
2638 subpattern 2 is defined later cause pcre_compile() to search forwards in
2639 the pattern for the relevant set of parentheses. This search went wrong
2640 when there were unescaped parentheses in a character class, parentheses
2641 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2642
2643 33. "Subroutine" calls and backreferences were previously restricted to
2644 referencing subpatterns earlier in the regex. This restriction has now
2645 been removed.
2646
2647 34. Added a number of extra features that are going to be in Perl 5.10. On the
2648 whole, these are just syntactic alternatives for features that PCRE had
2649 previously implemented using the Python syntax or my own invention. The
2650 other formats are all retained for compatibility.
2651
2652 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2653 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2654 also .NET compatible.
2655
2656 (b) A recursion or subroutine call to a named group can now be defined as
2657 (?&name) as well as (?P>name).
2658
2659 (c) A backreference to a named group can now be defined as \k<name> or
2660 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2661 5.10, are also .NET compatible.
2662
2663 (d) A conditional reference to a named group can now use the syntax
2664 (?(<name>) or (?('name') as well as (?(name).
2665
2666 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2667 groups (named and numbered) that are never evaluated inline, but can be
2668 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2669 is always false. There may be only one alternative in such a group.
2670
2671 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2672 as the simple (?(R). The condition is true only if the most recent
2673 recursion is that of the given number or name. It does not search out
2674 through the entire recursion stack.
2675
2676 (g) The escape \gN or \g{N} has been added, where N is a positive or
2677 negative number, specifying an absolute or relative reference.
2678
2679 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2680 some "unreachable code" warnings.
2681
2682 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2683 things, this adds five new scripts.
2684
2685 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2686 There were also incompatibilities regarding the handling of \Q..\E inside
2687 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2688 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2689
2690 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2691 matches an empty string, and forcibly breaks the loop. There were bugs in
2692 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2693 against aaaa the result was just "a" rather than "aaaa", for example. Two
2694 separate and independent bugs (that affected different cases) have been
2695 fixed.
2696
2697 39. Refactored the code to abolish the use of different opcodes for small
2698 capturing bracket numbers. This is a tidy that I avoided doing when I
2699 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2700 The new approach is not only tidier, it makes it possible to reduce the
2701 memory needed to fix the previous bug (38).
2702
2703 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2704 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2705 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2706 mode.
2707
2708 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2709 report.
2710
2711 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2712 copy construction and assignment in the C++ wrapper.
2713
2714 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2715 couple of bugs that could have given wrong results in the "--newline=crlf"
2716 case.
2717
2718 44. Added a number of casts and did some reorganization of signed/unsigned int
2719 variables following suggestions from Dair Grant. Also renamed the variable
2720 "this" as "item" because it is a C++ keyword.
2721
2722 45. Arranged for dftables to add
2723
2724 #include "pcre_internal.h"
2725
2726 to pcre_chartables.c because without it, gcc 4.x may remove the array
2727 definition from the final binary if PCRE is built into a static library and
2728 dead code stripping is activated.
2729
2730 46. For an unanchored pattern, if a match attempt fails at the start of a
2731 newline sequence, and the newline setting is CRLF or ANY, and the next two
2732 characters are CRLF, advance by two characters instead of one.
2733
2734
2735 Version 6.7 04-Jul-06
2736 ---------------------
2737
2738 1. In order to handle tests when input lines are enormously long, pcretest has
2739 been re-factored so that it automatically extends its buffers when
2740 necessary. The code is crude, but this _is_ just a test program. The
2741 default size has been increased from 32K to 50K.
2742
2743 2. The code in pcre_study() was using the value of the re argument before
2744 testing it for NULL. (Of course, in any sensible call of the function, it
2745 won't be NULL.)
2746
2747 3. The memmove() emulation function in pcre_internal.h, which is used on
2748 systems that lack both memmove() and bcopy() - that is, hardly ever -
2749 was missing a "static" storage class specifier.
2750
2751 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2752 containing an extended class (one that cannot be represented by a bitmap
2753 because it contains high-valued characters or Unicode property items, e.g.
2754 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2755 pattern, but PCRE should not loop if you do not (it no longer does).
2756 [Detail: two cases were found: (a) a repeated subpattern containing an
2757 extended class; (b) a recursive reference to a subpattern that followed a
2758 previous extended class. It wasn't skipping over the extended class
2759 correctly when UTF-8 mode was not set.]
2760
2761 5. A negated single-character class was not being recognized as fixed-length
2762 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2763 compile error "lookbehind assertion is not fixed length".
2764
2765 6. The RunPerlTest auxiliary script was showing an unexpected difference
2766 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2767 write a Perl script that can interpret lines of an input file either as
2768 byte characters or as UTF-8, which is what "perltest" was being required to
2769 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2770 can't do is switch easily at run time between having the "use utf8;" pragma
2771 or not. In the end, I fudged it by using the RunPerlTest script to insert
2772 "use utf8;" explicitly for the UTF-8 tests.
2773
2774 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2775 the end of the subject string, contrary to the documentation and to what
2776 Perl does. This was true of both matching functions. Now it matches only at
2777 the start of the subject and immediately after *internal* newlines.
2778
2779 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2780 a pointer to an int instead of a pointer to an unsigned long int. This
2781 caused problems on 64-bit systems.
2782
2783 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2784 instance of the 'standard' template library not being so standard".
2785
2786 10. There was no check on the number of named subpatterns nor the maximum
2787 length of a subpattern name. The product of these values is used to compute
2788 the size of the memory block for a compiled pattern. By supplying a very
2789 long subpattern name and a large number of named subpatterns, the size
2790 computation could be caused to overflow. This is now prevented by limiting
2791 the length of names to 32 characters, and the number of named subpatterns
2792 to 10,000.
2793
2794 11. Subpatterns that are repeated with specific counts have to be replicated in
2795 the compiled pattern. The size of memory for this was computed from the
2796 length of the subpattern and the repeat count. The latter is limited to
2797 65535, but there was no limit on the former, meaning that integer overflow
2798 could in principle occur. The compiled length of a repeated subpattern is
2799 now limited to 30,000 bytes in order to prevent this.
2800
2801 12. Added the optional facility to have named substrings with the same name.
2802
2803 13. Added the ability to use a named substring as a condition, using the
2804 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2805 are numbers (not recommended). Forward references are permitted.
2806
2807 14. Added forward references in named backreferences (if you see what I mean).
2808
2809 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2810 pattern could run off the end of the subject. For example, the pattern
2811 "(?s)(.{1,5})"8 did this with the subject "ab".
2812
2813 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2814 PCRE_CASELESS was set when matching characters that were quantified with ?
2815 or *.
2816
2817 17. A character class other than a single negated character that had a minimum
2818 but no maximum quantifier - for example [ab]{6,} - was not handled
2819 correctly by pce_dfa_exec(). It would match only one character.
2820
2821 18. A valid (though odd) pattern that looked like a POSIX character
2822 class but used an invalid character after [ (for example [[,abc,]]) caused
2823 pcre_compile() to give the error "Failed: internal error: code overflow" or
2824 in some cases to crash with a glibc free() error. This could even happen if
2825 the pattern terminated after [[ but there just happened to be a sequence of
2826 letters, a binary zero, and a closing ] in the memory that followed.
2827
2828 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2829 over the years. Originally (before any Unicode support), just the bottom 8
2830 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2831 output from "man perlunicode" includes this:
2832
2833 The regular expression compiler produces polymorphic opcodes. That
2834 is, the pattern adapts to the data and automatically switches to
2835 the Unicode character scheme when presented with Unicode data--or
2836 instead uses a traditional byte scheme when presented with byte
2837 data.
2838
2839 Sadly, a wide octal escape does not cause a switch, and in a string with
2840 no other multibyte characters, these octal escapes are treated as before.
2841 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2842 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2843 Unicode string.
2844
2845 I have not perpetrated such confusion in PCRE. Up till now, it took just
2846 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2847 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2848 translate to the appropriate multibyte character.
2849
2850 29. Applied some refactoring to reduce the number of warnings from Microsoft
2851 and Borland compilers. This has included removing the fudge introduced
2852 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2853 a warning about an unused variable.
2854
2855 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2856 characters since release 4.0, because Perl (from release 5.004) does not.
2857 [Or at least, is documented not to: some releases seem to be in conflict
2858 with the documentation.] However, when a pattern was studied with
2859 pcre_study() and all its branches started with \s, PCRE still included VT
2860 as a possible starting character. Of course, this did no harm; it just
2861 caused an unnecessary match attempt.
2862
2863 22. Removed a now-redundant internal flag bit that recorded the fact that case
2864 dependency changed within the pattern. This was once needed for "required
2865 byte" processing, but is no longer used. This recovers a now-scarce options
2866 bit. Also moved the least significant internal flag bit to the most-
2867 significant bit of the word, which was not previously used (hangover from
2868 the days when it was an int rather than a uint) to free up another bit for
2869 the future.
2870
2871 23. Added support for CRLF line endings as well as CR and LF. As well as the
2872 default being selectable at build time, it can now be changed at runtime
2873 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2874 specify that it is scanning data with non-default line endings.
2875
2876 24. Changed the definition of CXXLINK to make it agree with the definition of
2877 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2878
2879 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2880 recursions. This makes a big different to stack usage for some patterns.
2881
2882 26. If a subpattern containing a named recursion or subroutine reference such
2883 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2884 the space required for the compiled pattern went wrong and gave too small a
2885 value. Depending on the environment, this could lead to "Failed: internal
2886 error: code overflow at offset 49" or "glibc detected double free or
2887 corruption" errors.
2888
2889 27. Applied patches from Google (a) to support the new newline modes and (b) to
2890 advance over multibyte UTF-8 characters in GlobalReplace.
2891
2892 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2893 difference for some implementation of PCRE in some Windows version.
2894
2895 29. Added some extra testing facilities to pcretest:
2896
2897 \q<number> in a data line sets the "match limit" value
2898 \Q<number> in a data line sets the "match recursion limt" value
2899 -S <number> sets the stack size, where <number> is in megabytes
2900
2901 The -S option isn't available for Windows.
2902
2903
2904 Version 6.6 06-Feb-06
2905 ---------------------
2906
2907 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2908 in pcreposix.h. I have copied the definition from pcre.h.
2909
2910 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2911 because pcre.h is no longer a built file.
2912
2913 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2914 not normally included in the compiled code.
2915
2916
2917 Version 6.5 01-Feb-06
2918 ---------------------
2919
2920 1. When using the partial match feature with pcre_dfa_exec(), it was not
2921 anchoring the second and subsequent partial matches at the new starting
2922 point. This could lead to incorrect results. For example, with the pattern
2923 /1234/, partially matching against "123" and then "a4" gave a match.
2924
2925 2. Changes to pcregrep:
2926
2927 (a) All non-match returns from pcre_exec() were being treated as failures
2928 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2929 error message is output. Some extra information is given for the
2930 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2931 probably the only errors that are likely to be caused by users (by
2932 specifying a regex that has nested indefinite repeats, for instance).
2933 If there are more than 20 of these errors, pcregrep is abandoned.
2934
2935 (b) A binary zero was treated as data while matching, but terminated the
2936 output line if it was written out. This has been fixed: binary zeroes
2937 are now no different to any other data bytes.
2938
2939 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2940 used to set a locale for matching. The --locale=xxxx long option has
2941 been added (no short equivalent) to specify a locale explicitly on the
2942 pcregrep command, overriding the environment variables.
2943
2944 (d) When -B was used with -n, some line numbers in the output were one less
2945 than they should have been.
2946
2947 (e) Added the -o (--only-matching) option.
2948
2949 (f) If -A or -C was used with -c (count only), some lines of context were
2950 accidentally printed for the final match.
2951
2952 (g) Added the -H (--with-filename) option.
2953
2954 (h) The combination of options -rh failed to suppress file names for files
2955 that were found from directory arguments.
2956
2957 (i) Added the -D (--devices) and -d (--directories) options.
2958
2959 (j) Added the -F (--fixed-strings) option.
2960
2961 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2962
2963 (l) Added the --colo(u)r option.
2964
2965 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2966 is not present by default.
2967
2968 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2969 items such as (?R) or (?1), when the recursion could match a number of
2970 alternatives. If it matched one of the alternatives, but subsequently,
2971 outside the recursion, there was a failure, the code tried to back up into
2972 the recursion. However, because of the way PCRE is implemented, this is not
2973 possible, and the result was an incorrect result from the match.
2974
2975 In order to prevent this happening, the specification of recursion has
2976 been changed so that all such subpatterns are automatically treated as
2977 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2978
2979 4. I had overlooked the fact that, in some locales, there are characters for
2980 which isalpha() is true but neither isupper() nor islower() are true. In
2981 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2982 and ordfeminine) are like this. This affected the treatment of \w and \W
2983 when they appeared in character classes, but not when they appeared outside
2984 a character class. The bit map for "word" characters is now created
2985 separately from the results of isalnum() instead of just taking it from the
2986 upper, lower, and digit maps. (Plus the underscore character, of course.)
2987
2988 5. The above bug also affected the handling of POSIX character classes such as
2989 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2990 permanent tables. Instead, the bit maps for such a class were previously
2991 created as the appropriate unions of the upper, lower, and digit bitmaps.
2992 Now they are created by subtraction from the [[:word:]] class, which has
2993 its own bitmap.
2994
2995 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2996 It is created by subtracting the vertical space characters (\x09, \x0a,
2997 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2998 subtraction was done in the overall bitmap for a character class, meaning
2999 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3000 be recognized. This bug has been fixed.
3001
3002 7. Patches from the folks at Google:
3003
3004 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3005 real life, but is still worth protecting against".
3006
3007 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3008 regular expressions".
3009
3010 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3011 have it.
3012
3013 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3014 "configure" and the latter not, in order to fix a problem somebody had
3015 with compiling the Arg class on HP-UX.
3016
3017 (e) Improve the error-handling of the C++ wrapper a little bit.
3018
3019 (f) New tests for checking recursion limiting.
3020
3021 8. The pcre_memmove() function, which is used only if the environment does not
3022 have a standard memmove() function (and is therefore rarely compiled),
3023 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3024 returning a result (though PCRE never actually uses the result).
3025
3026 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3027 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3028 returned instead of calling malloc() with an overflowing number that would
3029 most likely cause subsequent chaos.
3030
3031 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3032
3033 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3034 with this option is matched, the nmatch and pmatch options of regexec() are
3035 ignored.
3036
3037 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3038 provided in case anyone wants to the the POSIX interface with UTF-8
3039 strings.
3040
3041 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3042 C++ linking (needed for some HP-UX environments).
3043
3044 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3045 (unused parameter) and in the pcre_printint() function (omitted "default"
3046 switch label when the default is to do nothing).
3047
3048 15. Added some code to make it possible, when PCRE is compiled as a C++
3049 library, to replace subject pointers for pcre_exec() with a smart pointer
3050 class, thus making it possible to process discontinuous strings.
3051
3052 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3053 much the same function. They were added by different people who were trying
3054 to make PCRE easy to compile on non-Unix systems. It has been suggested
3055 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3056 for compiling on Windows systems. I have therefore replaced it with
3057 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3058 defaults to "extern" for C or "extern C" for C++, which works fine on
3059 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3060 SCOPE with something explicit in config.h. In addition:
3061
3062 (a) pcreposix.h still had just "extern" instead of either of these macros;
3063 I have replaced it with PCRE_DATA_SCOPE.
3064
3065 (b) Functions such as _pcre_xclass(), which are internal to the library,
3066 but external in the C sense, all had PCRE_EXPORT in their definitions.
3067 This is apparently wrong for the Windows case, so I have removed it.
3068 (It makes no difference on Unix-like systems.)
3069
3070 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3071 of recursive calls to match(). This is different to MATCH_LIMIT because
3072 that limits the total number of calls to match(), not all of which increase
3073 the depth of recursion. Limiting the recursion depth limits the amount of
3074 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3075 when PCRE is compiled, and changed at run time. A patch from Google adds
3076 this functionality to the C++ interface.
3077
3078 18. Changes to the handling of Unicode character properties:
3079
3080 (a) Updated the table to Unicode 4.1.0.
3081
3082 (b) Recognize characters that are not in the table as "Cn" (undefined).
3083
3084 (c) I revised the way the table is implemented to a much improved format
3085 which includes recognition of ranges. It now supports the ranges that
3086 are defined in UnicodeData.txt, and it also amalgamates other
3087 characters into ranges. This has reduced the number of entries in the
3088 table from around 16,000 to around 3,000, thus reducing its size
3089 considerably. I realized I did not need to use a tree structure after
3090 all - a binary chop search is just as efficient. Having reduced the
3091 number of entries, I extended their size from 6 bytes to 8 bytes to
3092 allow for more data.
3093
3094 (d) Added support for Unicode script names via properties such as \p{Han}.
3095
3096 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3097 matching that character.
3098
3099 20. When matching a repeated Unicode property with a minimum greater than zero,
3100 (for example \pL{2,}), PCRE could look past the end of the subject if it
3101 reached it while seeking the minimum number of characters. This could
3102 happen only if some of the characters were more than one byte long, because
3103 there is a check for at least the minimum number of bytes.
3104
3105 21. Refactored the implementation of \p and \P so as to be more general, to
3106 allow for more different types of property in future. This has changed the
3107 compiled form incompatibly. Anybody with saved compiled patterns that use
3108 \p or \P will have to recompile them.
3109
3110 22. Added "Any" and "L&" to the supported property types.
3111
3112 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3113 but give a compile time error if the value is greater than 0xff.
3114
3115 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3116 accidentally not being installed or uninstalled.
3117
3118 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3119 made were to insert the current release number. This seemed silly, because
3120 it made things harder for people building PCRE on systems that don't run
3121 "configure". I have turned pcre.h into a distributed file, no longer built
3122 by "configure", with the version identification directly included. There is
3123 no longer a pcre.h.in file.
3124
3125 However, this change necessitated a change to the pcre-config script as
3126 well. It is built from pcre-config.in, and one of the substitutions was the
3127 release number. I have updated configure.ac so that ./configure now finds
3128 the release number by grepping pcre.h.
3129
3130 26. Added the ability to run the tests under valgrind.
3131
3132
3133 Version 6.4 05-Sep-05
3134 ---------------------
3135
3136 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3137 "--" to be printed when multiple files were scanned, even when none of the
3138 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3139 consider it to be a bug, and have restored the previous behaviour.
3140
3141 2. A couple of code tidies to get rid of compiler warnings.
3142
3143 3. The pcretest program used to cheat by referring to symbols in the library
3144 whose names begin with _pcre_. These are internal symbols that are not
3145 really supposed to be visible externally, and in some environments it is
3146 possible to suppress them. The cheating is now confined to including
3147 certain files from the library's source, which is a bit cleaner.
3148
3149 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3150 file's purpose clearer.
3151
3152 5. Reorganized pcre_ucp_findchar().
3153
3154
3155 Version 6.3 15-Aug-05
3156 ---------------------
3157
3158 1. The file libpcre.pc.in did not have general read permission in the tarball.
3159
3160 2. There were some problems when building without C++ support:
3161
3162 (a) If C++ support was not built, "make install" and "make test" still
3163 tried to test it.
3164
3165 (b) There were problems when the value of CXX was explicitly set. Some
3166 changes have been made to try to fix these, and ...
3167
3168 (c) --disable-cpp can now be used to explicitly disable C++ support.
3169
3170 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3171 backslash in a target when C++ was disabled. This confuses some
3172 versions of "make", apparently. Using an intermediate variable solves
3173 this. (Same for CPP_LOBJ.)
3174
3175 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3176 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3177 necessary on certain architectures.
3178
3179 4. Added a setting of -export-symbols-regex to the link command to remove
3180 those symbols that are exported in the C sense, but actually are local
3181 within the library, and not documented. Their names all begin with
3182 "_pcre_". This is not a perfect job, because (a) we have to except some
3183 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3184 available (and never for static libraries). I have made a note to try to
3185 find a way round (a) in the future.
3186
3187
3188 Version 6.2 01-Aug-05
3189 ---------------------
3190
3191 1. There was no test for integer overflow of quantifier values. A construction
3192 such as {1111111111111111} would give undefined results. What is worse, if
3193 a minimum quantifier for a parenthesized subpattern overflowed and became
3194 negative, the calculation of the memory size went wrong. This could have
3195 led to memory overwriting.
3196
3197 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3198
3199 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3200 operating environments where this matters.
3201
3202 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3203 PCRE options from within the C++ wrapper.
3204
3205 5. Named capturing subpatterns were not being correctly counted when a pattern
3206 was compiled. This caused two problems: (a) If there were more than 100
3207 such subpatterns, the calculation of the memory needed for the whole
3208 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3209 back references of the form \12, where the number was greater than 9, were
3210 not recognized as back references, even though there were sufficient
3211 previous subpatterns.
3212
3213 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3214 versions of gcc, e.g. 2.95.4.
3215
3216
3217 Version 6.1 21-Jun-05
3218 ---------------------
3219
3220 1. There was one reference to the variable "posix" in pcretest.c that was not
3221 surrounded by "#if !defined NOPOSIX".
3222
3223 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3224 the cross-check on the old pcre_info() function, for the benefit of the
3225 cut-down version of PCRE that is currently imported into Exim.
3226
3227 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3228 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3229 patterns that start (?i) but I don't think that matters. The use of (?i) is
3230 just an example; this all applies to the other options as well.
3231
3232 4. Since libtool seems to echo the compile commands it is issuing, the output
3233 from "make" can be reduced a bit by putting "@" in front of each libtool
3234 compile command.
3235
3236 5. Patch from the folks at Google for configure.in to be a bit more thorough
3237 in checking for a suitable C++ installation before trying to compile the
3238 C++ stuff. This should fix a reported problem when a compiler was present,
3239 but no suitable headers.
3240
3241 6. The man pages all had just "PCRE" as their title. I have changed them to
3242 be the relevant file name. I have also arranged that these names are
3243 retained in the file doc/pcre.txt, which is a concatenation in text format
3244 of all the man pages except the little individual ones for each function.
3245
3246 7. The NON-UNIX-USE file had not been updated for the different set of source
3247 files that come with release 6. I also added a few comments about the C++
3248 wrapper.
3249
3250
3251 Version 6.0 07-Jun-05
3252 ---------------------
3253
3254 1. Some minor internal re-organization to help with my DFA experiments.
3255
3256 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3257 didn't matter for the library itself when fully configured, but did matter
3258 when compiling without UCP support, or within Exim, where the ucp files are
3259 not imported.
3260
3261 3. Refactoring of the library code to split up the various functions into
3262 different source modules. The addition of the new DFA matching code (see
3263 below) to a single monolithic source would have made it really too
3264 unwieldy, quite apart from causing all the code to be include in a
3265 statically linked application, when only some functions are used. This is
3266 relevant even without the DFA addition now that patterns can be compiled in
3267 one application and matched in another.
3268
3269 The downside of splitting up is that there have to be some external
3270 functions and data tables that are used internally in different modules of
3271 the library but which are not part of the API. These have all had their
3272 names changed to start with "_pcre_" so that they are unlikely to clash
3273 with other external names.
3274
3275 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3276 a different (DFA) algorithm. Although it is slower than the original
3277 function, it does have some advantages for certain types of matching
3278 problem.
3279
3280 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3281 including restarting after a partial match.
3282
3283 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3284 defined when compiling for Windows was sent to me. I have put it into the
3285 code, though I have no means of testing or verifying it.
3286
3287 7. Added the pcre_refcount() auxiliary function.
3288
3289 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3290 match before or at the first newline in the subject string. In pcretest,
3291 the /f option on a pattern can be used to set this.
3292
3293 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3294 would behave wrongly. This has been present in PCRE since release 4.0.
3295
3296 10. A number of changes to the pcregrep command:
3297
3298 (a) Refactored how -x works; insert ^(...)$ instead of setting
3299 PCRE_ANCHORED and checking the length, in preparation for adding
3300 something similar for -w.
3301
3302 (b) Added the -w (match as a word) option.
3303
3304 (c) Refactored the way lines are read and buffered so as to have more
3305 than one at a time available.
3306
3307 (d) Implemented a pcregrep test script.
3308
3309 (e) Added the -M (multiline match) option. This allows patterns to match
3310 over several lines of the subject. The buffering ensures that at least
3311 8K, or the rest of the document (whichever is the shorter) is available
3312 for matching (and similarly the previous 8K for lookbehind assertions).
3313
3314 (f) Changed the --help output so that it now says
3315
3316 -w, --word-regex(p)
3317
3318 instead of two lines, one with "regex" and the other with "regexp"
3319 because that confused at least one person since the short forms are the
3320 same. (This required a bit of code, as the output is generated
3321 automatically from a table. It wasn't just a text change.)
3322
3323 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3324 option but starts with a hyphen. Could be a pattern or a path name
3325 starting with a hyphen, for instance.
3326
3327 (h) "-" can be given as a file name to represent stdin.
3328
3329 (i) When file names are being printed, "(standard input)" is used for
3330 the standard input, for compatibility with GNU grep. Previously
3331 "<stdin>" was used.
3332
3333 (j) The option --label=xxx can be used to supply a name to be used for
3334 stdin when file names are being printed. There is no short form.
3335
3336 (k) Re-factored the options decoding logic because we are going to add
3337 two more options that take data. Such options can now be given in four
3338 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3339
3340 (l) Added the -A, -B, and -C options for requesting that lines of context
3341 around matches be printed.
3342
3343 (m) Added the -L option to print the names of files that do not contain
3344 any matching lines, that is, the complement of -l.
3345
3346 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3347 continue to scan other files.
3348
3349 (o) The -s option was incorrectly implemented. For compatibility with other
3350 greps, it now suppresses the error message for a non-existent or non-
3351 accessible file (but not the return code). There is a new option called
3352 -q that suppresses the output of matching lines, which was what -s was
3353 previously doing.
3354
3355 (p) Added --include and --exclude options to specify files for inclusion
3356 and exclusion when recursing.
3357
3358 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3359 Hopefully, it now does.
3360
3361 12. Missing cast in pcre_study().
3362
3363 13. Added an "uninstall" target to the makefile.
3364
3365 14. Replaced "extern" in the function prototypes in Makefile.in with
3366 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3367 world, but is set differently for Windows.
3368
3369 15. Added a second compiling function called pcre_compile2(). The only
3370 difference is that it has an extra argument, which is a pointer to an
3371 integer error code. When there is a compile-time failure, this is set
3372 non-zero, in addition to the error test pointer being set to point to an
3373 error message. The new argument may be NULL if no error number is required
3374 (but then you may as well call pcre_compile(), which is now just a
3375 wrapper). This facility is provided because some applications need a
3376 numeric error indication, but it has also enabled me to tidy up the way
3377 compile-time errors are handled in the POSIX wrapper.
3378
3379 16. Added VPATH=.libs to the makefile; this should help when building with one
3380 prefix path and installing with another. (Or so I'm told by someone who
3381 knows more about this stuff than I do.)
3382
3383 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3384 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3385 match everything, including newlines. This is not POSIX-compatible, but
3386 somebody wanted the feature. From pcretest it can be activated by using
3387 both the P and the s flags.
3388
3389 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3390
3391 19. libpcre.pc was being incorrectly installed as executable.
3392
3393 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3394 it now also looks for '\r' so that it will work unmodified on Windows.
3395
3396 21. Added Google's contributed C++ wrapper to the distribution.
3397
3398 22. Added some untidy missing memory free() calls in pcretest, to keep
3399 Electric Fence happy when testing.
3400
3401
3402
3403 Version 5.0 13-Sep-04
3404 ---------------------
3405
3406 1. Internal change: literal characters are no longer packed up into items
3407 containing multiple characters in a single byte-string. Each character
3408 is now matched using a separate opcode. However, there may be more than one
3409 byte in the character in UTF-8 mode.
3410
3411 2. The pcre_callout_block structure has two new fields: pattern_position and
3412 next_item_length. These contain the offset in the pattern to the next match
3413 item, and its length, respectively.
3414
3415 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3416 insertion of callouts before each pattern item. Added the /C option to
3417 pcretest to make use of this.
3418
3419 4. On the advice of a Windows user, the lines
3420
3421 #if defined(_WIN32) || defined(WIN32)
3422 _setmode( _fileno( stdout ), 0x8000 );
3423 #endif /* defined(_WIN32) || defined(WIN32) */
3424
3425 have been added to the source of pcretest. This apparently does useful
3426 magic in relation to line terminators.
3427
3428 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3429 for the benefit of those environments where the "b" makes a difference.
3430
3431 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3432 to know about it. I have put a hack into configure.in that adds in code
3433 to set GCC=yes if CC=icc. This seems to end up at a point in the
3434 generated configure script that is early enough to affect the setting of
3435 compiler options, which is what is needed, but I have no means of testing
3436 whether it really works. (The user who reported this had patched the
3437 generated configure script, which of course I cannot do.)
3438
3439 LATER: After change 22 below (new libtool files), the configure script
3440 seems to know about icc (and also ecc). Therefore, I have commented out
3441 this hack in configure.in.
3442
3443 7. Added support for pkg-config (2 patches were sent in).
3444
3445 8. Negated POSIX character classes that used a combination of internal tables
3446 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3447 [[:^ascii]]. Typically, they would match almost any characters. The other
3448 POSIX classes were not broken in this way.
3449
3450 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3451 to find the match, as PCRE was deluded into thinking that the match had to
3452 start at the start point or following a newline. The same bug applied to
3453 patterns with negative forward assertions or any backward assertions
3454 preceding ".*" at the start, unless the pattern required a fixed first
3455 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3456
3457 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3458 starting at the last subject character, bytes beyond the end of the subject
3459 string were read.
3460
3461 11. Renamed the variable "class" as "classbits" to make life easier for C++
3462 users. (Previously there was a macro definition, but it apparently wasn't
3463 enough.)
3464
3465 12. Added the new field "tables" to the extra data so that tables can be passed
3466 in at exec time, or the internal tables can be re-selected. This allows
3467 a compiled regex to be saved and re-used at a later time by a different
3468 program that might have everything at different addresses.
3469
3470 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3471 -R library as well as a -L library.
3472
3473 14. The debugging options of pcretest (-d on the command line or D on a
3474 pattern) showed incorrect output for anything following an extended class
3475 that contained multibyte characters and which was followed by a quantifier.
3476
3477 15. Added optional support for general category Unicode character properties
3478 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3479 support. It adds about 90K to the size of the library. The meanings of the
3480 inbuilt class escapes such as \d and \s have NOT been changed.
3481
3482 16. Updated pcredemo.c to include calls to free() to release the memory for the
3483 compiled pattern.
3484
3485 17. The generated file chartables.c was being created in the source directory
3486 instead of in the building directory. This caused the build to fail if the
3487 source directory was different from the building directory, and was
3488 read-only.
3489
3490 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3491 file. No doubt somebody will tell me if they don't make sense... Also added
3492 Dan Mooney's comments about building on OpenVMS.
3493
3494 19. Added support for partial matching via the PCRE_PARTIAL option for
3495 pcre_exec() and the \P data escape in pcretest.
3496
3497 20. Extended pcretest with 3 new pattern features:
3498
3499 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3500 write the compiled pattern to the file whose name is "rest-of-line".
3501 This is a straight binary dump of the data, with the saved pointer to
3502 the character tables forced to be NULL. The study data, if any, is
3503 written too. After writing, pcretest reads a new pattern.
3504
3505 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3506 compiled pattern from the given file. There must not be any
3507 occurrences of "<" in the file name (pretty unlikely); if there are,
3508 pcretest will instead treat the initial "<" as a pattern delimiter.
3509 After reading in the pattern, pcretest goes on to read data lines as
3510 usual.
3511
3512 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3513 and 16-bit fields in a compiled pattern, to simulate a pattern that
3514 was compiled on a host of opposite endianness.
3515
3516 21. The pcre-exec() function can now cope with patterns that were compiled on
3517 hosts of opposite endianness, with this restriction:
3518
3519 As for any compiled expression that is saved and used later, the tables
3520 pointer field cannot be preserved; the extra_data field in the arguments
3521 to pcre_exec() should be used to pass in a tables address if a value
3522 other than the default internal tables were used at compile time.
3523
3524 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3525 now diagnosed as an error. Previously, most of the time, a negative number
3526 would have been treated as zero, but if in addition "ovector" was passed as
3527 NULL, a crash could occur.
3528
3529 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3530 new versions from the libtool 1.5 distribution (the last one is a copy of
3531 a file called libtool.m4). This seems to have fixed the need to patch
3532 "configure" to support Darwin 1.3 (which I used to do). However, I still
3533 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3534 workstation).
3535
3536 24. Changed the PCRE licence to be the more standard "BSD" licence.
3537
3538
3539 Version 4.5 01-Dec-03
3540 ---------------------
3541
3542 1. There has been some re-arrangement of the code for the match() function so
3543 that it can be compiled in a version that does not call itself recursively.
3544 Instead, it keeps those local variables that need separate instances for
3545 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3546 needs to "recurse". Keeping track of where control must go is done by means
3547 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3548 hide most of the details from the main code, and operates only if
3549 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3550 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3551 operating.
3552
3553 To make it easier for callers to provide specially tailored get/free
3554 functions for this usage, two new functions, pcre_stack_malloc, and
3555 pcre_stack_free, are used. They are always called in strict stacking order,
3556 and the size of block requested is always the same.
3557
3558 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3559 PCRE has been compiled to use the stack or the heap for recursion. The
3560 -C option of pcretest uses this to show which version is compiled.
3561
3562 A new data escape \S, is added to pcretest; it causes the amounts of store
3563 obtained and freed by both kinds of malloc/free at match time to be added
3564 to the output.
3565
3566 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3567 what's available on my current Linux desktop machine.
3568
3569 3. When matching a UTF-8 string, the test for a valid string at the start has
3570 been extended. If start_offset is not zero, PCRE now checks that it points
3571 to a byte that is the start of a UTF-8 character. If not, it returns
3572 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3573 this is necessary because there may be backward assertions in the pattern.
3574 When matching the same subject several times, it may save resources to use
3575 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3576
3577 4. The code for checking the validity of UTF-8 strings has been tightened so
3578 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3579 containing "overlong sequences".
3580
3581 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3582 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3583 should have been "&", but it just so happened that all the cases this let
3584 through by mistake were picked up later in the function.
3585
3586 6. I had used a variable called "isblank" - this is a C99 function, causing
3587 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3588
3589 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3590 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3591 so the version is shown; (c) stop "make test" echoing "./RunTest".
3592
3593 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3594
3595 9. The prototype for memmove() for systems that don't have it was using
3596 size_t, but the inclusion of the header that defines size_t was later. I've
3597 moved the #includes for the C headers earlier to avoid this.
3598
3599 10. Added some adjustments to the code to make it easier to compiler on certain
3600 special systems:
3601
3602 (a) Some "const" qualifiers were missing.
3603 (b) Added the macro EXPORT before all exported functions; by default this
3604 is defined to be empty.
3605 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3606 that it reads its output file name as an argument instead of writing
3607 to the standard output and assuming this can be redirected.
3608
3609 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3610 class containing characters with values greater than 255, PCRE compilation
3611 went into a loop.
3612
3613 12. A recursive reference to a subpattern that was within another subpattern
3614 that had a minimum quantifier of zero caused PCRE to crash. For example,
3615 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3616 recursion. If the recursively-called subpattern itself had a zero repeat,
3617 that was OK.
3618
3619 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3620 buffer into which it was copied (for escape processing) was still set at
3621 1024, so long lines caused crashes.
3622
3623 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3624 "internal error: code overflow...". This applied to any character class
3625 that was followed by a possessive quantifier.
3626
3627 15. Modified the Makefile to add libpcre.la as a prerequisite for
3628 libpcreposix.la because I was told this is needed for a parallel build to
3629 work.
3630
3631 16. If a pattern that contained .* following optional items at the start was
3632 studied, the wrong optimizing data was generated, leading to matching
3633 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3634 matching string must start with a or b or c. The correct conclusion for
3635 this pattern is that a match can start with any character.
3636
3637
3638 Version 4.4 13-Aug-03
3639 ---------------------
3640
3641 1. In UTF-8 mode, a character class containing characters with values between
3642 127 and 255 was not handled correctly if the compiled pattern was studied.
3643 In fixing this, I have also improved the studying algorithm for such
3644 classes (slightly).
3645
3646 2. Three internal functions had redundant arguments passed to them. Removal
3647 might give a very teeny performance improvement.
3648
3649 3. Documentation bug: the value of the capture_top field in a callout is *one
3650 more than* the number of the hightest numbered captured substring.
3651
3652 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3653 in incorrectly linking with a previously installed version. They now link
3654 explicitly with libpcre.la.
3655
3656 5. configure.in no longer needs to recognize Cygwin specially.
3657
3658 6. A problem in pcre.in for Windows platforms is fixed.
3659
3660 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3661 pcretest, it used to include the size of the study block as part of its
3662 output. Unfortunately, the structure contains a field that has a different
3663 size on different hardware architectures. This meant that the tests that
3664 showed this size failed. As the block is currently always of a fixed size,
3665 this information isn't actually particularly useful in pcretest output, so
3666 I have just removed it.
3667
3668 8. Three pre-processor statements accidentally did not start in column 1.
3669 Sadly, there are *still* compilers around that complain, even though
3670 standard C has not required this for well over a decade. Sigh.
3671
3672 9. In pcretest, the code for checking callouts passed small integers in the
3673 callout_data field, which is a void * field. However, some picky compilers
3674 complained about the casts involved for this on 64-bit systems. Now
3675 pcretest passes the address of the small integer instead, which should get
3676 rid of the warnings.
3677
3678 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3679 both compile and run time, and gives an error if an invalid UTF-8 sequence
3680 is found. There is a option for disabling this check in cases where the
3681 string is known to be correct and/or the maximum performance is wanted.
3682
3683 11. In response to a bug report, I changed one line in Makefile.in from
3684
3685 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3686 to
3687 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3688
3689 to look similar to other lines, but I have no way of telling whether this
3690 is the right thing to do, as I do not use Windows. No doubt I'll get told
3691 if it's wrong...
3692
3693
3694 Version 4.3 21-May-03
3695 ---------------------
3696
3697 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3698 Makefile.
3699
3700 2. Some refactoring to improve the quality of the code:
3701
3702 (i) The utf8_table... variables are now declared "const".
3703
3704 (ii) The code for \cx, which used the "case flipping" table to upper case
3705 lower case letters, now just substracts 32. This is ASCII-specific,
3706 but the whole concept of \cx is ASCII-specific, so it seems
3707 reasonable.
3708
3709 (iii) PCRE was using its character types table to recognize decimal and
3710 hexadecimal digits in the pattern. This is silly, because it handles
3711 only 0-9, a-f, and A-F, but the character types table is locale-
3712 specific, which means strange things might happen. A private
3713 table is now used for this - though it costs 256 bytes, a table is
3714 much faster than multiple explicit tests. Of course, the standard
3715 character types table is still used for matching digits in subject
3716 strings against \d.
3717
3718 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3719 ending in _t are). So I've renamed it as ESC_tee.
3720
3721 3. The first argument for regexec() in the POSIX wrapper should have been
3722 defined as "const".
3723
3724 4. Changed pcretest to use malloc() for its buffers so that they can be
3725 Electric Fenced for debugging.
3726
3727 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3728 to read one or more bytes before the start of the subject string. Often this
3729 had no effect on PCRE's behaviour, but in some circumstances it could
3730 provoke a segmentation fault.
3731
3732 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3733 to try to read one or more bytes before the start of the subject string.
3734
3735 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3736 UTF-8 support could misbehave in various ways if the subject string
3737 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3738 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3739 back over UTF-8 characters.)
3740
3741
3742 Version 4.2 14-Apr-03
3743 ---------------------
3744
3745 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3746
3747 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3748 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3749 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3750 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3751 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3752 and BUILD_EXEEXT
3753 Note: automatic setting of the BUILD variables is not yet working
3754 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3755 compile-time but not at link-time
3756 [LINK]: use for linking executables only
3757 make different versions for Windows and non-Windows
3758 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3759 libraries
3760 [LINK_FOR_BUILD]: new variable
3761 [OBJEXT]: use throughout
3762 [EXEEXT]: use throughout
3763 <winshared>: new target
3764 <wininstall>: new target
3765 <dftables.o>: use native compiler
3766 <dftables>: use native linker
3767 <install>: handle Windows platform correctly
3768 <clean>: ditto
3769 <check>: ditto
3770 copy DLL to top builddir before testing
3771
3772 As part of these changes, -no-undefined was removed again. This was reported
3773 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3774 in any case.
3775
3776 3. Some tidies to get rid of compiler warnings:
3777
3778 . In the match_data structure, match_limit was an unsigned long int, whereas
3779 match_call_count was an int. I've made them both unsigned long ints.
3780
3781 . In pcretest the fact that a const uschar * doesn't automatically cast to
3782 a void * provoked a warning.
3783
3784 . Turning on some more compiler warnings threw up some "shadow" variables
3785 and a few more missing casts.
3786
3787 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3788 option, a class that contained a single character with a value between 128
3789 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3790
3791 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3792 option, a class that contained several characters, but with at least one
3793 whose value was between 128 and 255 caused PCRE to crash.
3794
3795
3796 Version 4.1 12-Mar-03
3797 ---------------------
3798
3799 1. Compiling with gcc -pedantic found a couple of places where casts were
3800 needed, and a string in dftables.c that was longer than standard compilers are
3801 required to support.
3802
3803 2. Compiling with Sun's compiler found a few more places where the code could
3804 be tidied up in order to avoid warnings.
3805
3806 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3807 first of these names is deprecated in the latest Autoconf in favour of the name
3808 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3809 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3810 analogy I have changed it to CFLAGS_FOR_BUILD.
3811
3812 4. Added -no-undefined to the linking command in the Makefile, because this is
3813 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3814 linking step for the pcreposix library.
3815
3816 5. PCRE was failing to diagnose the case of two named groups with the same
3817 name.
3818
3819 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3820 literal character that is needed in the subject for a match, and scans along to
3821 ensure that it is present before embarking on the full matching process. This
3822 saves time in cases of nested unlimited repeats that are never going to match.
3823 Problem: the scan can take a lot of time if the subject is very long (e.g.
3824 megabytes), thus penalizing straightforward matches. It is now done only if the
3825 amount of subject to be scanned is less than 1000 bytes.
3826
3827 7. A lesser problem with the same optimization is that it was recording the
3828 first character of an anchored pattern as "needed", thus provoking a search
3829 right along the subject, even when the first match of the pattern was going to
3830 fail. The "needed" character is now not set for anchored patterns, unless it
3831 follows something in the pattern that is of non-fixed length. Thus, it still
3832 fulfils its original purpose of finding quick non-matches in cases of nested
3833 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3834
3835
3836 Version 4.0 17-Feb-03
3837 ---------------------
3838
3839 1. If a comment in an extended regex that started immediately after a meta-item
3840 extended to the end of string, PCRE compiled incorrect data. This could lead to
3841 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3842
3843 2. Moved to autoconf 2.53 and libtool 1.4.2.
3844
3845 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3846 the special perltest8 script is no longer needed - all the tests can be run
3847 from a single perltest script.
3848
3849 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3850 by \s. It has now been removed in PCRE. This means it isn't recognized as
3851 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3852 class [:space:] *does* include VT, thereby creating a mess.
3853
3854 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3855 space and tab.
3856
3857 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3858 its new features into the main test script, reducing the number of scripts.
3859
3860 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3861 were backward compatible, and made the (?i) apply to the whole pattern, as if
3862 /i were given. Now it behaves more logically, and applies the option setting
3863 only to what follows. PCRE has been changed to follow suit. However, if it
3864 finds options settings right at the start of the pattern, it extracts them into
3865 the global options, as before. Thus, they show up in the info data.
3866
3867 8. Added support for the \Q...\E escape sequence. Characters in between are
3868 treated as literals. This is slightly different from Perl in that $ and @ are
3869 also handled as literals inside the quotes. In Perl, they will cause variable
3870 interpolation. Note the following examples:
3871
3872 Pattern PCRE matches Perl matches
3873
3874 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3875 \Qabc\$xyz\E abc\$xyz abc\$xyz
3876 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3877
3878 For compatibility with Perl, \Q...\E sequences are recognized inside character
3879 classes as well as outside them.
3880
3881 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3882 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3883 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3884 signed/unsigned warnings.
3885
3886 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3887 option for pcretest, so I've replaced it by a simple function that does just
3888 that job.
3889
3890 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3891 "pcregrep -".
3892
3893 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3894 Java package. This provides some syntactic sugar for simple cases of what my
3895 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3896 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3897 item, you can use this simplified notation. Note that only makes sense with
3898 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3899 greediness, whatever the setting of the PCRE_UNGREEDY option.
3900
3901 13. A change of greediness default within a pattern was not taking effect at
3902 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3903 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3904 was abstracted outside.
3905
3906 14. PCRE now supports the \G assertion. It is true when the current matching
3907 position is at the start point of the match. This differs from \A when the
3908 starting offset is non-zero. Used with the /g option of pcretest (or similar
3909 code), it works in the same way as it does for Perl's /g option. If all
3910 alternatives of a regex begin with \G, the expression is anchored to the start
3911 match position, and the "anchored" flag is set in the compiled expression.
3912
3913 15. Some bugs concerning the handling of certain option changes within patterns
3914 have been fixed. These applied to options other than (?ims). For example,
3915 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3916 been the other way round. Some of this was related to change 7 above.
3917
3918 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3919 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3920 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3921 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3922
3923 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3924 mode. Unlike ".", it always matches newline, whatever the setting of
3925 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3926 assertions. Perl allows it, but it doesn't (in general) work because it can't
3927 calculate the length of the lookbehind. At least, that's the case for Perl
3928 5.8.0 - I've been told they are going to document that it doesn't work in
3929 future.
3930
3931 18. Added an error diagnosis for escapes that PCRE does not support: these are
3932 \L, \l, \N, \P, \p, \U, \u, and \X.
3933
3934 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3935 reading past the end of the pattern in cases such as /[abcd/.
3936
3937 20. PCRE was getting more memory than necessary for patterns with classes that
3938 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3939
3940 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3941 compiling PCRE for use with Virtual Pascal.
3942
3943 22. Small fix to the Makefile to make it work properly if the build is done
3944 outside the source tree.
3945
3946 23. Added a new extension: a condition to go with recursion. If a conditional
3947 subpattern starts with (?(R) the "true" branch is used if recursion has
3948 happened, whereas the "false" branch is used only at the top level.
3949
3950 24. When there was a very long string of literal characters (over 255 bytes
3951 without UTF support, over 250 bytes with UTF support), the computation of how
3952 much memory was required could be incorrect, leading to segfaults or other
3953 strange effects.
3954
3955 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3956 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3957 there was a subsequent back reference to those brackets. This meant that, for
3958 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3959 possible to check for precisely this case. All we can do is abandon the
3960 optimization if .* occurs inside capturing brackets when there are any back
3961 references whatsoever. (See below for a better fix that came later.)
3962
3963 26. The handling of the optimization for finding the first character of a
3964 non-anchored pattern, and for finding a character that is required later in the
3965 match were failing in some cases. This didn't break the matching; it just
3966 failed to optimize when it could. The way this is done has been re-implemented.
3967
3968 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3969
3970 28. Added a new feature that provides some of the functionality that Perl
3971 provides with (?{...}). The facility is termed a "callout". The way it is done
3972 in PCRE is for the caller to provide an optional function, by setting
3973 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3974 global variable. By default it is unset, which disables all calling out. To get
3975 the function called, the regex must include (?C) at appropriate points. This
3976 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3977 This provides a means of identifying different callout points. When PCRE
3978 reaches such a point in the regex, if pcre_callout has been set, the external
3979 function is called. It is provided with data in a structure called
3980 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3981 matching continues; if it returns a non-zero value, the match at the current
3982 point fails. However, backtracking will occur if possible. [This was changed
3983 later and other features added - see item 49 below.]
3984
3985 29. pcretest is upgraded to test the callout functionality. It provides a
3986 callout function that displays information. By default, it shows the start of
3987 the match and the current position in the text. There are some new data escapes
3988 to vary what happens:
3989
3990 \C+ in addition, show current contents of captured substrings
3991 \C- do not supply a callout function
3992 \C!n return 1 when callout number n is reached
3993 \C!n!m return 1 when callout number n is reached for the mth time
3994
3995 30. If pcregrep was called with the -l option and just a single file name, it
3996 output "<stdin>" if a match was found, instead of the file name.
3997
3998 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3999 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4000 pcre_exec(). This saves a malloc/free per call. The default value of
4001 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4002 when configuring.
4003
4004 32. The default maximum size of a compiled pattern is 64K. There have been a
4005 few cases of people hitting this limit. The code now uses macros to handle the
4006 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4007 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4008 configuring. Tests 2 and 5 work only with 2-byte links because they output
4009 debugging information about compiled patterns.
4010
4011 33. Internal code re-arrangements:
4012
4013 (a) Moved the debugging function for printing out a compiled regex into
4014 its own source file (printint.c) and used #include to pull it into
4015 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4016 separate copies.
4017
4018 (b) Defined the list of op-code names for debugging as a macro in
4019 internal.h so that it is next to the definition of the opcodes.
4020
4021 (c) Defined a table of op-code lengths for simpler skipping along compiled
4022 code. This is again a macro in internal.h so that it is next to the
4023 definition of the opcodes.
4024
4025 34. Added support for recursive calls to individual subpatterns, along the
4026 lines of Robin Houston's patch (but implemented somewhat differently).
4027
4028 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4029 allow it to read and process whole directories in Win32. This code was
4030 contributed by Lionel Fourquaux; it has not been tested by me.
4031
4032 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4033 used to name a group. Names consist of alphanumerics and underscores, and must
4034 be unique. Back references use the syntax (?P=name) and recursive calls use
4035 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4036 numbers. The function pcre_fullinfo() can be used after compilation to extract
4037 a name/number map. There are three relevant calls:
4038
4039 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4040 PCRE_INFO_NAMECOUNT yields the number of entries
4041 PCRE_INFO_NAMETABLE yields a pointer to the map.
4042
4043 The map is a vector of fixed-size entries. The size of each entry depends on
4044 the length of the longest name used. The first two bytes of each entry are the
4045 group number, most significant byte first. There follows the corresponding
4046 name, zero terminated. The names are in alphabetical order.
4047
4048 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4049 case instead of 255. Making it the same both with and without UTF-8 support
4050 means that the same test output works with both.
4051
4052 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4053 calling malloc() with a zero argument.
4054
4055 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4056 optimization. I've improved things by keeping a bitmap of backreferences with
4057 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4058 fact referenced, the optimization can be applied. It is unlikely that a
4059 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4060 the match to follow \n) will appear inside brackets with a number greater than
4061 31, but if it does, any back reference > 31 suppresses the optimization.
4062
4063 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4064 of disabling numbered capturing parentheses. Any opening parenthesis that is
4065 not followed by ? behaves as if it were followed by ?: but named parentheses
4066 can still be used for capturing (and they will acquire numbers in the usual
4067 way).
4068
4069 41. Redesigned the return codes from the match() function into yes/no/error so
4070 that errors can be passed back from deep inside the nested calls. A malloc
4071 failure while inside a recursive subpattern call now causes the
4072 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4073
4074 42. It is now possible to set a limit on the number of times the match()
4075 function is called in a call to pcre_exec(). This facility makes it possible to
4076 limit the amount of recursion and backtracking, though not in a directly
4077 obvious way, because the match() function is used in a number of different
4078 circumstances. The count starts from zero for each position in the subject
4079 string (for non-anchored patterns). The default limit is, for compatibility, a
4080 large number, namely 10 000 000. You can change this in two ways:
4081
4082 (a) When configuring PCRE before making, you can use --with-match-limit=n
4083 to set a default value for the compiled library.
4084
4085 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4086 a different value is set. See 45 below.
4087
4088 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4089
4090 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4091 of things that can be changed at compile time. The first argument specifies
4092 what is wanted and the second points to where the information is to be placed.
4093 The current list of available information is:
4094
4095 PCRE_CONFIG_UTF8
4096
4097 The output is an integer that is set to one if UTF-8 support is available;
4098 otherwise it is set to zero.
4099
4100 PCRE_CONFIG_NEWLINE
4101
4102 The output is an integer that it set to the value of the code that is used for
4103 newline. It is either LF (10) or CR (13).
4104
4105 PCRE_CONFIG_LINK_SIZE
4106
4107 The output is an integer that contains the number of bytes used for internal
4108 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4109
4110 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4111
4112 The output is an integer that contains the threshold above which the POSIX
4113 interface uses malloc() for output vectors. See item 31 above.
4114
4115 PCRE_CONFIG_MATCH_LIMIT
4116
4117 The output is an unsigned integer that contains the default limit of the number
4118 of match() calls in a pcre_exec() execution. See 42 above.
4119
4120 44. pcretest has been upgraded by the addition of the -C option. This causes it
4121 to extract all the available output from the new pcre_config() function, and to
4122 output it. The program then exits immediately.
4123
4124 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4125 order to support additional features. One way would have been to define
4126 pcre_exec2() (for example) with extra arguments, but this would not have been
4127 extensible, and would also have required all calls to the original function to
4128 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4129 is used for passing in "extra" data from pcre_study().
4130
4131 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4132 contains the following fields:
4133
4134 flags a bitmap indicating which of the following fields are set
4135 study_data opaque data from pcre_study()
4136 match_limit a way of specifying a limit on match() calls for a specific
4137 call to pcre_exec()
4138 callout_data data for callouts (see 49 below)
4139
4140 The flag bits are also defined in pcre.h, and are
4141
4142 PCRE_EXTRA_STUDY_DATA
4143 PCRE_EXTRA_MATCH_LIMIT
4144 PCRE_EXTRA_CALLOUT_DATA
4145
4146 The pcre_study() function now returns one of these new pcre_extra blocks, with
4147 the actual study data pointed to by the study_data field, and the
4148 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4149 before. That is, this change is entirely upwards-compatible and requires no
4150 change to existing code.
4151
4152 If you want to pass in additional data to pcre_exec(), you can either place it
4153 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4154 block.
4155
4156 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4157 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4158 times with different match limits, until it finds the minimum value needed for
4159 pcre_exec() to complete. The value is then output. This can be instructive; for
4160 most simple matches the number is quite small, but for pathological cases it
4161 gets very large very quickly.
4162
4163 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4164 returns the size of the data block pointed to by the study_data field in a
4165 pcre_extra block, that is, the value that was passed as the argument to
4166 pcre_malloc() when PCRE was getting memory in which to place the information
4167 created by pcre_study(). The fourth argument should point to a size_t variable.
4168 pcretest has been extended so that this information is shown after a successful
4169 pcre_study() call when information about the compiled regex is being displayed.
4170
4171 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4172 because what follows is always an absolute path. (Later: it turns out that this
4173 is more than cosmetic for MinGW, because it doesn't like empty path
4174 components.)
4175
4176 49. Some changes have been made to the callout feature (see 28 above):
4177
4178 (i) A callout function now has three choices for what it returns:
4179
4180 0 => success, carry on matching
4181 > 0 => failure at this point, but backtrack if possible
4182 < 0 => serious error, return this value from pcre_exec()
4183
4184 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4185 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4186 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4187 use by callout functions. It will never be used by PCRE itself.
4188
4189 (ii) The pcre_extra structure (see 45 above) has a void * field called
4190 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4191 pcre_callout_block structure has a field of the same name. The contents of
4192 the field passed in the pcre_extra structure are passed to the callout
4193 function in the corresponding field in the callout block. This makes it
4194 easier to use the same callout-containing regex from multiple threads. For
4195 testing, the pcretest program has a new data escape
4196
4197 \C*n pass the number n (may be negative) as callout_data
4198
4199 If the callout function in pcretest receives a non-zero value as
4200 callout_data, it returns that value.
4201
4202 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4203 there were some redundant $(CFLAGS) in commands that are now specified as
4204 $(LINK), which already includes $(CFLAGS).
4205
4206 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4207 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4208 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4209 one-byte characters throughout. Note that case-insensitive matching applies
4210 only to characters whose values are less than 256. PCRE doesn't support the
4211 notion of cases for higher-valued characters.
4212
4213 (i) A character class whose characters are all within 0-255 is handled as
4214 a bit map, and the map is inverted for negative classes. Previously, a
4215 character > 255 always failed to match such a class; however it should
4216 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4217
4218 (ii) A negated character class with a single character < 255 is coded as
4219 "not this character" (OP_NOT). This wasn't working properly when the test
4220 character was multibyte, either singly or repeated.
4221
4222 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4223 mode, for example: \x{100}{2,3}.
4224
4225 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4226 singly or repeated) now correctly test multibyte characters. However,
4227 PCRE doesn't recognize any characters with values greater than 255 as
4228 digits, spaces, or word characters. Such characters always match \D, \S,
4229 and \W, and never match \d, \s, or \w.
4230
4231 (v) Classes may now contain characters and character ranges with values
4232 greater than 255. For example: [ab\x{100}-\x{400}].
4233
4234 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4235 PCRE in UTF-8 mode.
4236
4237 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4238 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4239 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4240 value.)
4241
4242 53. The single man page has become too large. I have therefore split it up into
4243 a number of separate man pages. These also give rise to individual HTML pages;
4244 these are now put in a separate directory, and there is an index.html page that
4245 lists them all. Some hyperlinking between the pages has been installed.
4246
4247 54. Added convenience functions for handling named capturing parentheses.
4248
4249 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4250 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4251 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4252 are faulted.
4253
4254 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4255 calling configure. These values are used when compiling the dftables.c program
4256 which is run to generate the source of the default character tables. They
4257 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4258 you will need to set these values.
4259
4260 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4261
4262
4263 Version 3.9 02-Jan-02
4264 ---------------------
4265
4266 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4267
4268 2. If --disable-static was given, the building process failed when trying to
4269 build pcretest and pcregrep. (For some reason it was using libtool to compile
4270 them, which is not right, as they aren't part of the library.)
4271
4272
4273 Version 3.8 18-Dec-01
4274 ---------------------
4275
4276 1. The experimental UTF-8 code was completely screwed up. It was packing the
4277 bytes in the wrong order. How dumb can you get?
4278
4279
4280 Version 3.7 29-Oct-01
4281 ---------------------
4282
4283 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4284 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4285 this didn't happen under Solaris 8, where I normally test things.
4286
4287 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4288 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4289 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4290 this fix an hour or so after the initial 3.7 release.)
4291
4292
4293 Version 3.6 23-Oct-01
4294 ---------------------
4295
4296 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4297 offsets passed as NULL with zero offset count.
4298
4299 2. The config.guess and config.sub files had not been updated when I moved to
4300 the latest autoconf.
4301
4302
4303 Version 3.5 15-Aug-01
4304 ---------------------
4305
4306 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4307 had been forgotten.
4308
4309 2. By using declared but undefined structures, we can avoid using "void"
4310 definitions in pcre.h while keeping the internal definitions of the structures
4311 private.
4312
4313 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4314 user point of view, this means that both static and shared libraries are built
4315 by default, but this can be individually controlled. More of the work of
4316 handling this static/shared cases is now inside libtool instead of PCRE's make
4317 file.
4318
4319 4. The pcretest utility is now installed along with pcregrep because it is
4320 useful for users (to test regexs) and by doing this, it automatically gets
4321 relinked by libtool. The documentation has been turned into a man page, so
4322 there are now .1, .txt, and .html versions in /doc.
4323
4324 5. Upgrades to pcregrep:
4325 (i) Added long-form option names like gnu grep.
4326 (ii) Added --help to list all options with an explanatory phrase.
4327 (iii) Added -r, --recursive to recurse into sub-directories.
4328 (iv) Added -f, --file to read patterns from a file.
4329
4330 6. pcre_exec() was referring to its "code" argument before testing that
4331 argument for NULL (and giving an error if it was NULL).
4332
4333 7. Upgraded Makefile.in to allow for compiling in a different directory from
4334 the source directory.
4335
4336 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4337 options bits, the pointer it was passed was to an int instead of to an unsigned
4338 long int. This mattered only on 64-bit systems.
4339
4340 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4341 generated) instead of pcre.in, which it its source. Also made the same change
4342 in several of the .c files.
4343
4344 10. A new release of gcc defines printf() as a macro, which broke pcretest
4345 because it had an ifdef in the middle of a string argument for printf(). Fixed
4346 by using separate calls to printf().
4347
4348 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4349 script, to force use of CR or LF instead of \n in the source. On non-Unix
4350 systems, the value can be set in config.h.
4351
4352 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4353 absolute limit. Changed the text of the error message to make this clear, and
4354 likewise updated the man page.
4355
4356 13. The limit of 99 on the number of capturing subpatterns has been removed.
4357 The new limit is 65535, which I hope will not be a "real" limit.
4358
4359
4360 Version 3.4 22-Aug-00
4361 ---------------------
4362
4363 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4364
4365 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4366
4367
4368 Version 3.3 01-Aug-00
4369 ---------------------
4370
4371 1. If an octal character was given, but the value was greater than \377, it
4372 was not getting masked to the least significant bits, as documented. This could
4373 lead to crashes in some systems.
4374
4375 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4376 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4377
4378 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4379 These just pass their arguments on to (pcre_free)(), but they are provided
4380 because some uses of PCRE bind it to non-C systems that can call its functions,
4381 but cannot call free() or pcre_free() directly.
4382
4383 4. Add "make test" as a synonym for "make check". Corrected some comments in
4384 the Makefile.
4385
4386 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4387 Makefile.
4388
4389 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4390 command called pgrep for grepping around the active processes.
4391
4392 7. Added the beginnings of support for UTF-8 character strings.
4393
4394 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4395 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4396 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4397 out for the ar command.)
4398
4399
4400 Version 3.2 12-May-00
4401 ---------------------
4402
4403 This is purely a bug fixing release.
4404
4405 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4406 of ZA. This was just one example of several cases that could provoke this bug,
4407 which was introduced by change 9 of version 2.00. The code for breaking
4408 infinite loops after an iteration that matches an empty string was't working
4409 correctly.
4410
4411 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4412 when matched against abbab (for example). After matching an empty string, it
4413 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4414 caused it to match further down the string than it should.
4415
4416 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4417 was there because it doesn't seem to be needed, and it causes trouble on some
4418 systems, as it is not a Standard C header. It has been removed.
4419
4420 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4421 were reported on the Macintosh. The changes were from
4422
4423 while ((c = *(++ptr)) != 0 && c != '\n');
4424 to
4425 while ((c = *(++ptr)) != 0 && c != '\n') ;
4426
4427 Totally extraordinary, but if that's what it takes...
4428
4429 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4430 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4431 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4432 assumes the way PCRE uses memmove() (always moving upwards).
4433
4434 6. PCRE is being used in one environment where strchr() is not available. There
4435 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4436 faster code anyway.
4437
4438
4439 Version 3.1 09-Feb-00
4440 ---------------------
4441
4442 The only change in this release is the fixing of some bugs in Makefile.in for
4443 the "install" target:
4444
4445 (1) It was failing to install pcreposix.h.
4446
4447 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4448
4449
4450 Version 3.0 01-Feb-00
4451 ---------------------
4452
4453 1. Add support for the /+ modifier to perltest (to output $` like it does in
4454 pcretest).
4455
4456 2. Add support for the /g modifier to perltest.
4457
4458 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4459 matches null strings.
4460
4461 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4462 pattern. Perl treats empty patterns specially - it reuses the most recent
4463 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4464 effect.
4465
4466 5. The POSIX interface was broken in that it was just handing over the POSIX
4467 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4468 required a bigger vector, with some working space on the end. This means that
4469 the POSIX wrapper now has to get and free some memory, and copy the results.
4470
4471 6. Added some simple autoconf support, placing the test data and the
4472 documentation in separate directories, re-organizing some of the
4473 information files, and making it build pcre-config (a GNU standard). Also added
4474 libtool support for building PCRE as a shared library, which is now the
4475 default.
4476
4477 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4478 09 are not valid octal constants. Single digits will be used for minor values
4479 less than 10.
4480
4481 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4482 existing programs that set these in the POSIX interface can use PCRE without
4483 modification.
4484
4485 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4486 return all that pcre_info() returns, plus additional data. The pcre_info()
4487 function is retained for compatibility, but is considered to be obsolete.
4488
4489 10. Added experimental recursion feature (?R) to handle one common case that
4490 Perl 5.6 will be able to do with (?p{...}).
4491
4492 11. Added support for POSIX character classes like [:alpha:], which Perl is
4493 adopting.
4494
4495
4496 Version 2.08 31-Aug-99
4497 ----------------------
4498
4499 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4500 trying to match at the startoffset position, but instead was moving forward to
4501 the next newline as if a previous match had failed.
4502
4503 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4504 and could get into a loop if a null string was matched other than at the start
4505 of the subject.
4506
4507 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4508 be distinguished at compile time, and for completeness also added PCRE_DATE.
4509
4510 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4511 in GnuWin32 environments.
4512
4513
4514 Version 2.07 29-Jul-99
4515 ----------------------
4516
4517 1. The documentation is now supplied in plain text form and HTML as well as in
4518 the form of man page sources.
4519
4520 2. C++ compilers don't like assigning (void *) values to other pointer types.
4521 In particular this affects malloc(). Although there is no problem in Standard
4522 C, I've put in casts to keep C++ compilers happy.
4523
4524 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4525 should be (const char *).
4526
4527 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4528 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4529 However, I haven't made this a standard facility. The documentation doesn't
4530 mention it, and the Makefile doesn't support it.
4531
4532 5. The Makefile now contains an "install" target, with editable destinations at
4533 the top of the file. The pcretest program is not installed.
4534
4535 6. pgrep -V now gives the PCRE version number and date.
4536
4537 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4538 causing the entire string to be ignored, instead of just the last character.
4539
4540 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4541 non-matching string, it can take a very, very long time, even for strings of
4542 quite modest length, because of the nested recursion. PCRE now does better in
4543 some of these cases. It does this by remembering the last required literal
4544 character in the pattern, and pre-searching the subject to ensure it is present
4545 before running the real match. In other words, it applies a heuristic to detect
4546 some types of certain failure quickly, and in the above example, if presented
4547 with a string that has no trailing " it gives "no match" very quickly.
4548
4549 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4550 other alternatives are tried instead.
4551
4552
4553 Version 2.06 09-Jun-99
4554 ----------------------
4555
4556 1. Change pcretest's output for amount of store used to show just the code
4557 space, because the remainder (the data block) varies in size between 32-bit and
4558 64-bit systems.
4559
4560 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4561 start matching at. This allows lookbehinds to work when searching for multiple
4562 occurrences in a string.
4563
4564 3. Added additional options to pcretest for testing multiple occurrences:
4565
4566 /+ outputs the rest of the string that follows a match
4567 /g loops for multiple occurrences, using the new startoffset argument
4568 /G loops for multiple occurrences by passing an incremented pointer
4569
4570 4. PCRE wasn't doing the "first character" optimization for patterns starting
4571 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4572 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4573 the letter 'x'. On long subject strings, this gives a significant speed-up.
4574
4575
4576 Version 2.05 21-Apr-99
4577 ----------------------
4578
4579 1. Changed the type of magic_number from int to long int so that it works
4580 properly on 16-bit systems.
4581
4582 2. Fixed a bug which caused patterns starting with .* not to work correctly
4583 when the subject string contained newline characters. PCRE was assuming
4584 anchoring for such patterns in all cases, which is not correct because .* will
4585 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4586 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4587 must be retried after every newline in the subject.
4588
4589
4590 Version 2.04 18-Feb-99
4591 ----------------------
4592
4593 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4594 computation of the store needed to hold the pattern was incorrect (too large).
4595 If such patterns were nested a few deep, this could multiply and become a real
4596 problem.
4597
4598 2. Added /M option to pcretest to show the memory requirement of a specific
4599 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4600
4601 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4602 compiled in such a way that the backtracking after subsequent failure was
4603 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4604 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4605
4606
4607 Version 2.03 02-Feb-99
4608 ----------------------
4609
4610 1. Fixed typo and small mistake in man page.
4611
4612 2. Added 4th condition (GPL supersedes if conflict) and created separate
4613 LICENCE file containing the conditions.
4614
4615 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4616 Perl, that is the internal \ allows the delimiter to be included in the
4617 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4618 the final delimiter, add \ to the end of the pattern (to test the error).
4619
4620 4. Added the convenience functions for extracting substrings after a successful
4621 match. Updated pcretest to make it able to test these functions.
4622
4623
4624 Version 2.02 14-Jan-99
4625 ----------------------
4626
4627 1. Initialized the working variables associated with each extraction so that
4628 their saving and restoring doesn't refer to uninitialized store.
4629
4630 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4631 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4632 fix the problem.
4633
4634 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4635 calls, and wasn't printing the correct value for compiling calls. Increased the
4636 default value of LOOPREPEAT, and the number of significant figures in the
4637 times.
4638
4639 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4640
4641 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4642 a building problem on Windows NT with a FAT file system.
4643
4644
4645 Version 2.01 21-Oct-98
4646 ----------------------
4647
4648 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4649 to character tables built by pcre_maketables() in the current locale. If NULL
4650 is passed, the default tables are used.
4651
4652
4653 Version 2.00 24-Sep-98
4654 ----------------------
4655
4656 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4657 it any more.
4658
4659 2. Allow quantification of (?>) groups, and make it work correctly.
4660
4661 3. The first character computation wasn't working for (?>) groups.
4662
4663 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4664 end of the subject) and add 5.005's \z, which really does match only at the
4665 very end of the subject.
4666
4667 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4668
4669 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4670 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4671 localized options. All options to pcre_study() were also removed.
4672
4673 7. Add other new features from 5.005:
4674
4675 $(?<= positive lookbehind
4676 $(?<! negative lookbehind
4677 (?imsx-imsx) added the unsetting capability
4678 such a setting is global if at outer level; local otherwise
4679 (?imsx-imsx:) non-capturing groups with option setting
4680 (?(cond)re|re) conditional pattern matching
4681
4682 A backreference to itself in a repeated group matches the previous
4683 captured string.
4684
4685 8. General tidying up of studying (both automatic and via "study")
4686 consequential on the addition of new assertions.
4687
4688 9. As in 5.005, unlimited repeated groups that could match an empty substring
4689 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4690 runtime if any iteration does actually match an empty substring.
4691
4692 10. Include the RunTest script in the distribution.
4693
4694 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4695 discrepancies, some of which were old and were also with respect to 5.004. They
4696 have now been fixed.
4697
4698
4699 Version 1.09 28-Apr-98
4700 ----------------------
4701
4702 1. A negated single character class followed by a quantifier with a minimum
4703 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4704 program crashes, or just wrong answers. This did not apply to negated classes
4705 containing more than one character, or to minima other than one.
4706
4707
4708 Version 1.08 27-Mar-98
4709 ----------------------
4710
4711 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4712
4713 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4714 latter must appear before anything that relies on it in the pattern.
4715
4716
4717 Version 1.07 16-Feb-98
4718 ----------------------
4719
4720 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4721 repeat of a potentially empty string).
4722
4723
4724 Version 1.06 23-Jan-98
4725 ----------------------
4726
4727 1. Added Markus Oberhumer's little patches for C++.
4728
4729 2. Literal strings longer than 255 characters were broken.
4730
4731
4732 Version 1.05 23-Dec-97
4733 ----------------------
4734
4735 1. Negated character classes containing more than one character were failing if
4736 PCRE_CASELESS was set at run time.
4737
4738
4739 Version 1.04 19-Dec-97
4740 ----------------------
4741
4742 1. Corrected the man page, where some "const" qualifiers had been omitted.
4743
4744 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4745 input syntax.
4746
4747 3. Fixed memory leak which occurred when a regex with back references was
4748 matched with an offsets vector that wasn't big enough. The temporary memory
4749 that is used in this case wasn't being freed if the match failed.
4750
4751 4. Tidied pcretest to ensure it frees memory that it gets.
4752
4753 5. Temporary memory was being obtained in the case where the passed offsets
4754 vector was exactly big enough.
4755
4756 6. Corrected definition of offsetof() from change 5 below.
4757
4758 7. I had screwed up change 6 below and broken the rules for the use of
4759 setjmp(). Now fixed.
4760
4761
4762 Version 1.03 18-Dec-97
4763 ----------------------
4764
4765 1. A erroneous regex with a missing opening parenthesis was correctly
4766 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4767 on some systems.
4768
4769 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4770 it was reported that one broken compiler failed on the former because "code" is
4771 also an independent variable.
4772
4773 3. The erroneous regex a[]b caused an array overrun reference.
4774
4775 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4776 fail on data ending with that character. (It was going on too far, and checking
4777 the next character, typically a binary zero.) This was specific to the
4778 optimized code for single-character negative classes.
4779
4780 5. Added a contributed patch from the TIN world which does the following:
4781
4782 + Add an undef for memmove, in case the the system defines a macro for it.
4783
4784 + Add a definition of offsetof(), in case there isn't one. (I don't know
4785 the reason behind this - offsetof() is part of the ANSI standard - but
4786 it does no harm).
4787
4788 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4789 most of the places where whitespace preceded '#'. I have given up and
4790 allowed the remaining 2 cases to be at the margin.
4791
4792 + Rename some variables in pcre to eliminate shadowing. This seems very
4793 pedantic, but does no harm, of course.
4794
4795 6. Moved the call to setjmp() into its own function, to get rid of warnings
4796 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4797
4798 7. Constructs such as \d{8,} were compiling into the equivalent of
4799 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4800 outcome, but in this particular case used more store than had been allocated,
4801 which caused the bug to be discovered because it threw up an internal error.
4802
4803 8. The debugging code in both pcre and pcretest for outputting the compiled
4804 form of a regex was going wrong in the case of back references followed by
4805 curly-bracketed repeats.
4806
4807
4808 Version 1.02 12-Dec-97
4809 ----------------------
4810
4811 1. Typos in pcre.3 and comments in the source fixed.
4812
4813 2. Applied a contributed patch to get rid of places where it used to remove
4814 'const' from variables, and fixed some signed/unsigned and uninitialized
4815 variable warnings.
4816
4817 3. Added the "runtest" target to Makefile.
4818
4819 4. Set default compiler flag to -O2 rather than just -O.
4820
4821
4822 Version 1.01 19-Nov-97
4823 ----------------------
4824
4825 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4826 like /([ab]*)*/, that is, for classes with more than one character in them.
4827
4828 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4829 as /((?>a*))*/ (a PCRE_EXTRA facility).
4830
4831
4832 Version 1.00 18-Nov-97
4833 ----------------------
4834
4835 1. Added compile-time macros to support systems such as SunOS4 which don't have
4836 memmove() or strerror() but have other things that can be used instead.
4837
4838 2. Arranged that "make clean" removes the executables.
4839
4840
4841 Version 0.99 27-Oct-97
4842 ----------------------
4843
4844 1. Fixed bug in code for optimizing classes with only one character. It was
4845 initializing a 32-byte map regardless, which could cause it to run off the end
4846 of the memory it had got.
4847
4848 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4849
4850
4851 Version 0.98 22-Oct-97
4852 ----------------------
4853
4854 1. Fixed bug in code for handling temporary memory usage when there are more
4855 back references than supplied space in the ovector. This could cause segfaults.
4856
4857
4858 Version 0.97 21-Oct-97
4859 ----------------------
4860
4861 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4862
4863 2. Optimized negated single characters not to use a bit map.
4864
4865 3. Brought error texts together as macro definitions; clarified some of them;
4866 fixed one that was wrong - it said "range out of order" when it meant "invalid
4867 escape sequence".
4868
4869 4. Changed some char * arguments to const char *.
4870
4871 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4872
4873 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4874 pcretest.
4875
4876
4877 Version 0.96 16-Oct-97
4878 ----------------------
4879
4880 1. Added a simple "pgrep" utility to the distribution.
4881
4882 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4883 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4884 where "ddd" means "one or more decimal digits".
4885
4886 3. Fixed serious bug. If a pattern had a back reference, but the call to
4887 pcre_exec() didn't supply a large enough ovector to record the related
4888 identifying subpattern, the match always failed. PCRE now remembers the number
4889 of the largest back reference, and gets some temporary memory in which to save
4890 the offsets during matching if necessary, in order to ensure that
4891 backreferences always work.
4892
4893 4. Increased the compatibility with Perl in a number of ways:
4894
4895 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4896 to request this handling. The option can be set at compile or exec time.
4897
4898 (b) $ matches before a terminating newline by default; an option
4899 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4900 mode). The option can be set at compile or exec time.
4901
4902 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4903 the same as Perl's. If the decimal number it represents is less than 10
4904 or there aren't that many previous left capturing parentheses, an octal
4905 escape is read. Inside a character class, it's always an octal escape,
4906 even if it is a single digit.
4907
4908 (d) An escaped but undefined alphabetic character is taken as a literal,
4909 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4910 escapes.
4911
4912 (e) {0} is now permitted. (The previous item is removed from the compiled
4913 pattern).
4914
4915 5. Changed all the names of code files so that the basic parts are no longer
4916 than 10 characters, and abolished the teeny "globals.c" file.
4917
4918 6. Changed the handling of character classes; they are now done with a 32-byte
4919 bit map always.
4920
4921 7. Added the -d and /D options to pcretest to make it possible to look at the
4922 internals of compilation without having to recompile pcre.
4923
4924
4925 Version 0.95 23-Sep-97
4926 ----------------------
4927
4928 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4929 \x20 at the start of a run of normal characters. These were being treated as
4930 real characters, instead of the source characters being re-checked.
4931
4932
4933 Version 0.94 18-Sep-97
4934 ----------------------
4935
4936 1. The functions are now thread-safe, with the caveat that the global variables
4937 containing pointers to malloc() and free() or alternative functions are the
4938 same for all threads.
4939
4940 2. Get pcre_study() to generate a bitmap of initial characters for non-
4941 anchored patterns when this is possible, and use it if passed to pcre_exec().
4942
4943
4944 Version 0.93 15-Sep-97
4945 ----------------------
4946
4947 1. /(b)|(:+)/ was computing an incorrect first character.
4948
4949 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4950 but not actually doing anything yet.
4951
4952 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4953 as Perl does (e.g. [-az] or [az-]).
4954
4955 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4956 all possible positions.
4957
4958 5. Split up into different modules to avoid including unneeded functions in a
4959 compiled binary. However, compile and exec are still in one module. The "study"
4960 function is split off.
4961
4962 6. The character tables are now in a separate module whose source is generated
4963 by an auxiliary program - but can then be edited by hand if required. There are
4964 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4965 toupper() in the code.
4966
4967 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4968 make them global. Abolish the function for setting them, as the caller can now
4969 set them directly.
4970
4971
4972 Version 0.92 11-Sep-97
4973 ----------------------
4974
4975 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4976 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4977
4978 2. Caseless matching was not working in character classes if the characters in
4979 the pattern were in upper case.
4980
4981 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4982
4983 4. Make PCRE_ANCHORED public and accept as a compile option.
4984
4985 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4986 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4987 pass them.
4988
4989 6. Give an error if bad option bits passed at compile or run time.
4990
4991 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4992 pcretest to cause it to pass that flag.
4993
4994 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4995 options, and the first character, if set.
4996
4997 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4998
4999
5000 Version 0.91 10-Sep-97
5001 ----------------------
5002
5003 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5004 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5005
5006 2. PCRE was looping on encountering an indefinitely repeated back reference to
5007 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5008 Perl does - treats the match as successful.
5009
5010 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5