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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1245 - (show annotations)
Sat Feb 9 11:30:51 2013 UTC (6 years, 9 months ago) by zherczeg
File size: 241655 byte(s)
Error occurred while calculating annotation data.
Adding experimental support for callouts in JIT.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 WARNING: Callouts are not release ready! <- this line needs to be removed when it is.
47
48
49 Version 8.32 30-November-2012
50 -----------------------------
51
52 1. Improved JIT compiler optimizations for first character search and single
53 character iterators.
54
55 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
56 Patch by Daniel Richard G.
57
58 3. Single character iterator optimizations in the JIT compiler.
59
60 4. Improved JIT compiler optimizations for character ranges.
61
62 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
63 Reported by Giuseppe D'Angelo.
64
65 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
66 of a line, was being set incorrectly in cases where .* appeared inside
67 atomic brackets at the start of a pattern, or where there was a subsequent
68 *PRUNE or *SKIP.
69
70 7. Improved instruction cache flush for POWER/PowerPC.
71 Patch by Daniel Richard G.
72
73 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
74 grep:
75
76 (a) There is now no limit to the number of patterns to be matched.
77
78 (b) An error is given if a pattern is too long.
79
80 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
81 are now supported.
82
83 (d) --exclude-from and --include-from (multiple use) have been added.
84
85 (e) Exclusions and inclusions now apply to all files and directories, not
86 just to those obtained from scanning a directory recursively.
87
88 (f) Multiple uses of -f and --file-list are now supported.
89
90 (g) In a Windows environment, the default for -d has been changed from
91 "read" (the GNU grep default) to "skip", because otherwise the presence
92 of a directory in the file list provokes an error.
93
94 (h) The documentation has been revised and clarified in places.
95
96 9. Improve the matching speed of capturing brackets.
97
98 10. Changed the meaning of \X so that it now matches a Unicode extended
99 grapheme cluster.
100
101 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
102 out POSIX threads when JIT support is configured.
103
104 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
105
106 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
107 structure could go wrong in environments where size_t is not the same size
108 as int.
109
110 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
111 set.
112
113 15. The EBCDIC support had decayed; later updates to the code had included
114 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
115 general tidy up of EBCDIC-related issues, and the documentation was also
116 not quite right. There is now a test that can be run on ASCII systems to
117 check some of the EBCDIC-related things (but is it not a full test).
118
119 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
120 in a small tidy to the code.
121
122 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
123
124 18. If the --only-matching (-o) option in pcregrep is specified multiple
125 times, each one causes appropriate output. For example, -o1 -o2 outputs the
126 substrings matched by the 1st and 2nd capturing parentheses. A separating
127 string can be specified by --om-separator (default empty).
128
129 19. Improving the first n character searches.
130
131 20. Turn case lists for horizontal and vertical white space into macros so that
132 they are defined only once.
133
134 21. This set of changes together give more compatible Unicode case-folding
135 behaviour for characters that have more than one other case when UCP
136 support is available.
137
138 (a) The Unicode property table now has offsets into a new table of sets of
139 three or more characters that are case-equivalent. The MultiStage2.py
140 script that generates these tables (the pcre_ucd.c file) now scans
141 CaseFolding.txt instead of UnicodeData.txt for character case
142 information.
143
144 (b) The code for adding characters or ranges of characters to a character
145 class has been abstracted into a generalized function that also handles
146 case-independence. In UTF-mode with UCP support, this uses the new data
147 to handle characters with more than one other case.
148
149 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
150 whose other case is greater than 256 are now correctly matched
151 caselessly. Previously, the high codepoint matched the low one, but not
152 vice versa.
153
154 (d) The processing of \h, \H, \v, and \ in character classes now makes use
155 of the new class addition function, using character lists defined as
156 macros alongside the case definitions of 20 above.
157
158 (e) Caseless back references now work with characters that have more than
159 one other case.
160
161 (f) General caseless matching of characters with more than one other case
162 is supported.
163
164 22. Unicode character properties were updated from Unicode 6.2.0
165
166 23. Improved CMake support under Windows. Patch by Daniel Richard G.
167
168 24. Add support for 32-bit character strings, and UTF-32
169
170 25. Major JIT compiler update (code refactoring and bugfixing).
171 Experimental Sparc 32 support is added.
172
173 26. Applied a modified version of Daniel Richard G's patch to create
174 pcre.h.generic and config.h.generic by "make" instead of in the
175 PrepareRelease script.
176
177 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
178 pcre_compile.c when checking for a zero character.
179
180 28. Introducing a native interface for JIT. Through this interface, the compiled
181 machine code can be directly executed. The purpose of this interface is to
182 provide fast pattern matching, so several sanity checks are not performed.
183 However, feature tests are still performed. The new interface provides
184 1.4x speedup compared to the old one.
185
186 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
187 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
188 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
189
190 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
191 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
192 the "old" RFC 2279). Instead, it ended up passing a negative length to
193 pcre_exec().
194
195 31. Add support for GCC's visibility feature to hide internal functions.
196
197 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
198 "unknown -C option" after outputting 0 or 1.
199
200 33. There is now support for generating a code coverage report for the test
201 suite in environments where gcc is the compiler and lcov is installed. This
202 is mainly for the benefit of the developers.
203
204 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
205 unaddressable using valgrind annotations, allowing valgrind to detect
206 invalid memory accesses. This is mainly for the benefit of the developers.
207
208 25. (*UTF) can now be used to start a pattern in any of the three libraries.
209
210 26. Give configure error if --enable-cpp but no C++ compiler found.
211
212
213 Version 8.31 06-July-2012
214 -------------------------
215
216 1. Fixing a wrong JIT test case and some compiler warnings.
217
218 2. Removed a bashism from the RunTest script.
219
220 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
221 to unsigned type, result still unsigned" that was given by an MS compiler
222 on encountering the code "-sizeof(xxx)".
223
224 4. Partial matching support is added to the JIT compiler.
225
226 5. Fixed several bugs concerned with partial matching of items that consist
227 of more than one character:
228
229 (a) /^(..)\1/ did not partially match "aba" because checking references was
230 done on an "all or nothing" basis. This also applied to repeated
231 references.
232
233 (b) \R did not give a hard partial match if \r was found at the end of the
234 subject.
235
236 (c) \X did not give a hard partial match after matching one or more
237 characters at the end of the subject.
238
239 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
240 a partial match for the string "\r".
241
242 (e) When newline was set to CRLF, the metacharacter "." did not recognize
243 a partial match for a CR character at the end of the subject string.
244
245 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
246 running pcretest, the text "(JIT)" added to the output whenever JIT is
247 actually used to run the match.
248
249 7. Individual JIT compile options can be set in pcretest by following -s+[+]
250 or /S+[+] with a digit between 1 and 7.
251
252 8. OP_NOT now supports any UTF character not just single-byte ones.
253
254 9. (*MARK) control verb is now supported by the JIT compiler.
255
256 10. The command "./RunTest list" lists the available tests without actually
257 running any of them. (Because I keep forgetting what they all are.)
258
259 11. Add PCRE_INFO_MAXLOOKBEHIND.
260
261 12. Applied a (slightly modified) user-supplied patch that improves performance
262 when the heap is used for recursion (compiled with --disable-stack-for-
263 recursion). Instead of malloc and free for each heap frame each time a
264 logical recursion happens, frames are retained on a chain and re-used where
265 possible. This sometimes gives as much as 30% improvement.
266
267 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
268 call.
269
270 14. As documented, (*COMMIT) is now confined to within a positive assertion.
271
272 15. It is now possible to link pcretest with libedit as an alternative to
273 libreadline.
274
275 16. (*COMMIT) control verb is now supported by the JIT compiler.
276
277 17. The Unicode data tables have been updated to Unicode 6.1.0.
278
279 18. Added --file-list option to pcregrep.
280
281 19. Added binary file support to pcregrep, including the -a, --binary-files,
282 -I, and --text options.
283
284 20. The madvise function is renamed for posix_madvise for QNX compatibility
285 reasons. Fixed by Giuseppe D'Angelo.
286
287 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
288
289 22. Changed the option for creating symbolic links for 16-bit man pages from
290 -s to -sf so that re-installing does not cause issues.
291
292 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
293
294 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
295 matches in certain environments (the workspace was not being correctly
296 retained). Also added to pcre_dfa_exec() a simple plausibility check on
297 some of the workspace data at the beginning of a restart.
298
299 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
300 was not doing so when it should - probably a typo introduced by SVN 528
301 (change 8.10/14).
302
303 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
304 \w+ when the character tables indicated that \x{c4} was a word character.
305 There were several related cases, all because the tests for doing a table
306 lookup were testing for characters less than 127 instead of 255.
307
308 27. If a pattern contains capturing parentheses that are not used in a match,
309 their slots in the ovector are set to -1. For those that are higher than
310 any matched groups, this happens at the end of processing. In the case when
311 there were back references that the ovector was too small to contain
312 (causing temporary malloc'd memory to be used during matching), and the
313 highest capturing number was not used, memory off the end of the ovector
314 was incorrectly being set to -1. (It was using the size of the temporary
315 memory instead of the true size.)
316
317 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
318 ovector size, it uses memory at the end of the block that it has got.
319
320 29. Check for an overlong MARK name and give an error at compile time. The
321 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
322
323 30. JIT compiler update.
324
325 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
326 Rill for the patch.
327
328 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
329
330 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
331
332 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
333 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
334 was enabled.
335
336 35. Improve JIT code generation for greedy plus quantifier.
337
338 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
339 1 to "aa" instead of to an empty string. The bug affected repeated groups
340 that could potentially match an empty string.
341
342 37. Optimizing single character iterators in JIT.
343
344 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
345 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
346 codepoints that are too big for the mode are faulted, and in a UTF mode,
347 disallowed codepoints are also faulted.
348
349 39. If PCRE was compiled with UTF support, in three places in the DFA
350 matcher there was code that should only have been obeyed in UTF mode, but
351 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
352 processing when bytes with values greater than 127 were present. In 16-bit
353 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
354 both cases the values are those that cannot be the first data item in a UTF
355 character. The three items that might have provoked this were recursions,
356 possessively repeated groups, and atomic groups.
357
358 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
359 and pcregrep, because some OS require shared objects to be explicitly
360 passed to ld, causing the link step to fail if they are not.
361
362 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
363 mode, patterns that started with \h* or \R* might be incorrectly matched.
364
365
366 Version 8.30 04-February-2012
367 -----------------------------
368
369 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
370 name is defined in ctype.h.
371
372 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
373 only in quite long subpatterns.
374
375 3. Removed the function pcre_info(), which has been obsolete and deprecated
376 since it was replaced by pcre_fullinfo() in February 2000.
377
378 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
379 match a (*MARK), and the match failed at the start of the subject, a
380 reference to memory before the start of the subject could occur. This bug
381 was introduced by fix 17 of release 8.21.
382
383 5. A reference to an unset group with zero minimum repetition was giving
384 totally wrong answers (in non-JavaScript-compatibility mode). For example,
385 /(another)?(\1?)test/ matched against "hello world test". This bug was
386 introduced in release 8.13.
387
388 6. Add support for 16-bit character strings (a large amount of work involving
389 many changes and refactorings).
390
391 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
392 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
393 from a file.
394
395 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
396 rounding is not applied in this particular case).
397
398 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
399 if they appear, or are escaped, in patterns.
400
401 10. Get rid of a number of -Wunused-but-set-variable warnings.
402
403 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
404 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
405 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
406 also returns the mark "x". This bug applied to capturing parentheses,
407 non-capturing parentheses, and atomic parentheses. It also applied to some
408 assertions.
409
410 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
411 information out of configure.ac instead of relying on pcre.h.generic, which
412 is not stored in the repository.
413
414 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
415 -lreadline.
416
417 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
418
419 15. Applied Graycode's patch to put the top-level frame on the stack rather
420 than the heap when not using the stack for recursion. This gives a
421 performance improvement in many cases when recursion is not deep.
422
423 16. Experimental code added to "pcretest -C" to output the stack frame size.
424
425
426 Version 8.21 12-Dec-2011
427 ------------------------
428
429 1. Updating the JIT compiler.
430
431 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
432 are added as well.
433
434 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
435 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
436 calling _pcre_jit_exec. Some extra comments are added.
437
438 4. (*MARK) settings inside atomic groups that do not contain any capturing
439 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
440 was introduced by change 18 for 8.20.
441
442 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
443 ECMA-262 standard.
444
445 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
446 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
447 This bug was probably introduced by change 9 of 8.13.
448
449 7. While fixing 6 above, I noticed that a number of other items were being
450 incorrectly rejected as "not fixed length". This arose partly because newer
451 opcodes had not been added to the fixed-length checking code. I have (a)
452 corrected the bug and added tests for these items, and (b) arranged for an
453 error to occur if an unknown opcode is encountered while checking for fixed
454 length instead of just assuming "not fixed length". The items that were
455 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
456 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
457 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
458
459 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
460 being incorrectly compiled and would have given unpredicatble results.
461
462 9. A possessively repeated subpattern with minimum repeat count greater than
463 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
464 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
465 the first (A) could occur when it should not.
466
467 10. Add a cast and remove a redundant test from the code.
468
469 11. JIT should use pcre_malloc/pcre_free for allocation.
470
471 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
472 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
473 is anything other than /usr, -L is still shown).
474
475 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
476
477 14. Perl does not support \N without a following name in a [] class; PCRE now
478 also gives an error.
479
480 15. If a forward reference was repeated with an upper limit of around 2000,
481 it caused the error "internal error: overran compiling workspace". The
482 maximum number of forward references (including repeats) was limited by the
483 internal workspace, and dependent on the LINK_SIZE. The code has been
484 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
485 the default depends on LINK_SIZE. There is a new upper limit (for safety)
486 of around 200,000 forward references. While doing this, I also speeded up
487 the filling in of repeated forward references.
488
489 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
490 incorrectly expecting the subject to contain another "a" after the start.
491
492 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
493 in the match, the SKIP should be ignored. This was not happening; instead
494 the SKIP was being treated as NOMATCH. For patterns such as
495 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
496 tested.
497
498 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
499 now much more compatible with Perl, in particular in cases where the result
500 is a non-match for a non-anchored pattern. For example, if
501 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
502 "m", where previously it did not return a name. A side effect of this
503 change is that for partial matches, the last encountered mark name is
504 returned, as for non matches. A number of tests that were previously not
505 Perl-compatible have been moved into the Perl-compatible test files. The
506 refactoring has had the pleasing side effect of removing one argument from
507 the match() function, thus reducing its stack requirements.
508
509 19. If the /S+ option was used in pcretest to study a pattern using JIT,
510 subsequent uses of /S (without +) incorrectly behaved like /S+.
511
512 21. Retrieve executable code size support for the JIT compiler and fixing
513 some warnings.
514
515 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
516 not work when the shorter character appeared right at the end of the
517 subject string.
518
519 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
520 systems.
521
522 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
523 output it when the /M option is used in pcretest.
524
525 25. The CheckMan script was not being included in the distribution. Also, added
526 an explicit "perl" to run Perl scripts from the PrepareRelease script
527 because this is reportedly needed in Windows.
528
529 26. If study data was being save in a file and studying had not found a set of
530 "starts with" bytes for the pattern, the data written to the file (though
531 never used) was taken from uninitialized memory and so caused valgrind to
532 complain.
533
534 27. Updated RunTest.bat as provided by Sheri Pierce.
535
536 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
537
538 29. Computation of memory usage for the table of capturing group names was
539 giving an unnecessarily large value.
540
541
542 Version 8.20 21-Oct-2011
543 ------------------------
544
545 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
546 a POSIX class. After further experiments with Perl, which convinced me that
547 Perl has bugs and confusions, a closing square bracket is no longer allowed
548 in a POSIX name. This bug also affected patterns with classes that started
549 with full stops.
550
551 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
552 captured substring, but while checking the failing first alternative,
553 substring 1 is temporarily captured. If the output vector supplied to
554 pcre_exec() was not big enough for this capture, the yield of the function
555 was still zero ("insufficient space for captured substrings"). This cannot
556 be totally fixed without adding another stack variable, which seems a lot
557 of expense for a edge case. However, I have improved the situation in cases
558 such as /(a)(b)x|abc/ matched against "abc", where the return code
559 indicates that fewer than the maximum number of slots in the ovector have
560 been set.
561
562 3. Related to (2) above: when there are more back references in a pattern than
563 slots in the output vector, pcre_exec() uses temporary memory during
564 matching, and copies in the captures as far as possible afterwards. It was
565 using the entire output vector, but this conflicts with the specification
566 that only 2/3 is used for passing back captured substrings. Now it uses
567 only the first 2/3, for compatibility. This is, of course, another edge
568 case.
569
570 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
571 main code base, and can be used by building with --enable-jit. When this is
572 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
573 runtime --no-jit option is given.
574
575 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
576 ovector, the return from the function was zero, implying that there were
577 other matches that did not fit. The correct "exactly full" value is now
578 returned.
579
580 6. If a subpattern that was called recursively or as a subroutine contained
581 (*PRUNE) or any other control that caused it to give a non-standard return,
582 invalid errors such as "Error -26 (nested recursion at the same subject
583 position)" or even infinite loops could occur.
584
585 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
586 computing the minimum length on reaching *ACCEPT, and so ended up with the
587 wrong value of 1 rather than 0. Further investigation indicates that
588 computing a minimum subject length in the presence of *ACCEPT is difficult
589 (think back references, subroutine calls), and so I have changed the code
590 so that no minimum is registered for a pattern that contains *ACCEPT.
591
592 8. If (*THEN) was present in the first (true) branch of a conditional group,
593 it was not handled as intended. [But see 16 below.]
594
595 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
596 Sheri Pierce.
597
598 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
599 the first byte in a match must be "a".
600
601 11. Change 17 for 8.13 increased the recursion depth for patterns like
602 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
603 pattern contains any instances of (*THEN). If it does not, the old
604 optimizations are restored. It would be nice to do this on a per-group
605 basis, but at the moment that is not feasible.
606
607 12. In some environments, the output of pcretest -C is CRLF terminated. This
608 broke RunTest's code that checks for the link size. A single white space
609 character after the value is now allowed for.
610
611 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
612 For "fr", it uses the Windows-specific input and output files.
613
614 14. If (*THEN) appeared in a group that was called recursively or as a
615 subroutine, it did not work as intended. [But see next item.]
616
617 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
618 pattern fragments (but not containing any | characters). If A and B are
619 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
620 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
621 D. In other words, Perl considers parentheses that do not contain any |
622 characters to be part of a surrounding alternative, whereas PCRE was
623 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
624 differently. PCRE now behaves in the same way as Perl, except in the case
625 of subroutine/recursion calls such as (?1) which have in any case always
626 been different (but PCRE had them first :-).
627
628 16. Related to 15 above: Perl does not treat the | in a conditional group as
629 creating alternatives. Such a group is treated in the same way as an
630 ordinary group without any | characters when processing (*THEN). PCRE has
631 been changed to match Perl's behaviour.
632
633 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
634 RunGrepTest script failed.
635
636 18. Change 22 for version 13 caused atomic groups to use more stack. This is
637 inevitable for groups that contain captures, but it can lead to a lot of
638 stack use in large patterns. The old behaviour has been restored for atomic
639 groups that do not contain any capturing parentheses.
640
641 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
642 suppress the check for a minimum subject length at run time. (If it was
643 given to pcre_exec() or pcre_dfa_exec() it did work.)
644
645 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
646 fail to work when decoding hex characters in data strings in EBCDIC
647 environments.
648
649 21. It appears that in at least one Mac OS environment, the isxdigit() function
650 is implemented as a macro that evaluates to its argument more than once,
651 contravening the C 90 Standard (I haven't checked a later standard). There
652 was an instance in pcretest which caused it to go wrong when processing
653 \x{...} escapes in subject strings. The has been rewritten to avoid using
654 things like p++ in the argument of isxdigit().
655
656
657 Version 8.13 16-Aug-2011
658 ------------------------
659
660 1. The Unicode data tables have been updated to Unicode 6.0.0.
661
662 2. Two minor typos in pcre_internal.h have been fixed.
663
664 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
665 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
666 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
667
668 4. There were a number of related bugs in the code for matching backrefences
669 caselessly in UTF-8 mode when codes for the characters concerned were
670 different numbers of bytes. For example, U+023A and U+2C65 are an upper
671 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
672 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
673 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
674 2-byte code at the end of the subject (it thought there wasn't enough data
675 left).
676
677 5. Comprehensive information about what went wrong is now returned by
678 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
679 as the output vector has at least 2 elements. The offset of the start of
680 the failing character and a reason code are placed in the vector.
681
682 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
683 now returned is for the first byte of the failing character, instead of the
684 last byte inspected. This is an incompatible change, but I hope it is small
685 enough not to be a problem. It makes the returned offset consistent with
686 pcre_exec() and pcre_dfa_exec().
687
688 7. pcretest now gives a text phrase as well as the error number when
689 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
690 failure, the offset and reason code are output.
691
692 8. When \R was used with a maximizing quantifier it failed to skip backwards
693 over a \r\n pair if the subsequent match failed. Instead, it just skipped
694 back over a single character (\n). This seems wrong (because it treated the
695 two characters as a single entity when going forwards), conflicts with the
696 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
697 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
698 has been changed.
699
700 9. Some internal refactoring has changed the processing so that the handling
701 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
702 time (the PCRE_DOTALL option was changed this way some time ago: version
703 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
704 which was always a bit of a fudge. It also means that there is one less
705 argument for the match() function, which reduces its stack requirements
706 slightly. This change also fixes an incompatibility with Perl: the pattern
707 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
708
709 10. More internal refactoring has drastically reduced the number of recursive
710 calls to match() for possessively repeated groups such as (abc)++ when
711 using pcre_exec().
712
713 11. While implementing 10, a number of bugs in the handling of groups were
714 discovered and fixed:
715
716 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
717 (a|)*(?1) gave a compile-time internal error.
718 ((a|)+)+ did not notice that the outer group could match an empty string.
719 (^a|^)+ was not marked as anchored.
720 (.*a|.*)+ was not marked as matching at start or after a newline.
721
722 12. Yet more internal refactoring has removed another argument from the match()
723 function. Special calls to this function are now indicated by setting a
724 value in a variable in the "match data" data block.
725
726 13. Be more explicit in pcre_study() instead of relying on "default" for
727 opcodes that mean there is no starting character; this means that when new
728 ones are added and accidentally left out of pcre_study(), testing should
729 pick them up.
730
731 14. The -s option of pcretest has been documented for ages as being an old
732 synonym of -m (show memory usage). I have changed it to mean "force study
733 for every regex", that is, assume /S for every regex. This is similar to -i
734 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
735 using it. It makes it easier to run collections of tests with and without
736 study enabled, and thereby test pcre_study() more easily. All the standard
737 tests are now run with and without -s (but some patterns can be marked as
738 "never study" - see 20 below).
739
740 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
741 restoration of the capturing data to the outer values was not happening
742 correctly.
743
744 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
745 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
746 pattern had matched an empty string, and so incorrectly returned a no
747 match.
748
749 17. There was optimizing code for the last branch of non-capturing parentheses,
750 and also for the obeyed branch of a conditional subexpression, which used
751 tail recursion to cut down on stack usage. Unfortunately, now that there is
752 the possibility of (*THEN) occurring in these branches, tail recursion is
753 no longer possible because the return has to be checked for (*THEN). These
754 two optimizations have therefore been removed. [But see 8.20/11 above.]
755
756 18. If a pattern containing \R was studied, it was assumed that \R always
757 matched two bytes, thus causing the minimum subject length to be
758 incorrectly computed because \R can also match just one byte.
759
760 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
761 was incorrectly computed.
762
763 20. If /S is present twice on a test pattern in pcretest input, it now
764 *disables* studying, thereby overriding the use of -s on the command line
765 (see 14 above). This is necessary for one or two tests to keep the output
766 identical in both cases.
767
768 21. When (*ACCEPT) was used in an assertion that matched an empty string and
769 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
770
771 22. When an atomic group that contained a capturing parenthesis was
772 successfully matched, but the branch in which it appeared failed, the
773 capturing was not being forgotten if a higher numbered group was later
774 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
775 group 1 to "a", when in fact it should be unset. This applied to multi-
776 branched capturing and non-capturing groups, repeated or not, and also to
777 positive assertions (capturing in negative assertions does not happen
778 in PCRE) and also to nested atomic groups.
779
780 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
781 subject after a captured substring, to make it easier to tell which of a
782 number of identical substrings has been captured.
783
784 24. The way atomic groups are processed by pcre_exec() has been changed so that
785 if they are repeated, backtracking one repetition now resets captured
786 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
787 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
788 "aaa". Previously, it would have been "a". As part of this code
789 refactoring, the way recursive calls are handled has also been changed.
790
791 25. If an assertion condition captured any substrings, they were not passed
792 back unless some other capturing happened later. For example, if
793 (?(?=(a))a) was matched against "a", no capturing was returned.
794
795 26. When studying a pattern that contained subroutine calls or assertions,
796 the code for finding the minimum length of a possible match was handling
797 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
798 group 1 called group 2 while simultaneously a separate group 2 called group
799 1). A stack overflow occurred in this case. I have fixed this by limiting
800 the recursion depth to 10.
801
802 27. Updated RunTest.bat in the distribution to the version supplied by Tom
803 Fortmann. This supports explicit test numbers on the command line, and has
804 argument validation and error reporting.
805
806 28. An instance of \X with an unlimited repeat could fail if at any point the
807 first character it looked at was a mark character.
808
809 29. Some minor code refactoring concerning Unicode properties and scripts
810 should reduce the stack requirement of match() slightly.
811
812 30. Added the '=' option to pcretest to check the setting of unused capturing
813 slots at the end of the pattern, which are documented as being -1, but are
814 not included in the return count.
815
816 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
817 compiled something random. Now it gives a compile-time error (as does
818 Perl).
819
820 32. A *MARK encountered during the processing of a positive assertion is now
821 recorded and passed back (compatible with Perl).
822
823 33. If --only-matching or --colour was set on a pcregrep call whose pattern
824 had alternative anchored branches, the search for a second match in a line
825 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
826 matched the line "0102" twice. The same bug affected patterns that started
827 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
828 twice.
829
830 34. Previously, PCRE did not allow quantification of assertions. However, Perl
831 does, and because of capturing effects, quantifying parenthesized
832 assertions may at times be useful. Quantifiers are now allowed for
833 parenthesized assertions.
834
835 35. A minor code tidy in pcre_compile() when checking options for \R usage.
836
837 36. \g was being checked for fancy things in a character class, when it should
838 just be a literal "g".
839
840 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
841 appearance of a nested POSIX class supersedes an apparent external class.
842 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
843 unescaped square brackets may also appear as part of class names. For
844 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
845 more like Perl. (But see 8.20/1 above.)
846
847 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
848 was because it thought it was \N{name}, which is not supported).
849
850 39. Add minix to OS list not supporting the -S option in pcretest.
851
852 40. PCRE tries to detect cases of infinite recursion at compile time, but it
853 cannot analyze patterns in sufficient detail to catch mutual recursions
854 such as ((?1))((?2)). There is now a runtime test that gives an error if a
855 subgroup is called recursively as a subpattern for a second time at the
856 same position in the subject string. In previous releases this might have
857 been caught by the recursion limit, or it might have run out of stack.
858
859 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
860 happen only once. PCRE was, however incorrectly giving a compile time error
861 "recursive call could loop indefinitely" because it cannot analyze the
862 pattern in sufficient detail. The compile time test no longer happens when
863 PCRE is compiling a conditional subpattern, but actual runaway loops are
864 now caught at runtime (see 40 above).
865
866 42. It seems that Perl allows any characters other than a closing parenthesis
867 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
868 has been changed to be the same.
869
870 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
871 as not to get warnings when autogen.sh is called. Also changed
872 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
873
874 44. To help people who use pcregrep to scan files containing exceedingly long
875 lines, the following changes have been made:
876
877 (a) The default value of the buffer size parameter has been increased from
878 8K to 20K. (The actual buffer used is three times this size.)
879
880 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
881 PCRE is built.
882
883 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
884 to be set at run time.
885
886 (d) Numerical values in pcregrep options can be followed by K or M, for
887 example --buffer-size=50K.
888
889 (e) If a line being scanned overflows pcregrep's buffer, an error is now
890 given and the return code is set to 2.
891
892 45. Add a pointer to the latest mark to the callout data block.
893
894 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
895 partial match of an empty string instead of no match. This was specific to
896 the use of ".".
897
898 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
899 complete match instead of a partial match. This bug was dependent on both
900 the PCRE_UTF8 and PCRE_DOTALL options being set.
901
902 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
903 starting byte set, because \b was not being ignored.
904
905
906 Version 8.12 15-Jan-2011
907 ------------------------
908
909 1. Fixed some typos in the markup of the man pages, and wrote a script that
910 checks for such things as part of the documentation building process.
911
912 2. On a big-endian 64-bit system, pcregrep did not correctly process the
913 --match-limit and --recursion-limit options (added for 8.11). In
914 particular, this made one of the standard tests fail. (The integer value
915 went into the wrong half of a long int.)
916
917 3. If the --colour option was given to pcregrep with -v (invert match), it
918 did strange things, either producing crazy output, or crashing. It should,
919 of course, ignore a request for colour when reporting lines that do not
920 match.
921
922 4. Another pcregrep bug caused similar problems if --colour was specified with
923 -M (multiline) and the pattern match finished with a line ending.
924
925 5. In pcregrep, when a pattern that ended with a literal newline sequence was
926 matched in multiline mode, the following line was shown as part of the
927 match. This seems wrong, so I have changed it.
928
929 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
930 the check for further matches in the same line (so they could be coloured)
931 to overrun the end of the current line. If another match was found, it was
932 incorrectly shown (and then shown again when found in the next line).
933
934 7. If pcregrep was compiled under Windows, there was a reference to the
935 function pcregrep_exit() before it was defined. I am assuming this was
936 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
937 reported by a user. I've moved the definition above the reference.
938
939
940 Version 8.11 10-Dec-2010
941 ------------------------
942
943 1. (*THEN) was not working properly if there were untried alternatives prior
944 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
945 backtracked to try for "b" instead of moving to the next alternative branch
946 at the same level (in this case, to look for "c"). The Perl documentation
947 is clear that when (*THEN) is backtracked onto, it goes to the "next
948 alternative in the innermost enclosing group".
949
950 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
951 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
952 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
953 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
954 (*THEN).
955
956 3. If \s appeared in a character class, it removed the VT character from
957 the class, even if it had been included by some previous item, for example
958 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
959 of \s, but is part of the POSIX "space" class.)
960
961 4. A partial match never returns an empty string (because you can always
962 match an empty string at the end of the subject); however the checking for
963 an empty string was starting at the "start of match" point. This has been
964 changed to the "earliest inspected character" point, because the returned
965 data for a partial match starts at this character. This means that, for
966 example, /(?<=abc)def/ gives a partial match for the subject "abc"
967 (previously it gave "no match").
968
969 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
970 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
971 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
972 has an implication that the given string is incomplete (because a partial
973 match is preferred over a full match). For this reason, these items now
974 give a partial match in this situation. [Aside: previously, the one case
975 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
976 match rather than a full match, which was wrong by the old rules, but is
977 now correct.]
978
979 6. There was a bug in the handling of #-introduced comments, recognized when
980 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
981 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
982 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
983 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
984 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
985 places in pcre_compile().
986
987 7. Related to (6) above, when pcre_compile() was skipping #-introduced
988 comments when looking ahead for named forward references to subpatterns,
989 the only newline sequence it recognized was NL. It now handles newlines
990 according to the set newline convention.
991
992 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
993 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
994 cater for a lack of strerror(). These oversights have been fixed.
995
996 9. Added --match-limit and --recursion-limit to pcregrep.
997
998 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
999
1000 11. When the -o option was used, pcregrep was setting a return code of 1, even
1001 when matches were found, and --line-buffered was not being honoured.
1002
1003 12. Added an optional parentheses number to the -o and --only-matching options
1004 of pcregrep.
1005
1006 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1007 can match an empty string. The code to do it in pcretest and pcredemo
1008 needed fixing:
1009
1010 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1011 only one byte after an empty string match just before CRLF (this case
1012 just got forgotten; "any" and "anycrlf" were OK).
1013
1014 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1015 mode when an empty string match preceded an ASCII character followed by
1016 a non-ASCII character. (The code for advancing by one character rather
1017 than one byte was nonsense.)
1018
1019 (c) The pcredemo.c sample program did not have any code at all to handle
1020 the cases when CRLF is a valid newline sequence.
1021
1022 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1023 as a starting offset was within the subject string. There is now a new
1024 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1025 negative or greater than the length of the string. In order to test this,
1026 pcretest is extended to allow the setting of negative starting offsets.
1027
1028 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1029 starting offset points to the beginning of a UTF-8 character was
1030 unnecessarily clumsy. I tidied it up.
1031
1032 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1033 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1034
1035 17. Nobody had reported that the --include_dir option, which was added in
1036 release 7.7 should have been called --include-dir (hyphen, not underscore)
1037 for compatibility with GNU grep. I have changed it to --include-dir, but
1038 left --include_dir as an undocumented synonym, and the same for
1039 --exclude-dir, though that is not available in GNU grep, at least as of
1040 release 2.5.4.
1041
1042 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1043 characters from a string of bytes) have been redefined so as not to use
1044 loops, in order to improve performance in some environments. At the same
1045 time, I abstracted some of the common code into auxiliary macros to save
1046 repetition (this should not affect the compiled code).
1047
1048 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1049 compile-time error is now given if \c is not followed by an ASCII
1050 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1051 different, and any byte value is allowed.)
1052
1053 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1054 START_OPTIMIZE option, which is now allowed at compile time - but just
1055 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1056 to pcregrep and other applications that have no direct access to PCRE
1057 options. The new /Y option in pcretest sets this option when calling
1058 pcre_compile().
1059
1060 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1061 back references. Groups containing recursive back references were forced to
1062 be atomic by that change, but in the case of named groups, the amount of
1063 memory required was incorrectly computed, leading to "Failed: internal
1064 error: code overflow". This has been fixed.
1065
1066 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1067 pcretest.c, to avoid build problems in some Borland environments.
1068
1069
1070 Version 8.10 25-Jun-2010
1071 ------------------------
1072
1073 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1074 THEN.
1075
1076 2. (*ACCEPT) was not working when inside an atomic group.
1077
1078 3. Inside a character class, \B is treated as a literal by default, but
1079 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1080 causes the error). The code is unchanged, but I tidied the documentation.
1081
1082 4. Inside a character class, PCRE always treated \R and \X as literals,
1083 whereas Perl faults them if its -w option is set. I have changed PCRE so
1084 that it faults them when PCRE_EXTRA is set.
1085
1086 5. Added support for \N, which always matches any character other than
1087 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1088
1089 6. When compiling pcregrep with newer versions of gcc which may have
1090 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1091 declared with attribute warn_unused_result" were given. Just casting the
1092 result to (void) does not stop the warnings; a more elaborate fudge is
1093 needed. I've used a macro to implement this.
1094
1095 7. Minor change to pcretest.c to avoid a compiler warning.
1096
1097 8. Added four artifical Unicode properties to help with an option to make
1098 \s etc use properties (see next item). The new properties are: Xan
1099 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1100
1101 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1102 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1103 this option. Modified pcretest to add /W to test this facility. Added
1104 REG_UCP to make it available via the POSIX interface.
1105
1106 10. Added --line-buffered to pcregrep.
1107
1108 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1109 studied, and the match started with a letter with a code point greater than
1110 127 whose first byte was different to the first byte of the other case of
1111 the letter, the other case of this starting letter was not recognized
1112 (#976).
1113
1114 12. If a pattern that was studied started with a repeated Unicode property
1115 test, for example, \p{Nd}+, there was the theoretical possibility of
1116 setting up an incorrect bitmap of starting bytes, but fortunately it could
1117 not have actually happened in practice until change 8 above was made (it
1118 added property types that matched character-matching opcodes).
1119
1120 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1121 possible starting bytes for non-anchored patterns.
1122
1123 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1124 \R, and also a number of cases that involve Unicode properties, both
1125 explicit and implicit when PCRE_UCP is set.
1126
1127 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1128 input, it could crash or give wrong results if characters with values
1129 greater than 0xc0 were present in the subject string. (Detail: it assumed
1130 UTF-8 input when processing these items.)
1131
1132 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1133 size_t is 64-bit (#991).
1134
1135 17. Added a check for running out of memory when PCRE is compiled with
1136 --disable-stack-for-recursion (#990).
1137
1138 18. If the last data line in a file for pcretest does not have a newline on
1139 the end, a newline was missing in the output.
1140
1141 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1142 less than 128) in its various bitmaps. However, there is a facility for
1143 generating tables according to the current locale when PCRE is compiled. It
1144 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1145 characters, are recognized by isspace() and therefore were getting set in
1146 these tables, and indeed these tables seem to approximate to ISO 8859. This
1147 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1148 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1149 which of course cannot start UTF-8 characters. I have changed the code so
1150 that only real ASCII characters (less than 128) and the correct starting
1151 bytes for UTF-8 encodings are set for characters greater than 127 when in
1152 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1153 altogether.)
1154
1155 20. Added the /T option to pcretest so as to be able to run tests with non-
1156 standard character tables, thus making it possible to include the tests
1157 used for 19 above in the standard set of tests.
1158
1159 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1160 reference to a subpattern the other side of a comment that contains an
1161 opening parenthesis caused either an internal compiling error, or a
1162 reference to the wrong subpattern.
1163
1164
1165 Version 8.02 19-Mar-2010
1166 ------------------------
1167
1168 1. The Unicode data tables have been updated to Unicode 5.2.0.
1169
1170 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1171 configured.
1172
1173 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1174 original author of that file, following a query about its status.
1175
1176 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1177 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1178
1179 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1180 quantifier applied to a forward-referencing subroutine call, could compile
1181 incorrect code or give the error "internal error: previously-checked
1182 referenced subpattern not found".
1183
1184 6. Both MS Visual Studio and Symbian OS have problems with initializing
1185 variables to point to external functions. For these systems, therefore,
1186 pcre_malloc etc. are now initialized to local functions that call the
1187 relevant global functions.
1188
1189 7. There were two entries missing in the vectors called coptable and poptable
1190 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1191 I've fixed the data, and added a kludgy way of testing at compile time that
1192 the lengths are correct (equal to the number of opcodes).
1193
1194 8. Following on from 7, I added a similar kludge to check the length of the
1195 eint vector in pcreposix.c.
1196
1197 9. Error texts for pcre_compile() are held as one long string to avoid too
1198 much relocation at load time. To find a text, the string is searched,
1199 counting zeros. There was no check for running off the end of the string,
1200 which could happen if a new error number was added without updating the
1201 string.
1202
1203 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1204
1205 11. \K was not working if it appeared in an atomic group or in a group that
1206 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1207 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1208 the assertion is positive, but not if it is negative.
1209
1210 12. Change 11 fortuitously reduced the size of the stack frame used in the
1211 "match()" function of pcre_exec.c by one pointer. Forthcoming
1212 implementation of support for (*MARK) will need an extra pointer on the
1213 stack; I have reserved it now, so that the stack frame size does not
1214 decrease.
1215
1216 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1217 item in branch that calls a recursion is a subroutine call - as in the
1218 second branch in the above example - was incorrectly given the compile-
1219 time error "recursive call could loop indefinitely" because pcre_compile()
1220 was not correctly checking the subroutine for matching a non-empty string.
1221
1222 14. The checks for overrunning compiling workspace could trigger after an
1223 overrun had occurred. This is a "should never occur" error, but it can be
1224 triggered by pathological patterns such as hundreds of nested parentheses.
1225 The checks now trigger 100 bytes before the end of the workspace.
1226
1227 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1228
1229
1230 Version 8.01 19-Jan-2010
1231 ------------------------
1232
1233 1. If a pattern contained a conditional subpattern with only one branch (in
1234 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1235 computed the wrong minimum data length (which is of course zero for such
1236 subpatterns). This could cause incorrect "no match" results.
1237
1238 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1239 the pattern is reset in the first branch, pcre_compile() failed with
1240 "internal error: code overflow at offset...". This happened only when
1241 the reset was to the original external option setting. (An optimization
1242 abstracts leading options settings into an external setting, which was the
1243 cause of this.)
1244
1245 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1246 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1247 assertion pattern did not match (meaning that the assertion was true), it
1248 was incorrectly treated as false if the SKIP had been reached during the
1249 matching. This also applied to assertions used as conditions.
1250
1251 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1252 assertion subpattern, including such a pattern used as a condition,
1253 unpredictable results occurred, instead of the error return
1254 PCRE_ERROR_DFA_UITEM.
1255
1256 5. The C++ GlobalReplace function was not working like Perl for the special
1257 situation when an empty string is matched. It now does the fancy magic
1258 stuff that is necessary.
1259
1260 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1261 removed. (These were left over from very, very early versions of PCRE.)
1262
1263 7. Some cosmetic changes to the code to make life easier when compiling it
1264 as part of something else:
1265
1266 (a) Change DEBUG to PCRE_DEBUG.
1267
1268 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1269 called "current" as "current_branch", to prevent a collision with the
1270 Linux macro when compiled as a kernel module.
1271
1272 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1273 prevent a collision with the Linux macro when compiled as a kernel
1274 module.
1275
1276 8. In pcre_compile() there are some checks for integer overflows that used to
1277 cast potentially large values to (double). This has been changed to that
1278 when building, a check for int64_t is made, and if it is found, it is used
1279 instead, thus avoiding the use of floating point arithmetic. (There is no
1280 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1281 double.
1282
1283 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1284 2005 (difference between two addresses compared to an unsigned value).
1285
1286 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1287 custom one, because of the following reported problem in Windows:
1288
1289 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1290 under Win32.
1291 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1292 therefore missing the function definition.
1293 - The compiler thus generates a "C" signature for the test function.
1294 - The linker fails to find the "C" function.
1295 - PCRE fails to configure if asked to do so against libbz2.
1296
1297 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1298 messages were output:
1299
1300 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1301 rerunning libtoolize, to keep the correct libtool macros in-tree.
1302 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1303
1304 I have done both of these things.
1305
1306 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1307 most of the time, it *can* run out if it is given a pattern that contains a
1308 runaway infinite recursion. I updated the discussion in the pcrestack man
1309 page.
1310
1311 13. Now that we have gone to the x.xx style of version numbers, the minor
1312 version may start with zero. Using 08 or 09 is a bad idea because users
1313 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1314 interpreted as invalid octal numbers. I've updated the previous comment in
1315 configure.ac, and also added a check that gives an error if 08 or 09 are
1316 used.
1317
1318 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1319 causing partial matching to fail when the end of the subject matched \W
1320 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1321
1322 15. There were some discrepancies between the declarations in pcre_internal.h
1323 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1324 their definitions. The declarations used "const uschar *" and the
1325 definitions used USPTR. Even though USPTR is normally defined as "const
1326 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1327 reported that: "This difference in casting confuses some C++ compilers, for
1328 example, SunCC recognizes above declarations as different functions and
1329 generates broken code for hbpcre." I have changed the declarations to use
1330 USPTR.
1331
1332 16. GNU libtool is named differently on some systems. The autogen.sh script now
1333 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1334 (FreeBSD).
1335
1336 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1337 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1338 comment: "Figure out how to create a longlong from a string: strtoll and
1339 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1340 instance, but it only takes 2 args instead of 3!"
1341
1342 18. A subtle bug concerned with back references has been fixed by a change of
1343 specification, with a corresponding code fix. A pattern such as
1344 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1345 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1346 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1347 same bug. Such groups have to be quantified to be useful, or contained
1348 inside another quantified group. (If there's no repetition, the reference
1349 can never match.) The problem arises because, having left the group and
1350 moved on to the rest of the pattern, a later failure that backtracks into
1351 the group uses the captured value from the final iteration of the group
1352 rather than the correct earlier one. I have fixed this in PCRE by forcing
1353 any group that contains a reference to itself to be an atomic group; that
1354 is, there cannot be any backtracking into it once it has completed. This is
1355 similar to recursive and subroutine calls.
1356
1357
1358 Version 8.00 19-Oct-09
1359 ----------------------
1360
1361 1. The table for translating pcre_compile() error codes into POSIX error codes
1362 was out-of-date, and there was no check on the pcre_compile() error code
1363 being within the table. This could lead to an OK return being given in
1364 error.
1365
1366 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1367 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1368 in a Windows environment.
1369
1370 3. The pcregrep --count option prints the count for each file even when it is
1371 zero, as does GNU grep. However, pcregrep was also printing all files when
1372 --files-with-matches was added. Now, when both options are given, it prints
1373 counts only for those files that have at least one match. (GNU grep just
1374 prints the file name in this circumstance, but including the count seems
1375 more useful - otherwise, why use --count?) Also ensured that the
1376 combination -clh just lists non-zero counts, with no names.
1377
1378 4. The long form of the pcregrep -F option was incorrectly implemented as
1379 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1380 but it seems right to fix it, and I didn't think it was worth preserving
1381 the old behaviour.
1382
1383 5. The command line items --regex=pattern and --regexp=pattern were not
1384 recognized by pcregrep, which required --regex pattern or --regexp pattern
1385 (with a space rather than an '='). The man page documented the '=' forms,
1386 which are compatible with GNU grep; these now work.
1387
1388 6. No libpcreposix.pc file was created for pkg-config; there was just
1389 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1390
1391 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1392 when UCP support is not needed, by modifying the Python script that
1393 generates it from Unicode data files. This should not matter if the module
1394 is correctly used as a library, but I received one complaint about 50K of
1395 unwanted data. My guess is that the person linked everything into his
1396 program rather than using a library. Anyway, it does no harm.
1397
1398 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1399 was a minimum greater than 1 for a wide character in a possessive
1400 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1401 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1402 character. Chaos in the form of incorrect output or a compiling loop could
1403 result.
1404
1405 9. The restrictions on what a pattern can contain when partial matching is
1406 requested for pcre_exec() have been removed. All patterns can now be
1407 partially matched by this function. In addition, if there are at least two
1408 slots in the offset vector, the offset of the earliest inspected character
1409 for the match and the offset of the end of the subject are set in them when
1410 PCRE_ERROR_PARTIAL is returned.
1411
1412 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1413 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1414 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1415 and may be more useful for multi-segment matching.
1416
1417 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1418 used to be given if ever the end of the subject was reached; now it is
1419 given only if matching could not proceed because another character was
1420 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1421 string "Z", which now yields "no match" instead of "partial match". In the
1422 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1423 final character ended with (*FAIL).
1424
1425 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1426 if the pattern had a "must contain" character that was already found in the
1427 earlier partial match, unless partial matching was again requested. For
1428 example, with the pattern /dog.(body)?/, the "must contain" character is
1429 "g". If the first part-match was for the string "dog", restarting with
1430 "sbody" failed. This bug has been fixed.
1431
1432 13. The string returned by pcre_dfa_exec() after a partial match has been
1433 changed so that it starts at the first inspected character rather than the
1434 first character of the match. This makes a difference only if the pattern
1435 starts with a lookbehind assertion or \b or \B (\K is not supported by
1436 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1437 matching functions compatible, and I think it's the right thing to do.
1438
1439 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1440 so that the demonstration program is easily available in environments where
1441 PCRE has not been installed from source.
1442
1443 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1444 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1445 library.
1446
1447 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1448 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1449 is not the first non-POSIX option to be added. Clearly some people find
1450 these options useful.
1451
1452 17. If a caller to the POSIX matching function regexec() passes a non-zero
1453 value for nmatch with a NULL value for pmatch, the value of
1454 nmatch is forced to zero.
1455
1456 18. RunGrepTest did not have a test for the availability of the -u option of
1457 the diff command, as RunTest does. It now checks in the same way as
1458 RunTest, and also checks for the -b option.
1459
1460 19. If an odd number of negated classes containing just a single character
1461 interposed, within parentheses, between a forward reference to a named
1462 subpattern and the definition of the subpattern, compilation crashed with
1463 an internal error, complaining that it could not find the referenced
1464 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1465 [The bug was that it was starting one character too far in when skipping
1466 over the character class, thus treating the ] as data rather than
1467 terminating the class. This meant it could skip too much.]
1468
1469 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1470 /g option in pcretest when the pattern contains \K, which makes it possible
1471 to have an empty string match not at the start, even when the pattern is
1472 anchored. Updated pcretest and pcredemo to use this option.
1473
1474 21. If the maximum number of capturing subpatterns in a recursion was greater
1475 than the maximum at the outer level, the higher number was returned, but
1476 with unset values at the outer level. The correct (outer level) value is
1477 now given.
1478
1479 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1480 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1481 make it do so. The string so far is captured, making this feature
1482 compatible with Perl.
1483
1484 23. The tests have been re-organized, adding tests 11 and 12, to make it
1485 possible to check the Perl 5.10 features against Perl 5.10.
1486
1487 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1488 pattern matches a fixed length string. PCRE did not allow this; now it
1489 does. Neither allows recursion.
1490
1491 25. I finally figured out how to implement a request to provide the minimum
1492 length of subject string that was needed in order to match a given pattern.
1493 (It was back references and recursion that I had previously got hung up
1494 on.) This code has now been added to pcre_study(); it finds a lower bound
1495 to the length of subject needed. It is not necessarily the greatest lower
1496 bound, but using it to avoid searching strings that are too short does give
1497 some useful speed-ups. The value is available to calling programs via
1498 pcre_fullinfo().
1499
1500 26. While implementing 25, I discovered to my embarrassment that pcretest had
1501 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1502 study optimizations had never been tested with that matching function.
1503 Oops. What is worse, even when it was passed study data, there was a bug in
1504 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1505 were also very few tests of studied patterns with pcre_dfa_exec().
1506
1507 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1508 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1509 on the other side of the coin, they are no longer allowed to have different
1510 names, because these cannot be distinguished in PCRE, and this has caused
1511 confusion. (This is a difference from Perl.)
1512
1513 28. When duplicate subpattern names are present (necessarily with different
1514 numbers, as required by 27 above), and a test is made by name in a
1515 conditional pattern, either for a subpattern having been matched, or for
1516 recursion in such a pattern, all the associated numbered subpatterns are
1517 tested, and the overall condition is true if the condition is true for any
1518 one of them. This is the way Perl works, and is also more like the way
1519 testing by number works.
1520
1521
1522 Version 7.9 11-Apr-09
1523 ---------------------
1524
1525 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1526 (pcretest), all targets were linked against these libraries. This included
1527 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1528 libraries. This caused unwanted dependencies to be created. This problem
1529 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1530 pcretest is linked with readline.
1531
1532 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1533 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1534 moved outside it again, because FALSE and TRUE are already defined in AIX,
1535 but BOOL is not.
1536
1537 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1538 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1539
1540 4. The pcregrep documentation said spaces were inserted as well as colons (or
1541 hyphens) following file names and line numbers when outputting matching
1542 lines. This is not true; no spaces are inserted. I have also clarified the
1543 wording for the --colour (or --color) option.
1544
1545 5. In pcregrep, when --colour was used with -o, the list of matching strings
1546 was not coloured; this is different to GNU grep, so I have changed it to be
1547 the same.
1548
1549 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1550 each matching line was coloured. Now it goes on to look for further matches
1551 of any of the test patterns, which is the same behaviour as GNU grep.
1552
1553 7. A pattern that could match an empty string could cause pcregrep to loop; it
1554 doesn't make sense to accept an empty string match in pcregrep, so I have
1555 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1556 seems to be how GNU grep behaves.
1557
1558 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1559 start or after a newline", because the conditional assertion was not being
1560 correctly handled. The rule now is that both the assertion and what follows
1561 in the first alternative must satisfy the test.
1562
1563 9. If auto-callout was enabled in a pattern with a conditional group whose
1564 condition was an assertion, PCRE could crash during matching, both with
1565 pcre_exec() and pcre_dfa_exec().
1566
1567 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1568 used for matching.
1569
1570 11. Unicode property support in character classes was not working for
1571 characters (bytes) greater than 127 when not in UTF-8 mode.
1572
1573 12. Added the -M command line option to pcretest.
1574
1575 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1576
1577 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1578
1579 16. Added comments and documentation about mis-use of no_arg in the C++
1580 wrapper.
1581
1582 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1583 from Martin Jerabek that uses macro names for all relevant character and
1584 string constants.
1585
1586 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1587 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1588 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1589 these, but not everybody uses configure.
1590
1591 19. A conditional group that had only one branch was not being correctly
1592 recognized as an item that could match an empty string. This meant that an
1593 enclosing group might also not be so recognized, causing infinite looping
1594 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1595 with the subject "ab", where knowledge that the repeated group can match
1596 nothing is needed in order to break the loop.
1597
1598 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1599 exec(), but without supplying a callout function, matching went wrong.
1600
1601 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1602 leak if the size of the offset vector was greater than 30. When the vector
1603 is smaller, the saved offsets during recursion go onto a local stack
1604 vector, but for larger vectors malloc() is used. It was failing to free
1605 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1606 error, in fact).
1607
1608 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1609 heapframe that is used only when UTF-8 support is enabled. This caused no
1610 problem, but was untidy.
1611
1612 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1613 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1614 included within another project.
1615
1616 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1617 slightly modified by me:
1618
1619 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1620 not building pcregrep.
1621
1622 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1623 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1624
1625 25. Forward references, both numeric and by name, in patterns that made use of
1626 duplicate group numbers, could behave incorrectly or give incorrect errors,
1627 because when scanning forward to find the reference group, PCRE was not
1628 taking into account the duplicate group numbers. A pattern such as
1629 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1630
1631 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1632 the feature of a custom pointer more persuasive (as requested by a user).
1633
1634 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1635 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1636 pre-defined.
1637
1638 28. Added support for (*UTF8) at the start of a pattern.
1639
1640 29. Arrange for flags added by the "release type" setting in CMake to be shown
1641 in the configuration summary.
1642
1643
1644 Version 7.8 05-Sep-08
1645 ---------------------
1646
1647 1. Replaced UCP searching code with optimized version as implemented for Ad
1648 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1649 stage table and inline lookup instead of a function, giving speed ups of 2
1650 to 5 times on some simple patterns that I tested. Permission was given to
1651 distribute the MultiStage2.py script that generates the tables (it's not in
1652 the tarball, but is in the Subversion repository).
1653
1654 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1655 scripts.
1656
1657 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1658 a group with a zero qualifier. The result of the study could be incorrect,
1659 or the function might crash, depending on the pattern.
1660
1661 4. Caseless matching was not working for non-ASCII characters in back
1662 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1663 It now works when Unicode Property Support is available.
1664
1665 5. In pcretest, an escape such as \x{de} in the data was always generating
1666 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1667 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1668 truncation.
1669
1670 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1671
1672 7. Added two (int) casts to pcregrep when printing the difference of two
1673 pointers, in case they are 64-bit values.
1674
1675 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1676 test 2 if it fails.
1677
1678 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1679 and a #define of that name to empty if it is not externally set. This is to
1680 allow users of MSVC to set it if necessary.
1681
1682 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1683 the convenience functions in the pcre_get.c source file.
1684
1685 11. An option change at the start of a pattern that had top-level alternatives
1686 could cause overwriting and/or a crash. This command provoked a crash in
1687 some environments:
1688
1689 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1690
1691 This potential security problem was recorded as CVE-2008-2371.
1692
1693 12. For a pattern where the match had to start at the beginning or immediately
1694 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1695 pcre_dfa_exec() could read past the end of the passed subject if there was
1696 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1697 pcretest so that it places the subject at the end of its malloc-ed buffer.
1698
1699 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1700 exec() might read past the end of the data buffer in UTF-8 mode.
1701
1702 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1703 the data contained the byte 0x85 as part of a UTF-8 character within its
1704 first line. This applied both to normal and DFA matching.
1705
1706 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1707 /^[^d]*?$/8 failed to match "abc".
1708
1709 16. Added a missing copyright notice to pcrecpp_internal.h.
1710
1711 17. Make it more clear in the documentation that values returned from
1712 pcre_exec() in ovector are byte offsets, not character counts.
1713
1714 18. Tidied a few places to stop certain compilers from issuing warnings.
1715
1716 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1717 supplied by Stefan Weber. I made a further small update for 7.8 because
1718 there is a change of source arrangements: the pcre_searchfuncs.c module is
1719 replaced by pcre_ucd.c.
1720
1721
1722 Version 7.7 07-May-08
1723 ---------------------
1724
1725 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1726 a string to a long long, pretend we don't even have a long long." This is
1727 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1728
1729 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1730 pre-7.6 versions, which defined a global no_arg variable instead of putting
1731 it in the RE class. (See also #8 below.)
1732
1733 3. Remove a line of dead code, identified by coverity and reported by Nuno
1734 Lopes.
1735
1736 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1737
1738 (1) The include/exclude patterns were being applied to the whole pathnames
1739 of files, instead of just to the final components.
1740
1741 (2) If there was more than one level of directory, the subdirectories were
1742 skipped unless they satisfied the include/exclude conditions. This is
1743 inconsistent with GNU grep (and could even be seen as contrary to the
1744 pcregrep specification - which I improved to make it absolutely clear).
1745 The action now is always to scan all levels of directory, and just
1746 apply the include/exclude patterns to regular files.
1747
1748 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1749 --exclude_dir in the tests to avoid scanning .svn directories.
1750
1751 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1752 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1753 doesn't support NULs in patterns.
1754
1755 7. Added some missing "const"s to declarations of static tables in
1756 pcre_compile.c and pcre_dfa_exec.c.
1757
1758 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1759 caused by fix #2 above. (Subsequently also a second patch to fix the
1760 first patch. And a third patch - this was a messy problem.)
1761
1762 9. Applied Craig's patch to remove the use of push_back().
1763
1764 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1765 matching function regexec().
1766
1767 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1768 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1769 references. PCRE supports relative numbers with this syntax (I don't think
1770 Oniguruma does).
1771
1772 12. Previously, a group with a zero repeat such as (...){0} was completely
1773 omitted from the compiled regex. However, this means that if the group
1774 was called as a subroutine from elsewhere in the pattern, things went wrong
1775 (an internal error was given). Such groups are now left in the compiled
1776 pattern, with a new opcode that causes them to be skipped at execution
1777 time.
1778
1779 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1780 to the way PCRE behaves:
1781
1782 (a) A lone ] character is dis-allowed (Perl treats it as data).
1783
1784 (b) A back reference to an unmatched subpattern matches an empty string
1785 (Perl fails the current match path).
1786
1787 (c) A data ] in a character class must be notated as \] because if the
1788 first data character in a class is ], it defines an empty class. (In
1789 Perl it is not possible to have an empty class.) The empty class []
1790 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1791 The negative empty class [^] matches any one character, independently
1792 of the DOTALL setting.
1793
1794 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1795 non-existent subpattern following a character class starting with ']' and
1796 containing () gave an internal compiling error instead of "reference to
1797 non-existent subpattern". Fortunately, when the pattern did exist, the
1798 compiled code was correct. (When scanning forwards to check for the
1799 existencd of the subpattern, it was treating the data ']' as terminating
1800 the class, so got the count wrong. When actually compiling, the reference
1801 was subsequently set up correctly.)
1802
1803 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1804 it was being rejected as not supported by pcre_dfa_exec(), even though
1805 other assertions are supported. I have made pcre_dfa_exec() support
1806 (*FAIL).
1807
1808 16. The implementation of 13c above involved the invention of a new opcode,
1809 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1810 cannot be changed at match time, I realized I could make a small
1811 improvement to matching performance by compiling OP_ALLANY instead of
1812 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1813 on the OP_ANY path.
1814
1815 17. Compiling pcretest on Windows with readline support failed without the
1816 following two fixes: (1) Make the unistd.h include conditional on
1817 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1818
1819 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1820 ncurses library to be included for pcretest when ReadLine support is
1821 requested, but also to allow for it to be overridden. This patch came from
1822 Daniel Bergström.
1823
1824 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1825 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1826 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1827 spotting this.
1828
1829
1830 Version 7.6 28-Jan-08
1831 ---------------------
1832
1833 1. A character class containing a very large number of characters with
1834 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1835 overflow.
1836
1837 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1838 HAVE_LONG_LONG is not defined.
1839
1840 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1841 bring them up to date and include new features. This patch includes:
1842
1843 - Fixed PH's badly added libz and libbz2 support.
1844 - Fixed a problem with static linking.
1845 - Added pcredemo. [But later removed - see 7 below.]
1846 - Fixed dftables problem and added an option.
1847 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1848 HAVE_LONG_LONG.
1849 - Added readline support for pcretest.
1850 - Added an listing of the option settings after cmake has run.
1851
1852 4. A user submitted a patch to Makefile that makes it easy to create
1853 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1854 Makefile.am that cause it to include this special target, without
1855 affecting anything else. Note that the same mingw target plus all
1856 the other distribution libraries and programs are now supported
1857 when configuring with CMake (see 6 below) instead of with
1858 Configure/Make.
1859
1860 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1861 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1862 exported in the Windows port". It has not yet been confirmed that the patch
1863 solves the problem, but it does no harm.
1864
1865 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1866 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1867 with CMake, and also correct the comment about stack recursion.
1868
1869 7. Remove the automatic building of pcredemo from the ./configure system and
1870 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1871 of a program that users should build themselves after PCRE is installed, so
1872 building it automatically is not really right. What is more, it gave
1873 trouble in some build environments.
1874
1875 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1876
1877
1878 Version 7.5 10-Jan-08
1879 ---------------------
1880
1881 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1882 values in parens when parsing an RE using the C++ wrapper."
1883
1884 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1885 Characters greater than 255 were excluded from the class instead of being
1886 included.
1887
1888 3. The same bug as (2) above applied to negated POSIX classes such as
1889 [:^space:].
1890
1891 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1892 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1893 I have changed it.
1894
1895 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1896 first named subpattern) and a construct such as (?&a) would reference the
1897 first named subpattern whose name started with "a" (in other words, the
1898 length check was missing). Both these problems are fixed. "Subpattern name
1899 expected" is now given for (?&) (a zero-length name), and this patch also
1900 makes it give the same error for \k'' (previously it complained that that
1901 was a reference to a non-existent subpattern).
1902
1903 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1904 this is right because (?- can be followed by option settings as well as by
1905 digits. I have, however, made the messages clearer.
1906
1907 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1908 than the number used in the conditional) now cause a compile-time error.
1909 This is actually not compatible with Perl, which accepts such patterns, but
1910 treats the conditional as always being FALSE (as PCRE used to), but it
1911 seems to me that giving a diagnostic is better.
1912
1913 8. Change "alphameric" to the more common word "alphanumeric" in comments
1914 and messages.
1915
1916 9. Fix two occurrences of "backslash" in comments that should have been
1917 "backspace".
1918
1919 10. Remove two redundant lines of code that can never be obeyed (their function
1920 was moved elsewhere).
1921
1922 11. The program that makes PCRE's Unicode character property table had a bug
1923 which caused it to generate incorrect table entries for sequences of
1924 characters that have the same character type, but are in different scripts.
1925 It amalgamated them into a single range, with the script of the first of
1926 them. In other words, some characters were in the wrong script. There were
1927 thirteen such cases, affecting characters in the following ranges:
1928
1929 U+002b0 - U+002c1
1930 U+0060c - U+0060d
1931 U+0061e - U+00612
1932 U+0064b - U+0065e
1933 U+0074d - U+0076d
1934 U+01800 - U+01805
1935 U+01d00 - U+01d77
1936 U+01d9b - U+01dbf
1937 U+0200b - U+0200f
1938 U+030fc - U+030fe
1939 U+03260 - U+0327f
1940 U+0fb46 - U+0fbb1
1941 U+10450 - U+1049d
1942
1943 12. The -o option (show only the matching part of a line) for pcregrep was not
1944 compatible with GNU grep in that, if there was more than one match in a
1945 line, it showed only the first of them. It now behaves in the same way as
1946 GNU grep.
1947
1948 13. If the -o and -v options were combined for pcregrep, it printed a blank
1949 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1950 does the same. The return code can be used to tell if there were any
1951 non-matching lines.
1952
1953 14. Added --file-offsets and --line-offsets to pcregrep.
1954
1955 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1956 infinitely looping recursion. The bug was that positive lookaheads were not
1957 being skipped when checking for a possible empty match (negative lookaheads
1958 and both kinds of lookbehind were skipped).
1959
1960 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1961 inclusion of <windows.h> to before rather than after the definition of
1962 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1963
1964 17. Specifying a possessive quantifier with a specific limit for a Unicode
1965 character property caused pcre_compile() to compile bad code, which led at
1966 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1967 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1968 caused the error; without that there was no problem.
1969
1970 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1971
1972 19. Added --enable-pcretest-libreadline.
1973
1974 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1975 RE::GlobalReplace(). As a result, the number of replacements returned was
1976 double what it should be. I removed one of the increments, but Craig sent a
1977 later patch that removed the other one (the right fix) and added unit tests
1978 that check the return values (which was not done before).
1979
1980 21. Several CMake things:
1981
1982 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1983 the names libpcre and libpcreposix, not just pcre and pcreposix.
1984
1985 (2) The above change means that pcretest and pcregrep are now correctly
1986 linked with the newly-built libraries, not previously installed ones.
1987
1988 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1989
1990 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1991 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1992 UTF-8 newline character). The key issue is that the pattern starts .*;
1993 this means that the match must be either at the beginning, or after a
1994 newline. The bug was in the code for advancing after a failed match and
1995 checking that the new position followed a newline. It was not taking
1996 account of UTF-8 characters correctly.
1997
1998 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1999 character classes. PCRE was not treating the sequence [:...:] as a
2000 character class unless the ... were all letters. Perl, however, seems to
2001 allow any characters between [: and :], though of course it rejects as
2002 unknown any "names" that contain non-letters, because all the known class
2003 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2004 for example, whereas PCRE did not - it did not recognize a POSIX character
2005 class. This seemed a bit dangerous, so the code has been changed to be
2006 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2007 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2008 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2009 Perl does, and where it didn't before.
2010
2011 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2012 Windows environments %n is disabled by default.
2013
2014
2015 Version 7.4 21-Sep-07
2016 ---------------------
2017
2018 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2019 means that a class such as [\s] counted as "explicit reference to CR or
2020 LF". That isn't really right - the whole point of the change was to try to
2021 help when there was an actual mention of one of the two characters. So now
2022 the change happens only if \r or \n (or a literal CR or LF) character is
2023 encountered.
2024
2025 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2026 of both had grown to the point where there were only 3 bits left.
2027 Fortunately, there was spare space in the data structure, and so I have
2028 moved the internal flags into a new 16-bit field to free up more option
2029 bits.
2030
2031 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2032 but did not set the internal JCHANGED flag - either of these is enough to
2033 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2034 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2035 start sets both bits.
2036
2037 4. Added options (at build time, compile time, exec time) to change \R from
2038 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2039
2040 5. doc/pcresyntax.html was missing from the distribution.
2041
2042 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2043 compatibility, even though it is no longer used.
2044
2045 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2046 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2047 windows.h file is present (where different names are used). [This was
2048 reversed later after testing - see 16 below.]
2049
2050 8. Changed all #include <config.h> to #include "config.h". There were also
2051 some further <pcre.h> cases that I changed to "pcre.h".
2052
2053 9. When pcregrep was used with the --colour option, it missed the line ending
2054 sequence off the lines that it output.
2055
2056 10. It was pointed out to me that arrays of string pointers cause lots of
2057 relocations when a shared library is dynamically loaded. A technique of
2058 using a single long string with a table of offsets can drastically reduce
2059 these. I have refactored PCRE in four places to do this. The result is
2060 dramatic:
2061
2062 Originally: 290
2063 After changing UCP table: 187
2064 After changing error message table: 43
2065 After changing table of "verbs" 36
2066 After changing table of Posix names 22
2067
2068 Thanks to the folks working on Gregex for glib for this insight.
2069
2070 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2071 unicode-properties was also set.
2072
2073 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2074
2075 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2076 checked only for CRLF.
2077
2078 14. Added casts to pcretest.c to avoid compiler warnings.
2079
2080 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2081
2082 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2083 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2084 entirely. This removes changes made in 7 above.
2085
2086 17. The CMake files have been updated, and there is now more information about
2087 building with CMake in the NON-UNIX-USE document.
2088
2089
2090 Version 7.3 28-Aug-07
2091 ---------------------
2092
2093 1. In the rejigging of the build system that eventually resulted in 7.1, the
2094 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2095 brackets there is not right, since it causes compilers to look for an
2096 installed pcre.h, not the version that is in the source that is being
2097 compiled (which of course may be different). I have changed it back to:
2098
2099 #include "pcre.h"
2100
2101 I have a vague recollection that the change was concerned with compiling in
2102 different directories, but in the new build system, that is taken care of
2103 by the VPATH setting the Makefile.
2104
2105 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2106 when the subject happened to end in the byte 0x85 (e.g. if the last
2107 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2108 characters but of course it shouldn't be taken as a newline when it is part
2109 of another character. The bug was that, for an unlimited repeat of . in
2110 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2111 characters when looking for a newline.
2112
2113 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2114
2115 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2116 in debug output.
2117
2118 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2119 long printing in the pcrecpp unittest when running under MinGW.
2120
2121 6. ESC_K was left out of the EBCDIC table.
2122
2123 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2124 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2125 limit also applies to "virtual nesting" when a pattern is recursive, and in
2126 this case 1000 isn't so big. I have been able to remove this limit at the
2127 expense of backing off one optimization in certain circumstances. Normally,
2128 when pcre_exec() would call its internal match() function recursively and
2129 immediately return the result unconditionally, it uses a "tail recursion"
2130 feature to save stack. However, when a subpattern that can match an empty
2131 string has an unlimited repetition quantifier, it no longer makes this
2132 optimization. That gives it a stack frame in which to save the data for
2133 checking that an empty string has been matched. Previously this was taken
2134 from the 1000-entry workspace that had been reserved. So now there is no
2135 explicit limit, but more stack is used.
2136
2137 8. Applied Daniel's patches to solve problems with the import/export magic
2138 syntax that is required for Windows, and which was going wrong for the
2139 pcreposix and pcrecpp parts of the library. These were overlooked when this
2140 problem was solved for the main library.
2141
2142 9. There were some crude static tests to avoid integer overflow when computing
2143 the size of patterns that contain repeated groups with explicit upper
2144 limits. As the maximum quantifier is 65535, the maximum group length was
2145 set at 30,000 so that the product of these two numbers did not overflow a
2146 32-bit integer. However, it turns out that people want to use groups that
2147 are longer than 30,000 bytes (though not repeat them that many times).
2148 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2149 made it possible to implement the integer overflow checks in a much more
2150 dynamic way, which I have now done. The artificial limitation on group
2151 length has been removed - we now have only the limit on the total length of
2152 the compiled pattern, which depends on the LINK_SIZE setting.
2153
2154 10. Fixed a bug in the documentation for get/copy named substring when
2155 duplicate names are permitted. If none of the named substrings are set, the
2156 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2157 empty string.
2158
2159 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2160 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2161 because the ] is interpreted as the first data character and the
2162 terminating ] is not found. PCRE has been made compatible with Perl in this
2163 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2164 cause memory overwriting.
2165
2166 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2167 string has been matched (to stop an infinite loop). It was not recognizing
2168 a conditional subpattern that could match an empty string if that
2169 subpattern was within another subpattern. For example, it looped when
2170 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2171 condition was not nested. This bug has been fixed.
2172
2173 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2174 past the start of the subject in the presence of bytes with the top bit
2175 set, for example "\x8aBCD".
2176
2177 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2178 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2179
2180 14. Optimized (?!) to (*FAIL).
2181
2182 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2183 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2184 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2185 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2186 does: it's just the validity check that is more restrictive.
2187
2188 16. Inserted checks for integer overflows during escape sequence (backslash)
2189 processing, and also fixed erroneous offset values for syntax errors during
2190 backslash processing.
2191
2192 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2193 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2194
2195 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2196 caused an overrun.
2197
2198 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2199 something other than just ASCII characters) inside a group that had an
2200 unlimited repeat caused a loop at compile time (while checking to see
2201 whether the group could match an empty string).
2202
2203 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2204 [\P{Any}] did so. (Error in the code for printing property names.)
2205
2206 21. An orphan \E inside a character class could cause a crash.
2207
2208 22. A repeated capturing bracket such as (A)? could cause a wild memory
2209 reference during compilation.
2210
2211 23. There are several functions in pcre_compile() that scan along a compiled
2212 expression for various reasons (e.g. to see if it's fixed length for look
2213 behind). There were bugs in these functions when a repeated \p or \P was
2214 present in the pattern. These operators have additional parameters compared
2215 with \d, etc, and these were not being taken into account when moving along
2216 the compiled data. Specifically:
2217
2218 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2219 length.
2220
2221 (b) An item such as \pL+ within a repeated group could cause crashes or
2222 loops.
2223
2224 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2225 "reference to non-existent subpattern" error.
2226
2227 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2228
2229 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2230 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2231
2232 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2233
2234 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2235 character were causing crashes (broken optimization).
2236
2237 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2238 \p or \P) caused a compile-time loop.
2239
2240 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2241 break. For example, the unstudied pattern [\r\n]A does not match the string
2242 "\r\nA" because change 7.0/46 below moves the current point on by two
2243 characters after failing to match at the start. However, the pattern \nA
2244 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2245 the same is true. There doesn't seem any very clean way out of this, but
2246 what I have chosen to do makes the common cases work: PCRE now takes note
2247 of whether there can be an explicit match for \r or \n anywhere in the
2248 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2249 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2250 pattern has explicit CR or LF references.
2251
2252 29. Added (*CR) etc for changing newline setting at start of pattern.
2253
2254
2255 Version 7.2 19-Jun-07
2256 ---------------------
2257
2258 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2259 which is apparently normally available under Windows.
2260
2261 2. Re-jig the pcregrep tests with different newline settings in an attempt
2262 to make them independent of the local environment's newline setting.
2263
2264 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2265
2266 4. Some of the "internals" tests were previously cut out when the link size
2267 was not 2, because the output contained actual offsets. The recent new
2268 "Z" feature of pcretest means that these can be cut out, making the tests
2269 usable with all link sizes.
2270
2271 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2272 stack recursion. This gives a massive performance boost under BSD, but just
2273 a small improvement under Linux. However, it saves one field in the frame
2274 in all cases.
2275
2276 6. Added more features from the forthcoming Perl 5.10:
2277
2278 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2279 recursion call. It refers to the nth most recently opened parentheses.
2280
2281 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2282 to be opened parentheses.
2283
2284 (c) Conditions that refer to capturing parentheses can be specified
2285 relatively, for example, (?(-2)... or (?(+3)...
2286
2287 (d) \K resets the start of the current match so that everything before
2288 is not part of it.
2289
2290 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2291
2292 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2293 reference syntax.
2294
2295 (g) (?| introduces a group in which the numbering of parentheses in each
2296 alternative starts with the same number.
2297
2298 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2299
2300 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2301 PCRE_INFO_JCHANGED.
2302
2303 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2304 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2305 for detecting groups that can match an empty string.
2306
2307 9. A pattern with a very large number of alternatives (more than several
2308 hundred) was running out of internal workspace during the pre-compile
2309 phase, where pcre_compile() figures out how much memory will be needed. A
2310 bit of new cunning has reduced the workspace needed for groups with
2311 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2312 workspace instead of running out of the 4096 that are available.
2313
2314 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2315
2316 11. Applied patch from Google to remove an optimization that didn't quite work.
2317 The report of the bug said:
2318
2319 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2320 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2321 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2322
2323 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2324 it matched the wrong number of bytes.
2325
2326
2327 Version 7.1 24-Apr-07
2328 ---------------------
2329
2330 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2331 that is more "standard", making use of automake and other Autotools. There
2332 is some re-arrangement of the files and adjustment of comments consequent
2333 on this.
2334
2335 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2336 for recursive directory scanning broke on some systems because the files
2337 are not scanned in any specific order and on different systems the order
2338 was different. A call to "sort" has been inserted into RunGrepTest for the
2339 approprate test as a short-term fix. In the longer term there may be an
2340 alternative.
2341
2342 3. I had an email from Eric Raymond about problems translating some of PCRE's
2343 man pages to HTML (despite the fact that I distribute HTML pages, some
2344 people do their own conversions for various reasons). The problems
2345 concerned the use of low-level troff macros .br and .in. I have therefore
2346 removed all such uses from the man pages (some were redundant, some could
2347 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2348 HTML has been updated to handle .nf/.fi and to complain if it encounters
2349 .br or .in.
2350
2351 4. Updated comments in configure.ac that get placed in config.h.in and also
2352 arranged for config.h to be included in the distribution, with the name
2353 config.h.generic, for the benefit of those who have to compile without
2354 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2355
2356 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2357 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2358 makevp.bat for the current PCRE, using the additional files
2359 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2360
2361 6. A Windows user reported a minor discrepancy with test 2, which turned out
2362 to be caused by a trailing space on an input line that had got lost in his
2363 copy. The trailing space was an accident, so I've just removed it.
2364
2365 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2366 that is needed.
2367
2368 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2369 as "const" (a) because they are and (b) because it helps the PHP
2370 maintainers who have recently made a script to detect big data structures
2371 in the php code that should be moved to the .rodata section. I remembered
2372 to update Builducptable as well, so it won't revert if ucptable.h is ever
2373 re-created.
2374
2375 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2376 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2377 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2378 support is not required. This saves 1.5-2K of code, which is important in
2379 some applications.
2380
2381 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2382 so as not to refer to the tables, even though these functions will never be
2383 called when UTF-8 support is disabled. Otherwise there are problems with a
2384 shared library.
2385
2386 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2387
2388 (a) It was defining its arguments as char * instead of void *.
2389
2390 (b) It was assuming that all moves were upwards in memory; this was true
2391 a long time ago when I wrote it, but is no longer the case.
2392
2393 The emulated memove() is provided for those environments that have neither
2394 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2395 is clearly not the case, as these two bugs were recently reported.
2396
2397 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2398 and Detrail to create the HTML documentation, the .txt form of the man
2399 pages, and it removes trailing spaces from listed files. It also creates
2400 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2401 case, it wraps all the #defines with #ifndefs. This script should be run
2402 before "make dist".
2403
2404 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2405 with Unicode property support.
2406
2407 (a) For a maximizing quantifier, if the two different cases of the
2408 character were of different lengths in their UTF-8 codings (there are
2409 some cases like this - I found 11), and the matching function had to
2410 back up over a mixture of the two cases, it incorrectly assumed they
2411 were both the same length.
2412
2413 (b) When PCRE was configured to use the heap rather than the stack for
2414 recursion during matching, it was not correctly preserving the data for
2415 the other case of a UTF-8 character when checking ahead for a match
2416 while processing a minimizing repeat. If the check also involved
2417 matching a wide character, but failed, corruption could cause an
2418 erroneous result when trying to check for a repeat of the original
2419 character.
2420
2421 13. Some tidying changes to the testing mechanism:
2422
2423 (a) The RunTest script now detects the internal link size and whether there
2424 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2425 values substituted by "configure". (The RunGrepTest script already did
2426 this for UTF-8.) The configure.ac script no longer substitutes the
2427 relevant variables.
2428
2429 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2430 with length and offset values. This means that the output is different
2431 for different internal link sizes. Test 2 is skipped for link sizes
2432 other than 2 because of this, bypassing the problem. Unfortunately,
2433 there was also a test in test 3 (the locale tests) that used /B and
2434 failed for link sizes other than 2. Rather than cut the whole test out,
2435 I have added a new /Z option to pcretest that replaces the length and
2436 offset values with spaces. This is now used to make test 3 independent
2437 of link size. (Test 2 will be tidied up later.)
2438
2439 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2440 segmentation fault instead of returning the appropriate error message.
2441
2442 15. In multiline mode when the newline sequence was set to "any", the pattern
2443 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2444 This doesn't seem right; it now treats the CRLF combination as the line
2445 ending, and so does not match in that case. It's only a pattern such as ^$
2446 that would hit this one: something like ^ABC$ would have failed after \r
2447 and then tried again after \r\n.
2448
2449 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2450 in an attempt to make files that differ only in their line terminators
2451 compare equal. This works on Linux.
2452
2453 17. Under certain error circumstances pcregrep might try to free random memory
2454 as it exited. This is now fixed, thanks to valgrind.
2455
2456 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2457 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2458 was because its rules for how to advance for /g after matching an empty
2459 string at the end of a line did not allow for this case. They now check for
2460 it specially.
2461
2462 20. pcretest is supposed to handle patterns and data of any length, by
2463 extending its buffers when necessary. It was getting this wrong when the
2464 buffer for a data line had to be extended.
2465
2466 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2467 CRLF as a newline sequence.
2468
2469 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2470 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2471 I have nevertheless tidied it up.
2472
2473 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2474
2475 24. Added a man page for pcre-config.
2476
2477
2478 Version 7.0 19-Dec-06
2479 ---------------------
2480
2481 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2482 moving to gcc 4.1.1.
2483
2484 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2485 sys/time.h, which is documented as needed for this function. It doesn't
2486 seem to matter on Linux, but it showed up on some releases of OS X.
2487
2488 3. It seems that there are systems where bytes whose values are greater than
2489 127 match isprint() in the "C" locale. The "C" locale should be the
2490 default when a C program starts up. In most systems, only ASCII printing
2491 characters match isprint(). This difference caused the output from pcretest
2492 to vary, making some of the tests fail. I have changed pcretest so that:
2493
2494 (a) When it is outputting text in the compiled version of a pattern, bytes
2495 other than 32-126 are always shown as hex escapes.
2496
2497 (b) When it is outputting text that is a matched part of a subject string,
2498 it does the same, unless a different locale has been set for the match
2499 (using the /L modifier). In this case, it uses isprint() to decide.
2500
2501 4. Fixed a major bug that caused incorrect computation of the amount of memory
2502 required for a compiled pattern when options that changed within the
2503 pattern affected the logic of the preliminary scan that determines the
2504 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2505 that the computed length was too small. The symptoms of this bug were
2506 either the PCRE error "internal error: code overflow" from pcre_compile(),
2507 or a glibc crash with a message such as "pcretest: free(): invalid next
2508 size (fast)". Examples of patterns that provoked this bug (shown in
2509 pcretest format) are:
2510
2511 /(?-x: )/x
2512 /(?x)(?-x: \s*#\s*)/
2513 /((?i)[\x{c0}])/8
2514 /(?i:[\x{c0}])/8
2515
2516 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2517 is now done differently.
2518
2519 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2520 wrapper classes; (b) implement a new function in the C++ scanner that is
2521 more efficient than the old way of doing things because it avoids levels of
2522 recursion in the regex matching; (c) add a paragraph to the documentation
2523 for the FullMatch() function.
2524
2525 6. The escape sequence \n was being treated as whatever was defined as
2526 "newline". Not only was this contrary to the documentation, which states
2527 that \n is character 10 (hex 0A), but it also went horribly wrong when
2528 "newline" was defined as CRLF. This has been fixed.
2529
2530 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2531 was being set to -1 for the "end of line" case (supposedly a value that no
2532 character can have). Though this value is never used (the check for end of
2533 line is "zero bytes in current character"), it caused compiler complaints.
2534 I've changed it to 0xffffffff.
2535
2536 8. In pcre_version.c, the version string was being built by a sequence of
2537 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2538 string (as it is for production releases) called a macro with an empty
2539 argument. The C standard says the result of this is undefined. The gcc
2540 compiler treats it as an empty string (which was what was wanted) but it is
2541 reported that Visual C gives an error. The source has been hacked around to
2542 avoid this problem.
2543
2544 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2545 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2546 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2547 of them did).
2548
2549 10. Originally, pcretest opened its input and output without "b"; then I was
2550 told that "b" was needed in some environments, so it was added for release
2551 5.0 to both the input and output. (It makes no difference on Unix-like
2552 systems.) Later I was told that it is wrong for the input on Windows. I've
2553 now abstracted the modes into two macros, to make it easier to fiddle with
2554 them, and removed "b" from the input mode under Windows.
2555
2556 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2557
2558 12. Added -help and --help to pcretest as an official way of being reminded
2559 of the options.
2560
2561 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2562 and pcrecpp.cc because they annoy compilers at high warning levels.
2563
2564 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2565
2566 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2567 scripts are not C programs :-) and which was not noticed because it works
2568 on Linux.
2569
2570 16. pcretest is supposed to handle any length of pattern and data line (as one
2571 line or as a continued sequence of lines) by extending its input buffer if
2572 necessary. This feature was broken for very long pattern lines, leading to
2573 a string of junk being passed to pcre_compile() if the pattern was longer
2574 than about 50K.
2575
2576 17. I have done a major re-factoring of the way pcre_compile() computes the
2577 amount of memory needed for a compiled pattern. Previously, there was code
2578 that made a preliminary scan of the pattern in order to do this. That was
2579 OK when PCRE was new, but as the facilities have expanded, it has become
2580 harder and harder to keep it in step with the real compile phase, and there
2581 have been a number of bugs (see for example, 4 above). I have now found a
2582 cunning way of running the real compile function in a "fake" mode that
2583 enables it to compute how much memory it would need, while actually only
2584 ever using a few hundred bytes of working memory and without too many
2585 tests of the mode. This should make future maintenance and development
2586 easier. A side effect of this work is that the limit of 200 on the nesting
2587 depth of parentheses has been removed (though this was never a serious
2588 limitation, I suspect). However, there is a downside: pcre_compile() now
2589 runs more slowly than before (30% or more, depending on the pattern). I
2590 hope this isn't a big issue. There is no effect on runtime performance.
2591
2592 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2593 newline (only possible for the last line of a file) and it was a
2594 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2595
2596 19. Added additional timing features to pcretest. (1) The -tm option now times
2597 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2598 separate command line item, by a number that specifies the number of
2599 repeats to use when timing. The default is 50000; this gives better
2600 precision, but takes uncomfortably long for very large patterns.
2601
2602 20. Extended pcre_study() to be more clever in cases where a branch of a
2603 subpattern has no definite first character. For example, (a*|b*)[cd] would
2604 previously give no result from pcre_study(). Now it recognizes that the
2605 first character must be a, b, c, or d.
2606
2607 21. There was an incorrect error "recursive call could loop indefinitely" if
2608 a subpattern (or the entire pattern) that was being tested for matching an
2609 empty string contained only one non-empty item after a nested subpattern.
2610 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2611 incorrectly, because the \d was being skipped in the check.
2612
2613 22. The pcretest program now has a new pattern option /B and a command line
2614 option -b, which is equivalent to adding /B to every pattern. This causes
2615 it to show the compiled bytecode, without the additional information that
2616 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2617 is the same as /B/I).
2618
2619 23. A new optimization is now able automatically to treat some sequences such
2620 as a*b as a*+b. More specifically, if something simple (such as a character
2621 or a simple class like \d) has an unlimited quantifier, and is followed by
2622 something that cannot possibly match the quantified thing, the quantifier
2623 is automatically "possessified".
2624
2625 24. A recursive reference to a subpattern whose number was greater than 39
2626 went wrong under certain circumstances in UTF-8 mode. This bug could also
2627 have affected the operation of pcre_study().
2628
2629 25. Realized that a little bit of performance could be had by replacing
2630 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2631
2632 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2633
2634 27. Possessive quantifiers such as a++ were previously implemented by turning
2635 them into atomic groups such as ($>a+). Now they have their own opcodes,
2636 which improves performance. This includes the automatically created ones
2637 from 23 above.
2638
2639 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2640 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2641 the first matched character to be a colon. This applied both to named and
2642 numbered groups.
2643
2644 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2645
2646 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2647 building PCRE on Windows easier, so I have included it in the distribution.
2648
2649 31. There is now a check in pcretest against a ridiculously large number being
2650 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2651 loop, the loop is abandoned.
2652
2653 32. Forward references to subpatterns in conditions such as (?(2)...) where
2654 subpattern 2 is defined later cause pcre_compile() to search forwards in
2655 the pattern for the relevant set of parentheses. This search went wrong
2656 when there were unescaped parentheses in a character class, parentheses
2657 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2658
2659 33. "Subroutine" calls and backreferences were previously restricted to
2660 referencing subpatterns earlier in the regex. This restriction has now
2661 been removed.
2662
2663 34. Added a number of extra features that are going to be in Perl 5.10. On the
2664 whole, these are just syntactic alternatives for features that PCRE had
2665 previously implemented using the Python syntax or my own invention. The
2666 other formats are all retained for compatibility.
2667
2668 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2669 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2670 also .NET compatible.
2671
2672 (b) A recursion or subroutine call to a named group can now be defined as
2673 (?&name) as well as (?P>name).
2674
2675 (c) A backreference to a named group can now be defined as \k<name> or
2676 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2677 5.10, are also .NET compatible.
2678
2679 (d) A conditional reference to a named group can now use the syntax
2680 (?(<name>) or (?('name') as well as (?(name).
2681
2682 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2683 groups (named and numbered) that are never evaluated inline, but can be
2684 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2685 is always false. There may be only one alternative in such a group.
2686
2687 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2688 as the simple (?(R). The condition is true only if the most recent
2689 recursion is that of the given number or name. It does not search out
2690 through the entire recursion stack.
2691
2692 (g) The escape \gN or \g{N} has been added, where N is a positive or
2693 negative number, specifying an absolute or relative reference.
2694
2695 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2696 some "unreachable code" warnings.
2697
2698 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2699 things, this adds five new scripts.
2700
2701 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2702 There were also incompatibilities regarding the handling of \Q..\E inside
2703 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2704 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2705
2706 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2707 matches an empty string, and forcibly breaks the loop. There were bugs in
2708 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2709 against aaaa the result was just "a" rather than "aaaa", for example. Two
2710 separate and independent bugs (that affected different cases) have been
2711 fixed.
2712
2713 39. Refactored the code to abolish the use of different opcodes for small
2714 capturing bracket numbers. This is a tidy that I avoided doing when I
2715 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2716 The new approach is not only tidier, it makes it possible to reduce the
2717 memory needed to fix the previous bug (38).
2718
2719 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2720 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2721 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2722 mode.
2723
2724 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2725 report.
2726
2727 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2728 copy construction and assignment in the C++ wrapper.
2729
2730 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2731 couple of bugs that could have given wrong results in the "--newline=crlf"
2732 case.
2733
2734 44. Added a number of casts and did some reorganization of signed/unsigned int
2735 variables following suggestions from Dair Grant. Also renamed the variable
2736 "this" as "item" because it is a C++ keyword.
2737
2738 45. Arranged for dftables to add
2739
2740 #include "pcre_internal.h"
2741
2742 to pcre_chartables.c because without it, gcc 4.x may remove the array
2743 definition from the final binary if PCRE is built into a static library and
2744 dead code stripping is activated.
2745
2746 46. For an unanchored pattern, if a match attempt fails at the start of a
2747 newline sequence, and the newline setting is CRLF or ANY, and the next two
2748 characters are CRLF, advance by two characters instead of one.
2749
2750
2751 Version 6.7 04-Jul-06
2752 ---------------------
2753
2754 1. In order to handle tests when input lines are enormously long, pcretest has
2755 been re-factored so that it automatically extends its buffers when
2756 necessary. The code is crude, but this _is_ just a test program. The
2757 default size has been increased from 32K to 50K.
2758
2759 2. The code in pcre_study() was using the value of the re argument before
2760 testing it for NULL. (Of course, in any sensible call of the function, it
2761 won't be NULL.)
2762
2763 3. The memmove() emulation function in pcre_internal.h, which is used on
2764 systems that lack both memmove() and bcopy() - that is, hardly ever -
2765 was missing a "static" storage class specifier.
2766
2767 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2768 containing an extended class (one that cannot be represented by a bitmap
2769 because it contains high-valued characters or Unicode property items, e.g.
2770 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2771 pattern, but PCRE should not loop if you do not (it no longer does).
2772 [Detail: two cases were found: (a) a repeated subpattern containing an
2773 extended class; (b) a recursive reference to a subpattern that followed a
2774 previous extended class. It wasn't skipping over the extended class
2775 correctly when UTF-8 mode was not set.]
2776
2777 5. A negated single-character class was not being recognized as fixed-length
2778 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2779 compile error "lookbehind assertion is not fixed length".
2780
2781 6. The RunPerlTest auxiliary script was showing an unexpected difference
2782 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2783 write a Perl script that can interpret lines of an input file either as
2784 byte characters or as UTF-8, which is what "perltest" was being required to
2785 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2786 can't do is switch easily at run time between having the "use utf8;" pragma
2787 or not. In the end, I fudged it by using the RunPerlTest script to insert
2788 "use utf8;" explicitly for the UTF-8 tests.
2789
2790 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2791 the end of the subject string, contrary to the documentation and to what
2792 Perl does. This was true of both matching functions. Now it matches only at
2793 the start of the subject and immediately after *internal* newlines.
2794
2795 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2796 a pointer to an int instead of a pointer to an unsigned long int. This
2797 caused problems on 64-bit systems.
2798
2799 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2800 instance of the 'standard' template library not being so standard".
2801
2802 10. There was no check on the number of named subpatterns nor the maximum
2803 length of a subpattern name. The product of these values is used to compute
2804 the size of the memory block for a compiled pattern. By supplying a very
2805 long subpattern name and a large number of named subpatterns, the size
2806 computation could be caused to overflow. This is now prevented by limiting
2807 the length of names to 32 characters, and the number of named subpatterns
2808 to 10,000.
2809
2810 11. Subpatterns that are repeated with specific counts have to be replicated in
2811 the compiled pattern. The size of memory for this was computed from the
2812 length of the subpattern and the repeat count. The latter is limited to
2813 65535, but there was no limit on the former, meaning that integer overflow
2814 could in principle occur. The compiled length of a repeated subpattern is
2815 now limited to 30,000 bytes in order to prevent this.
2816
2817 12. Added the optional facility to have named substrings with the same name.
2818
2819 13. Added the ability to use a named substring as a condition, using the
2820 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2821 are numbers (not recommended). Forward references are permitted.
2822
2823 14. Added forward references in named backreferences (if you see what I mean).
2824
2825 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2826 pattern could run off the end of the subject. For example, the pattern
2827 "(?s)(.{1,5})"8 did this with the subject "ab".
2828
2829 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2830 PCRE_CASELESS was set when matching characters that were quantified with ?
2831 or *.
2832
2833 17. A character class other than a single negated character that had a minimum
2834 but no maximum quantifier - for example [ab]{6,} - was not handled
2835 correctly by pce_dfa_exec(). It would match only one character.
2836
2837 18. A valid (though odd) pattern that looked like a POSIX character
2838 class but used an invalid character after [ (for example [[,abc,]]) caused
2839 pcre_compile() to give the error "Failed: internal error: code overflow" or
2840 in some cases to crash with a glibc free() error. This could even happen if
2841 the pattern terminated after [[ but there just happened to be a sequence of
2842 letters, a binary zero, and a closing ] in the memory that followed.
2843
2844 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2845 over the years. Originally (before any Unicode support), just the bottom 8
2846 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2847 output from "man perlunicode" includes this:
2848
2849 The regular expression compiler produces polymorphic opcodes. That
2850 is, the pattern adapts to the data and automatically switches to
2851 the Unicode character scheme when presented with Unicode data--or
2852 instead uses a traditional byte scheme when presented with byte
2853 data.
2854
2855 Sadly, a wide octal escape does not cause a switch, and in a string with
2856 no other multibyte characters, these octal escapes are treated as before.
2857 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2858 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2859 Unicode string.
2860
2861 I have not perpetrated such confusion in PCRE. Up till now, it took just
2862 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2863 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2864 translate to the appropriate multibyte character.
2865
2866 29. Applied some refactoring to reduce the number of warnings from Microsoft
2867 and Borland compilers. This has included removing the fudge introduced
2868 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2869 a warning about an unused variable.
2870
2871 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2872 characters since release 4.0, because Perl (from release 5.004) does not.
2873 [Or at least, is documented not to: some releases seem to be in conflict
2874 with the documentation.] However, when a pattern was studied with
2875 pcre_study() and all its branches started with \s, PCRE still included VT
2876 as a possible starting character. Of course, this did no harm; it just
2877 caused an unnecessary match attempt.
2878
2879 22. Removed a now-redundant internal flag bit that recorded the fact that case
2880 dependency changed within the pattern. This was once needed for "required
2881 byte" processing, but is no longer used. This recovers a now-scarce options
2882 bit. Also moved the least significant internal flag bit to the most-
2883 significant bit of the word, which was not previously used (hangover from
2884 the days when it was an int rather than a uint) to free up another bit for
2885 the future.
2886
2887 23. Added support for CRLF line endings as well as CR and LF. As well as the
2888 default being selectable at build time, it can now be changed at runtime
2889 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2890 specify that it is scanning data with non-default line endings.
2891
2892 24. Changed the definition of CXXLINK to make it agree with the definition of
2893 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2894
2895 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2896 recursions. This makes a big different to stack usage for some patterns.
2897
2898 26. If a subpattern containing a named recursion or subroutine reference such
2899 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2900 the space required for the compiled pattern went wrong and gave too small a
2901 value. Depending on the environment, this could lead to "Failed: internal
2902 error: code overflow at offset 49" or "glibc detected double free or
2903 corruption" errors.
2904
2905 27. Applied patches from Google (a) to support the new newline modes and (b) to
2906 advance over multibyte UTF-8 characters in GlobalReplace.
2907
2908 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2909 difference for some implementation of PCRE in some Windows version.
2910
2911 29. Added some extra testing facilities to pcretest:
2912
2913 \q<number> in a data line sets the "match limit" value
2914 \Q<number> in a data line sets the "match recursion limt" value
2915 -S <number> sets the stack size, where <number> is in megabytes
2916
2917 The -S option isn't available for Windows.
2918
2919
2920 Version 6.6 06-Feb-06
2921 ---------------------
2922
2923 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2924 in pcreposix.h. I have copied the definition from pcre.h.
2925
2926 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2927 because pcre.h is no longer a built file.
2928
2929 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2930 not normally included in the compiled code.
2931
2932
2933 Version 6.5 01-Feb-06
2934 ---------------------
2935
2936 1. When using the partial match feature with pcre_dfa_exec(), it was not
2937 anchoring the second and subsequent partial matches at the new starting
2938 point. This could lead to incorrect results. For example, with the pattern
2939 /1234/, partially matching against "123" and then "a4" gave a match.
2940
2941 2. Changes to pcregrep:
2942
2943 (a) All non-match returns from pcre_exec() were being treated as failures
2944 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2945 error message is output. Some extra information is given for the
2946 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2947 probably the only errors that are likely to be caused by users (by
2948 specifying a regex that has nested indefinite repeats, for instance).
2949 If there are more than 20 of these errors, pcregrep is abandoned.
2950
2951 (b) A binary zero was treated as data while matching, but terminated the
2952 output line if it was written out. This has been fixed: binary zeroes
2953 are now no different to any other data bytes.
2954
2955 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2956 used to set a locale for matching. The --locale=xxxx long option has
2957 been added (no short equivalent) to specify a locale explicitly on the
2958 pcregrep command, overriding the environment variables.
2959
2960 (d) When -B was used with -n, some line numbers in the output were one less
2961 than they should have been.
2962
2963 (e) Added the -o (--only-matching) option.
2964
2965 (f) If -A or -C was used with -c (count only), some lines of context were
2966 accidentally printed for the final match.
2967
2968 (g) Added the -H (--with-filename) option.
2969
2970 (h) The combination of options -rh failed to suppress file names for files
2971 that were found from directory arguments.
2972
2973 (i) Added the -D (--devices) and -d (--directories) options.
2974
2975 (j) Added the -F (--fixed-strings) option.
2976
2977 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2978
2979 (l) Added the --colo(u)r option.
2980
2981 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2982 is not present by default.
2983
2984 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2985 items such as (?R) or (?1), when the recursion could match a number of
2986 alternatives. If it matched one of the alternatives, but subsequently,
2987 outside the recursion, there was a failure, the code tried to back up into
2988 the recursion. However, because of the way PCRE is implemented, this is not
2989 possible, and the result was an incorrect result from the match.
2990
2991 In order to prevent this happening, the specification of recursion has
2992 been changed so that all such subpatterns are automatically treated as
2993 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2994
2995 4. I had overlooked the fact that, in some locales, there are characters for
2996 which isalpha() is true but neither isupper() nor islower() are true. In
2997 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2998 and ordfeminine) are like this. This affected the treatment of \w and \W
2999 when they appeared in character classes, but not when they appeared outside
3000 a character class. The bit map for "word" characters is now created
3001 separately from the results of isalnum() instead of just taking it from the
3002 upper, lower, and digit maps. (Plus the underscore character, of course.)
3003
3004 5. The above bug also affected the handling of POSIX character classes such as
3005 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3006 permanent tables. Instead, the bit maps for such a class were previously
3007 created as the appropriate unions of the upper, lower, and digit bitmaps.
3008 Now they are created by subtraction from the [[:word:]] class, which has
3009 its own bitmap.
3010
3011 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3012 It is created by subtracting the vertical space characters (\x09, \x0a,
3013 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3014 subtraction was done in the overall bitmap for a character class, meaning
3015 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3016 be recognized. This bug has been fixed.
3017
3018 7. Patches from the folks at Google:
3019
3020 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3021 real life, but is still worth protecting against".
3022
3023 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3024 regular expressions".
3025
3026 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3027 have it.
3028
3029 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3030 "configure" and the latter not, in order to fix a problem somebody had
3031 with compiling the Arg class on HP-UX.
3032
3033 (e) Improve the error-handling of the C++ wrapper a little bit.
3034
3035 (f) New tests for checking recursion limiting.
3036
3037 8. The pcre_memmove() function, which is used only if the environment does not
3038 have a standard memmove() function (and is therefore rarely compiled),
3039 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3040 returning a result (though PCRE never actually uses the result).
3041
3042 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3043 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3044 returned instead of calling malloc() with an overflowing number that would
3045 most likely cause subsequent chaos.
3046
3047 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3048
3049 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3050 with this option is matched, the nmatch and pmatch options of regexec() are
3051 ignored.
3052
3053 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3054 provided in case anyone wants to the the POSIX interface with UTF-8
3055 strings.
3056
3057 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3058 C++ linking (needed for some HP-UX environments).
3059
3060 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3061 (unused parameter) and in the pcre_printint() function (omitted "default"
3062 switch label when the default is to do nothing).
3063
3064 15. Added some code to make it possible, when PCRE is compiled as a C++
3065 library, to replace subject pointers for pcre_exec() with a smart pointer
3066 class, thus making it possible to process discontinuous strings.
3067
3068 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3069 much the same function. They were added by different people who were trying
3070 to make PCRE easy to compile on non-Unix systems. It has been suggested
3071 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3072 for compiling on Windows systems. I have therefore replaced it with
3073 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3074 defaults to "extern" for C or "extern C" for C++, which works fine on
3075 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3076 SCOPE with something explicit in config.h. In addition:
3077
3078 (a) pcreposix.h still had just "extern" instead of either of these macros;
3079 I have replaced it with PCRE_DATA_SCOPE.
3080
3081 (b) Functions such as _pcre_xclass(), which are internal to the library,
3082 but external in the C sense, all had PCRE_EXPORT in their definitions.
3083 This is apparently wrong for the Windows case, so I have removed it.
3084 (It makes no difference on Unix-like systems.)
3085
3086 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3087 of recursive calls to match(). This is different to MATCH_LIMIT because
3088 that limits the total number of calls to match(), not all of which increase
3089 the depth of recursion. Limiting the recursion depth limits the amount of
3090 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3091 when PCRE is compiled, and changed at run time. A patch from Google adds
3092 this functionality to the C++ interface.
3093
3094 18. Changes to the handling of Unicode character properties:
3095
3096 (a) Updated the table to Unicode 4.1.0.
3097
3098 (b) Recognize characters that are not in the table as "Cn" (undefined).
3099
3100 (c) I revised the way the table is implemented to a much improved format
3101 which includes recognition of ranges. It now supports the ranges that
3102 are defined in UnicodeData.txt, and it also amalgamates other
3103 characters into ranges. This has reduced the number of entries in the
3104 table from around 16,000 to around 3,000, thus reducing its size
3105 considerably. I realized I did not need to use a tree structure after
3106 all - a binary chop search is just as efficient. Having reduced the
3107 number of entries, I extended their size from 6 bytes to 8 bytes to
3108 allow for more data.
3109
3110 (d) Added support for Unicode script names via properties such as \p{Han}.
3111
3112 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3113 matching that character.
3114
3115 20. When matching a repeated Unicode property with a minimum greater than zero,
3116 (for example \pL{2,}), PCRE could look past the end of the subject if it
3117 reached it while seeking the minimum number of characters. This could
3118 happen only if some of the characters were more than one byte long, because
3119 there is a check for at least the minimum number of bytes.
3120
3121 21. Refactored the implementation of \p and \P so as to be more general, to
3122 allow for more different types of property in future. This has changed the
3123 compiled form incompatibly. Anybody with saved compiled patterns that use
3124 \p or \P will have to recompile them.
3125
3126 22. Added "Any" and "L&" to the supported property types.
3127
3128 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3129 but give a compile time error if the value is greater than 0xff.
3130
3131 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3132 accidentally not being installed or uninstalled.
3133
3134 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3135 made were to insert the current release number. This seemed silly, because
3136 it made things harder for people building PCRE on systems that don't run
3137 "configure". I have turned pcre.h into a distributed file, no longer built
3138 by "configure", with the version identification directly included. There is
3139 no longer a pcre.h.in file.
3140
3141 However, this change necessitated a change to the pcre-config script as
3142 well. It is built from pcre-config.in, and one of the substitutions was the
3143 release number. I have updated configure.ac so that ./configure now finds
3144 the release number by grepping pcre.h.
3145
3146 26. Added the ability to run the tests under valgrind.
3147
3148
3149 Version 6.4 05-Sep-05
3150 ---------------------
3151
3152 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3153 "--" to be printed when multiple files were scanned, even when none of the
3154 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3155 consider it to be a bug, and have restored the previous behaviour.
3156
3157 2. A couple of code tidies to get rid of compiler warnings.
3158
3159 3. The pcretest program used to cheat by referring to symbols in the library
3160 whose names begin with _pcre_. These are internal symbols that are not
3161 really supposed to be visible externally, and in some environments it is
3162 possible to suppress them. The cheating is now confined to including
3163 certain files from the library's source, which is a bit cleaner.
3164
3165 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3166 file's purpose clearer.
3167
3168 5. Reorganized pcre_ucp_findchar().
3169
3170
3171 Version 6.3 15-Aug-05
3172 ---------------------
3173
3174 1. The file libpcre.pc.in did not have general read permission in the tarball.
3175
3176 2. There were some problems when building without C++ support:
3177
3178 (a) If C++ support was not built, "make install" and "make test" still
3179 tried to test it.
3180
3181 (b) There were problems when the value of CXX was explicitly set. Some
3182 changes have been made to try to fix these, and ...
3183
3184 (c) --disable-cpp can now be used to explicitly disable C++ support.
3185
3186 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3187 backslash in a target when C++ was disabled. This confuses some
3188 versions of "make", apparently. Using an intermediate variable solves
3189 this. (Same for CPP_LOBJ.)
3190
3191 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3192 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3193 necessary on certain architectures.
3194
3195 4. Added a setting of -export-symbols-regex to the link command to remove
3196 those symbols that are exported in the C sense, but actually are local
3197 within the library, and not documented. Their names all begin with
3198 "_pcre_". This is not a perfect job, because (a) we have to except some
3199 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3200 available (and never for static libraries). I have made a note to try to
3201 find a way round (a) in the future.
3202
3203
3204 Version 6.2 01-Aug-05
3205 ---------------------
3206
3207 1. There was no test for integer overflow of quantifier values. A construction
3208 such as {1111111111111111} would give undefined results. What is worse, if
3209 a minimum quantifier for a parenthesized subpattern overflowed and became
3210 negative, the calculation of the memory size went wrong. This could have
3211 led to memory overwriting.
3212
3213 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3214
3215 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3216 operating environments where this matters.
3217
3218 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3219 PCRE options from within the C++ wrapper.
3220
3221 5. Named capturing subpatterns were not being correctly counted when a pattern
3222 was compiled. This caused two problems: (a) If there were more than 100
3223 such subpatterns, the calculation of the memory needed for the whole
3224 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3225 back references of the form \12, where the number was greater than 9, were
3226 not recognized as back references, even though there were sufficient
3227 previous subpatterns.
3228
3229 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3230 versions of gcc, e.g. 2.95.4.
3231
3232
3233 Version 6.1 21-Jun-05
3234 ---------------------
3235
3236 1. There was one reference to the variable "posix" in pcretest.c that was not
3237 surrounded by "#if !defined NOPOSIX".
3238
3239 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3240 the cross-check on the old pcre_info() function, for the benefit of the
3241 cut-down version of PCRE that is currently imported into Exim.
3242
3243 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3244 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3245 patterns that start (?i) but I don't think that matters. The use of (?i) is
3246 just an example; this all applies to the other options as well.
3247
3248 4. Since libtool seems to echo the compile commands it is issuing, the output
3249 from "make" can be reduced a bit by putting "@" in front of each libtool
3250 compile command.
3251
3252 5. Patch from the folks at Google for configure.in to be a bit more thorough
3253 in checking for a suitable C++ installation before trying to compile the
3254 C++ stuff. This should fix a reported problem when a compiler was present,
3255 but no suitable headers.
3256
3257 6. The man pages all had just "PCRE" as their title. I have changed them to
3258 be the relevant file name. I have also arranged that these names are
3259 retained in the file doc/pcre.txt, which is a concatenation in text format
3260 of all the man pages except the little individual ones for each function.
3261
3262 7. The NON-UNIX-USE file had not been updated for the different set of source
3263 files that come with release 6. I also added a few comments about the C++
3264 wrapper.
3265
3266
3267 Version 6.0 07-Jun-05
3268 ---------------------
3269
3270 1. Some minor internal re-organization to help with my DFA experiments.
3271
3272 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3273 didn't matter for the library itself when fully configured, but did matter
3274 when compiling without UCP support, or within Exim, where the ucp files are
3275 not imported.
3276
3277 3. Refactoring of the library code to split up the various functions into
3278 different source modules. The addition of the new DFA matching code (see
3279 below) to a single monolithic source would have made it really too
3280 unwieldy, quite apart from causing all the code to be include in a
3281 statically linked application, when only some functions are used. This is
3282 relevant even without the DFA addition now that patterns can be compiled in
3283 one application and matched in another.
3284
3285 The downside of splitting up is that there have to be some external
3286 functions and data tables that are used internally in different modules of
3287 the library but which are not part of the API. These have all had their
3288 names changed to start with "_pcre_" so that they are unlikely to clash
3289 with other external names.
3290
3291 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3292 a different (DFA) algorithm. Although it is slower than the original
3293 function, it does have some advantages for certain types of matching
3294 problem.
3295
3296 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3297 including restarting after a partial match.
3298
3299 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3300 defined when compiling for Windows was sent to me. I have put it into the
3301 code, though I have no means of testing or verifying it.
3302
3303 7. Added the pcre_refcount() auxiliary function.
3304
3305 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3306 match before or at the first newline in the subject string. In pcretest,
3307 the /f option on a pattern can be used to set this.
3308
3309 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3310 would behave wrongly. This has been present in PCRE since release 4.0.
3311
3312 10. A number of changes to the pcregrep command:
3313
3314 (a) Refactored how -x works; insert ^(...)$ instead of setting
3315 PCRE_ANCHORED and checking the length, in preparation for adding
3316 something similar for -w.
3317
3318 (b) Added the -w (match as a word) option.
3319
3320 (c) Refactored the way lines are read and buffered so as to have more
3321 than one at a time available.
3322
3323 (d) Implemented a pcregrep test script.
3324
3325 (e) Added the -M (multiline match) option. This allows patterns to match
3326 over several lines of the subject. The buffering ensures that at least
3327 8K, or the rest of the document (whichever is the shorter) is available
3328 for matching (and similarly the previous 8K for lookbehind assertions).
3329
3330 (f) Changed the --help output so that it now says
3331
3332 -w, --word-regex(p)
3333
3334 instead of two lines, one with "regex" and the other with "regexp"
3335 because that confused at least one person since the short forms are the
3336 same. (This required a bit of code, as the output is generated
3337 automatically from a table. It wasn't just a text change.)
3338
3339 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3340 option but starts with a hyphen. Could be a pattern or a path name
3341 starting with a hyphen, for instance.
3342
3343 (h) "-" can be given as a file name to represent stdin.
3344
3345 (i) When file names are being printed, "(standard input)" is used for
3346 the standard input, for compatibility with GNU grep. Previously
3347 "<stdin>" was used.
3348
3349 (j) The option --label=xxx can be used to supply a name to be used for
3350 stdin when file names are being printed. There is no short form.
3351
3352 (k) Re-factored the options decoding logic because we are going to add
3353 two more options that take data. Such options can now be given in four
3354 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3355
3356 (l) Added the -A, -B, and -C options for requesting that lines of context
3357 around matches be printed.
3358
3359 (m) Added the -L option to print the names of files that do not contain
3360 any matching lines, that is, the complement of -l.
3361
3362 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3363 continue to scan other files.
3364
3365 (o) The -s option was incorrectly implemented. For compatibility with other
3366 greps, it now suppresses the error message for a non-existent or non-
3367 accessible file (but not the return code). There is a new option called
3368 -q that suppresses the output of matching lines, which was what -s was
3369 previously doing.
3370
3371 (p) Added --include and --exclude options to specify files for inclusion
3372 and exclusion when recursing.
3373
3374 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3375 Hopefully, it now does.
3376
3377 12. Missing cast in pcre_study().
3378
3379 13. Added an "uninstall" target to the makefile.
3380
3381 14. Replaced "extern" in the function prototypes in Makefile.in with
3382 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3383 world, but is set differently for Windows.
3384
3385 15. Added a second compiling function called pcre_compile2(). The only
3386 difference is that it has an extra argument, which is a pointer to an
3387 integer error code. When there is a compile-time failure, this is set
3388 non-zero, in addition to the error test pointer being set to point to an
3389 error message. The new argument may be NULL if no error number is required
3390 (but then you may as well call pcre_compile(), which is now just a
3391 wrapper). This facility is provided because some applications need a
3392 numeric error indication, but it has also enabled me to tidy up the way
3393 compile-time errors are handled in the POSIX wrapper.
3394
3395 16. Added VPATH=.libs to the makefile; this should help when building with one
3396 prefix path and installing with another. (Or so I'm told by someone who
3397 knows more about this stuff than I do.)
3398
3399 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3400 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3401 match everything, including newlines. This is not POSIX-compatible, but
3402 somebody wanted the feature. From pcretest it can be activated by using
3403 both the P and the s flags.
3404
3405 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3406
3407 19. libpcre.pc was being incorrectly installed as executable.
3408
3409 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3410 it now also looks for '\r' so that it will work unmodified on Windows.
3411
3412 21. Added Google's contributed C++ wrapper to the distribution.
3413
3414 22. Added some untidy missing memory free() calls in pcretest, to keep
3415 Electric Fence happy when testing.
3416
3417
3418
3419 Version 5.0 13-Sep-04
3420 ---------------------
3421
3422 1. Internal change: literal characters are no longer packed up into items
3423 containing multiple characters in a single byte-string. Each character
3424 is now matched using a separate opcode. However, there may be more than one
3425 byte in the character in UTF-8 mode.
3426
3427 2. The pcre_callout_block structure has two new fields: pattern_position and
3428 next_item_length. These contain the offset in the pattern to the next match
3429 item, and its length, respectively.
3430
3431 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3432 insertion of callouts before each pattern item. Added the /C option to
3433 pcretest to make use of this.
3434
3435 4. On the advice of a Windows user, the lines
3436
3437 #if defined(_WIN32) || defined(WIN32)
3438 _setmode( _fileno( stdout ), 0x8000 );
3439 #endif /* defined(_WIN32) || defined(WIN32) */
3440
3441 have been added to the source of pcretest. This apparently does useful
3442 magic in relation to line terminators.
3443
3444 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3445 for the benefit of those environments where the "b" makes a difference.
3446
3447 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3448 to know about it. I have put a hack into configure.in that adds in code
3449 to set GCC=yes if CC=icc. This seems to end up at a point in the
3450 generated configure script that is early enough to affect the setting of
3451 compiler options, which is what is needed, but I have no means of testing
3452 whether it really works. (The user who reported this had patched the
3453 generated configure script, which of course I cannot do.)
3454
3455 LATER: After change 22 below (new libtool files), the configure script
3456 seems to know about icc (and also ecc). Therefore, I have commented out
3457 this hack in configure.in.
3458
3459 7. Added support for pkg-config (2 patches were sent in).
3460
3461 8. Negated POSIX character classes that used a combination of internal tables
3462 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3463 [[:^ascii]]. Typically, they would match almost any characters. The other
3464 POSIX classes were not broken in this way.
3465
3466 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3467 to find the match, as PCRE was deluded into thinking that the match had to
3468 start at the start point or following a newline. The same bug applied to
3469 patterns with negative forward assertions or any backward assertions
3470 preceding ".*" at the start, unless the pattern required a fixed first
3471 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3472
3473 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3474 starting at the last subject character, bytes beyond the end of the subject
3475 string were read.
3476
3477 11. Renamed the variable "class" as "classbits" to make life easier for C++
3478 users. (Previously there was a macro definition, but it apparently wasn't
3479 enough.)
3480
3481 12. Added the new field "tables" to the extra data so that tables can be passed
3482 in at exec time, or the internal tables can be re-selected. This allows
3483 a compiled regex to be saved and re-used at a later time by a different
3484 program that might have everything at different addresses.
3485
3486 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3487 -R library as well as a -L library.
3488
3489 14. The debugging options of pcretest (-d on the command line or D on a
3490 pattern) showed incorrect output for anything following an extended class
3491 that contained multibyte characters and which was followed by a quantifier.
3492
3493 15. Added optional support for general category Unicode character properties
3494 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3495 support. It adds about 90K to the size of the library. The meanings of the
3496 inbuilt class escapes such as \d and \s have NOT been changed.
3497
3498 16. Updated pcredemo.c to include calls to free() to release the memory for the
3499 compiled pattern.
3500
3501 17. The generated file chartables.c was being created in the source directory
3502 instead of in the building directory. This caused the build to fail if the
3503 source directory was different from the building directory, and was
3504 read-only.
3505
3506 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3507 file. No doubt somebody will tell me if they don't make sense... Also added
3508 Dan Mooney's comments about building on OpenVMS.
3509
3510 19. Added support for partial matching via the PCRE_PARTIAL option for
3511 pcre_exec() and the \P data escape in pcretest.
3512
3513 20. Extended pcretest with 3 new pattern features:
3514
3515 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3516 write the compiled pattern to the file whose name is "rest-of-line".
3517 This is a straight binary dump of the data, with the saved pointer to
3518 the character tables forced to be NULL. The study data, if any, is
3519 written too. After writing, pcretest reads a new pattern.
3520
3521 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3522 compiled pattern from the given file. There must not be any
3523 occurrences of "<" in the file name (pretty unlikely); if there are,
3524 pcretest will instead treat the initial "<" as a pattern delimiter.
3525 After reading in the pattern, pcretest goes on to read data lines as
3526 usual.
3527
3528 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3529 and 16-bit fields in a compiled pattern, to simulate a pattern that
3530 was compiled on a host of opposite endianness.
3531
3532 21. The pcre-exec() function can now cope with patterns that were compiled on
3533 hosts of opposite endianness, with this restriction:
3534
3535 As for any compiled expression that is saved and used later, the tables
3536 pointer field cannot be preserved; the extra_data field in the arguments
3537 to pcre_exec() should be used to pass in a tables address if a value
3538 other than the default internal tables were used at compile time.
3539
3540 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3541 now diagnosed as an error. Previously, most of the time, a negative number
3542 would have been treated as zero, but if in addition "ovector" was passed as
3543 NULL, a crash could occur.
3544
3545 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3546 new versions from the libtool 1.5 distribution (the last one is a copy of
3547 a file called libtool.m4). This seems to have fixed the need to patch
3548 "configure" to support Darwin 1.3 (which I used to do). However, I still
3549 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3550 workstation).
3551
3552 24. Changed the PCRE licence to be the more standard "BSD" licence.
3553
3554
3555 Version 4.5 01-Dec-03
3556 ---------------------
3557
3558 1. There has been some re-arrangement of the code for the match() function so
3559 that it can be compiled in a version that does not call itself recursively.
3560 Instead, it keeps those local variables that need separate instances for
3561 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3562 needs to "recurse". Keeping track of where control must go is done by means
3563 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3564 hide most of the details from the main code, and operates only if
3565 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3566 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3567 operating.
3568
3569 To make it easier for callers to provide specially tailored get/free
3570 functions for this usage, two new functions, pcre_stack_malloc, and
3571 pcre_stack_free, are used. They are always called in strict stacking order,
3572 and the size of block requested is always the same.
3573
3574 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3575 PCRE has been compiled to use the stack or the heap for recursion. The
3576 -C option of pcretest uses this to show which version is compiled.
3577
3578 A new data escape \S, is added to pcretest; it causes the amounts of store
3579 obtained and freed by both kinds of malloc/free at match time to be added
3580 to the output.
3581
3582 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3583 what's available on my current Linux desktop machine.
3584
3585 3. When matching a UTF-8 string, the test for a valid string at the start has
3586 been extended. If start_offset is not zero, PCRE now checks that it points
3587 to a byte that is the start of a UTF-8 character. If not, it returns
3588 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3589 this is necessary because there may be backward assertions in the pattern.
3590 When matching the same subject several times, it may save resources to use
3591 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3592
3593 4. The code for checking the validity of UTF-8 strings has been tightened so
3594 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3595 containing "overlong sequences".
3596
3597 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3598 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3599 should have been "&", but it just so happened that all the cases this let
3600 through by mistake were picked up later in the function.
3601
3602 6. I had used a variable called "isblank" - this is a C99 function, causing
3603 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3604
3605 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3606 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3607 so the version is shown; (c) stop "make test" echoing "./RunTest".
3608
3609 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3610
3611 9. The prototype for memmove() for systems that don't have it was using
3612 size_t, but the inclusion of the header that defines size_t was later. I've
3613 moved the #includes for the C headers earlier to avoid this.
3614
3615 10. Added some adjustments to the code to make it easier to compiler on certain
3616 special systems:
3617
3618 (a) Some "const" qualifiers were missing.
3619 (b) Added the macro EXPORT before all exported functions; by default this
3620 is defined to be empty.
3621 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3622 that it reads its output file name as an argument instead of writing
3623 to the standard output and assuming this can be redirected.
3624
3625 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3626 class containing characters with values greater than 255, PCRE compilation
3627 went into a loop.
3628
3629 12. A recursive reference to a subpattern that was within another subpattern
3630 that had a minimum quantifier of zero caused PCRE to crash. For example,
3631 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3632 recursion. If the recursively-called subpattern itself had a zero repeat,
3633 that was OK.
3634
3635 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3636 buffer into which it was copied (for escape processing) was still set at
3637 1024, so long lines caused crashes.
3638
3639 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3640 "internal error: code overflow...". This applied to any character class
3641 that was followed by a possessive quantifier.
3642
3643 15. Modified the Makefile to add libpcre.la as a prerequisite for
3644 libpcreposix.la because I was told this is needed for a parallel build to
3645 work.
3646
3647 16. If a pattern that contained .* following optional items at the start was
3648 studied, the wrong optimizing data was generated, leading to matching
3649 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3650 matching string must start with a or b or c. The correct conclusion for
3651 this pattern is that a match can start with any character.
3652
3653
3654 Version 4.4 13-Aug-03
3655 ---------------------
3656
3657 1. In UTF-8 mode, a character class containing characters with values between
3658 127 and 255 was not handled correctly if the compiled pattern was studied.
3659 In fixing this, I have also improved the studying algorithm for such
3660 classes (slightly).
3661
3662 2. Three internal functions had redundant arguments passed to them. Removal
3663 might give a very teeny performance improvement.
3664
3665 3. Documentation bug: the value of the capture_top field in a callout is *one
3666 more than* the number of the hightest numbered captured substring.
3667
3668 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3669 in incorrectly linking with a previously installed version. They now link
3670 explicitly with libpcre.la.
3671
3672 5. configure.in no longer needs to recognize Cygwin specially.
3673
3674 6. A problem in pcre.in for Windows platforms is fixed.
3675
3676 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3677 pcretest, it used to include the size of the study block as part of its
3678 output. Unfortunately, the structure contains a field that has a different
3679 size on different hardware architectures. This meant that the tests that
3680 showed this size failed. As the block is currently always of a fixed size,
3681 this information isn't actually particularly useful in pcretest output, so
3682 I have just removed it.
3683
3684 8. Three pre-processor statements accidentally did not start in column 1.
3685 Sadly, there are *still* compilers around that complain, even though
3686 standard C has not required this for well over a decade. Sigh.
3687
3688 9. In pcretest, the code for checking callouts passed small integers in the
3689 callout_data field, which is a void * field. However, some picky compilers
3690 complained about the casts involved for this on 64-bit systems. Now
3691 pcretest passes the address of the small integer instead, which should get
3692 rid of the warnings.
3693
3694 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3695 both compile and run time, and gives an error if an invalid UTF-8 sequence
3696 is found. There is a option for disabling this check in cases where the
3697 string is known to be correct and/or the maximum performance is wanted.
3698
3699 11. In response to a bug report, I changed one line in Makefile.in from
3700
3701 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3702 to
3703 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3704
3705 to look similar to other lines, but I have no way of telling whether this
3706 is the right thing to do, as I do not use Windows. No doubt I'll get told
3707 if it's wrong...
3708
3709
3710 Version 4.3 21-May-03
3711 ---------------------
3712
3713 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3714 Makefile.
3715
3716 2. Some refactoring to improve the quality of the code:
3717
3718 (i) The utf8_table... variables are now declared "const".
3719
3720 (ii) The code for \cx, which used the "case flipping" table to upper case
3721 lower case letters, now just substracts 32. This is ASCII-specific,
3722 but the whole concept of \cx is ASCII-specific, so it seems
3723 reasonable.
3724
3725 (iii) PCRE was using its character types table to recognize decimal and
3726 hexadecimal digits in the pattern. This is silly, because it handles
3727 only 0-9, a-f, and A-F, but the character types table is locale-
3728 specific, which means strange things might happen. A private
3729 table is now used for this - though it costs 256 bytes, a table is
3730 much faster than multiple explicit tests. Of course, the standard
3731 character types table is still used for matching digits in subject
3732 strings against \d.
3733
3734 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3735 ending in _t are). So I've renamed it as ESC_tee.
3736
3737 3. The first argument for regexec() in the POSIX wrapper should have been
3738 defined as "const".
3739
3740 4. Changed pcretest to use malloc() for its buffers so that they can be
3741 Electric Fenced for debugging.
3742
3743 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3744 to read one or more bytes before the start of the subject string. Often this
3745 had no effect on PCRE's behaviour, but in some circumstances it could
3746 provoke a segmentation fault.
3747
3748 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3749 to try to read one or more bytes before the start of the subject string.
3750
3751 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3752 UTF-8 support could misbehave in various ways if the subject string
3753 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3754 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3755 back over UTF-8 characters.)
3756
3757
3758 Version 4.2 14-Apr-03
3759 ---------------------
3760
3761 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3762
3763 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3764 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3765 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3766 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3767 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3768 and BUILD_EXEEXT
3769 Note: automatic setting of the BUILD variables is not yet working
3770 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3771 compile-time but not at link-time
3772 [LINK]: use for linking executables only
3773 make different versions for Windows and non-Windows
3774 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3775 libraries
3776 [LINK_FOR_BUILD]: new variable
3777 [OBJEXT]: use throughout
3778 [EXEEXT]: use throughout
3779 <winshared>: new target
3780 <wininstall>: new target
3781 <dftables.o>: use native compiler
3782 <dftables>: use native linker
3783 <install>: handle Windows platform correctly
3784 <clean>: ditto
3785 <check>: ditto
3786 copy DLL to top builddir before testing
3787
3788 As part of these changes, -no-undefined was removed again. This was reported
3789 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3790 in any case.
3791
3792 3. Some tidies to get rid of compiler warnings:
3793
3794 . In the match_data structure, match_limit was an unsigned long int, whereas
3795 match_call_count was an int. I've made them both unsigned long ints.
3796
3797 . In pcretest the fact that a const uschar * doesn't automatically cast to
3798 a void * provoked a warning.
3799
3800 . Turning on some more compiler warnings threw up some "shadow" variables
3801 and a few more missing casts.
3802
3803 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3804 option, a class that contained a single character with a value between 128
3805 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3806
3807 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3808 option, a class that contained several characters, but with at least one
3809 whose value was between 128 and 255 caused PCRE to crash.
3810
3811
3812 Version 4.1 12-Mar-03
3813 ---------------------
3814
3815 1. Compiling with gcc -pedantic found a couple of places where casts were
3816 needed, and a string in dftables.c that was longer than standard compilers are
3817 required to support.
3818
3819 2. Compiling with Sun's compiler found a few more places where the code could
3820 be tidied up in order to avoid warnings.
3821
3822 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3823 first of these names is deprecated in the latest Autoconf in favour of the name
3824 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3825 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3826 analogy I have changed it to CFLAGS_FOR_BUILD.
3827
3828 4. Added -no-undefined to the linking command in the Makefile, because this is
3829 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3830 linking step for the pcreposix library.
3831
3832 5. PCRE was failing to diagnose the case of two named groups with the same
3833 name.
3834
3835 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3836 literal character that is needed in the subject for a match, and scans along to
3837 ensure that it is present before embarking on the full matching process. This
3838 saves time in cases of nested unlimited repeats that are never going to match.
3839 Problem: the scan can take a lot of time if the subject is very long (e.g.
3840 megabytes), thus penalizing straightforward matches. It is now done only if the
3841 amount of subject to be scanned is less than 1000 bytes.
3842
3843 7. A lesser problem with the same optimization is that it was recording the
3844 first character of an anchored pattern as "needed", thus provoking a search
3845 right along the subject, even when the first match of the pattern was going to
3846 fail. The "needed" character is now not set for anchored patterns, unless it
3847 follows something in the pattern that is of non-fixed length. Thus, it still
3848 fulfils its original purpose of finding quick non-matches in cases of nested
3849 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3850
3851
3852 Version 4.0 17-Feb-03
3853 ---------------------
3854
3855 1. If a comment in an extended regex that started immediately after a meta-item
3856 extended to the end of string, PCRE compiled incorrect data. This could lead to
3857 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3858
3859 2. Moved to autoconf 2.53 and libtool 1.4.2.
3860
3861 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3862 the special perltest8 script is no longer needed - all the tests can be run
3863 from a single perltest script.
3864
3865 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3866 by \s. It has now been removed in PCRE. This means it isn't recognized as
3867 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3868 class [:space:] *does* include VT, thereby creating a mess.
3869
3870 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3871 space and tab.
3872
3873 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3874 its new features into the main test script, reducing the number of scripts.
3875
3876 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3877 were backward compatible, and made the (?i) apply to the whole pattern, as if
3878 /i were given. Now it behaves more logically, and applies the option setting
3879 only to what follows. PCRE has been changed to follow suit. However, if it
3880 finds options settings right at the start of the pattern, it extracts them into
3881 the global options, as before. Thus, they show up in the info data.
3882
3883 8. Added support for the \Q...\E escape sequence. Characters in between are
3884 treated as literals. This is slightly different from Perl in that $ and @ are
3885 also handled as literals inside the quotes. In Perl, they will cause variable
3886 interpolation. Note the following examples:
3887
3888 Pattern PCRE matches Perl matches
3889
3890 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3891 \Qabc\$xyz\E abc\$xyz abc\$xyz
3892 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3893
3894 For compatibility with Perl, \Q...\E sequences are recognized inside character
3895 classes as well as outside them.
3896
3897 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3898 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3899 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3900 signed/unsigned warnings.
3901
3902 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3903 option for pcretest, so I've replaced it by a simple function that does just
3904 that job.
3905
3906 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3907 "pcregrep -".
3908
3909 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3910 Java package. This provides some syntactic sugar for simple cases of what my
3911 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3912 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3913 item, you can use this simplified notation. Note that only makes sense with
3914 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3915 greediness, whatever the setting of the PCRE_UNGREEDY option.
3916
3917 13. A change of greediness default within a pattern was not taking effect at
3918 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3919 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3920 was abstracted outside.
3921
3922 14. PCRE now supports the \G assertion. It is true when the current matching
3923 position is at the start point of the match. This differs from \A when the
3924 starting offset is non-zero. Used with the /g option of pcretest (or similar
3925 code), it works in the same way as it does for Perl's /g option. If all
3926 alternatives of a regex begin with \G, the expression is anchored to the start
3927 match position, and the "anchored" flag is set in the compiled expression.
3928
3929 15. Some bugs concerning the handling of certain option changes within patterns
3930 have been fixed. These applied to options other than (?ims). For example,
3931 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3932 been the other way round. Some of this was related to change 7 above.
3933
3934 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3935 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3936 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3937 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3938
3939 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3940 mode. Unlike ".", it always matches newline, whatever the setting of
3941 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3942 assertions. Perl allows it, but it doesn't (in general) work because it can't
3943 calculate the length of the lookbehind. At least, that's the case for Perl
3944 5.8.0 - I've been told they are going to document that it doesn't work in
3945 future.
3946
3947 18. Added an error diagnosis for escapes that PCRE does not support: these are
3948 \L, \l, \N, \P, \p, \U, \u, and \X.
3949
3950 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3951 reading past the end of the pattern in cases such as /[abcd/.
3952
3953 20. PCRE was getting more memory than necessary for patterns with classes that
3954 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3955
3956 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3957 compiling PCRE for use with Virtual Pascal.
3958
3959 22. Small fix to the Makefile to make it work properly if the build is done
3960 outside the source tree.
3961
3962 23. Added a new extension: a condition to go with recursion. If a conditional
3963 subpattern starts with (?(R) the "true" branch is used if recursion has
3964 happened, whereas the "false" branch is used only at the top level.
3965
3966 24. When there was a very long string of literal characters (over 255 bytes
3967 without UTF support, over 250 bytes with UTF support), the computation of how
3968 much memory was required could be incorrect, leading to segfaults or other
3969 strange effects.
3970
3971 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3972 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3973 there was a subsequent back reference to those brackets. This meant that, for
3974 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3975 possible to check for precisely this case. All we can do is abandon the
3976 optimization if .* occurs inside capturing brackets when there are any back
3977 references whatsoever. (See below for a better fix that came later.)
3978
3979 26. The handling of the optimization for finding the first character of a
3980 non-anchored pattern, and for finding a character that is required later in the
3981 match were failing in some cases. This didn't break the matching; it just
3982 failed to optimize when it could. The way this is done has been re-implemented.
3983
3984 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3985
3986 28. Added a new feature that provides some of the functionality that Perl
3987 provides with (?{...}). The facility is termed a "callout". The way it is done
3988 in PCRE is for the caller to provide an optional function, by setting
3989 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3990 global variable. By default it is unset, which disables all calling out. To get
3991 the function called, the regex must include (?C) at appropriate points. This
3992 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3993 This provides a means of identifying different callout points. When PCRE
3994 reaches such a point in the regex, if pcre_callout has been set, the external
3995 function is called. It is provided with data in a structure called
3996 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3997 matching continues; if it returns a non-zero value, the match at the current
3998 point fails. However, backtracking will occur if possible. [This was changed
3999 later and other features added - see item 49 below.]
4000
4001 29. pcretest is upgraded to test the callout functionality. It provides a
4002 callout function that displays information. By default, it shows the start of
4003 the match and the current position in the text. There are some new data escapes
4004 to vary what happens:
4005
4006 \C+ in addition, show current contents of captured substrings
4007 \C- do not supply a callout function
4008 \C!n return 1 when callout number n is reached
4009 \C!n!m return 1 when callout number n is reached for the mth time
4010
4011 30. If pcregrep was called with the -l option and just a single file name, it
4012 output "<stdin>" if a match was found, instead of the file name.
4013
4014 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4015 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4016 pcre_exec(). This saves a malloc/free per call. The default value of
4017 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4018 when configuring.
4019
4020 32. The default maximum size of a compiled pattern is 64K. There have been a
4021 few cases of people hitting this limit. The code now uses macros to handle the
4022 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4023 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4024 configuring. Tests 2 and 5 work only with 2-byte links because they output
4025 debugging information about compiled patterns.
4026
4027 33. Internal code re-arrangements:
4028
4029 (a) Moved the debugging function for printing out a compiled regex into
4030 its own source file (printint.c) and used #include to pull it into
4031 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4032 separate copies.
4033
4034 (b) Defined the list of op-code names for debugging as a macro in
4035 internal.h so that it is next to the definition of the opcodes.
4036
4037 (c) Defined a table of op-code lengths for simpler skipping along compiled
4038 code. This is again a macro in internal.h so that it is next to the
4039 definition of the opcodes.
4040
4041 34. Added support for recursive calls to individual subpatterns, along the
4042 lines of Robin Houston's patch (but implemented somewhat differently).
4043
4044 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4045 allow it to read and process whole directories in Win32. This code was
4046 contributed by Lionel Fourquaux; it has not been tested by me.
4047
4048 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4049 used to name a group. Names consist of alphanumerics and underscores, and must
4050 be unique. Back references use the syntax (?P=name) and recursive calls use
4051 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4052 numbers. The function pcre_fullinfo() can be used after compilation to extract
4053 a name/number map. There are three relevant calls:
4054
4055 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4056 PCRE_INFO_NAMECOUNT yields the number of entries
4057 PCRE_INFO_NAMETABLE yields a pointer to the map.
4058
4059 The map is a vector of fixed-size entries. The size of each entry depends on
4060 the length of the longest name used. The first two bytes of each entry are the
4061 group number, most significant byte first. There follows the corresponding
4062 name, zero terminated. The names are in alphabetical order.
4063
4064 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4065 case instead of 255. Making it the same both with and without UTF-8 support
4066 means that the same test output works with both.
4067
4068 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4069 calling malloc() with a zero argument.
4070
4071 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4072 optimization. I've improved things by keeping a bitmap of backreferences with
4073 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4074 fact referenced, the optimization can be applied. It is unlikely that a
4075 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4076 the match to follow \n) will appear inside brackets with a number greater than
4077 31, but if it does, any back reference > 31 suppresses the optimization.
4078
4079 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4080 of disabling numbered capturing parentheses. Any opening parenthesis that is
4081 not followed by ? behaves as if it were followed by ?: but named parentheses
4082 can still be used for capturing (and they will acquire numbers in the usual
4083 way).
4084
4085 41. Redesigned the return codes from the match() function into yes/no/error so
4086 that errors can be passed back from deep inside the nested calls. A malloc
4087 failure while inside a recursive subpattern call now causes the
4088 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4089
4090 42. It is now possible to set a limit on the number of times the match()
4091 function is called in a call to pcre_exec(). This facility makes it possible to
4092 limit the amount of recursion and backtracking, though not in a directly
4093 obvious way, because the match() function is used in a number of different
4094 circumstances. The count starts from zero for each position in the subject
4095 string (for non-anchored patterns). The default limit is, for compatibility, a
4096 large number, namely 10 000 000. You can change this in two ways:
4097
4098 (a) When configuring PCRE before making, you can use --with-match-limit=n
4099 to set a default value for the compiled library.
4100
4101 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4102 a different value is set. See 45 below.
4103
4104 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4105
4106 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4107 of things that can be changed at compile time. The first argument specifies
4108 what is wanted and the second points to where the information is to be placed.
4109 The current list of available information is:
4110
4111 PCRE_CONFIG_UTF8
4112
4113 The output is an integer that is set to one if UTF-8 support is available;
4114 otherwise it is set to zero.
4115
4116 PCRE_CONFIG_NEWLINE
4117
4118 The output is an integer that it set to the value of the code that is used for
4119 newline. It is either LF (10) or CR (13).
4120
4121 PCRE_CONFIG_LINK_SIZE
4122
4123 The output is an integer that contains the number of bytes used for internal
4124 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4125
4126 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4127
4128 The output is an integer that contains the threshold above which the POSIX
4129 interface uses malloc() for output vectors. See item 31 above.
4130
4131 PCRE_CONFIG_MATCH_LIMIT
4132
4133 The output is an unsigned integer that contains the default limit of the number
4134 of match() calls in a pcre_exec() execution. See 42 above.
4135
4136 44. pcretest has been upgraded by the addition of the -C option. This causes it
4137 to extract all the available output from the new pcre_config() function, and to
4138 output it. The program then exits immediately.
4139
4140 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4141 order to support additional features. One way would have been to define
4142 pcre_exec2() (for example) with extra arguments, but this would not have been
4143 extensible, and would also have required all calls to the original function to
4144 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4145 is used for passing in "extra" data from pcre_study().
4146
4147 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4148 contains the following fields:
4149
4150 flags a bitmap indicating which of the following fields are set
4151 study_data opaque data from pcre_study()
4152 match_limit a way of specifying a limit on match() calls for a specific
4153 call to pcre_exec()
4154 callout_data data for callouts (see 49 below)
4155
4156 The flag bits are also defined in pcre.h, and are
4157
4158 PCRE_EXTRA_STUDY_DATA
4159 PCRE_EXTRA_MATCH_LIMIT
4160 PCRE_EXTRA_CALLOUT_DATA
4161
4162 The pcre_study() function now returns one of these new pcre_extra blocks, with
4163 the actual study data pointed to by the study_data field, and the
4164 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4165 before. That is, this change is entirely upwards-compatible and requires no
4166 change to existing code.
4167
4168 If you want to pass in additional data to pcre_exec(), you can either place it
4169 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4170 block.
4171
4172 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4173 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4174 times with different match limits, until it finds the minimum value needed for
4175 pcre_exec() to complete. The value is then output. This can be instructive; for
4176 most simple matches the number is quite small, but for pathological cases it
4177 gets very large very quickly.
4178
4179 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4180 returns the size of the data block pointed to by the study_data field in a
4181 pcre_extra block, that is, the value that was passed as the argument to
4182 pcre_malloc() when PCRE was getting memory in which to place the information
4183 created by pcre_study(). The fourth argument should point to a size_t variable.
4184 pcretest has been extended so that this information is shown after a successful
4185 pcre_study() call when information about the compiled regex is being displayed.
4186
4187 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4188 because what follows is always an absolute path. (Later: it turns out that this
4189 is more than cosmetic for MinGW, because it doesn't like empty path
4190 components.)
4191
4192 49. Some changes have been made to the callout feature (see 28 above):
4193
4194 (i) A callout function now has three choices for what it returns:
4195
4196 0 => success, carry on matching
4197 > 0 => failure at this point, but backtrack if possible
4198 < 0 => serious error, return this value from pcre_exec()
4199
4200 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4201 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4202 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4203 use by callout functions. It will never be used by PCRE itself.
4204
4205 (ii) The pcre_extra structure (see 45 above) has a void * field called
4206 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4207 pcre_callout_block structure has a field of the same name. The contents of
4208 the field passed in the pcre_extra structure are passed to the callout
4209 function in the corresponding field in the callout block. This makes it
4210 easier to use the same callout-containing regex from multiple threads. For
4211 testing, the pcretest program has a new data escape
4212
4213 \C*n pass the number n (may be negative) as callout_data
4214
4215 If the callout function in pcretest receives a non-zero value as
4216 callout_data, it returns that value.
4217
4218 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4219 there were some redundant $(CFLAGS) in commands that are now specified as
4220 $(LINK), which already includes $(CFLAGS).
4221
4222 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4223 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4224 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4225 one-byte characters throughout. Note that case-insensitive matching applies
4226 only to characters whose values are less than 256. PCRE doesn't support the
4227 notion of cases for higher-valued characters.
4228
4229 (i) A character class whose characters are all within 0-255 is handled as
4230 a bit map, and the map is inverted for negative classes. Previously, a
4231 character > 255 always failed to match such a class; however it should
4232 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4233
4234 (ii) A negated character class with a single character < 255 is coded as
4235 "not this character" (OP_NOT). This wasn't working properly when the test
4236 character was multibyte, either singly or repeated.
4237
4238 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4239 mode, for example: \x{100}{2,3}.
4240
4241 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4242 singly or repeated) now correctly test multibyte characters. However,
4243 PCRE doesn't recognize any characters with values greater than 255 as
4244 digits, spaces, or word characters. Such characters always match \D, \S,
4245 and \W, and never match \d, \s, or \w.
4246
4247 (v) Classes may now contain characters and character ranges with values
4248 greater than 255. For example: [ab\x{100}-\x{400}].
4249
4250 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4251 PCRE in UTF-8 mode.
4252
4253 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4254 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4255 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4256 value.)
4257
4258 53. The single man page has become too large. I have therefore split it up into
4259 a number of separate man pages. These also give rise to individual HTML pages;
4260 these are now put in a separate directory, and there is an index.html page that
4261 lists them all. Some hyperlinking between the pages has been installed.
4262
4263 54. Added convenience functions for handling named capturing parentheses.
4264
4265 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4266 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4267 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4268 are faulted.
4269
4270 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4271 calling configure. These values are used when compiling the dftables.c program
4272 which is run to generate the source of the default character tables. They
4273 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4274 you will need to set these values.
4275
4276 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4277
4278
4279 Version 3.9 02-Jan-02
4280 ---------------------
4281
4282 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4283
4284 2. If --disable-static was given, the building process failed when trying to
4285 build pcretest and pcregrep. (For some reason it was using libtool to compile
4286 them, which is not right, as they aren't part of the library.)
4287
4288
4289 Version 3.8 18-Dec-01
4290 ---------------------
4291
4292 1. The experimental UTF-8 code was completely screwed up. It was packing the
4293 bytes in the wrong order. How dumb can you get?
4294
4295
4296 Version 3.7 29-Oct-01
4297 ---------------------
4298
4299 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4300 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4301 this didn't happen under Solaris 8, where I normally test things.
4302
4303 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4304 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4305 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4306 this fix an hour or so after the initial 3.7 release.)
4307
4308
4309 Version 3.6 23-Oct-01
4310 ---------------------
4311
4312 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4313 offsets passed as NULL with zero offset count.
4314
4315 2. The config.guess and config.sub files had not been updated when I moved to
4316 the latest autoconf.
4317
4318
4319 Version 3.5 15-Aug-01
4320 ---------------------
4321
4322 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4323 had been forgotten.
4324
4325 2. By using declared but undefined structures, we can avoid using "void"
4326 definitions in pcre.h while keeping the internal definitions of the structures
4327 private.
4328
4329 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4330 user point of view, this means that both static and shared libraries are built
4331 by default, but this can be individually controlled. More of the work of
4332 handling this static/shared cases is now inside libtool instead of PCRE's make
4333 file.
4334
4335 4. The pcretest utility is now installed along with pcregrep because it is
4336 useful for users (to test regexs) and by doing this, it automatically gets
4337 relinked by libtool. The documentation has been turned into a man page, so
4338 there are now .1, .txt, and .html versions in /doc.
4339
4340 5. Upgrades to pcregrep:
4341 (i) Added long-form option names like gnu grep.
4342 (ii) Added --help to list all options with an explanatory phrase.
4343 (iii) Added -r, --recursive to recurse into sub-directories.
4344 (iv) Added -f, --file to read patterns from a file.
4345
4346 6. pcre_exec() was referring to its "code" argument before testing that
4347 argument for NULL (and giving an error if it was NULL).
4348
4349 7. Upgraded Makefile.in to allow for compiling in a different directory from
4350 the source directory.
4351
4352 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4353 options bits, the pointer it was passed was to an int instead of to an unsigned
4354 long int. This mattered only on 64-bit systems.
4355
4356 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4357 generated) instead of pcre.in, which it its source. Also made the same change
4358 in several of the .c files.
4359
4360 10. A new release of gcc defines printf() as a macro, which broke pcretest
4361 because it had an ifdef in the middle of a string argument for printf(). Fixed
4362 by using separate calls to printf().
4363
4364 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4365 script, to force use of CR or LF instead of \n in the source. On non-Unix
4366 systems, the value can be set in config.h.
4367
4368 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4369 absolute limit. Changed the text of the error message to make this clear, and
4370 likewise updated the man page.
4371
4372 13. The limit of 99 on the number of capturing subpatterns has been removed.
4373 The new limit is 65535, which I hope will not be a "real" limit.
4374
4375
4376 Version 3.4 22-Aug-00
4377 ---------------------
4378
4379 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4380
4381 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4382
4383
4384 Version 3.3 01-Aug-00
4385 ---------------------
4386
4387 1. If an octal character was given, but the value was greater than \377, it
4388 was not getting masked to the least significant bits, as documented. This could
4389 lead to crashes in some systems.
4390
4391 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4392 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4393
4394 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4395 These just pass their arguments on to (pcre_free)(), but they are provided
4396 because some uses of PCRE bind it to non-C systems that can call its functions,
4397 but cannot call free() or pcre_free() directly.
4398
4399 4. Add "make test" as a synonym for "make check". Corrected some comments in
4400 the Makefile.
4401
4402 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4403 Makefile.
4404
4405 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4406 command called pgrep for grepping around the active processes.
4407
4408 7. Added the beginnings of support for UTF-8 character strings.
4409
4410 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4411 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4412 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4413 out for the ar command.)
4414
4415
4416 Version 3.2 12-May-00
4417 ---------------------
4418
4419 This is purely a bug fixing release.
4420
4421 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4422 of ZA. This was just one example of several cases that could provoke this bug,
4423 which was introduced by change 9 of version 2.00. The code for breaking
4424 infinite loops after an iteration that matches an empty string was't working
4425 correctly.
4426
4427 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4428 when matched against abbab (for example). After matching an empty string, it
4429 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4430 caused it to match further down the string than it should.
4431
4432 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4433 was there because it doesn't seem to be needed, and it causes trouble on some
4434 systems, as it is not a Standard C header. It has been removed.
4435
4436 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4437 were reported on the Macintosh. The changes were from
4438
4439 while ((c = *(++ptr)) != 0 && c != '\n');
4440 to
4441 while ((c = *(++ptr)) != 0 && c != '\n') ;
4442
4443 Totally extraordinary, but if that's what it takes...
4444
4445 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4446 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4447 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4448 assumes the way PCRE uses memmove() (always moving upwards).
4449
4450 6. PCRE is being used in one environment where strchr() is not available. There
4451 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4452 faster code anyway.
4453
4454
4455 Version 3.1 09-Feb-00
4456 ---------------------
4457
4458 The only change in this release is the fixing of some bugs in Makefile.in for
4459 the "install" target:
4460
4461 (1) It was failing to install pcreposix.h.
4462
4463 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4464
4465
4466 Version 3.0 01-Feb-00
4467 ---------------------
4468
4469 1. Add support for the /+ modifier to perltest (to output $` like it does in
4470 pcretest).
4471
4472 2. Add support for the /g modifier to perltest.
4473
4474 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4475 matches null strings.
4476
4477 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4478 pattern. Perl treats empty patterns specially - it reuses the most recent
4479 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4480 effect.
4481
4482 5. The POSIX interface was broken in that it was just handing over the POSIX
4483 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4484 required a bigger vector, with some working space on the end. This means that
4485 the POSIX wrapper now has to get and free some memory, and copy the results.
4486
4487 6. Added some simple autoconf support, placing the test data and the
4488 documentation in separate directories, re-organizing some of the
4489 information files, and making it build pcre-config (a GNU standard). Also added
4490 libtool support for building PCRE as a shared library, which is now the
4491 default.
4492
4493 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4494 09 are not valid octal constants. Single digits will be used for minor values
4495 less than 10.
4496
4497 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4498 existing programs that set these in the POSIX interface can use PCRE without
4499 modification.
4500
4501 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4502 return all that pcre_info() returns, plus additional data. The pcre_info()
4503 function is retained for compatibility, but is considered to be obsolete.
4504
4505 10. Added experimental recursion feature (?R) to handle one common case that
4506 Perl 5.6 will be able to do with (?p{...}).
4507
4508 11. Added support for POSIX character classes like [:alpha:], which Perl is
4509 adopting.
4510
4511
4512 Version 2.08 31-Aug-99
4513 ----------------------
4514
4515 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4516 trying to match at the startoffset position, but instead was moving forward to
4517 the next newline as if a previous match had failed.
4518
4519 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4520 and could get into a loop if a null string was matched other than at the start
4521 of the subject.
4522
4523 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4524 be distinguished at compile time, and for completeness also added PCRE_DATE.
4525
4526 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4527 in GnuWin32 environments.
4528
4529
4530 Version 2.07 29-Jul-99
4531 ----------------------
4532
4533 1. The documentation is now supplied in plain text form and HTML as well as in
4534 the form of man page sources.
4535
4536 2. C++ compilers don't like assigning (void *) values to other pointer types.
4537 In particular this affects malloc(). Although there is no problem in Standard
4538 C, I've put in casts to keep C++ compilers happy.
4539
4540 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4541 should be (const char *).
4542
4543 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4544 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4545 However, I haven't made this a standard facility. The documentation doesn't
4546 mention it, and the Makefile doesn't support it.
4547
4548 5. The Makefile now contains an "install" target, with editable destinations at
4549 the top of the file. The pcretest program is not installed.
4550
4551 6. pgrep -V now gives the PCRE version number and date.
4552
4553 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4554 causing the entire string to be ignored, instead of just the last character.
4555
4556 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4557 non-matching string, it can take a very, very long time, even for strings of
4558 quite modest length, because of the nested recursion. PCRE now does better in
4559 some of these cases. It does this by remembering the last required literal
4560 character in the pattern, and pre-searching the subject to ensure it is present
4561 before running the real match. In other words, it applies a heuristic to detect
4562 some types of certain failure quickly, and in the above example, if presented
4563 with a string that has no trailing " it gives "no match" very quickly.
4564
4565 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4566 other alternatives are tried instead.
4567
4568
4569 Version 2.06 09-Jun-99
4570 ----------------------
4571
4572 1. Change pcretest's output for amount of store used to show just the code
4573 space, because the remainder (the data block) varies in size between 32-bit and
4574 64-bit systems.
4575
4576 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4577 start matching at. This allows lookbehinds to work when searching for multiple
4578 occurrences in a string.
4579
4580 3. Added additional options to pcretest for testing multiple occurrences:
4581
4582 /+ outputs the rest of the string that follows a match
4583 /g loops for multiple occurrences, using the new startoffset argument
4584 /G loops for multiple occurrences by passing an incremented pointer
4585
4586 4. PCRE wasn't doing the "first character" optimization for patterns starting
4587 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4588 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4589 the letter 'x'. On long subject strings, this gives a significant speed-up.
4590
4591
4592 Version 2.05 21-Apr-99
4593 ----------------------
4594
4595 1. Changed the type of magic_number from int to long int so that it works
4596 properly on 16-bit systems.
4597
4598 2. Fixed a bug which caused patterns starting with .* not to work correctly
4599 when the subject string contained newline characters. PCRE was assuming
4600 anchoring for such patterns in all cases, which is not correct because .* will
4601 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4602 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4603 must be retried after every newline in the subject.
4604
4605
4606 Version 2.04 18-Feb-99
4607 ----------------------
4608
4609 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4610 computation of the store needed to hold the pattern was incorrect (too large).
4611 If such patterns were nested a few deep, this could multiply and become a real
4612 problem.
4613
4614 2. Added /M option to pcretest to show the memory requirement of a specific
4615 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4616
4617 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4618 compiled in such a way that the backtracking after subsequent failure was
4619 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4620 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4621
4622
4623 Version 2.03 02-Feb-99
4624 ----------------------
4625
4626 1. Fixed typo and small mistake in man page.
4627
4628 2. Added 4th condition (GPL supersedes if conflict) and created separate
4629 LICENCE file containing the conditions.
4630
4631 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4632 Perl, that is the internal \ allows the delimiter to be included in the
4633 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4634 the final delimiter, add \ to the end of the pattern (to test the error).
4635
4636 4. Added the convenience functions for extracting substrings after a successful
4637 match. Updated pcretest to make it able to test these functions.
4638
4639
4640 Version 2.02 14-Jan-99
4641 ----------------------
4642
4643 1. Initialized the working variables associated with each extraction so that
4644 their saving and restoring doesn't refer to uninitialized store.
4645
4646 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4647 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4648 fix the problem.
4649
4650 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4651 calls, and wasn't printing the correct value for compiling calls. Increased the
4652 default value of LOOPREPEAT, and the number of significant figures in the
4653 times.
4654
4655 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4656
4657 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4658 a building problem on Windows NT with a FAT file system.
4659
4660
4661 Version 2.01 21-Oct-98
4662 ----------------------
4663
4664 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4665 to character tables built by pcre_maketables() in the current locale. If NULL
4666 is passed, the default tables are used.
4667
4668
4669 Version 2.00 24-Sep-98
4670 ----------------------
4671
4672 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4673 it any more.
4674
4675 2. Allow quantification of (?>) groups, and make it work correctly.
4676
4677 3. The first character computation wasn't working for (?>) groups.
4678
4679 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4680 end of the subject) and add 5.005's \z, which really does match only at the
4681 very end of the subject.
4682
4683 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4684
4685 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4686 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4687 localized options. All options to pcre_study() were also removed.
4688
4689 7. Add other new features from 5.005:
4690
4691 $(?<= positive lookbehind
4692 $(?<! negative lookbehind
4693 (?imsx-imsx) added the unsetting capability
4694 such a setting is global if at outer level; local otherwise
4695 (?imsx-imsx:) non-capturing groups with option setting
4696 (?(cond)re|re) conditional pattern matching
4697
4698 A backreference to itself in a repeated group matches the previous
4699 captured string.
4700
4701 8. General tidying up of studying (both automatic and via "study")
4702 consequential on the addition of new assertions.
4703
4704 9. As in 5.005, unlimited repeated groups that could match an empty substring
4705 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4706 runtime if any iteration does actually match an empty substring.
4707
4708 10. Include the RunTest script in the distribution.
4709
4710 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4711 discrepancies, some of which were old and were also with respect to 5.004. They
4712 have now been fixed.
4713
4714
4715 Version 1.09 28-Apr-98
4716 ----------------------
4717
4718 1. A negated single character class followed by a quantifier with a minimum
4719 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4720 program crashes, or just wrong answers. This did not apply to negated classes
4721 containing more than one character, or to minima other than one.
4722
4723
4724 Version 1.08 27-Mar-98
4725 ----------------------
4726
4727 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4728
4729 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4730 latter must appear before anything that relies on it in the pattern.
4731
4732
4733 Version 1.07 16-Feb-98
4734 ----------------------
4735
4736 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4737 repeat of a potentially empty string).
4738
4739
4740 Version 1.06 23-Jan-98
4741 ----------------------
4742
4743 1. Added Markus Oberhumer's little patches for C++.
4744
4745 2. Literal strings longer than 255 characters were broken.
4746
4747
4748 Version 1.05 23-Dec-97
4749 ----------------------
4750
4751 1. Negated character classes containing more than one character were failing if
4752 PCRE_CASELESS was set at run time.
4753
4754
4755 Version 1.04 19-Dec-97
4756 ----------------------
4757
4758 1. Corrected the man page, where some "const" qualifiers had been omitted.
4759
4760 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4761 input syntax.
4762
4763 3. Fixed memory leak which occurred when a regex with back references was
4764 matched with an offsets vector that wasn't big enough. The temporary memory
4765 that is used in this case wasn't being freed if the match failed.
4766
4767 4. Tidied pcretest to ensure it frees memory that it gets.
4768
4769 5. Temporary memory was being obtained in the case where the passed offsets
4770 vector was exactly big enough.
4771
4772 6. Corrected definition of offsetof() from change 5 below.
4773
4774 7. I had screwed up change 6 below and broken the rules for the use of
4775 setjmp(). Now fixed.
4776
4777
4778 Version 1.03 18-Dec-97
4779 ----------------------
4780
4781 1. A erroneous regex with a missing opening parenthesis was correctly
4782 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4783 on some systems.
4784
4785 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4786 it was reported that one broken compiler failed on the former because "code" is
4787 also an independent variable.
4788
4789 3. The erroneous regex a[]b caused an array overrun reference.
4790
4791 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4792 fail on data ending with that character. (It was going on too far, and checking
4793 the next character, typically a binary zero.) This was specific to the
4794 optimized code for single-character negative classes.
4795
4796 5. Added a contributed patch from the TIN world which does the following:
4797
4798 + Add an undef for memmove, in case the the system defines a macro for it.
4799
4800 + Add a definition of offsetof(), in case there isn't one. (I don't know
4801 the reason behind this - offsetof() is part of the ANSI standard - but
4802 it does no harm).
4803
4804 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4805 most of the places where whitespace preceded '#'. I have given up and
4806 allowed the remaining 2 cases to be at the margin.
4807
4808 + Rename some variables in pcre to eliminate shadowing. This seems very
4809 pedantic, but does no harm, of course.
4810
4811 6. Moved the call to setjmp() into its own function, to get rid of warnings
4812 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4813
4814 7. Constructs such as \d{8,} were compiling into the equivalent of
4815 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4816 outcome, but in this particular case used more store than had been allocated,
4817 which caused the bug to be discovered because it threw up an internal error.
4818
4819 8. The debugging code in both pcre and pcretest for outputting the compiled
4820 form of a regex was going wrong in the case of back references followed by
4821 curly-bracketed repeats.
4822
4823
4824 Version 1.02 12-Dec-97
4825 ----------------------
4826
4827 1. Typos in pcre.3 and comments in the source fixed.
4828
4829 2. Applied a contributed patch to get rid of places where it used to remove
4830 'const' from variables, and fixed some signed/unsigned and uninitialized
4831 variable warnings.
4832
4833 3. Added the "runtest" target to Makefile.
4834
4835 4. Set default compiler flag to -O2 rather than just -O.
4836
4837
4838 Version 1.01 19-Nov-97
4839 ----------------------
4840
4841 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4842 like /([ab]*)*/, that is, for classes with more than one character in them.
4843
4844 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4845 as /((?>a*))*/ (a PCRE_EXTRA facility).
4846
4847
4848 Version 1.00 18-Nov-97
4849 ----------------------
4850
4851 1. Added compile-time macros to support systems such as SunOS4 which don't have
4852 memmove() or strerror() but have other things that can be used instead.
4853
4854 2. Arranged that "make clean" removes the executables.
4855
4856
4857 Version 0.99 27-Oct-97
4858 ----------------------
4859
4860 1. Fixed bug in code for optimizing classes with only one character. It was
4861 initializing a 32-byte map regardless, which could cause it to run off the end
4862 of the memory it had got.
4863
4864 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4865
4866
4867 Version 0.98 22-Oct-97
4868 ----------------------
4869
4870 1. Fixed bug in code for handling temporary memory usage when there are more
4871 back references than supplied space in the ovector. This could cause segfaults.
4872
4873
4874 Version 0.97 21-Oct-97
4875 ----------------------
4876
4877 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4878
4879 2. Optimized negated single characters not to use a bit map.
4880
4881 3. Brought error texts together as macro definitions; clarified some of them;
4882 fixed one that was wrong - it said "range out of order" when it meant "invalid
4883 escape sequence".
4884
4885 4. Changed some char * arguments to const char *.
4886
4887 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4888
4889 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4890 pcretest.
4891
4892
4893 Version 0.96 16-Oct-97
4894 ----------------------
4895
4896 1. Added a simple "pgrep" utility to the distribution.
4897
4898 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4899 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4900 where "ddd" means "one or more decimal digits".
4901
4902 3. Fixed serious bug. If a pattern had a back reference, but the call to
4903 pcre_exec() didn't supply a large enough ovector to record the related
4904 identifying subpattern, the match always failed. PCRE now remembers the number
4905 of the largest back reference, and gets some temporary memory in which to save
4906 the offsets during matching if necessary, in order to ensure that
4907 backreferences always work.
4908
4909 4. Increased the compatibility with Perl in a number of ways:
4910
4911 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4912 to request this handling. The option can be set at compile or exec time.
4913
4914 (b) $ matches before a terminating newline by default; an option
4915 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4916 mode). The option can be set at compile or exec time.
4917
4918 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4919 the same as Perl's. If the decimal number it represents is less than 10
4920 or there aren't that many previous left capturing parentheses, an octal
4921 escape is read. Inside a character class, it's always an octal escape,
4922 even if it is a single digit.
4923
4924 (d) An escaped but undefined alphabetic character is taken as a literal,
4925 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4926 escapes.
4927
4928 (e) {0} is now permitted. (The previous item is removed from the compiled
4929 pattern).
4930
4931 5. Changed all the names of code files so that the basic parts are no longer
4932 than 10 characters, and abolished the teeny "globals.c" file.
4933
4934 6. Changed the handling of character classes; they are now done with a 32-byte
4935 bit map always.
4936
4937 7. Added the -d and /D options to pcretest to make it possible to look at the
4938 internals of compilation without having to recompile pcre.
4939
4940
4941 Version 0.95 23-Sep-97
4942 ----------------------
4943
4944 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4945 \x20 at the start of a run of normal characters. These were being treated as
4946 real characters, instead of the source characters being re-checked.
4947
4948
4949 Version 0.94 18-Sep-97
4950 ----------------------
4951
4952 1. The functions are now thread-safe, with the caveat that the global variables
4953 containing pointers to malloc() and free() or alternative functions are the
4954 same for all threads.
4955
4956 2. Get pcre_study() to generate a bitmap of initial characters for non-
4957 anchored patterns when this is possible, and use it if passed to pcre_exec().
4958
4959
4960 Version 0.93 15-Sep-97
4961 ----------------------
4962
4963 1. /(b)|(:+)/ was computing an incorrect first character.
4964
4965 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4966 but not actually doing anything yet.
4967
4968 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4969 as Perl does (e.g. [-az] or [az-]).
4970
4971 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4972 all possible positions.
4973
4974 5. Split up into different modules to avoid including unneeded functions in a
4975 compiled binary. However, compile and exec are still in one module. The "study"
4976 function is split off.
4977
4978 6. The character tables are now in a separate module whose source is generated
4979 by an auxiliary program - but can then be edited by hand if required. There are
4980 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4981 toupper() in the code.
4982
4983 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4984 make them global. Abolish the function for setting them, as the caller can now
4985 set them directly.
4986
4987
4988 Version 0.92 11-Sep-97
4989 ----------------------
4990
4991 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4992 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4993
4994 2. Caseless matching was not working in character classes if the characters in
4995 the pattern were in upper case.
4996
4997 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4998
4999 4. Make PCRE_ANCHORED public and accept as a compile option.
5000
5001 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5002 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5003 pass them.
5004
5005 6. Give an error if bad option bits passed at compile or run time.
5006
5007 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5008 pcretest to cause it to pass that flag.
5009
5010 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5011 options, and the first character, if set.
5012
5013 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5014
5015
5016 Version 0.91 10-Sep-97
5017 ----------------------
5018
5019 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5020 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5021
5022 2. PCRE was looping on encountering an indefinitely repeated back reference to
5023 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5024 Perl does - treats the match as successful.
5025
5026 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5