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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1248 - (show annotations)
Wed Feb 13 17:36:38 2013 UTC (6 years, 8 months ago) by ph10
File size: 242290 byte(s)
Fix various save/revert cases for capture_last in recursion and also don't 
diagnose overflow when it has reverted.
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 11. The value of capture_last that is passed to callouts was incorrect in some
49 cases when there was a capture on one path that was subsequently abandoned
50 after a backtrack. Also, the capture_last value is now reset after a
51 recursion, since all captures are also reset in this case.
52
53 12. The interpreter no longer returns the "too many substrings" error in the
54 case when an overflowing capture is in a branch that is subsequently
55 abandoned after a backtrack.
56
57 13. In the pathological case when an offset vector of size 2 is used, pcretest
58 now prints out the matched string after a yield of 0 or 1.
59
60
61 Version 8.32 30-November-2012
62 -----------------------------
63
64 1. Improved JIT compiler optimizations for first character search and single
65 character iterators.
66
67 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
68 Patch by Daniel Richard G.
69
70 3. Single character iterator optimizations in the JIT compiler.
71
72 4. Improved JIT compiler optimizations for character ranges.
73
74 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
75 Reported by Giuseppe D'Angelo.
76
77 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
78 of a line, was being set incorrectly in cases where .* appeared inside
79 atomic brackets at the start of a pattern, or where there was a subsequent
80 *PRUNE or *SKIP.
81
82 7. Improved instruction cache flush for POWER/PowerPC.
83 Patch by Daniel Richard G.
84
85 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
86 grep:
87
88 (a) There is now no limit to the number of patterns to be matched.
89
90 (b) An error is given if a pattern is too long.
91
92 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
93 are now supported.
94
95 (d) --exclude-from and --include-from (multiple use) have been added.
96
97 (e) Exclusions and inclusions now apply to all files and directories, not
98 just to those obtained from scanning a directory recursively.
99
100 (f) Multiple uses of -f and --file-list are now supported.
101
102 (g) In a Windows environment, the default for -d has been changed from
103 "read" (the GNU grep default) to "skip", because otherwise the presence
104 of a directory in the file list provokes an error.
105
106 (h) The documentation has been revised and clarified in places.
107
108 9. Improve the matching speed of capturing brackets.
109
110 10. Changed the meaning of \X so that it now matches a Unicode extended
111 grapheme cluster.
112
113 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
114 out POSIX threads when JIT support is configured.
115
116 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
117
118 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
119 structure could go wrong in environments where size_t is not the same size
120 as int.
121
122 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
123 set.
124
125 15. The EBCDIC support had decayed; later updates to the code had included
126 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
127 general tidy up of EBCDIC-related issues, and the documentation was also
128 not quite right. There is now a test that can be run on ASCII systems to
129 check some of the EBCDIC-related things (but is it not a full test).
130
131 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
132 in a small tidy to the code.
133
134 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
135
136 18. If the --only-matching (-o) option in pcregrep is specified multiple
137 times, each one causes appropriate output. For example, -o1 -o2 outputs the
138 substrings matched by the 1st and 2nd capturing parentheses. A separating
139 string can be specified by --om-separator (default empty).
140
141 19. Improving the first n character searches.
142
143 20. Turn case lists for horizontal and vertical white space into macros so that
144 they are defined only once.
145
146 21. This set of changes together give more compatible Unicode case-folding
147 behaviour for characters that have more than one other case when UCP
148 support is available.
149
150 (a) The Unicode property table now has offsets into a new table of sets of
151 three or more characters that are case-equivalent. The MultiStage2.py
152 script that generates these tables (the pcre_ucd.c file) now scans
153 CaseFolding.txt instead of UnicodeData.txt for character case
154 information.
155
156 (b) The code for adding characters or ranges of characters to a character
157 class has been abstracted into a generalized function that also handles
158 case-independence. In UTF-mode with UCP support, this uses the new data
159 to handle characters with more than one other case.
160
161 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
162 whose other case is greater than 256 are now correctly matched
163 caselessly. Previously, the high codepoint matched the low one, but not
164 vice versa.
165
166 (d) The processing of \h, \H, \v, and \ in character classes now makes use
167 of the new class addition function, using character lists defined as
168 macros alongside the case definitions of 20 above.
169
170 (e) Caseless back references now work with characters that have more than
171 one other case.
172
173 (f) General caseless matching of characters with more than one other case
174 is supported.
175
176 22. Unicode character properties were updated from Unicode 6.2.0
177
178 23. Improved CMake support under Windows. Patch by Daniel Richard G.
179
180 24. Add support for 32-bit character strings, and UTF-32
181
182 25. Major JIT compiler update (code refactoring and bugfixing).
183 Experimental Sparc 32 support is added.
184
185 26. Applied a modified version of Daniel Richard G's patch to create
186 pcre.h.generic and config.h.generic by "make" instead of in the
187 PrepareRelease script.
188
189 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
190 pcre_compile.c when checking for a zero character.
191
192 28. Introducing a native interface for JIT. Through this interface, the compiled
193 machine code can be directly executed. The purpose of this interface is to
194 provide fast pattern matching, so several sanity checks are not performed.
195 However, feature tests are still performed. The new interface provides
196 1.4x speedup compared to the old one.
197
198 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
199 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
200 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
201
202 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
203 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
204 the "old" RFC 2279). Instead, it ended up passing a negative length to
205 pcre_exec().
206
207 31. Add support for GCC's visibility feature to hide internal functions.
208
209 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
210 "unknown -C option" after outputting 0 or 1.
211
212 33. There is now support for generating a code coverage report for the test
213 suite in environments where gcc is the compiler and lcov is installed. This
214 is mainly for the benefit of the developers.
215
216 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
217 unaddressable using valgrind annotations, allowing valgrind to detect
218 invalid memory accesses. This is mainly for the benefit of the developers.
219
220 25. (*UTF) can now be used to start a pattern in any of the three libraries.
221
222 26. Give configure error if --enable-cpp but no C++ compiler found.
223
224
225 Version 8.31 06-July-2012
226 -------------------------
227
228 1. Fixing a wrong JIT test case and some compiler warnings.
229
230 2. Removed a bashism from the RunTest script.
231
232 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
233 to unsigned type, result still unsigned" that was given by an MS compiler
234 on encountering the code "-sizeof(xxx)".
235
236 4. Partial matching support is added to the JIT compiler.
237
238 5. Fixed several bugs concerned with partial matching of items that consist
239 of more than one character:
240
241 (a) /^(..)\1/ did not partially match "aba" because checking references was
242 done on an "all or nothing" basis. This also applied to repeated
243 references.
244
245 (b) \R did not give a hard partial match if \r was found at the end of the
246 subject.
247
248 (c) \X did not give a hard partial match after matching one or more
249 characters at the end of the subject.
250
251 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
252 a partial match for the string "\r".
253
254 (e) When newline was set to CRLF, the metacharacter "." did not recognize
255 a partial match for a CR character at the end of the subject string.
256
257 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
258 running pcretest, the text "(JIT)" added to the output whenever JIT is
259 actually used to run the match.
260
261 7. Individual JIT compile options can be set in pcretest by following -s+[+]
262 or /S+[+] with a digit between 1 and 7.
263
264 8. OP_NOT now supports any UTF character not just single-byte ones.
265
266 9. (*MARK) control verb is now supported by the JIT compiler.
267
268 10. The command "./RunTest list" lists the available tests without actually
269 running any of them. (Because I keep forgetting what they all are.)
270
271 11. Add PCRE_INFO_MAXLOOKBEHIND.
272
273 12. Applied a (slightly modified) user-supplied patch that improves performance
274 when the heap is used for recursion (compiled with --disable-stack-for-
275 recursion). Instead of malloc and free for each heap frame each time a
276 logical recursion happens, frames are retained on a chain and re-used where
277 possible. This sometimes gives as much as 30% improvement.
278
279 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
280 call.
281
282 14. As documented, (*COMMIT) is now confined to within a positive assertion.
283
284 15. It is now possible to link pcretest with libedit as an alternative to
285 libreadline.
286
287 16. (*COMMIT) control verb is now supported by the JIT compiler.
288
289 17. The Unicode data tables have been updated to Unicode 6.1.0.
290
291 18. Added --file-list option to pcregrep.
292
293 19. Added binary file support to pcregrep, including the -a, --binary-files,
294 -I, and --text options.
295
296 20. The madvise function is renamed for posix_madvise for QNX compatibility
297 reasons. Fixed by Giuseppe D'Angelo.
298
299 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
300
301 22. Changed the option for creating symbolic links for 16-bit man pages from
302 -s to -sf so that re-installing does not cause issues.
303
304 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
305
306 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
307 matches in certain environments (the workspace was not being correctly
308 retained). Also added to pcre_dfa_exec() a simple plausibility check on
309 some of the workspace data at the beginning of a restart.
310
311 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
312 was not doing so when it should - probably a typo introduced by SVN 528
313 (change 8.10/14).
314
315 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
316 \w+ when the character tables indicated that \x{c4} was a word character.
317 There were several related cases, all because the tests for doing a table
318 lookup were testing for characters less than 127 instead of 255.
319
320 27. If a pattern contains capturing parentheses that are not used in a match,
321 their slots in the ovector are set to -1. For those that are higher than
322 any matched groups, this happens at the end of processing. In the case when
323 there were back references that the ovector was too small to contain
324 (causing temporary malloc'd memory to be used during matching), and the
325 highest capturing number was not used, memory off the end of the ovector
326 was incorrectly being set to -1. (It was using the size of the temporary
327 memory instead of the true size.)
328
329 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
330 ovector size, it uses memory at the end of the block that it has got.
331
332 29. Check for an overlong MARK name and give an error at compile time. The
333 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
334
335 30. JIT compiler update.
336
337 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
338 Rill for the patch.
339
340 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
341
342 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
343
344 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
345 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
346 was enabled.
347
348 35. Improve JIT code generation for greedy plus quantifier.
349
350 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
351 1 to "aa" instead of to an empty string. The bug affected repeated groups
352 that could potentially match an empty string.
353
354 37. Optimizing single character iterators in JIT.
355
356 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
357 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
358 codepoints that are too big for the mode are faulted, and in a UTF mode,
359 disallowed codepoints are also faulted.
360
361 39. If PCRE was compiled with UTF support, in three places in the DFA
362 matcher there was code that should only have been obeyed in UTF mode, but
363 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
364 processing when bytes with values greater than 127 were present. In 16-bit
365 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
366 both cases the values are those that cannot be the first data item in a UTF
367 character. The three items that might have provoked this were recursions,
368 possessively repeated groups, and atomic groups.
369
370 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
371 and pcregrep, because some OS require shared objects to be explicitly
372 passed to ld, causing the link step to fail if they are not.
373
374 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
375 mode, patterns that started with \h* or \R* might be incorrectly matched.
376
377
378 Version 8.30 04-February-2012
379 -----------------------------
380
381 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
382 name is defined in ctype.h.
383
384 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
385 only in quite long subpatterns.
386
387 3. Removed the function pcre_info(), which has been obsolete and deprecated
388 since it was replaced by pcre_fullinfo() in February 2000.
389
390 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
391 match a (*MARK), and the match failed at the start of the subject, a
392 reference to memory before the start of the subject could occur. This bug
393 was introduced by fix 17 of release 8.21.
394
395 5. A reference to an unset group with zero minimum repetition was giving
396 totally wrong answers (in non-JavaScript-compatibility mode). For example,
397 /(another)?(\1?)test/ matched against "hello world test". This bug was
398 introduced in release 8.13.
399
400 6. Add support for 16-bit character strings (a large amount of work involving
401 many changes and refactorings).
402
403 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
404 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
405 from a file.
406
407 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
408 rounding is not applied in this particular case).
409
410 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
411 if they appear, or are escaped, in patterns.
412
413 10. Get rid of a number of -Wunused-but-set-variable warnings.
414
415 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
416 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
417 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
418 also returns the mark "x". This bug applied to capturing parentheses,
419 non-capturing parentheses, and atomic parentheses. It also applied to some
420 assertions.
421
422 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
423 information out of configure.ac instead of relying on pcre.h.generic, which
424 is not stored in the repository.
425
426 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
427 -lreadline.
428
429 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
430
431 15. Applied Graycode's patch to put the top-level frame on the stack rather
432 than the heap when not using the stack for recursion. This gives a
433 performance improvement in many cases when recursion is not deep.
434
435 16. Experimental code added to "pcretest -C" to output the stack frame size.
436
437
438 Version 8.21 12-Dec-2011
439 ------------------------
440
441 1. Updating the JIT compiler.
442
443 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
444 are added as well.
445
446 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
447 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
448 calling _pcre_jit_exec. Some extra comments are added.
449
450 4. (*MARK) settings inside atomic groups that do not contain any capturing
451 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
452 was introduced by change 18 for 8.20.
453
454 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
455 ECMA-262 standard.
456
457 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
458 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
459 This bug was probably introduced by change 9 of 8.13.
460
461 7. While fixing 6 above, I noticed that a number of other items were being
462 incorrectly rejected as "not fixed length". This arose partly because newer
463 opcodes had not been added to the fixed-length checking code. I have (a)
464 corrected the bug and added tests for these items, and (b) arranged for an
465 error to occur if an unknown opcode is encountered while checking for fixed
466 length instead of just assuming "not fixed length". The items that were
467 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
468 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
469 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
470
471 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
472 being incorrectly compiled and would have given unpredicatble results.
473
474 9. A possessively repeated subpattern with minimum repeat count greater than
475 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
476 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
477 the first (A) could occur when it should not.
478
479 10. Add a cast and remove a redundant test from the code.
480
481 11. JIT should use pcre_malloc/pcre_free for allocation.
482
483 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
484 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
485 is anything other than /usr, -L is still shown).
486
487 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
488
489 14. Perl does not support \N without a following name in a [] class; PCRE now
490 also gives an error.
491
492 15. If a forward reference was repeated with an upper limit of around 2000,
493 it caused the error "internal error: overran compiling workspace". The
494 maximum number of forward references (including repeats) was limited by the
495 internal workspace, and dependent on the LINK_SIZE. The code has been
496 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
497 the default depends on LINK_SIZE. There is a new upper limit (for safety)
498 of around 200,000 forward references. While doing this, I also speeded up
499 the filling in of repeated forward references.
500
501 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
502 incorrectly expecting the subject to contain another "a" after the start.
503
504 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
505 in the match, the SKIP should be ignored. This was not happening; instead
506 the SKIP was being treated as NOMATCH. For patterns such as
507 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
508 tested.
509
510 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
511 now much more compatible with Perl, in particular in cases where the result
512 is a non-match for a non-anchored pattern. For example, if
513 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
514 "m", where previously it did not return a name. A side effect of this
515 change is that for partial matches, the last encountered mark name is
516 returned, as for non matches. A number of tests that were previously not
517 Perl-compatible have been moved into the Perl-compatible test files. The
518 refactoring has had the pleasing side effect of removing one argument from
519 the match() function, thus reducing its stack requirements.
520
521 19. If the /S+ option was used in pcretest to study a pattern using JIT,
522 subsequent uses of /S (without +) incorrectly behaved like /S+.
523
524 21. Retrieve executable code size support for the JIT compiler and fixing
525 some warnings.
526
527 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
528 not work when the shorter character appeared right at the end of the
529 subject string.
530
531 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
532 systems.
533
534 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
535 output it when the /M option is used in pcretest.
536
537 25. The CheckMan script was not being included in the distribution. Also, added
538 an explicit "perl" to run Perl scripts from the PrepareRelease script
539 because this is reportedly needed in Windows.
540
541 26. If study data was being save in a file and studying had not found a set of
542 "starts with" bytes for the pattern, the data written to the file (though
543 never used) was taken from uninitialized memory and so caused valgrind to
544 complain.
545
546 27. Updated RunTest.bat as provided by Sheri Pierce.
547
548 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
549
550 29. Computation of memory usage for the table of capturing group names was
551 giving an unnecessarily large value.
552
553
554 Version 8.20 21-Oct-2011
555 ------------------------
556
557 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
558 a POSIX class. After further experiments with Perl, which convinced me that
559 Perl has bugs and confusions, a closing square bracket is no longer allowed
560 in a POSIX name. This bug also affected patterns with classes that started
561 with full stops.
562
563 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
564 captured substring, but while checking the failing first alternative,
565 substring 1 is temporarily captured. If the output vector supplied to
566 pcre_exec() was not big enough for this capture, the yield of the function
567 was still zero ("insufficient space for captured substrings"). This cannot
568 be totally fixed without adding another stack variable, which seems a lot
569 of expense for a edge case. However, I have improved the situation in cases
570 such as /(a)(b)x|abc/ matched against "abc", where the return code
571 indicates that fewer than the maximum number of slots in the ovector have
572 been set.
573
574 3. Related to (2) above: when there are more back references in a pattern than
575 slots in the output vector, pcre_exec() uses temporary memory during
576 matching, and copies in the captures as far as possible afterwards. It was
577 using the entire output vector, but this conflicts with the specification
578 that only 2/3 is used for passing back captured substrings. Now it uses
579 only the first 2/3, for compatibility. This is, of course, another edge
580 case.
581
582 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
583 main code base, and can be used by building with --enable-jit. When this is
584 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
585 runtime --no-jit option is given.
586
587 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
588 ovector, the return from the function was zero, implying that there were
589 other matches that did not fit. The correct "exactly full" value is now
590 returned.
591
592 6. If a subpattern that was called recursively or as a subroutine contained
593 (*PRUNE) or any other control that caused it to give a non-standard return,
594 invalid errors such as "Error -26 (nested recursion at the same subject
595 position)" or even infinite loops could occur.
596
597 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
598 computing the minimum length on reaching *ACCEPT, and so ended up with the
599 wrong value of 1 rather than 0. Further investigation indicates that
600 computing a minimum subject length in the presence of *ACCEPT is difficult
601 (think back references, subroutine calls), and so I have changed the code
602 so that no minimum is registered for a pattern that contains *ACCEPT.
603
604 8. If (*THEN) was present in the first (true) branch of a conditional group,
605 it was not handled as intended. [But see 16 below.]
606
607 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
608 Sheri Pierce.
609
610 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
611 the first byte in a match must be "a".
612
613 11. Change 17 for 8.13 increased the recursion depth for patterns like
614 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
615 pattern contains any instances of (*THEN). If it does not, the old
616 optimizations are restored. It would be nice to do this on a per-group
617 basis, but at the moment that is not feasible.
618
619 12. In some environments, the output of pcretest -C is CRLF terminated. This
620 broke RunTest's code that checks for the link size. A single white space
621 character after the value is now allowed for.
622
623 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
624 For "fr", it uses the Windows-specific input and output files.
625
626 14. If (*THEN) appeared in a group that was called recursively or as a
627 subroutine, it did not work as intended. [But see next item.]
628
629 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
630 pattern fragments (but not containing any | characters). If A and B are
631 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
632 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
633 D. In other words, Perl considers parentheses that do not contain any |
634 characters to be part of a surrounding alternative, whereas PCRE was
635 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
636 differently. PCRE now behaves in the same way as Perl, except in the case
637 of subroutine/recursion calls such as (?1) which have in any case always
638 been different (but PCRE had them first :-).
639
640 16. Related to 15 above: Perl does not treat the | in a conditional group as
641 creating alternatives. Such a group is treated in the same way as an
642 ordinary group without any | characters when processing (*THEN). PCRE has
643 been changed to match Perl's behaviour.
644
645 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
646 RunGrepTest script failed.
647
648 18. Change 22 for version 13 caused atomic groups to use more stack. This is
649 inevitable for groups that contain captures, but it can lead to a lot of
650 stack use in large patterns. The old behaviour has been restored for atomic
651 groups that do not contain any capturing parentheses.
652
653 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
654 suppress the check for a minimum subject length at run time. (If it was
655 given to pcre_exec() or pcre_dfa_exec() it did work.)
656
657 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
658 fail to work when decoding hex characters in data strings in EBCDIC
659 environments.
660
661 21. It appears that in at least one Mac OS environment, the isxdigit() function
662 is implemented as a macro that evaluates to its argument more than once,
663 contravening the C 90 Standard (I haven't checked a later standard). There
664 was an instance in pcretest which caused it to go wrong when processing
665 \x{...} escapes in subject strings. The has been rewritten to avoid using
666 things like p++ in the argument of isxdigit().
667
668
669 Version 8.13 16-Aug-2011
670 ------------------------
671
672 1. The Unicode data tables have been updated to Unicode 6.0.0.
673
674 2. Two minor typos in pcre_internal.h have been fixed.
675
676 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
677 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
678 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
679
680 4. There were a number of related bugs in the code for matching backrefences
681 caselessly in UTF-8 mode when codes for the characters concerned were
682 different numbers of bytes. For example, U+023A and U+2C65 are an upper
683 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
684 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
685 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
686 2-byte code at the end of the subject (it thought there wasn't enough data
687 left).
688
689 5. Comprehensive information about what went wrong is now returned by
690 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
691 as the output vector has at least 2 elements. The offset of the start of
692 the failing character and a reason code are placed in the vector.
693
694 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
695 now returned is for the first byte of the failing character, instead of the
696 last byte inspected. This is an incompatible change, but I hope it is small
697 enough not to be a problem. It makes the returned offset consistent with
698 pcre_exec() and pcre_dfa_exec().
699
700 7. pcretest now gives a text phrase as well as the error number when
701 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
702 failure, the offset and reason code are output.
703
704 8. When \R was used with a maximizing quantifier it failed to skip backwards
705 over a \r\n pair if the subsequent match failed. Instead, it just skipped
706 back over a single character (\n). This seems wrong (because it treated the
707 two characters as a single entity when going forwards), conflicts with the
708 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
709 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
710 has been changed.
711
712 9. Some internal refactoring has changed the processing so that the handling
713 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
714 time (the PCRE_DOTALL option was changed this way some time ago: version
715 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
716 which was always a bit of a fudge. It also means that there is one less
717 argument for the match() function, which reduces its stack requirements
718 slightly. This change also fixes an incompatibility with Perl: the pattern
719 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
720
721 10. More internal refactoring has drastically reduced the number of recursive
722 calls to match() for possessively repeated groups such as (abc)++ when
723 using pcre_exec().
724
725 11. While implementing 10, a number of bugs in the handling of groups were
726 discovered and fixed:
727
728 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
729 (a|)*(?1) gave a compile-time internal error.
730 ((a|)+)+ did not notice that the outer group could match an empty string.
731 (^a|^)+ was not marked as anchored.
732 (.*a|.*)+ was not marked as matching at start or after a newline.
733
734 12. Yet more internal refactoring has removed another argument from the match()
735 function. Special calls to this function are now indicated by setting a
736 value in a variable in the "match data" data block.
737
738 13. Be more explicit in pcre_study() instead of relying on "default" for
739 opcodes that mean there is no starting character; this means that when new
740 ones are added and accidentally left out of pcre_study(), testing should
741 pick them up.
742
743 14. The -s option of pcretest has been documented for ages as being an old
744 synonym of -m (show memory usage). I have changed it to mean "force study
745 for every regex", that is, assume /S for every regex. This is similar to -i
746 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
747 using it. It makes it easier to run collections of tests with and without
748 study enabled, and thereby test pcre_study() more easily. All the standard
749 tests are now run with and without -s (but some patterns can be marked as
750 "never study" - see 20 below).
751
752 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
753 restoration of the capturing data to the outer values was not happening
754 correctly.
755
756 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
757 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
758 pattern had matched an empty string, and so incorrectly returned a no
759 match.
760
761 17. There was optimizing code for the last branch of non-capturing parentheses,
762 and also for the obeyed branch of a conditional subexpression, which used
763 tail recursion to cut down on stack usage. Unfortunately, now that there is
764 the possibility of (*THEN) occurring in these branches, tail recursion is
765 no longer possible because the return has to be checked for (*THEN). These
766 two optimizations have therefore been removed. [But see 8.20/11 above.]
767
768 18. If a pattern containing \R was studied, it was assumed that \R always
769 matched two bytes, thus causing the minimum subject length to be
770 incorrectly computed because \R can also match just one byte.
771
772 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
773 was incorrectly computed.
774
775 20. If /S is present twice on a test pattern in pcretest input, it now
776 *disables* studying, thereby overriding the use of -s on the command line
777 (see 14 above). This is necessary for one or two tests to keep the output
778 identical in both cases.
779
780 21. When (*ACCEPT) was used in an assertion that matched an empty string and
781 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
782
783 22. When an atomic group that contained a capturing parenthesis was
784 successfully matched, but the branch in which it appeared failed, the
785 capturing was not being forgotten if a higher numbered group was later
786 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
787 group 1 to "a", when in fact it should be unset. This applied to multi-
788 branched capturing and non-capturing groups, repeated or not, and also to
789 positive assertions (capturing in negative assertions does not happen
790 in PCRE) and also to nested atomic groups.
791
792 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
793 subject after a captured substring, to make it easier to tell which of a
794 number of identical substrings has been captured.
795
796 24. The way atomic groups are processed by pcre_exec() has been changed so that
797 if they are repeated, backtracking one repetition now resets captured
798 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
799 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
800 "aaa". Previously, it would have been "a". As part of this code
801 refactoring, the way recursive calls are handled has also been changed.
802
803 25. If an assertion condition captured any substrings, they were not passed
804 back unless some other capturing happened later. For example, if
805 (?(?=(a))a) was matched against "a", no capturing was returned.
806
807 26. When studying a pattern that contained subroutine calls or assertions,
808 the code for finding the minimum length of a possible match was handling
809 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
810 group 1 called group 2 while simultaneously a separate group 2 called group
811 1). A stack overflow occurred in this case. I have fixed this by limiting
812 the recursion depth to 10.
813
814 27. Updated RunTest.bat in the distribution to the version supplied by Tom
815 Fortmann. This supports explicit test numbers on the command line, and has
816 argument validation and error reporting.
817
818 28. An instance of \X with an unlimited repeat could fail if at any point the
819 first character it looked at was a mark character.
820
821 29. Some minor code refactoring concerning Unicode properties and scripts
822 should reduce the stack requirement of match() slightly.
823
824 30. Added the '=' option to pcretest to check the setting of unused capturing
825 slots at the end of the pattern, which are documented as being -1, but are
826 not included in the return count.
827
828 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
829 compiled something random. Now it gives a compile-time error (as does
830 Perl).
831
832 32. A *MARK encountered during the processing of a positive assertion is now
833 recorded and passed back (compatible with Perl).
834
835 33. If --only-matching or --colour was set on a pcregrep call whose pattern
836 had alternative anchored branches, the search for a second match in a line
837 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
838 matched the line "0102" twice. The same bug affected patterns that started
839 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
840 twice.
841
842 34. Previously, PCRE did not allow quantification of assertions. However, Perl
843 does, and because of capturing effects, quantifying parenthesized
844 assertions may at times be useful. Quantifiers are now allowed for
845 parenthesized assertions.
846
847 35. A minor code tidy in pcre_compile() when checking options for \R usage.
848
849 36. \g was being checked for fancy things in a character class, when it should
850 just be a literal "g".
851
852 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
853 appearance of a nested POSIX class supersedes an apparent external class.
854 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
855 unescaped square brackets may also appear as part of class names. For
856 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
857 more like Perl. (But see 8.20/1 above.)
858
859 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
860 was because it thought it was \N{name}, which is not supported).
861
862 39. Add minix to OS list not supporting the -S option in pcretest.
863
864 40. PCRE tries to detect cases of infinite recursion at compile time, but it
865 cannot analyze patterns in sufficient detail to catch mutual recursions
866 such as ((?1))((?2)). There is now a runtime test that gives an error if a
867 subgroup is called recursively as a subpattern for a second time at the
868 same position in the subject string. In previous releases this might have
869 been caught by the recursion limit, or it might have run out of stack.
870
871 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
872 happen only once. PCRE was, however incorrectly giving a compile time error
873 "recursive call could loop indefinitely" because it cannot analyze the
874 pattern in sufficient detail. The compile time test no longer happens when
875 PCRE is compiling a conditional subpattern, but actual runaway loops are
876 now caught at runtime (see 40 above).
877
878 42. It seems that Perl allows any characters other than a closing parenthesis
879 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
880 has been changed to be the same.
881
882 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
883 as not to get warnings when autogen.sh is called. Also changed
884 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
885
886 44. To help people who use pcregrep to scan files containing exceedingly long
887 lines, the following changes have been made:
888
889 (a) The default value of the buffer size parameter has been increased from
890 8K to 20K. (The actual buffer used is three times this size.)
891
892 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
893 PCRE is built.
894
895 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
896 to be set at run time.
897
898 (d) Numerical values in pcregrep options can be followed by K or M, for
899 example --buffer-size=50K.
900
901 (e) If a line being scanned overflows pcregrep's buffer, an error is now
902 given and the return code is set to 2.
903
904 45. Add a pointer to the latest mark to the callout data block.
905
906 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
907 partial match of an empty string instead of no match. This was specific to
908 the use of ".".
909
910 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
911 complete match instead of a partial match. This bug was dependent on both
912 the PCRE_UTF8 and PCRE_DOTALL options being set.
913
914 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
915 starting byte set, because \b was not being ignored.
916
917
918 Version 8.12 15-Jan-2011
919 ------------------------
920
921 1. Fixed some typos in the markup of the man pages, and wrote a script that
922 checks for such things as part of the documentation building process.
923
924 2. On a big-endian 64-bit system, pcregrep did not correctly process the
925 --match-limit and --recursion-limit options (added for 8.11). In
926 particular, this made one of the standard tests fail. (The integer value
927 went into the wrong half of a long int.)
928
929 3. If the --colour option was given to pcregrep with -v (invert match), it
930 did strange things, either producing crazy output, or crashing. It should,
931 of course, ignore a request for colour when reporting lines that do not
932 match.
933
934 4. Another pcregrep bug caused similar problems if --colour was specified with
935 -M (multiline) and the pattern match finished with a line ending.
936
937 5. In pcregrep, when a pattern that ended with a literal newline sequence was
938 matched in multiline mode, the following line was shown as part of the
939 match. This seems wrong, so I have changed it.
940
941 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
942 the check for further matches in the same line (so they could be coloured)
943 to overrun the end of the current line. If another match was found, it was
944 incorrectly shown (and then shown again when found in the next line).
945
946 7. If pcregrep was compiled under Windows, there was a reference to the
947 function pcregrep_exit() before it was defined. I am assuming this was
948 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
949 reported by a user. I've moved the definition above the reference.
950
951
952 Version 8.11 10-Dec-2010
953 ------------------------
954
955 1. (*THEN) was not working properly if there were untried alternatives prior
956 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
957 backtracked to try for "b" instead of moving to the next alternative branch
958 at the same level (in this case, to look for "c"). The Perl documentation
959 is clear that when (*THEN) is backtracked onto, it goes to the "next
960 alternative in the innermost enclosing group".
961
962 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
963 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
964 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
965 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
966 (*THEN).
967
968 3. If \s appeared in a character class, it removed the VT character from
969 the class, even if it had been included by some previous item, for example
970 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
971 of \s, but is part of the POSIX "space" class.)
972
973 4. A partial match never returns an empty string (because you can always
974 match an empty string at the end of the subject); however the checking for
975 an empty string was starting at the "start of match" point. This has been
976 changed to the "earliest inspected character" point, because the returned
977 data for a partial match starts at this character. This means that, for
978 example, /(?<=abc)def/ gives a partial match for the subject "abc"
979 (previously it gave "no match").
980
981 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
982 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
983 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
984 has an implication that the given string is incomplete (because a partial
985 match is preferred over a full match). For this reason, these items now
986 give a partial match in this situation. [Aside: previously, the one case
987 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
988 match rather than a full match, which was wrong by the old rules, but is
989 now correct.]
990
991 6. There was a bug in the handling of #-introduced comments, recognized when
992 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
993 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
994 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
995 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
996 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
997 places in pcre_compile().
998
999 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1000 comments when looking ahead for named forward references to subpatterns,
1001 the only newline sequence it recognized was NL. It now handles newlines
1002 according to the set newline convention.
1003
1004 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1005 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1006 cater for a lack of strerror(). These oversights have been fixed.
1007
1008 9. Added --match-limit and --recursion-limit to pcregrep.
1009
1010 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1011
1012 11. When the -o option was used, pcregrep was setting a return code of 1, even
1013 when matches were found, and --line-buffered was not being honoured.
1014
1015 12. Added an optional parentheses number to the -o and --only-matching options
1016 of pcregrep.
1017
1018 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1019 can match an empty string. The code to do it in pcretest and pcredemo
1020 needed fixing:
1021
1022 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1023 only one byte after an empty string match just before CRLF (this case
1024 just got forgotten; "any" and "anycrlf" were OK).
1025
1026 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1027 mode when an empty string match preceded an ASCII character followed by
1028 a non-ASCII character. (The code for advancing by one character rather
1029 than one byte was nonsense.)
1030
1031 (c) The pcredemo.c sample program did not have any code at all to handle
1032 the cases when CRLF is a valid newline sequence.
1033
1034 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1035 as a starting offset was within the subject string. There is now a new
1036 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1037 negative or greater than the length of the string. In order to test this,
1038 pcretest is extended to allow the setting of negative starting offsets.
1039
1040 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1041 starting offset points to the beginning of a UTF-8 character was
1042 unnecessarily clumsy. I tidied it up.
1043
1044 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1045 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1046
1047 17. Nobody had reported that the --include_dir option, which was added in
1048 release 7.7 should have been called --include-dir (hyphen, not underscore)
1049 for compatibility with GNU grep. I have changed it to --include-dir, but
1050 left --include_dir as an undocumented synonym, and the same for
1051 --exclude-dir, though that is not available in GNU grep, at least as of
1052 release 2.5.4.
1053
1054 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1055 characters from a string of bytes) have been redefined so as not to use
1056 loops, in order to improve performance in some environments. At the same
1057 time, I abstracted some of the common code into auxiliary macros to save
1058 repetition (this should not affect the compiled code).
1059
1060 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1061 compile-time error is now given if \c is not followed by an ASCII
1062 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1063 different, and any byte value is allowed.)
1064
1065 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1066 START_OPTIMIZE option, which is now allowed at compile time - but just
1067 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1068 to pcregrep and other applications that have no direct access to PCRE
1069 options. The new /Y option in pcretest sets this option when calling
1070 pcre_compile().
1071
1072 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1073 back references. Groups containing recursive back references were forced to
1074 be atomic by that change, but in the case of named groups, the amount of
1075 memory required was incorrectly computed, leading to "Failed: internal
1076 error: code overflow". This has been fixed.
1077
1078 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1079 pcretest.c, to avoid build problems in some Borland environments.
1080
1081
1082 Version 8.10 25-Jun-2010
1083 ------------------------
1084
1085 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1086 THEN.
1087
1088 2. (*ACCEPT) was not working when inside an atomic group.
1089
1090 3. Inside a character class, \B is treated as a literal by default, but
1091 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1092 causes the error). The code is unchanged, but I tidied the documentation.
1093
1094 4. Inside a character class, PCRE always treated \R and \X as literals,
1095 whereas Perl faults them if its -w option is set. I have changed PCRE so
1096 that it faults them when PCRE_EXTRA is set.
1097
1098 5. Added support for \N, which always matches any character other than
1099 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1100
1101 6. When compiling pcregrep with newer versions of gcc which may have
1102 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1103 declared with attribute warn_unused_result" were given. Just casting the
1104 result to (void) does not stop the warnings; a more elaborate fudge is
1105 needed. I've used a macro to implement this.
1106
1107 7. Minor change to pcretest.c to avoid a compiler warning.
1108
1109 8. Added four artifical Unicode properties to help with an option to make
1110 \s etc use properties (see next item). The new properties are: Xan
1111 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1112
1113 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1114 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1115 this option. Modified pcretest to add /W to test this facility. Added
1116 REG_UCP to make it available via the POSIX interface.
1117
1118 10. Added --line-buffered to pcregrep.
1119
1120 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1121 studied, and the match started with a letter with a code point greater than
1122 127 whose first byte was different to the first byte of the other case of
1123 the letter, the other case of this starting letter was not recognized
1124 (#976).
1125
1126 12. If a pattern that was studied started with a repeated Unicode property
1127 test, for example, \p{Nd}+, there was the theoretical possibility of
1128 setting up an incorrect bitmap of starting bytes, but fortunately it could
1129 not have actually happened in practice until change 8 above was made (it
1130 added property types that matched character-matching opcodes).
1131
1132 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1133 possible starting bytes for non-anchored patterns.
1134
1135 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1136 \R, and also a number of cases that involve Unicode properties, both
1137 explicit and implicit when PCRE_UCP is set.
1138
1139 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1140 input, it could crash or give wrong results if characters with values
1141 greater than 0xc0 were present in the subject string. (Detail: it assumed
1142 UTF-8 input when processing these items.)
1143
1144 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1145 size_t is 64-bit (#991).
1146
1147 17. Added a check for running out of memory when PCRE is compiled with
1148 --disable-stack-for-recursion (#990).
1149
1150 18. If the last data line in a file for pcretest does not have a newline on
1151 the end, a newline was missing in the output.
1152
1153 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1154 less than 128) in its various bitmaps. However, there is a facility for
1155 generating tables according to the current locale when PCRE is compiled. It
1156 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1157 characters, are recognized by isspace() and therefore were getting set in
1158 these tables, and indeed these tables seem to approximate to ISO 8859. This
1159 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1160 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1161 which of course cannot start UTF-8 characters. I have changed the code so
1162 that only real ASCII characters (less than 128) and the correct starting
1163 bytes for UTF-8 encodings are set for characters greater than 127 when in
1164 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1165 altogether.)
1166
1167 20. Added the /T option to pcretest so as to be able to run tests with non-
1168 standard character tables, thus making it possible to include the tests
1169 used for 19 above in the standard set of tests.
1170
1171 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1172 reference to a subpattern the other side of a comment that contains an
1173 opening parenthesis caused either an internal compiling error, or a
1174 reference to the wrong subpattern.
1175
1176
1177 Version 8.02 19-Mar-2010
1178 ------------------------
1179
1180 1. The Unicode data tables have been updated to Unicode 5.2.0.
1181
1182 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1183 configured.
1184
1185 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1186 original author of that file, following a query about its status.
1187
1188 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1189 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1190
1191 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1192 quantifier applied to a forward-referencing subroutine call, could compile
1193 incorrect code or give the error "internal error: previously-checked
1194 referenced subpattern not found".
1195
1196 6. Both MS Visual Studio and Symbian OS have problems with initializing
1197 variables to point to external functions. For these systems, therefore,
1198 pcre_malloc etc. are now initialized to local functions that call the
1199 relevant global functions.
1200
1201 7. There were two entries missing in the vectors called coptable and poptable
1202 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1203 I've fixed the data, and added a kludgy way of testing at compile time that
1204 the lengths are correct (equal to the number of opcodes).
1205
1206 8. Following on from 7, I added a similar kludge to check the length of the
1207 eint vector in pcreposix.c.
1208
1209 9. Error texts for pcre_compile() are held as one long string to avoid too
1210 much relocation at load time. To find a text, the string is searched,
1211 counting zeros. There was no check for running off the end of the string,
1212 which could happen if a new error number was added without updating the
1213 string.
1214
1215 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1216
1217 11. \K was not working if it appeared in an atomic group or in a group that
1218 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1219 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1220 the assertion is positive, but not if it is negative.
1221
1222 12. Change 11 fortuitously reduced the size of the stack frame used in the
1223 "match()" function of pcre_exec.c by one pointer. Forthcoming
1224 implementation of support for (*MARK) will need an extra pointer on the
1225 stack; I have reserved it now, so that the stack frame size does not
1226 decrease.
1227
1228 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1229 item in branch that calls a recursion is a subroutine call - as in the
1230 second branch in the above example - was incorrectly given the compile-
1231 time error "recursive call could loop indefinitely" because pcre_compile()
1232 was not correctly checking the subroutine for matching a non-empty string.
1233
1234 14. The checks for overrunning compiling workspace could trigger after an
1235 overrun had occurred. This is a "should never occur" error, but it can be
1236 triggered by pathological patterns such as hundreds of nested parentheses.
1237 The checks now trigger 100 bytes before the end of the workspace.
1238
1239 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1240
1241
1242 Version 8.01 19-Jan-2010
1243 ------------------------
1244
1245 1. If a pattern contained a conditional subpattern with only one branch (in
1246 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1247 computed the wrong minimum data length (which is of course zero for such
1248 subpatterns). This could cause incorrect "no match" results.
1249
1250 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1251 the pattern is reset in the first branch, pcre_compile() failed with
1252 "internal error: code overflow at offset...". This happened only when
1253 the reset was to the original external option setting. (An optimization
1254 abstracts leading options settings into an external setting, which was the
1255 cause of this.)
1256
1257 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1258 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1259 assertion pattern did not match (meaning that the assertion was true), it
1260 was incorrectly treated as false if the SKIP had been reached during the
1261 matching. This also applied to assertions used as conditions.
1262
1263 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1264 assertion subpattern, including such a pattern used as a condition,
1265 unpredictable results occurred, instead of the error return
1266 PCRE_ERROR_DFA_UITEM.
1267
1268 5. The C++ GlobalReplace function was not working like Perl for the special
1269 situation when an empty string is matched. It now does the fancy magic
1270 stuff that is necessary.
1271
1272 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1273 removed. (These were left over from very, very early versions of PCRE.)
1274
1275 7. Some cosmetic changes to the code to make life easier when compiling it
1276 as part of something else:
1277
1278 (a) Change DEBUG to PCRE_DEBUG.
1279
1280 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1281 called "current" as "current_branch", to prevent a collision with the
1282 Linux macro when compiled as a kernel module.
1283
1284 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1285 prevent a collision with the Linux macro when compiled as a kernel
1286 module.
1287
1288 8. In pcre_compile() there are some checks for integer overflows that used to
1289 cast potentially large values to (double). This has been changed to that
1290 when building, a check for int64_t is made, and if it is found, it is used
1291 instead, thus avoiding the use of floating point arithmetic. (There is no
1292 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1293 double.
1294
1295 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1296 2005 (difference between two addresses compared to an unsigned value).
1297
1298 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1299 custom one, because of the following reported problem in Windows:
1300
1301 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1302 under Win32.
1303 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1304 therefore missing the function definition.
1305 - The compiler thus generates a "C" signature for the test function.
1306 - The linker fails to find the "C" function.
1307 - PCRE fails to configure if asked to do so against libbz2.
1308
1309 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1310 messages were output:
1311
1312 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1313 rerunning libtoolize, to keep the correct libtool macros in-tree.
1314 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1315
1316 I have done both of these things.
1317
1318 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1319 most of the time, it *can* run out if it is given a pattern that contains a
1320 runaway infinite recursion. I updated the discussion in the pcrestack man
1321 page.
1322
1323 13. Now that we have gone to the x.xx style of version numbers, the minor
1324 version may start with zero. Using 08 or 09 is a bad idea because users
1325 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1326 interpreted as invalid octal numbers. I've updated the previous comment in
1327 configure.ac, and also added a check that gives an error if 08 or 09 are
1328 used.
1329
1330 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1331 causing partial matching to fail when the end of the subject matched \W
1332 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1333
1334 15. There were some discrepancies between the declarations in pcre_internal.h
1335 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1336 their definitions. The declarations used "const uschar *" and the
1337 definitions used USPTR. Even though USPTR is normally defined as "const
1338 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1339 reported that: "This difference in casting confuses some C++ compilers, for
1340 example, SunCC recognizes above declarations as different functions and
1341 generates broken code for hbpcre." I have changed the declarations to use
1342 USPTR.
1343
1344 16. GNU libtool is named differently on some systems. The autogen.sh script now
1345 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1346 (FreeBSD).
1347
1348 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1349 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1350 comment: "Figure out how to create a longlong from a string: strtoll and
1351 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1352 instance, but it only takes 2 args instead of 3!"
1353
1354 18. A subtle bug concerned with back references has been fixed by a change of
1355 specification, with a corresponding code fix. A pattern such as
1356 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1357 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1358 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1359 same bug. Such groups have to be quantified to be useful, or contained
1360 inside another quantified group. (If there's no repetition, the reference
1361 can never match.) The problem arises because, having left the group and
1362 moved on to the rest of the pattern, a later failure that backtracks into
1363 the group uses the captured value from the final iteration of the group
1364 rather than the correct earlier one. I have fixed this in PCRE by forcing
1365 any group that contains a reference to itself to be an atomic group; that
1366 is, there cannot be any backtracking into it once it has completed. This is
1367 similar to recursive and subroutine calls.
1368
1369
1370 Version 8.00 19-Oct-09
1371 ----------------------
1372
1373 1. The table for translating pcre_compile() error codes into POSIX error codes
1374 was out-of-date, and there was no check on the pcre_compile() error code
1375 being within the table. This could lead to an OK return being given in
1376 error.
1377
1378 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1379 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1380 in a Windows environment.
1381
1382 3. The pcregrep --count option prints the count for each file even when it is
1383 zero, as does GNU grep. However, pcregrep was also printing all files when
1384 --files-with-matches was added. Now, when both options are given, it prints
1385 counts only for those files that have at least one match. (GNU grep just
1386 prints the file name in this circumstance, but including the count seems
1387 more useful - otherwise, why use --count?) Also ensured that the
1388 combination -clh just lists non-zero counts, with no names.
1389
1390 4. The long form of the pcregrep -F option was incorrectly implemented as
1391 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1392 but it seems right to fix it, and I didn't think it was worth preserving
1393 the old behaviour.
1394
1395 5. The command line items --regex=pattern and --regexp=pattern were not
1396 recognized by pcregrep, which required --regex pattern or --regexp pattern
1397 (with a space rather than an '='). The man page documented the '=' forms,
1398 which are compatible with GNU grep; these now work.
1399
1400 6. No libpcreposix.pc file was created for pkg-config; there was just
1401 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1402
1403 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1404 when UCP support is not needed, by modifying the Python script that
1405 generates it from Unicode data files. This should not matter if the module
1406 is correctly used as a library, but I received one complaint about 50K of
1407 unwanted data. My guess is that the person linked everything into his
1408 program rather than using a library. Anyway, it does no harm.
1409
1410 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1411 was a minimum greater than 1 for a wide character in a possessive
1412 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1413 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1414 character. Chaos in the form of incorrect output or a compiling loop could
1415 result.
1416
1417 9. The restrictions on what a pattern can contain when partial matching is
1418 requested for pcre_exec() have been removed. All patterns can now be
1419 partially matched by this function. In addition, if there are at least two
1420 slots in the offset vector, the offset of the earliest inspected character
1421 for the match and the offset of the end of the subject are set in them when
1422 PCRE_ERROR_PARTIAL is returned.
1423
1424 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1425 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1426 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1427 and may be more useful for multi-segment matching.
1428
1429 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1430 used to be given if ever the end of the subject was reached; now it is
1431 given only if matching could not proceed because another character was
1432 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1433 string "Z", which now yields "no match" instead of "partial match". In the
1434 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1435 final character ended with (*FAIL).
1436
1437 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1438 if the pattern had a "must contain" character that was already found in the
1439 earlier partial match, unless partial matching was again requested. For
1440 example, with the pattern /dog.(body)?/, the "must contain" character is
1441 "g". If the first part-match was for the string "dog", restarting with
1442 "sbody" failed. This bug has been fixed.
1443
1444 13. The string returned by pcre_dfa_exec() after a partial match has been
1445 changed so that it starts at the first inspected character rather than the
1446 first character of the match. This makes a difference only if the pattern
1447 starts with a lookbehind assertion or \b or \B (\K is not supported by
1448 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1449 matching functions compatible, and I think it's the right thing to do.
1450
1451 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1452 so that the demonstration program is easily available in environments where
1453 PCRE has not been installed from source.
1454
1455 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1456 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1457 library.
1458
1459 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1460 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1461 is not the first non-POSIX option to be added. Clearly some people find
1462 these options useful.
1463
1464 17. If a caller to the POSIX matching function regexec() passes a non-zero
1465 value for nmatch with a NULL value for pmatch, the value of
1466 nmatch is forced to zero.
1467
1468 18. RunGrepTest did not have a test for the availability of the -u option of
1469 the diff command, as RunTest does. It now checks in the same way as
1470 RunTest, and also checks for the -b option.
1471
1472 19. If an odd number of negated classes containing just a single character
1473 interposed, within parentheses, between a forward reference to a named
1474 subpattern and the definition of the subpattern, compilation crashed with
1475 an internal error, complaining that it could not find the referenced
1476 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1477 [The bug was that it was starting one character too far in when skipping
1478 over the character class, thus treating the ] as data rather than
1479 terminating the class. This meant it could skip too much.]
1480
1481 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1482 /g option in pcretest when the pattern contains \K, which makes it possible
1483 to have an empty string match not at the start, even when the pattern is
1484 anchored. Updated pcretest and pcredemo to use this option.
1485
1486 21. If the maximum number of capturing subpatterns in a recursion was greater
1487 than the maximum at the outer level, the higher number was returned, but
1488 with unset values at the outer level. The correct (outer level) value is
1489 now given.
1490
1491 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1492 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1493 make it do so. The string so far is captured, making this feature
1494 compatible with Perl.
1495
1496 23. The tests have been re-organized, adding tests 11 and 12, to make it
1497 possible to check the Perl 5.10 features against Perl 5.10.
1498
1499 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1500 pattern matches a fixed length string. PCRE did not allow this; now it
1501 does. Neither allows recursion.
1502
1503 25. I finally figured out how to implement a request to provide the minimum
1504 length of subject string that was needed in order to match a given pattern.
1505 (It was back references and recursion that I had previously got hung up
1506 on.) This code has now been added to pcre_study(); it finds a lower bound
1507 to the length of subject needed. It is not necessarily the greatest lower
1508 bound, but using it to avoid searching strings that are too short does give
1509 some useful speed-ups. The value is available to calling programs via
1510 pcre_fullinfo().
1511
1512 26. While implementing 25, I discovered to my embarrassment that pcretest had
1513 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1514 study optimizations had never been tested with that matching function.
1515 Oops. What is worse, even when it was passed study data, there was a bug in
1516 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1517 were also very few tests of studied patterns with pcre_dfa_exec().
1518
1519 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1520 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1521 on the other side of the coin, they are no longer allowed to have different
1522 names, because these cannot be distinguished in PCRE, and this has caused
1523 confusion. (This is a difference from Perl.)
1524
1525 28. When duplicate subpattern names are present (necessarily with different
1526 numbers, as required by 27 above), and a test is made by name in a
1527 conditional pattern, either for a subpattern having been matched, or for
1528 recursion in such a pattern, all the associated numbered subpatterns are
1529 tested, and the overall condition is true if the condition is true for any
1530 one of them. This is the way Perl works, and is also more like the way
1531 testing by number works.
1532
1533
1534 Version 7.9 11-Apr-09
1535 ---------------------
1536
1537 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1538 (pcretest), all targets were linked against these libraries. This included
1539 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1540 libraries. This caused unwanted dependencies to be created. This problem
1541 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1542 pcretest is linked with readline.
1543
1544 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1545 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1546 moved outside it again, because FALSE and TRUE are already defined in AIX,
1547 but BOOL is not.
1548
1549 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1550 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1551
1552 4. The pcregrep documentation said spaces were inserted as well as colons (or
1553 hyphens) following file names and line numbers when outputting matching
1554 lines. This is not true; no spaces are inserted. I have also clarified the
1555 wording for the --colour (or --color) option.
1556
1557 5. In pcregrep, when --colour was used with -o, the list of matching strings
1558 was not coloured; this is different to GNU grep, so I have changed it to be
1559 the same.
1560
1561 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1562 each matching line was coloured. Now it goes on to look for further matches
1563 of any of the test patterns, which is the same behaviour as GNU grep.
1564
1565 7. A pattern that could match an empty string could cause pcregrep to loop; it
1566 doesn't make sense to accept an empty string match in pcregrep, so I have
1567 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1568 seems to be how GNU grep behaves.
1569
1570 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1571 start or after a newline", because the conditional assertion was not being
1572 correctly handled. The rule now is that both the assertion and what follows
1573 in the first alternative must satisfy the test.
1574
1575 9. If auto-callout was enabled in a pattern with a conditional group whose
1576 condition was an assertion, PCRE could crash during matching, both with
1577 pcre_exec() and pcre_dfa_exec().
1578
1579 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1580 used for matching.
1581
1582 11. Unicode property support in character classes was not working for
1583 characters (bytes) greater than 127 when not in UTF-8 mode.
1584
1585 12. Added the -M command line option to pcretest.
1586
1587 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1588
1589 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1590
1591 16. Added comments and documentation about mis-use of no_arg in the C++
1592 wrapper.
1593
1594 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1595 from Martin Jerabek that uses macro names for all relevant character and
1596 string constants.
1597
1598 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1599 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1600 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1601 these, but not everybody uses configure.
1602
1603 19. A conditional group that had only one branch was not being correctly
1604 recognized as an item that could match an empty string. This meant that an
1605 enclosing group might also not be so recognized, causing infinite looping
1606 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1607 with the subject "ab", where knowledge that the repeated group can match
1608 nothing is needed in order to break the loop.
1609
1610 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1611 exec(), but without supplying a callout function, matching went wrong.
1612
1613 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1614 leak if the size of the offset vector was greater than 30. When the vector
1615 is smaller, the saved offsets during recursion go onto a local stack
1616 vector, but for larger vectors malloc() is used. It was failing to free
1617 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1618 error, in fact).
1619
1620 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1621 heapframe that is used only when UTF-8 support is enabled. This caused no
1622 problem, but was untidy.
1623
1624 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1625 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1626 included within another project.
1627
1628 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1629 slightly modified by me:
1630
1631 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1632 not building pcregrep.
1633
1634 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1635 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1636
1637 25. Forward references, both numeric and by name, in patterns that made use of
1638 duplicate group numbers, could behave incorrectly or give incorrect errors,
1639 because when scanning forward to find the reference group, PCRE was not
1640 taking into account the duplicate group numbers. A pattern such as
1641 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1642
1643 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1644 the feature of a custom pointer more persuasive (as requested by a user).
1645
1646 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1647 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1648 pre-defined.
1649
1650 28. Added support for (*UTF8) at the start of a pattern.
1651
1652 29. Arrange for flags added by the "release type" setting in CMake to be shown
1653 in the configuration summary.
1654
1655
1656 Version 7.8 05-Sep-08
1657 ---------------------
1658
1659 1. Replaced UCP searching code with optimized version as implemented for Ad
1660 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1661 stage table and inline lookup instead of a function, giving speed ups of 2
1662 to 5 times on some simple patterns that I tested. Permission was given to
1663 distribute the MultiStage2.py script that generates the tables (it's not in
1664 the tarball, but is in the Subversion repository).
1665
1666 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1667 scripts.
1668
1669 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1670 a group with a zero qualifier. The result of the study could be incorrect,
1671 or the function might crash, depending on the pattern.
1672
1673 4. Caseless matching was not working for non-ASCII characters in back
1674 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1675 It now works when Unicode Property Support is available.
1676
1677 5. In pcretest, an escape such as \x{de} in the data was always generating
1678 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1679 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1680 truncation.
1681
1682 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1683
1684 7. Added two (int) casts to pcregrep when printing the difference of two
1685 pointers, in case they are 64-bit values.
1686
1687 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1688 test 2 if it fails.
1689
1690 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1691 and a #define of that name to empty if it is not externally set. This is to
1692 allow users of MSVC to set it if necessary.
1693
1694 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1695 the convenience functions in the pcre_get.c source file.
1696
1697 11. An option change at the start of a pattern that had top-level alternatives
1698 could cause overwriting and/or a crash. This command provoked a crash in
1699 some environments:
1700
1701 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1702
1703 This potential security problem was recorded as CVE-2008-2371.
1704
1705 12. For a pattern where the match had to start at the beginning or immediately
1706 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1707 pcre_dfa_exec() could read past the end of the passed subject if there was
1708 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1709 pcretest so that it places the subject at the end of its malloc-ed buffer.
1710
1711 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1712 exec() might read past the end of the data buffer in UTF-8 mode.
1713
1714 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1715 the data contained the byte 0x85 as part of a UTF-8 character within its
1716 first line. This applied both to normal and DFA matching.
1717
1718 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1719 /^[^d]*?$/8 failed to match "abc".
1720
1721 16. Added a missing copyright notice to pcrecpp_internal.h.
1722
1723 17. Make it more clear in the documentation that values returned from
1724 pcre_exec() in ovector are byte offsets, not character counts.
1725
1726 18. Tidied a few places to stop certain compilers from issuing warnings.
1727
1728 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1729 supplied by Stefan Weber. I made a further small update for 7.8 because
1730 there is a change of source arrangements: the pcre_searchfuncs.c module is
1731 replaced by pcre_ucd.c.
1732
1733
1734 Version 7.7 07-May-08
1735 ---------------------
1736
1737 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1738 a string to a long long, pretend we don't even have a long long." This is
1739 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1740
1741 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1742 pre-7.6 versions, which defined a global no_arg variable instead of putting
1743 it in the RE class. (See also #8 below.)
1744
1745 3. Remove a line of dead code, identified by coverity and reported by Nuno
1746 Lopes.
1747
1748 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1749
1750 (1) The include/exclude patterns were being applied to the whole pathnames
1751 of files, instead of just to the final components.
1752
1753 (2) If there was more than one level of directory, the subdirectories were
1754 skipped unless they satisfied the include/exclude conditions. This is
1755 inconsistent with GNU grep (and could even be seen as contrary to the
1756 pcregrep specification - which I improved to make it absolutely clear).
1757 The action now is always to scan all levels of directory, and just
1758 apply the include/exclude patterns to regular files.
1759
1760 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1761 --exclude_dir in the tests to avoid scanning .svn directories.
1762
1763 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1764 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1765 doesn't support NULs in patterns.
1766
1767 7. Added some missing "const"s to declarations of static tables in
1768 pcre_compile.c and pcre_dfa_exec.c.
1769
1770 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1771 caused by fix #2 above. (Subsequently also a second patch to fix the
1772 first patch. And a third patch - this was a messy problem.)
1773
1774 9. Applied Craig's patch to remove the use of push_back().
1775
1776 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1777 matching function regexec().
1778
1779 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1780 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1781 references. PCRE supports relative numbers with this syntax (I don't think
1782 Oniguruma does).
1783
1784 12. Previously, a group with a zero repeat such as (...){0} was completely
1785 omitted from the compiled regex. However, this means that if the group
1786 was called as a subroutine from elsewhere in the pattern, things went wrong
1787 (an internal error was given). Such groups are now left in the compiled
1788 pattern, with a new opcode that causes them to be skipped at execution
1789 time.
1790
1791 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1792 to the way PCRE behaves:
1793
1794 (a) A lone ] character is dis-allowed (Perl treats it as data).
1795
1796 (b) A back reference to an unmatched subpattern matches an empty string
1797 (Perl fails the current match path).
1798
1799 (c) A data ] in a character class must be notated as \] because if the
1800 first data character in a class is ], it defines an empty class. (In
1801 Perl it is not possible to have an empty class.) The empty class []
1802 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1803 The negative empty class [^] matches any one character, independently
1804 of the DOTALL setting.
1805
1806 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1807 non-existent subpattern following a character class starting with ']' and
1808 containing () gave an internal compiling error instead of "reference to
1809 non-existent subpattern". Fortunately, when the pattern did exist, the
1810 compiled code was correct. (When scanning forwards to check for the
1811 existencd of the subpattern, it was treating the data ']' as terminating
1812 the class, so got the count wrong. When actually compiling, the reference
1813 was subsequently set up correctly.)
1814
1815 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1816 it was being rejected as not supported by pcre_dfa_exec(), even though
1817 other assertions are supported. I have made pcre_dfa_exec() support
1818 (*FAIL).
1819
1820 16. The implementation of 13c above involved the invention of a new opcode,
1821 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1822 cannot be changed at match time, I realized I could make a small
1823 improvement to matching performance by compiling OP_ALLANY instead of
1824 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1825 on the OP_ANY path.
1826
1827 17. Compiling pcretest on Windows with readline support failed without the
1828 following two fixes: (1) Make the unistd.h include conditional on
1829 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1830
1831 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1832 ncurses library to be included for pcretest when ReadLine support is
1833 requested, but also to allow for it to be overridden. This patch came from
1834 Daniel Bergström.
1835
1836 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1837 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1838 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1839 spotting this.
1840
1841
1842 Version 7.6 28-Jan-08
1843 ---------------------
1844
1845 1. A character class containing a very large number of characters with
1846 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1847 overflow.
1848
1849 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1850 HAVE_LONG_LONG is not defined.
1851
1852 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1853 bring them up to date and include new features. This patch includes:
1854
1855 - Fixed PH's badly added libz and libbz2 support.
1856 - Fixed a problem with static linking.
1857 - Added pcredemo. [But later removed - see 7 below.]
1858 - Fixed dftables problem and added an option.
1859 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1860 HAVE_LONG_LONG.
1861 - Added readline support for pcretest.
1862 - Added an listing of the option settings after cmake has run.
1863
1864 4. A user submitted a patch to Makefile that makes it easy to create
1865 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1866 Makefile.am that cause it to include this special target, without
1867 affecting anything else. Note that the same mingw target plus all
1868 the other distribution libraries and programs are now supported
1869 when configuring with CMake (see 6 below) instead of with
1870 Configure/Make.
1871
1872 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1873 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1874 exported in the Windows port". It has not yet been confirmed that the patch
1875 solves the problem, but it does no harm.
1876
1877 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1878 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1879 with CMake, and also correct the comment about stack recursion.
1880
1881 7. Remove the automatic building of pcredemo from the ./configure system and
1882 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1883 of a program that users should build themselves after PCRE is installed, so
1884 building it automatically is not really right. What is more, it gave
1885 trouble in some build environments.
1886
1887 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1888
1889
1890 Version 7.5 10-Jan-08
1891 ---------------------
1892
1893 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1894 values in parens when parsing an RE using the C++ wrapper."
1895
1896 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1897 Characters greater than 255 were excluded from the class instead of being
1898 included.
1899
1900 3. The same bug as (2) above applied to negated POSIX classes such as
1901 [:^space:].
1902
1903 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1904 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1905 I have changed it.
1906
1907 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1908 first named subpattern) and a construct such as (?&a) would reference the
1909 first named subpattern whose name started with "a" (in other words, the
1910 length check was missing). Both these problems are fixed. "Subpattern name
1911 expected" is now given for (?&) (a zero-length name), and this patch also
1912 makes it give the same error for \k'' (previously it complained that that
1913 was a reference to a non-existent subpattern).
1914
1915 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1916 this is right because (?- can be followed by option settings as well as by
1917 digits. I have, however, made the messages clearer.
1918
1919 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1920 than the number used in the conditional) now cause a compile-time error.
1921 This is actually not compatible with Perl, which accepts such patterns, but
1922 treats the conditional as always being FALSE (as PCRE used to), but it
1923 seems to me that giving a diagnostic is better.
1924
1925 8. Change "alphameric" to the more common word "alphanumeric" in comments
1926 and messages.
1927
1928 9. Fix two occurrences of "backslash" in comments that should have been
1929 "backspace".
1930
1931 10. Remove two redundant lines of code that can never be obeyed (their function
1932 was moved elsewhere).
1933
1934 11. The program that makes PCRE's Unicode character property table had a bug
1935 which caused it to generate incorrect table entries for sequences of
1936 characters that have the same character type, but are in different scripts.
1937 It amalgamated them into a single range, with the script of the first of
1938 them. In other words, some characters were in the wrong script. There were
1939 thirteen such cases, affecting characters in the following ranges:
1940
1941 U+002b0 - U+002c1
1942 U+0060c - U+0060d
1943 U+0061e - U+00612
1944 U+0064b - U+0065e
1945 U+0074d - U+0076d
1946 U+01800 - U+01805
1947 U+01d00 - U+01d77
1948 U+01d9b - U+01dbf
1949 U+0200b - U+0200f
1950 U+030fc - U+030fe
1951 U+03260 - U+0327f
1952 U+0fb46 - U+0fbb1
1953 U+10450 - U+1049d
1954
1955 12. The -o option (show only the matching part of a line) for pcregrep was not
1956 compatible with GNU grep in that, if there was more than one match in a
1957 line, it showed only the first of them. It now behaves in the same way as
1958 GNU grep.
1959
1960 13. If the -o and -v options were combined for pcregrep, it printed a blank
1961 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1962 does the same. The return code can be used to tell if there were any
1963 non-matching lines.
1964
1965 14. Added --file-offsets and --line-offsets to pcregrep.
1966
1967 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1968 infinitely looping recursion. The bug was that positive lookaheads were not
1969 being skipped when checking for a possible empty match (negative lookaheads
1970 and both kinds of lookbehind were skipped).
1971
1972 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1973 inclusion of <windows.h> to before rather than after the definition of
1974 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1975
1976 17. Specifying a possessive quantifier with a specific limit for a Unicode
1977 character property caused pcre_compile() to compile bad code, which led at
1978 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1979 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1980 caused the error; without that there was no problem.
1981
1982 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1983
1984 19. Added --enable-pcretest-libreadline.
1985
1986 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1987 RE::GlobalReplace(). As a result, the number of replacements returned was
1988 double what it should be. I removed one of the increments, but Craig sent a
1989 later patch that removed the other one (the right fix) and added unit tests
1990 that check the return values (which was not done before).
1991
1992 21. Several CMake things:
1993
1994 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1995 the names libpcre and libpcreposix, not just pcre and pcreposix.
1996
1997 (2) The above change means that pcretest and pcregrep are now correctly
1998 linked with the newly-built libraries, not previously installed ones.
1999
2000 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2001
2002 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2003 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2004 UTF-8 newline character). The key issue is that the pattern starts .*;
2005 this means that the match must be either at the beginning, or after a
2006 newline. The bug was in the code for advancing after a failed match and
2007 checking that the new position followed a newline. It was not taking
2008 account of UTF-8 characters correctly.
2009
2010 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2011 character classes. PCRE was not treating the sequence [:...:] as a
2012 character class unless the ... were all letters. Perl, however, seems to
2013 allow any characters between [: and :], though of course it rejects as
2014 unknown any "names" that contain non-letters, because all the known class
2015 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2016 for example, whereas PCRE did not - it did not recognize a POSIX character
2017 class. This seemed a bit dangerous, so the code has been changed to be
2018 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2019 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2020 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2021 Perl does, and where it didn't before.
2022
2023 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2024 Windows environments %n is disabled by default.
2025
2026
2027 Version 7.4 21-Sep-07
2028 ---------------------
2029
2030 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2031 means that a class such as [\s] counted as "explicit reference to CR or
2032 LF". That isn't really right - the whole point of the change was to try to
2033 help when there was an actual mention of one of the two characters. So now
2034 the change happens only if \r or \n (or a literal CR or LF) character is
2035 encountered.
2036
2037 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2038 of both had grown to the point where there were only 3 bits left.
2039 Fortunately, there was spare space in the data structure, and so I have
2040 moved the internal flags into a new 16-bit field to free up more option
2041 bits.
2042
2043 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2044 but did not set the internal JCHANGED flag - either of these is enough to
2045 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2046 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2047 start sets both bits.
2048
2049 4. Added options (at build time, compile time, exec time) to change \R from
2050 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2051
2052 5. doc/pcresyntax.html was missing from the distribution.
2053
2054 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2055 compatibility, even though it is no longer used.
2056
2057 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2058 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2059 windows.h file is present (where different names are used). [This was
2060 reversed later after testing - see 16 below.]
2061
2062 8. Changed all #include <config.h> to #include "config.h". There were also
2063 some further <pcre.h> cases that I changed to "pcre.h".
2064
2065 9. When pcregrep was used with the --colour option, it missed the line ending
2066 sequence off the lines that it output.
2067
2068 10. It was pointed out to me that arrays of string pointers cause lots of
2069 relocations when a shared library is dynamically loaded. A technique of
2070 using a single long string with a table of offsets can drastically reduce
2071 these. I have refactored PCRE in four places to do this. The result is
2072 dramatic:
2073
2074 Originally: 290
2075 After changing UCP table: 187
2076 After changing error message table: 43
2077 After changing table of "verbs" 36
2078 After changing table of Posix names 22
2079
2080 Thanks to the folks working on Gregex for glib for this insight.
2081
2082 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2083 unicode-properties was also set.
2084
2085 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2086
2087 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2088 checked only for CRLF.
2089
2090 14. Added casts to pcretest.c to avoid compiler warnings.
2091
2092 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2093
2094 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2095 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2096 entirely. This removes changes made in 7 above.
2097
2098 17. The CMake files have been updated, and there is now more information about
2099 building with CMake in the NON-UNIX-USE document.
2100
2101
2102 Version 7.3 28-Aug-07
2103 ---------------------
2104
2105 1. In the rejigging of the build system that eventually resulted in 7.1, the
2106 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2107 brackets there is not right, since it causes compilers to look for an
2108 installed pcre.h, not the version that is in the source that is being
2109 compiled (which of course may be different). I have changed it back to:
2110
2111 #include "pcre.h"
2112
2113 I have a vague recollection that the change was concerned with compiling in
2114 different directories, but in the new build system, that is taken care of
2115 by the VPATH setting the Makefile.
2116
2117 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2118 when the subject happened to end in the byte 0x85 (e.g. if the last
2119 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2120 characters but of course it shouldn't be taken as a newline when it is part
2121 of another character. The bug was that, for an unlimited repeat of . in
2122 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2123 characters when looking for a newline.
2124
2125 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2126
2127 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2128 in debug output.
2129
2130 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2131 long printing in the pcrecpp unittest when running under MinGW.
2132
2133 6. ESC_K was left out of the EBCDIC table.
2134
2135 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2136 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2137 limit also applies to "virtual nesting" when a pattern is recursive, and in
2138 this case 1000 isn't so big. I have been able to remove this limit at the
2139 expense of backing off one optimization in certain circumstances. Normally,
2140 when pcre_exec() would call its internal match() function recursively and
2141 immediately return the result unconditionally, it uses a "tail recursion"
2142 feature to save stack. However, when a subpattern that can match an empty
2143 string has an unlimited repetition quantifier, it no longer makes this
2144 optimization. That gives it a stack frame in which to save the data for
2145 checking that an empty string has been matched. Previously this was taken
2146 from the 1000-entry workspace that had been reserved. So now there is no
2147 explicit limit, but more stack is used.
2148
2149 8. Applied Daniel's patches to solve problems with the import/export magic
2150 syntax that is required for Windows, and which was going wrong for the
2151 pcreposix and pcrecpp parts of the library. These were overlooked when this
2152 problem was solved for the main library.
2153
2154 9. There were some crude static tests to avoid integer overflow when computing
2155 the size of patterns that contain repeated groups with explicit upper
2156 limits. As the maximum quantifier is 65535, the maximum group length was
2157 set at 30,000 so that the product of these two numbers did not overflow a
2158 32-bit integer. However, it turns out that people want to use groups that
2159 are longer than 30,000 bytes (though not repeat them that many times).
2160 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2161 made it possible to implement the integer overflow checks in a much more
2162 dynamic way, which I have now done. The artificial limitation on group
2163 length has been removed - we now have only the limit on the total length of
2164 the compiled pattern, which depends on the LINK_SIZE setting.
2165
2166 10. Fixed a bug in the documentation for get/copy named substring when
2167 duplicate names are permitted. If none of the named substrings are set, the
2168 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2169 empty string.
2170
2171 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2172 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2173 because the ] is interpreted as the first data character and the
2174 terminating ] is not found. PCRE has been made compatible with Perl in this
2175 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2176 cause memory overwriting.
2177
2178 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2179 string has been matched (to stop an infinite loop). It was not recognizing
2180 a conditional subpattern that could match an empty string if that
2181 subpattern was within another subpattern. For example, it looped when
2182 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2183 condition was not nested. This bug has been fixed.
2184
2185 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2186 past the start of the subject in the presence of bytes with the top bit
2187 set, for example "\x8aBCD".
2188
2189 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2190 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2191
2192 14. Optimized (?!) to (*FAIL).
2193
2194 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2195 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2196 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2197 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2198 does: it's just the validity check that is more restrictive.
2199
2200 16. Inserted checks for integer overflows during escape sequence (backslash)
2201 processing, and also fixed erroneous offset values for syntax errors during
2202 backslash processing.
2203
2204 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2205 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2206
2207 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2208 caused an overrun.
2209
2210 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2211 something other than just ASCII characters) inside a group that had an
2212 unlimited repeat caused a loop at compile time (while checking to see
2213 whether the group could match an empty string).
2214
2215 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2216 [\P{Any}] did so. (Error in the code for printing property names.)
2217
2218 21. An orphan \E inside a character class could cause a crash.
2219
2220 22. A repeated capturing bracket such as (A)? could cause a wild memory
2221 reference during compilation.
2222
2223 23. There are several functions in pcre_compile() that scan along a compiled
2224 expression for various reasons (e.g. to see if it's fixed length for look
2225 behind). There were bugs in these functions when a repeated \p or \P was
2226 present in the pattern. These operators have additional parameters compared
2227 with \d, etc, and these were not being taken into account when moving along
2228 the compiled data. Specifically:
2229
2230 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2231 length.
2232
2233 (b) An item such as \pL+ within a repeated group could cause crashes or
2234 loops.
2235
2236 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2237 "reference to non-existent subpattern" error.
2238
2239 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2240
2241 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2242 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2243
2244 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2245
2246 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2247 character were causing crashes (broken optimization).
2248
2249 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2250 \p or \P) caused a compile-time loop.
2251
2252 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2253 break. For example, the unstudied pattern [\r\n]A does not match the string
2254 "\r\nA" because change 7.0/46 below moves the current point on by two
2255 characters after failing to match at the start. However, the pattern \nA
2256 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2257 the same is true. There doesn't seem any very clean way out of this, but
2258 what I have chosen to do makes the common cases work: PCRE now takes note
2259 of whether there can be an explicit match for \r or \n anywhere in the
2260 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2261 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2262 pattern has explicit CR or LF references.
2263
2264 29. Added (*CR) etc for changing newline setting at start of pattern.
2265
2266
2267 Version 7.2 19-Jun-07
2268 ---------------------
2269
2270 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2271 which is apparently normally available under Windows.
2272
2273 2. Re-jig the pcregrep tests with different newline settings in an attempt
2274 to make them independent of the local environment's newline setting.
2275
2276 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2277
2278 4. Some of the "internals" tests were previously cut out when the link size
2279 was not 2, because the output contained actual offsets. The recent new
2280 "Z" feature of pcretest means that these can be cut out, making the tests
2281 usable with all link sizes.
2282
2283 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2284 stack recursion. This gives a massive performance boost under BSD, but just
2285 a small improvement under Linux. However, it saves one field in the frame
2286 in all cases.
2287
2288 6. Added more features from the forthcoming Perl 5.10:
2289
2290 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2291 recursion call. It refers to the nth most recently opened parentheses.
2292
2293 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2294 to be opened parentheses.
2295
2296 (c) Conditions that refer to capturing parentheses can be specified
2297 relatively, for example, (?(-2)... or (?(+3)...
2298
2299 (d) \K resets the start of the current match so that everything before
2300 is not part of it.
2301
2302 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2303
2304 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2305 reference syntax.
2306
2307 (g) (?| introduces a group in which the numbering of parentheses in each
2308 alternative starts with the same number.
2309
2310 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2311
2312 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2313 PCRE_INFO_JCHANGED.
2314
2315 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2316 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2317 for detecting groups that can match an empty string.
2318
2319 9. A pattern with a very large number of alternatives (more than several
2320 hundred) was running out of internal workspace during the pre-compile
2321 phase, where pcre_compile() figures out how much memory will be needed. A
2322 bit of new cunning has reduced the workspace needed for groups with
2323 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2324 workspace instead of running out of the 4096 that are available.
2325
2326 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2327
2328 11. Applied patch from Google to remove an optimization that didn't quite work.
2329 The report of the bug said:
2330
2331 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2332 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2333 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2334
2335 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2336 it matched the wrong number of bytes.
2337
2338
2339 Version 7.1 24-Apr-07
2340 ---------------------
2341
2342 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2343 that is more "standard", making use of automake and other Autotools. There
2344 is some re-arrangement of the files and adjustment of comments consequent
2345 on this.
2346
2347 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2348 for recursive directory scanning broke on some systems because the files
2349 are not scanned in any specific order and on different systems the order
2350 was different. A call to "sort" has been inserted into RunGrepTest for the
2351 approprate test as a short-term fix. In the longer term there may be an
2352 alternative.
2353
2354 3. I had an email from Eric Raymond about problems translating some of PCRE's
2355 man pages to HTML (despite the fact that I distribute HTML pages, some
2356 people do their own conversions for various reasons). The problems
2357 concerned the use of low-level troff macros .br and .in. I have therefore
2358 removed all such uses from the man pages (some were redundant, some could
2359 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2360 HTML has been updated to handle .nf/.fi and to complain if it encounters
2361 .br or .in.
2362
2363 4. Updated comments in configure.ac that get placed in config.h.in and also
2364 arranged for config.h to be included in the distribution, with the name
2365 config.h.generic, for the benefit of those who have to compile without
2366 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2367
2368 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2369 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2370 makevp.bat for the current PCRE, using the additional files
2371 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2372
2373 6. A Windows user reported a minor discrepancy with test 2, which turned out
2374 to be caused by a trailing space on an input line that had got lost in his
2375 copy. The trailing space was an accident, so I've just removed it.
2376
2377 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2378 that is needed.
2379
2380 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2381 as "const" (a) because they are and (b) because it helps the PHP
2382 maintainers who have recently made a script to detect big data structures
2383 in the php code that should be moved to the .rodata section. I remembered
2384 to update Builducptable as well, so it won't revert if ucptable.h is ever
2385 re-created.
2386
2387 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2388 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2389 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2390 support is not required. This saves 1.5-2K of code, which is important in
2391 some applications.
2392
2393 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2394 so as not to refer to the tables, even though these functions will never be
2395 called when UTF-8 support is disabled. Otherwise there are problems with a
2396 shared library.
2397
2398 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2399
2400 (a) It was defining its arguments as char * instead of void *.
2401
2402 (b) It was assuming that all moves were upwards in memory; this was true
2403 a long time ago when I wrote it, but is no longer the case.
2404
2405 The emulated memove() is provided for those environments that have neither
2406 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2407 is clearly not the case, as these two bugs were recently reported.
2408
2409 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2410 and Detrail to create the HTML documentation, the .txt form of the man
2411 pages, and it removes trailing spaces from listed files. It also creates
2412 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2413 case, it wraps all the #defines with #ifndefs. This script should be run
2414 before "make dist".
2415
2416 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2417 with Unicode property support.
2418
2419 (a) For a maximizing quantifier, if the two different cases of the
2420 character were of different lengths in their UTF-8 codings (there are
2421 some cases like this - I found 11), and the matching function had to
2422 back up over a mixture of the two cases, it incorrectly assumed they
2423 were both the same length.
2424
2425 (b) When PCRE was configured to use the heap rather than the stack for
2426 recursion during matching, it was not correctly preserving the data for
2427 the other case of a UTF-8 character when checking ahead for a match
2428 while processing a minimizing repeat. If the check also involved
2429 matching a wide character, but failed, corruption could cause an
2430 erroneous result when trying to check for a repeat of the original
2431 character.
2432
2433 13. Some tidying changes to the testing mechanism:
2434
2435 (a) The RunTest script now detects the internal link size and whether there
2436 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2437 values substituted by "configure". (The RunGrepTest script already did
2438 this for UTF-8.) The configure.ac script no longer substitutes the
2439 relevant variables.
2440
2441 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2442 with length and offset values. This means that the output is different
2443 for different internal link sizes. Test 2 is skipped for link sizes
2444 other than 2 because of this, bypassing the problem. Unfortunately,
2445 there was also a test in test 3 (the locale tests) that used /B and
2446 failed for link sizes other than 2. Rather than cut the whole test out,
2447 I have added a new /Z option to pcretest that replaces the length and
2448 offset values with spaces. This is now used to make test 3 independent
2449 of link size. (Test 2 will be tidied up later.)
2450
2451 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2452 segmentation fault instead of returning the appropriate error message.
2453
2454 15. In multiline mode when the newline sequence was set to "any", the pattern
2455 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2456 This doesn't seem right; it now treats the CRLF combination as the line
2457 ending, and so does not match in that case. It's only a pattern such as ^$
2458 that would hit this one: something like ^ABC$ would have failed after \r
2459 and then tried again after \r\n.
2460
2461 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2462 in an attempt to make files that differ only in their line terminators
2463 compare equal. This works on Linux.
2464
2465 17. Under certain error circumstances pcregrep might try to free random memory
2466 as it exited. This is now fixed, thanks to valgrind.
2467
2468 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2469 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2470 was because its rules for how to advance for /g after matching an empty
2471 string at the end of a line did not allow for this case. They now check for
2472 it specially.
2473
2474 20. pcretest is supposed to handle patterns and data of any length, by
2475 extending its buffers when necessary. It was getting this wrong when the
2476 buffer for a data line had to be extended.
2477
2478 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2479 CRLF as a newline sequence.
2480
2481 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2482 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2483 I have nevertheless tidied it up.
2484
2485 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2486
2487 24. Added a man page for pcre-config.
2488
2489
2490 Version 7.0 19-Dec-06
2491 ---------------------
2492
2493 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2494 moving to gcc 4.1.1.
2495
2496 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2497 sys/time.h, which is documented as needed for this function. It doesn't
2498 seem to matter on Linux, but it showed up on some releases of OS X.
2499
2500 3. It seems that there are systems where bytes whose values are greater than
2501 127 match isprint() in the "C" locale. The "C" locale should be the
2502 default when a C program starts up. In most systems, only ASCII printing
2503 characters match isprint(). This difference caused the output from pcretest
2504 to vary, making some of the tests fail. I have changed pcretest so that:
2505
2506 (a) When it is outputting text in the compiled version of a pattern, bytes
2507 other than 32-126 are always shown as hex escapes.
2508
2509 (b) When it is outputting text that is a matched part of a subject string,
2510 it does the same, unless a different locale has been set for the match
2511 (using the /L modifier). In this case, it uses isprint() to decide.
2512
2513 4. Fixed a major bug that caused incorrect computation of the amount of memory
2514 required for a compiled pattern when options that changed within the
2515 pattern affected the logic of the preliminary scan that determines the
2516 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2517 that the computed length was too small. The symptoms of this bug were
2518 either the PCRE error "internal error: code overflow" from pcre_compile(),
2519 or a glibc crash with a message such as "pcretest: free(): invalid next
2520 size (fast)". Examples of patterns that provoked this bug (shown in
2521 pcretest format) are:
2522
2523 /(?-x: )/x
2524 /(?x)(?-x: \s*#\s*)/
2525 /((?i)[\x{c0}])/8
2526 /(?i:[\x{c0}])/8
2527
2528 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2529 is now done differently.
2530
2531 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2532 wrapper classes; (b) implement a new function in the C++ scanner that is
2533 more efficient than the old way of doing things because it avoids levels of
2534 recursion in the regex matching; (c) add a paragraph to the documentation
2535 for the FullMatch() function.
2536
2537 6. The escape sequence \n was being treated as whatever was defined as
2538 "newline". Not only was this contrary to the documentation, which states
2539 that \n is character 10 (hex 0A), but it also went horribly wrong when
2540 "newline" was defined as CRLF. This has been fixed.
2541
2542 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2543 was being set to -1 for the "end of line" case (supposedly a value that no
2544 character can have). Though this value is never used (the check for end of
2545 line is "zero bytes in current character"), it caused compiler complaints.
2546 I've changed it to 0xffffffff.
2547
2548 8. In pcre_version.c, the version string was being built by a sequence of
2549 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2550 string (as it is for production releases) called a macro with an empty
2551 argument. The C standard says the result of this is undefined. The gcc
2552 compiler treats it as an empty string (which was what was wanted) but it is
2553 reported that Visual C gives an error. The source has been hacked around to
2554 avoid this problem.
2555
2556 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2557 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2558 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2559 of them did).
2560
2561 10. Originally, pcretest opened its input and output without "b"; then I was
2562 told that "b" was needed in some environments, so it was added for release
2563 5.0 to both the input and output. (It makes no difference on Unix-like
2564 systems.) Later I was told that it is wrong for the input on Windows. I've
2565 now abstracted the modes into two macros, to make it easier to fiddle with
2566 them, and removed "b" from the input mode under Windows.
2567
2568 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2569
2570 12. Added -help and --help to pcretest as an official way of being reminded
2571 of the options.
2572
2573 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2574 and pcrecpp.cc because they annoy compilers at high warning levels.
2575
2576 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2577
2578 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2579 scripts are not C programs :-) and which was not noticed because it works
2580 on Linux.
2581
2582 16. pcretest is supposed to handle any length of pattern and data line (as one
2583 line or as a continued sequence of lines) by extending its input buffer if
2584 necessary. This feature was broken for very long pattern lines, leading to
2585 a string of junk being passed to pcre_compile() if the pattern was longer
2586 than about 50K.
2587
2588 17. I have done a major re-factoring of the way pcre_compile() computes the
2589 amount of memory needed for a compiled pattern. Previously, there was code
2590 that made a preliminary scan of the pattern in order to do this. That was
2591 OK when PCRE was new, but as the facilities have expanded, it has become
2592 harder and harder to keep it in step with the real compile phase, and there
2593 have been a number of bugs (see for example, 4 above). I have now found a
2594 cunning way of running the real compile function in a "fake" mode that
2595 enables it to compute how much memory it would need, while actually only
2596 ever using a few hundred bytes of working memory and without too many
2597 tests of the mode. This should make future maintenance and development
2598 easier. A side effect of this work is that the limit of 200 on the nesting
2599 depth of parentheses has been removed (though this was never a serious
2600 limitation, I suspect). However, there is a downside: pcre_compile() now
2601 runs more slowly than before (30% or more, depending on the pattern). I
2602 hope this isn't a big issue. There is no effect on runtime performance.
2603
2604 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2605 newline (only possible for the last line of a file) and it was a
2606 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2607
2608 19. Added additional timing features to pcretest. (1) The -tm option now times
2609 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2610 separate command line item, by a number that specifies the number of
2611 repeats to use when timing. The default is 50000; this gives better
2612 precision, but takes uncomfortably long for very large patterns.
2613
2614 20. Extended pcre_study() to be more clever in cases where a branch of a
2615 subpattern has no definite first character. For example, (a*|b*)[cd] would
2616 previously give no result from pcre_study(). Now it recognizes that the
2617 first character must be a, b, c, or d.
2618
2619 21. There was an incorrect error "recursive call could loop indefinitely" if
2620 a subpattern (or the entire pattern) that was being tested for matching an
2621 empty string contained only one non-empty item after a nested subpattern.
2622 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2623 incorrectly, because the \d was being skipped in the check.
2624
2625 22. The pcretest program now has a new pattern option /B and a command line
2626 option -b, which is equivalent to adding /B to every pattern. This causes
2627 it to show the compiled bytecode, without the additional information that
2628 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2629 is the same as /B/I).
2630
2631 23. A new optimization is now able automatically to treat some sequences such
2632 as a*b as a*+b. More specifically, if something simple (such as a character
2633 or a simple class like \d) has an unlimited quantifier, and is followed by
2634 something that cannot possibly match the quantified thing, the quantifier
2635 is automatically "possessified".
2636
2637 24. A recursive reference to a subpattern whose number was greater than 39
2638 went wrong under certain circumstances in UTF-8 mode. This bug could also
2639 have affected the operation of pcre_study().
2640
2641 25. Realized that a little bit of performance could be had by replacing
2642 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2643
2644 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2645
2646 27. Possessive quantifiers such as a++ were previously implemented by turning
2647 them into atomic groups such as ($>a+). Now they have their own opcodes,
2648 which improves performance. This includes the automatically created ones
2649 from 23 above.
2650
2651 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2652 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2653 the first matched character to be a colon. This applied both to named and
2654 numbered groups.
2655
2656 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2657
2658 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2659 building PCRE on Windows easier, so I have included it in the distribution.
2660
2661 31. There is now a check in pcretest against a ridiculously large number being
2662 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2663 loop, the loop is abandoned.
2664
2665 32. Forward references to subpatterns in conditions such as (?(2)...) where
2666 subpattern 2 is defined later cause pcre_compile() to search forwards in
2667 the pattern for the relevant set of parentheses. This search went wrong
2668 when there were unescaped parentheses in a character class, parentheses
2669 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2670
2671 33. "Subroutine" calls and backreferences were previously restricted to
2672 referencing subpatterns earlier in the regex. This restriction has now
2673 been removed.
2674
2675 34. Added a number of extra features that are going to be in Perl 5.10. On the
2676 whole, these are just syntactic alternatives for features that PCRE had
2677 previously implemented using the Python syntax or my own invention. The
2678 other formats are all retained for compatibility.
2679
2680 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2681 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2682 also .NET compatible.
2683
2684 (b) A recursion or subroutine call to a named group can now be defined as
2685 (?&name) as well as (?P>name).
2686
2687 (c) A backreference to a named group can now be defined as \k<name> or
2688 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2689 5.10, are also .NET compatible.
2690
2691 (d) A conditional reference to a named group can now use the syntax
2692 (?(<name>) or (?('name') as well as (?(name).
2693
2694 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2695 groups (named and numbered) that are never evaluated inline, but can be
2696 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2697 is always false. There may be only one alternative in such a group.
2698
2699 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2700 as the simple (?(R). The condition is true only if the most recent
2701 recursion is that of the given number or name. It does not search out
2702 through the entire recursion stack.
2703
2704 (g) The escape \gN or \g{N} has been added, where N is a positive or
2705 negative number, specifying an absolute or relative reference.
2706
2707 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2708 some "unreachable code" warnings.
2709
2710 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2711 things, this adds five new scripts.
2712
2713 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2714 There were also incompatibilities regarding the handling of \Q..\E inside
2715 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2716 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2717
2718 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2719 matches an empty string, and forcibly breaks the loop. There were bugs in
2720 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2721 against aaaa the result was just "a" rather than "aaaa", for example. Two
2722 separate and independent bugs (that affected different cases) have been
2723 fixed.
2724
2725 39. Refactored the code to abolish the use of different opcodes for small
2726 capturing bracket numbers. This is a tidy that I avoided doing when I
2727 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2728 The new approach is not only tidier, it makes it possible to reduce the
2729 memory needed to fix the previous bug (38).
2730
2731 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2732 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2733 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2734 mode.
2735
2736 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2737 report.
2738
2739 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2740 copy construction and assignment in the C++ wrapper.
2741
2742 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2743 couple of bugs that could have given wrong results in the "--newline=crlf"
2744 case.
2745
2746 44. Added a number of casts and did some reorganization of signed/unsigned int
2747 variables following suggestions from Dair Grant. Also renamed the variable
2748 "this" as "item" because it is a C++ keyword.
2749
2750 45. Arranged for dftables to add
2751
2752 #include "pcre_internal.h"
2753
2754 to pcre_chartables.c because without it, gcc 4.x may remove the array
2755 definition from the final binary if PCRE is built into a static library and
2756 dead code stripping is activated.
2757
2758 46. For an unanchored pattern, if a match attempt fails at the start of a
2759 newline sequence, and the newline setting is CRLF or ANY, and the next two
2760 characters are CRLF, advance by two characters instead of one.
2761
2762
2763 Version 6.7 04-Jul-06
2764 ---------------------
2765
2766 1. In order to handle tests when input lines are enormously long, pcretest has
2767 been re-factored so that it automatically extends its buffers when
2768 necessary. The code is crude, but this _is_ just a test program. The
2769 default size has been increased from 32K to 50K.
2770
2771 2. The code in pcre_study() was using the value of the re argument before
2772 testing it for NULL. (Of course, in any sensible call of the function, it
2773 won't be NULL.)
2774
2775 3. The memmove() emulation function in pcre_internal.h, which is used on
2776 systems that lack both memmove() and bcopy() - that is, hardly ever -
2777 was missing a "static" storage class specifier.
2778
2779 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2780 containing an extended class (one that cannot be represented by a bitmap
2781 because it contains high-valued characters or Unicode property items, e.g.
2782 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2783 pattern, but PCRE should not loop if you do not (it no longer does).
2784 [Detail: two cases were found: (a) a repeated subpattern containing an
2785 extended class; (b) a recursive reference to a subpattern that followed a
2786 previous extended class. It wasn't skipping over the extended class
2787 correctly when UTF-8 mode was not set.]
2788
2789 5. A negated single-character class was not being recognized as fixed-length
2790 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2791 compile error "lookbehind assertion is not fixed length".
2792
2793 6. The RunPerlTest auxiliary script was showing an unexpected difference
2794 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2795 write a Perl script that can interpret lines of an input file either as
2796 byte characters or as UTF-8, which is what "perltest" was being required to
2797 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2798 can't do is switch easily at run time between having the "use utf8;" pragma
2799 or not. In the end, I fudged it by using the RunPerlTest script to insert
2800 "use utf8;" explicitly for the UTF-8 tests.
2801
2802 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2803 the end of the subject string, contrary to the documentation and to what
2804 Perl does. This was true of both matching functions. Now it matches only at
2805 the start of the subject and immediately after *internal* newlines.
2806
2807 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2808 a pointer to an int instead of a pointer to an unsigned long int. This
2809 caused problems on 64-bit systems.
2810
2811 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2812 instance of the 'standard' template library not being so standard".
2813
2814 10. There was no check on the number of named subpatterns nor the maximum
2815 length of a subpattern name. The product of these values is used to compute
2816 the size of the memory block for a compiled pattern. By supplying a very
2817 long subpattern name and a large number of named subpatterns, the size
2818 computation could be caused to overflow. This is now prevented by limiting
2819 the length of names to 32 characters, and the number of named subpatterns
2820 to 10,000.
2821
2822 11. Subpatterns that are repeated with specific counts have to be replicated in
2823 the compiled pattern. The size of memory for this was computed from the
2824 length of the subpattern and the repeat count. The latter is limited to
2825 65535, but there was no limit on the former, meaning that integer overflow
2826 could in principle occur. The compiled length of a repeated subpattern is
2827 now limited to 30,000 bytes in order to prevent this.
2828
2829 12. Added the optional facility to have named substrings with the same name.
2830
2831 13. Added the ability to use a named substring as a condition, using the
2832 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2833 are numbers (not recommended). Forward references are permitted.
2834
2835 14. Added forward references in named backreferences (if you see what I mean).
2836
2837 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2838 pattern could run off the end of the subject. For example, the pattern
2839 "(?s)(.{1,5})"8 did this with the subject "ab".
2840
2841 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2842 PCRE_CASELESS was set when matching characters that were quantified with ?
2843 or *.
2844
2845 17. A character class other than a single negated character that had a minimum
2846 but no maximum quantifier - for example [ab]{6,} - was not handled
2847 correctly by pce_dfa_exec(). It would match only one character.
2848
2849 18. A valid (though odd) pattern that looked like a POSIX character
2850 class but used an invalid character after [ (for example [[,abc,]]) caused
2851 pcre_compile() to give the error "Failed: internal error: code overflow" or
2852 in some cases to crash with a glibc free() error. This could even happen if
2853 the pattern terminated after [[ but there just happened to be a sequence of
2854 letters, a binary zero, and a closing ] in the memory that followed.
2855
2856 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2857 over the years. Originally (before any Unicode support), just the bottom 8
2858 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2859 output from "man perlunicode" includes this:
2860
2861 The regular expression compiler produces polymorphic opcodes. That
2862 is, the pattern adapts to the data and automatically switches to
2863 the Unicode character scheme when presented with Unicode data--or
2864 instead uses a traditional byte scheme when presented with byte
2865 data.
2866
2867 Sadly, a wide octal escape does not cause a switch, and in a string with
2868 no other multibyte characters, these octal escapes are treated as before.
2869 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2870 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2871 Unicode string.
2872
2873 I have not perpetrated such confusion in PCRE. Up till now, it took just
2874 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2875 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2876 translate to the appropriate multibyte character.
2877
2878 29. Applied some refactoring to reduce the number of warnings from Microsoft
2879 and Borland compilers. This has included removing the fudge introduced
2880 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2881 a warning about an unused variable.
2882
2883 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2884 characters since release 4.0, because Perl (from release 5.004) does not.
2885 [Or at least, is documented not to: some releases seem to be in conflict
2886 with the documentation.] However, when a pattern was studied with
2887 pcre_study() and all its branches started with \s, PCRE still included VT
2888 as a possible starting character. Of course, this did no harm; it just
2889 caused an unnecessary match attempt.
2890
2891 22. Removed a now-redundant internal flag bit that recorded the fact that case
2892 dependency changed within the pattern. This was once needed for "required
2893 byte" processing, but is no longer used. This recovers a now-scarce options
2894 bit. Also moved the least significant internal flag bit to the most-
2895 significant bit of the word, which was not previously used (hangover from
2896 the days when it was an int rather than a uint) to free up another bit for
2897 the future.
2898
2899 23. Added support for CRLF line endings as well as CR and LF. As well as the
2900 default being selectable at build time, it can now be changed at runtime
2901 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2902 specify that it is scanning data with non-default line endings.
2903
2904 24. Changed the definition of CXXLINK to make it agree with the definition of
2905 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2906
2907 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2908 recursions. This makes a big different to stack usage for some patterns.
2909
2910 26. If a subpattern containing a named recursion or subroutine reference such
2911 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2912 the space required for the compiled pattern went wrong and gave too small a
2913 value. Depending on the environment, this could lead to "Failed: internal
2914 error: code overflow at offset 49" or "glibc detected double free or
2915 corruption" errors.
2916
2917 27. Applied patches from Google (a) to support the new newline modes and (b) to
2918 advance over multibyte UTF-8 characters in GlobalReplace.
2919
2920 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2921 difference for some implementation of PCRE in some Windows version.
2922
2923 29. Added some extra testing facilities to pcretest:
2924
2925 \q<number> in a data line sets the "match limit" value
2926 \Q<number> in a data line sets the "match recursion limt" value
2927 -S <number> sets the stack size, where <number> is in megabytes
2928
2929 The -S option isn't available for Windows.
2930
2931
2932 Version 6.6 06-Feb-06
2933 ---------------------
2934
2935 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2936 in pcreposix.h. I have copied the definition from pcre.h.
2937
2938 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2939 because pcre.h is no longer a built file.
2940
2941 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2942 not normally included in the compiled code.
2943
2944
2945 Version 6.5 01-Feb-06
2946 ---------------------
2947
2948 1. When using the partial match feature with pcre_dfa_exec(), it was not
2949 anchoring the second and subsequent partial matches at the new starting
2950 point. This could lead to incorrect results. For example, with the pattern
2951 /1234/, partially matching against "123" and then "a4" gave a match.
2952
2953 2. Changes to pcregrep:
2954
2955 (a) All non-match returns from pcre_exec() were being treated as failures
2956 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2957 error message is output. Some extra information is given for the
2958 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2959 probably the only errors that are likely to be caused by users (by
2960 specifying a regex that has nested indefinite repeats, for instance).
2961 If there are more than 20 of these errors, pcregrep is abandoned.
2962
2963 (b) A binary zero was treated as data while matching, but terminated the
2964 output line if it was written out. This has been fixed: binary zeroes
2965 are now no different to any other data bytes.
2966
2967 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2968 used to set a locale for matching. The --locale=xxxx long option has
2969 been added (no short equivalent) to specify a locale explicitly on the
2970 pcregrep command, overriding the environment variables.
2971
2972 (d) When -B was used with -n, some line numbers in the output were one less
2973 than they should have been.
2974
2975 (e) Added the -o (--only-matching) option.
2976
2977 (f) If -A or -C was used with -c (count only), some lines of context were
2978 accidentally printed for the final match.
2979
2980 (g) Added the -H (--with-filename) option.
2981
2982 (h) The combination of options -rh failed to suppress file names for files
2983 that were found from directory arguments.
2984
2985 (i) Added the -D (--devices) and -d (--directories) options.
2986
2987 (j) Added the -F (--fixed-strings) option.
2988
2989 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2990
2991 (l) Added the --colo(u)r option.
2992
2993 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2994 is not present by default.
2995
2996 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2997 items such as (?R) or (?1), when the recursion could match a number of
2998 alternatives. If it matched one of the alternatives, but subsequently,
2999 outside the recursion, there was a failure, the code tried to back up into
3000 the recursion. However, because of the way PCRE is implemented, this is not
3001 possible, and the result was an incorrect result from the match.
3002
3003 In order to prevent this happening, the specification of recursion has
3004 been changed so that all such subpatterns are automatically treated as
3005 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3006
3007 4. I had overlooked the fact that, in some locales, there are characters for
3008 which isalpha() is true but neither isupper() nor islower() are true. In
3009 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3010 and ordfeminine) are like this. This affected the treatment of \w and \W
3011 when they appeared in character classes, but not when they appeared outside
3012 a character class. The bit map for "word" characters is now created
3013 separately from the results of isalnum() instead of just taking it from the
3014 upper, lower, and digit maps. (Plus the underscore character, of course.)
3015
3016 5. The above bug also affected the handling of POSIX character classes such as
3017 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3018 permanent tables. Instead, the bit maps for such a class were previously
3019 created as the appropriate unions of the upper, lower, and digit bitmaps.
3020 Now they are created by subtraction from the [[:word:]] class, which has
3021 its own bitmap.
3022
3023 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3024 It is created by subtracting the vertical space characters (\x09, \x0a,
3025 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3026 subtraction was done in the overall bitmap for a character class, meaning
3027 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3028 be recognized. This bug has been fixed.
3029
3030 7. Patches from the folks at Google:
3031
3032 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3033 real life, but is still worth protecting against".
3034
3035 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3036 regular expressions".
3037
3038 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3039 have it.
3040
3041 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3042 "configure" and the latter not, in order to fix a problem somebody had
3043 with compiling the Arg class on HP-UX.
3044
3045 (e) Improve the error-handling of the C++ wrapper a little bit.
3046
3047 (f) New tests for checking recursion limiting.
3048
3049 8. The pcre_memmove() function, which is used only if the environment does not
3050 have a standard memmove() function (and is therefore rarely compiled),
3051 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3052 returning a result (though PCRE never actually uses the result).
3053
3054 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3055 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3056 returned instead of calling malloc() with an overflowing number that would
3057 most likely cause subsequent chaos.
3058
3059 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3060
3061 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3062 with this option is matched, the nmatch and pmatch options of regexec() are
3063 ignored.
3064
3065 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3066 provided in case anyone wants to the the POSIX interface with UTF-8
3067 strings.
3068
3069 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3070 C++ linking (needed for some HP-UX environments).
3071
3072 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3073 (unused parameter) and in the pcre_printint() function (omitted "default"
3074 switch label when the default is to do nothing).
3075
3076 15. Added some code to make it possible, when PCRE is compiled as a C++
3077 library, to replace subject pointers for pcre_exec() with a smart pointer
3078 class, thus making it possible to process discontinuous strings.
3079
3080 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3081 much the same function. They were added by different people who were trying
3082 to make PCRE easy to compile on non-Unix systems. It has been suggested
3083 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3084 for compiling on Windows systems. I have therefore replaced it with
3085 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3086 defaults to "extern" for C or "extern C" for C++, which works fine on
3087 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3088 SCOPE with something explicit in config.h. In addition:
3089
3090 (a) pcreposix.h still had just "extern" instead of either of these macros;
3091 I have replaced it with PCRE_DATA_SCOPE.
3092
3093 (b) Functions such as _pcre_xclass(), which are internal to the library,
3094 but external in the C sense, all had PCRE_EXPORT in their definitions.
3095 This is apparently wrong for the Windows case, so I have removed it.
3096 (It makes no difference on Unix-like systems.)
3097
3098 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3099 of recursive calls to match(). This is different to MATCH_LIMIT because
3100 that limits the total number of calls to match(), not all of which increase
3101 the depth of recursion. Limiting the recursion depth limits the amount of
3102 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3103 when PCRE is compiled, and changed at run time. A patch from Google adds
3104 this functionality to the C++ interface.
3105
3106 18. Changes to the handling of Unicode character properties:
3107
3108 (a) Updated the table to Unicode 4.1.0.
3109
3110 (b) Recognize characters that are not in the table as "Cn" (undefined).
3111
3112 (c) I revised the way the table is implemented to a much improved format
3113 which includes recognition of ranges. It now supports the ranges that
3114 are defined in UnicodeData.txt, and it also amalgamates other
3115 characters into ranges. This has reduced the number of entries in the
3116 table from around 16,000 to around 3,000, thus reducing its size
3117 considerably. I realized I did not need to use a tree structure after
3118 all - a binary chop search is just as efficient. Having reduced the
3119 number of entries, I extended their size from 6 bytes to 8 bytes to
3120 allow for more data.
3121
3122 (d) Added support for Unicode script names via properties such as \p{Han}.
3123
3124 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3125 matching that character.
3126
3127 20. When matching a repeated Unicode property with a minimum greater than zero,
3128 (for example \pL{2,}), PCRE could look past the end of the subject if it
3129 reached it while seeking the minimum number of characters. This could
3130 happen only if some of the characters were more than one byte long, because
3131 there is a check for at least the minimum number of bytes.
3132
3133 21. Refactored the implementation of \p and \P so as to be more general, to
3134 allow for more different types of property in future. This has changed the
3135 compiled form incompatibly. Anybody with saved compiled patterns that use
3136 \p or \P will have to recompile them.
3137
3138 22. Added "Any" and "L&" to the supported property types.
3139
3140 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3141 but give a compile time error if the value is greater than 0xff.
3142
3143 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3144 accidentally not being installed or uninstalled.
3145
3146 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3147 made were to insert the current release number. This seemed silly, because
3148 it made things harder for people building PCRE on systems that don't run
3149 "configure". I have turned pcre.h into a distributed file, no longer built
3150 by "configure", with the version identification directly included. There is
3151 no longer a pcre.h.in file.
3152
3153 However, this change necessitated a change to the pcre-config script as
3154 well. It is built from pcre-config.in, and one of the substitutions was the
3155 release number. I have updated configure.ac so that ./configure now finds
3156 the release number by grepping pcre.h.
3157
3158 26. Added the ability to run the tests under valgrind.
3159
3160
3161 Version 6.4 05-Sep-05
3162 ---------------------
3163
3164 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3165 "--" to be printed when multiple files were scanned, even when none of the
3166 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3167 consider it to be a bug, and have restored the previous behaviour.
3168
3169 2. A couple of code tidies to get rid of compiler warnings.
3170
3171 3. The pcretest program used to cheat by referring to symbols in the library
3172 whose names begin with _pcre_. These are internal symbols that are not
3173 really supposed to be visible externally, and in some environments it is
3174 possible to suppress them. The cheating is now confined to including
3175 certain files from the library's source, which is a bit cleaner.
3176
3177 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3178 file's purpose clearer.
3179
3180 5. Reorganized pcre_ucp_findchar().
3181
3182
3183 Version 6.3 15-Aug-05
3184 ---------------------
3185
3186 1. The file libpcre.pc.in did not have general read permission in the tarball.
3187
3188 2. There were some problems when building without C++ support:
3189
3190 (a) If C++ support was not built, "make install" and "make test" still
3191 tried to test it.
3192
3193 (b) There were problems when the value of CXX was explicitly set. Some
3194 changes have been made to try to fix these, and ...
3195
3196 (c) --disable-cpp can now be used to explicitly disable C++ support.
3197
3198 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3199 backslash in a target when C++ was disabled. This confuses some
3200 versions of "make", apparently. Using an intermediate variable solves
3201 this. (Same for CPP_LOBJ.)
3202
3203 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3204 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3205 necessary on certain architectures.
3206
3207 4. Added a setting of -export-symbols-regex to the link command to remove
3208 those symbols that are exported in the C sense, but actually are local
3209 within the library, and not documented. Their names all begin with
3210 "_pcre_". This is not a perfect job, because (a) we have to except some
3211 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3212 available (and never for static libraries). I have made a note to try to
3213 find a way round (a) in the future.
3214
3215
3216 Version 6.2 01-Aug-05
3217 ---------------------
3218
3219 1. There was no test for integer overflow of quantifier values. A construction
3220 such as {1111111111111111} would give undefined results. What is worse, if
3221 a minimum quantifier for a parenthesized subpattern overflowed and became
3222 negative, the calculation of the memory size went wrong. This could have
3223 led to memory overwriting.
3224
3225 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3226
3227 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3228 operating environments where this matters.
3229
3230 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3231 PCRE options from within the C++ wrapper.
3232
3233 5. Named capturing subpatterns were not being correctly counted when a pattern
3234 was compiled. This caused two problems: (a) If there were more than 100
3235 such subpatterns, the calculation of the memory needed for the whole
3236 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3237 back references of the form \12, where the number was greater than 9, were
3238 not recognized as back references, even though there were sufficient
3239 previous subpatterns.
3240
3241 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3242 versions of gcc, e.g. 2.95.4.
3243
3244
3245 Version 6.1 21-Jun-05
3246 ---------------------
3247
3248 1. There was one reference to the variable "posix" in pcretest.c that was not
3249 surrounded by "#if !defined NOPOSIX".
3250
3251 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3252 the cross-check on the old pcre_info() function, for the benefit of the
3253 cut-down version of PCRE that is currently imported into Exim.
3254
3255 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3256 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3257 patterns that start (?i) but I don't think that matters. The use of (?i) is
3258 just an example; this all applies to the other options as well.
3259
3260 4. Since libtool seems to echo the compile commands it is issuing, the output
3261 from "make" can be reduced a bit by putting "@" in front of each libtool
3262 compile command.
3263
3264 5. Patch from the folks at Google for configure.in to be a bit more thorough
3265 in checking for a suitable C++ installation before trying to compile the
3266 C++ stuff. This should fix a reported problem when a compiler was present,
3267 but no suitable headers.
3268
3269 6. The man pages all had just "PCRE" as their title. I have changed them to
3270 be the relevant file name. I have also arranged that these names are
3271 retained in the file doc/pcre.txt, which is a concatenation in text format
3272 of all the man pages except the little individual ones for each function.
3273
3274 7. The NON-UNIX-USE file had not been updated for the different set of source
3275 files that come with release 6. I also added a few comments about the C++
3276 wrapper.
3277
3278
3279 Version 6.0 07-Jun-05
3280 ---------------------
3281
3282 1. Some minor internal re-organization to help with my DFA experiments.
3283
3284 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3285 didn't matter for the library itself when fully configured, but did matter
3286 when compiling without UCP support, or within Exim, where the ucp files are
3287 not imported.
3288
3289 3. Refactoring of the library code to split up the various functions into
3290 different source modules. The addition of the new DFA matching code (see
3291 below) to a single monolithic source would have made it really too
3292 unwieldy, quite apart from causing all the code to be include in a
3293 statically linked application, when only some functions are used. This is
3294 relevant even without the DFA addition now that patterns can be compiled in
3295 one application and matched in another.
3296
3297 The downside of splitting up is that there have to be some external
3298 functions and data tables that are used internally in different modules of
3299 the library but which are not part of the API. These have all had their
3300 names changed to start with "_pcre_" so that they are unlikely to clash
3301 with other external names.
3302
3303 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3304 a different (DFA) algorithm. Although it is slower than the original
3305 function, it does have some advantages for certain types of matching
3306 problem.
3307
3308 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3309 including restarting after a partial match.
3310
3311 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3312 defined when compiling for Windows was sent to me. I have put it into the
3313 code, though I have no means of testing or verifying it.
3314
3315 7. Added the pcre_refcount() auxiliary function.
3316
3317 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3318 match before or at the first newline in the subject string. In pcretest,
3319 the /f option on a pattern can be used to set this.
3320
3321 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3322 would behave wrongly. This has been present in PCRE since release 4.0.
3323
3324 10. A number of changes to the pcregrep command:
3325
3326 (a) Refactored how -x works; insert ^(...)$ instead of setting
3327 PCRE_ANCHORED and checking the length, in preparation for adding
3328 something similar for -w.
3329
3330 (b) Added the -w (match as a word) option.
3331
3332 (c) Refactored the way lines are read and buffered so as to have more
3333 than one at a time available.
3334
3335 (d) Implemented a pcregrep test script.
3336
3337 (e) Added the -M (multiline match) option. This allows patterns to match
3338 over several lines of the subject. The buffering ensures that at least
3339 8K, or the rest of the document (whichever is the shorter) is available
3340 for matching (and similarly the previous 8K for lookbehind assertions).
3341
3342 (f) Changed the --help output so that it now says
3343
3344 -w, --word-regex(p)
3345
3346 instead of two lines, one with "regex" and the other with "regexp"
3347 because that confused at least one person since the short forms are the
3348 same. (This required a bit of code, as the output is generated
3349 automatically from a table. It wasn't just a text change.)
3350
3351 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3352 option but starts with a hyphen. Could be a pattern or a path name
3353 starting with a hyphen, for instance.
3354
3355 (h) "-" can be given as a file name to represent stdin.
3356
3357 (i) When file names are being printed, "(standard input)" is used for
3358 the standard input, for compatibility with GNU grep. Previously
3359 "<stdin>" was used.
3360
3361 (j) The option --label=xxx can be used to supply a name to be used for
3362 stdin when file names are being printed. There is no short form.
3363
3364 (k) Re-factored the options decoding logic because we are going to add
3365 two more options that take data. Such options can now be given in four
3366 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3367
3368 (l) Added the -A, -B, and -C options for requesting that lines of context
3369 around matches be printed.
3370
3371 (m) Added the -L option to print the names of files that do not contain
3372 any matching lines, that is, the complement of -l.
3373
3374 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3375 continue to scan other files.
3376
3377 (o) The -s option was incorrectly implemented. For compatibility with other
3378 greps, it now suppresses the error message for a non-existent or non-
3379 accessible file (but not the return code). There is a new option called
3380 -q that suppresses the output of matching lines, which was what -s was
3381 previously doing.
3382
3383 (p) Added --include and --exclude options to specify files for inclusion
3384 and exclusion when recursing.
3385
3386 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3387 Hopefully, it now does.
3388
3389 12. Missing cast in pcre_study().
3390
3391 13. Added an "uninstall" target to the makefile.
3392
3393 14. Replaced "extern" in the function prototypes in Makefile.in with
3394 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3395 world, but is set differently for Windows.
3396
3397 15. Added a second compiling function called pcre_compile2(). The only
3398 difference is that it has an extra argument, which is a pointer to an
3399 integer error code. When there is a compile-time failure, this is set
3400 non-zero, in addition to the error test pointer being set to point to an
3401 error message. The new argument may be NULL if no error number is required
3402 (but then you may as well call pcre_compile(), which is now just a
3403 wrapper). This facility is provided because some applications need a
3404 numeric error indication, but it has also enabled me to tidy up the way
3405 compile-time errors are handled in the POSIX wrapper.
3406
3407 16. Added VPATH=.libs to the makefile; this should help when building with one
3408 prefix path and installing with another. (Or so I'm told by someone who
3409 knows more about this stuff than I do.)
3410
3411 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3412 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3413 match everything, including newlines. This is not POSIX-compatible, but
3414 somebody wanted the feature. From pcretest it can be activated by using
3415 both the P and the s flags.
3416
3417 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3418
3419 19. libpcre.pc was being incorrectly installed as executable.
3420
3421 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3422 it now also looks for '\r' so that it will work unmodified on Windows.
3423
3424 21. Added Google's contributed C++ wrapper to the distribution.
3425
3426 22. Added some untidy missing memory free() calls in pcretest, to keep
3427 Electric Fence happy when testing.
3428
3429
3430
3431 Version 5.0 13-Sep-04
3432 ---------------------
3433
3434 1. Internal change: literal characters are no longer packed up into items
3435 containing multiple characters in a single byte-string. Each character
3436 is now matched using a separate opcode. However, there may be more than one
3437 byte in the character in UTF-8 mode.
3438
3439 2. The pcre_callout_block structure has two new fields: pattern_position and
3440 next_item_length. These contain the offset in the pattern to the next match
3441 item, and its length, respectively.
3442
3443 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3444 insertion of callouts before each pattern item. Added the /C option to
3445 pcretest to make use of this.
3446
3447 4. On the advice of a Windows user, the lines
3448
3449 #if defined(_WIN32) || defined(WIN32)
3450 _setmode( _fileno( stdout ), 0x8000 );
3451 #endif /* defined(_WIN32) || defined(WIN32) */
3452
3453 have been added to the source of pcretest. This apparently does useful
3454 magic in relation to line terminators.
3455
3456 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3457 for the benefit of those environments where the "b" makes a difference.
3458
3459 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3460 to know about it. I have put a hack into configure.in that adds in code
3461 to set GCC=yes if CC=icc. This seems to end up at a point in the
3462 generated configure script that is early enough to affect the setting of
3463 compiler options, which is what is needed, but I have no means of testing
3464 whether it really works. (The user who reported this had patched the
3465 generated configure script, which of course I cannot do.)
3466
3467 LATER: After change 22 below (new libtool files), the configure script
3468 seems to know about icc (and also ecc). Therefore, I have commented out
3469 this hack in configure.in.
3470
3471 7. Added support for pkg-config (2 patches were sent in).
3472
3473 8. Negated POSIX character classes that used a combination of internal tables
3474 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3475 [[:^ascii]]. Typically, they would match almost any characters. The other
3476 POSIX classes were not broken in this way.
3477
3478 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3479 to find the match, as PCRE was deluded into thinking that the match had to
3480 start at the start point or following a newline. The same bug applied to
3481 patterns with negative forward assertions or any backward assertions
3482 preceding ".*" at the start, unless the pattern required a fixed first
3483 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3484
3485 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3486 starting at the last subject character, bytes beyond the end of the subject
3487 string were read.
3488
3489 11. Renamed the variable "class" as "classbits" to make life easier for C++
3490 users. (Previously there was a macro definition, but it apparently wasn't
3491 enough.)
3492
3493 12. Added the new field "tables" to the extra data so that tables can be passed
3494 in at exec time, or the internal tables can be re-selected. This allows
3495 a compiled regex to be saved and re-used at a later time by a different
3496 program that might have everything at different addresses.
3497
3498 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3499 -R library as well as a -L library.
3500
3501 14. The debugging options of pcretest (-d on the command line or D on a
3502 pattern) showed incorrect output for anything following an extended class
3503 that contained multibyte characters and which was followed by a quantifier.
3504
3505 15. Added optional support for general category Unicode character properties
3506 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3507 support. It adds about 90K to the size of the library. The meanings of the
3508 inbuilt class escapes such as \d and \s have NOT been changed.
3509
3510 16. Updated pcredemo.c to include calls to free() to release the memory for the
3511 compiled pattern.
3512
3513 17. The generated file chartables.c was being created in the source directory
3514 instead of in the building directory. This caused the build to fail if the
3515 source directory was different from the building directory, and was
3516 read-only.
3517
3518 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3519 file. No doubt somebody will tell me if they don't make sense... Also added
3520 Dan Mooney's comments about building on OpenVMS.
3521
3522 19. Added support for partial matching via the PCRE_PARTIAL option for
3523 pcre_exec() and the \P data escape in pcretest.
3524
3525 20. Extended pcretest with 3 new pattern features:
3526
3527 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3528 write the compiled pattern to the file whose name is "rest-of-line".
3529 This is a straight binary dump of the data, with the saved pointer to
3530 the character tables forced to be NULL. The study data, if any, is
3531 written too. After writing, pcretest reads a new pattern.
3532
3533 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3534 compiled pattern from the given file. There must not be any
3535 occurrences of "<" in the file name (pretty unlikely); if there are,
3536 pcretest will instead treat the initial "<" as a pattern delimiter.
3537 After reading in the pattern, pcretest goes on to read data lines as
3538 usual.
3539
3540 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3541 and 16-bit fields in a compiled pattern, to simulate a pattern that
3542 was compiled on a host of opposite endianness.
3543
3544 21. The pcre-exec() function can now cope with patterns that were compiled on
3545 hosts of opposite endianness, with this restriction:
3546
3547 As for any compiled expression that is saved and used later, the tables
3548 pointer field cannot be preserved; the extra_data field in the arguments
3549 to pcre_exec() should be used to pass in a tables address if a value
3550 other than the default internal tables were used at compile time.
3551
3552 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3553 now diagnosed as an error. Previously, most of the time, a negative number
3554 would have been treated as zero, but if in addition "ovector" was passed as
3555 NULL, a crash could occur.
3556
3557 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3558 new versions from the libtool 1.5 distribution (the last one is a copy of
3559 a file called libtool.m4). This seems to have fixed the need to patch
3560 "configure" to support Darwin 1.3 (which I used to do). However, I still
3561 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3562 workstation).
3563
3564 24. Changed the PCRE licence to be the more standard "BSD" licence.
3565
3566
3567 Version 4.5 01-Dec-03
3568 ---------------------
3569
3570 1. There has been some re-arrangement of the code for the match() function so
3571 that it can be compiled in a version that does not call itself recursively.
3572 Instead, it keeps those local variables that need separate instances for
3573 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3574 needs to "recurse". Keeping track of where control must go is done by means
3575 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3576 hide most of the details from the main code, and operates only if
3577 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3578 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3579 operating.
3580
3581 To make it easier for callers to provide specially tailored get/free
3582 functions for this usage, two new functions, pcre_stack_malloc, and
3583 pcre_stack_free, are used. They are always called in strict stacking order,
3584 and the size of block requested is always the same.
3585
3586 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3587 PCRE has been compiled to use the stack or the heap for recursion. The
3588 -C option of pcretest uses this to show which version is compiled.
3589
3590 A new data escape \S, is added to pcretest; it causes the amounts of store
3591 obtained and freed by both kinds of malloc/free at match time to be added
3592 to the output.
3593
3594 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3595 what's available on my current Linux desktop machine.
3596
3597 3. When matching a UTF-8 string, the test for a valid string at the start has
3598 been extended. If start_offset is not zero, PCRE now checks that it points
3599 to a byte that is the start of a UTF-8 character. If not, it returns
3600 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3601 this is necessary because there may be backward assertions in the pattern.
3602 When matching the same subject several times, it may save resources to use
3603 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3604
3605 4. The code for checking the validity of UTF-8 strings has been tightened so
3606 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3607 containing "overlong sequences".
3608
3609 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3610 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3611 should have been "&", but it just so happened that all the cases this let
3612 through by mistake were picked up later in the function.
3613
3614 6. I had used a variable called "isblank" - this is a C99 function, causing
3615 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3616
3617 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3618 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3619 so the version is shown; (c) stop "make test" echoing "./RunTest".
3620
3621 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3622
3623 9. The prototype for memmove() for systems that don't have it was using
3624 size_t, but the inclusion of the header that defines size_t was later. I've
3625 moved the #includes for the C headers earlier to avoid this.
3626
3627 10. Added some adjustments to the code to make it easier to compiler on certain
3628 special systems:
3629
3630 (a) Some "const" qualifiers were missing.
3631 (b) Added the macro EXPORT before all exported functions; by default this
3632 is defined to be empty.
3633 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3634 that it reads its output file name as an argument instead of writing
3635 to the standard output and assuming this can be redirected.
3636
3637 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3638 class containing characters with values greater than 255, PCRE compilation
3639 went into a loop.
3640
3641 12. A recursive reference to a subpattern that was within another subpattern
3642 that had a minimum quantifier of zero caused PCRE to crash. For example,
3643 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3644 recursion. If the recursively-called subpattern itself had a zero repeat,
3645 that was OK.
3646
3647 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3648 buffer into which it was copied (for escape processing) was still set at
3649 1024, so long lines caused crashes.
3650
3651 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3652 "internal error: code overflow...". This applied to any character class
3653 that was followed by a possessive quantifier.
3654
3655 15. Modified the Makefile to add libpcre.la as a prerequisite for
3656 libpcreposix.la because I was told this is needed for a parallel build to
3657 work.
3658
3659 16. If a pattern that contained .* following optional items at the start was
3660 studied, the wrong optimizing data was generated, leading to matching
3661 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3662 matching string must start with a or b or c. The correct conclusion for
3663 this pattern is that a match can start with any character.
3664
3665
3666 Version 4.4 13-Aug-03
3667 ---------------------
3668
3669 1. In UTF-8 mode, a character class containing characters with values between
3670 127 and 255 was not handled correctly if the compiled pattern was studied.
3671 In fixing this, I have also improved the studying algorithm for such
3672 classes (slightly).
3673
3674 2. Three internal functions had redundant arguments passed to them. Removal
3675 might give a very teeny performance improvement.
3676
3677 3. Documentation bug: the value of the capture_top field in a callout is *one
3678 more than* the number of the hightest numbered captured substring.
3679
3680 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3681 in incorrectly linking with a previously installed version. They now link
3682 explicitly with libpcre.la.
3683
3684 5. configure.in no longer needs to recognize Cygwin specially.
3685
3686 6. A problem in pcre.in for Windows platforms is fixed.
3687
3688 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3689 pcretest, it used to include the size of the study block as part of its
3690 output. Unfortunately, the structure contains a field that has a different
3691 size on different hardware architectures. This meant that the tests that
3692 showed this size failed. As the block is currently always of a fixed size,
3693 this information isn't actually particularly useful in pcretest output, so
3694 I have just removed it.
3695
3696 8. Three pre-processor statements accidentally did not start in column 1.
3697 Sadly, there are *still* compilers around that complain, even though
3698 standard C has not required this for well over a decade. Sigh.
3699
3700 9. In pcretest, the code for checking callouts passed small integers in the
3701 callout_data field, which is a void * field. However, some picky compilers
3702 complained about the casts involved for this on 64-bit systems. Now
3703 pcretest passes the address of the small integer instead, which should get
3704 rid of the warnings.
3705
3706 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3707 both compile and run time, and gives an error if an invalid UTF-8 sequence
3708 is found. There is a option for disabling this check in cases where the
3709 string is known to be correct and/or the maximum performance is wanted.
3710
3711 11. In response to a bug report, I changed one line in Makefile.in from
3712
3713 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3714 to
3715 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3716
3717 to look similar to other lines, but I have no way of telling whether this
3718 is the right thing to do, as I do not use Windows. No doubt I'll get told
3719 if it's wrong...
3720
3721
3722 Version 4.3 21-May-03
3723 ---------------------
3724
3725 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3726 Makefile.
3727
3728 2. Some refactoring to improve the quality of the code:
3729
3730 (i) The utf8_table... variables are now declared "const".
3731
3732 (ii) The code for \cx, which used the "case flipping" table to upper case
3733 lower case letters, now just substracts 32. This is ASCII-specific,
3734 but the whole concept of \cx is ASCII-specific, so it seems
3735 reasonable.
3736
3737 (iii) PCRE was using its character types table to recognize decimal and
3738 hexadecimal digits in the pattern. This is silly, because it handles
3739 only 0-9, a-f, and A-F, but the character types table is locale-
3740 specific, which means strange things might happen. A private
3741 table is now used for this - though it costs 256 bytes, a table is
3742 much faster than multiple explicit tests. Of course, the standard
3743 character types table is still used for matching digits in subject
3744 strings against \d.
3745
3746 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3747 ending in _t are). So I've renamed it as ESC_tee.
3748
3749 3. The first argument for regexec() in the POSIX wrapper should have been
3750 defined as "const".
3751
3752 4. Changed pcretest to use malloc() for its buffers so that they can be
3753 Electric Fenced for debugging.
3754
3755 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3756 to read one or more bytes before the start of the subject string. Often this
3757 had no effect on PCRE's behaviour, but in some circumstances it could
3758 provoke a segmentation fault.
3759
3760 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3761 to try to read one or more bytes before the start of the subject string.
3762
3763 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3764 UTF-8 support could misbehave in various ways if the subject string
3765 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3766 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3767 back over UTF-8 characters.)
3768
3769
3770 Version 4.2 14-Apr-03
3771 ---------------------
3772
3773 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3774
3775 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3776 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3777 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3778 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3779 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3780 and BUILD_EXEEXT
3781 Note: automatic setting of the BUILD variables is not yet working
3782 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3783 compile-time but not at link-time
3784 [LINK]: use for linking executables only
3785 make different versions for Windows and non-Windows
3786 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3787 libraries
3788 [LINK_FOR_BUILD]: new variable
3789 [OBJEXT]: use throughout
3790 [EXEEXT]: use throughout
3791 <winshared>: new target
3792 <wininstall>: new target
3793 <dftables.o>: use native compiler
3794 <dftables>: use native linker
3795 <install>: handle Windows platform correctly
3796 <clean>: ditto
3797 <check>: ditto
3798 copy DLL to top builddir before testing
3799
3800 As part of these changes, -no-undefined was removed again. This was reported
3801 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3802 in any case.
3803
3804 3. Some tidies to get rid of compiler warnings:
3805
3806 . In the match_data structure, match_limit was an unsigned long int, whereas
3807 match_call_count was an int. I've made them both unsigned long ints.
3808
3809 . In pcretest the fact that a const uschar * doesn't automatically cast to
3810 a void * provoked a warning.
3811
3812 . Turning on some more compiler warnings threw up some "shadow" variables
3813 and a few more missing casts.
3814
3815 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3816 option, a class that contained a single character with a value between 128
3817 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3818
3819 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3820 option, a class that contained several characters, but with at least one
3821 whose value was between 128 and 255 caused PCRE to crash.
3822
3823
3824 Version 4.1 12-Mar-03
3825 ---------------------
3826
3827 1. Compiling with gcc -pedantic found a couple of places where casts were
3828 needed, and a string in dftables.c that was longer than standard compilers are
3829 required to support.
3830
3831 2. Compiling with Sun's compiler found a few more places where the code could
3832 be tidied up in order to avoid warnings.
3833
3834 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3835 first of these names is deprecated in the latest Autoconf in favour of the name
3836 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3837 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3838 analogy I have changed it to CFLAGS_FOR_BUILD.
3839
3840 4. Added -no-undefined to the linking command in the Makefile, because this is
3841 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3842 linking step for the pcreposix library.
3843
3844 5. PCRE was failing to diagnose the case of two named groups with the same
3845 name.
3846
3847 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3848 literal character that is needed in the subject for a match, and scans along to
3849 ensure that it is present before embarking on the full matching process. This
3850 saves time in cases of nested unlimited repeats that are never going to match.
3851 Problem: the scan can take a lot of time if the subject is very long (e.g.
3852 megabytes), thus penalizing straightforward matches. It is now done only if the
3853 amount of subject to be scanned is less than 1000 bytes.
3854
3855 7. A lesser problem with the same optimization is that it was recording the
3856 first character of an anchored pattern as "needed", thus provoking a search
3857 right along the subject, even when the first match of the pattern was going to
3858 fail. The "needed" character is now not set for anchored patterns, unless it
3859 follows something in the pattern that is of non-fixed length. Thus, it still
3860 fulfils its original purpose of finding quick non-matches in cases of nested
3861 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3862
3863
3864 Version 4.0 17-Feb-03
3865 ---------------------
3866
3867 1. If a comment in an extended regex that started immediately after a meta-item
3868 extended to the end of string, PCRE compiled incorrect data. This could lead to
3869 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3870
3871 2. Moved to autoconf 2.53 and libtool 1.4.2.
3872
3873 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3874 the special perltest8 script is no longer needed - all the tests can be run
3875 from a single perltest script.
3876
3877 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3878 by \s. It has now been removed in PCRE. This means it isn't recognized as
3879 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3880 class [:space:] *does* include VT, thereby creating a mess.
3881
3882 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3883 space and tab.
3884
3885 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3886 its new features into the main test script, reducing the number of scripts.
3887
3888 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3889 were backward compatible, and made the (?i) apply to the whole pattern, as if
3890 /i were given. Now it behaves more logically, and applies the option setting
3891 only to what follows. PCRE has been changed to follow suit. However, if it
3892 finds options settings right at the start of the pattern, it extracts them into
3893 the global options, as before. Thus, they show up in the info data.
3894
3895 8. Added support for the \Q...\E escape sequence. Characters in between are
3896 treated as literals. This is slightly different from Perl in that $ and @ are
3897 also handled as literals inside the quotes. In Perl, they will cause variable
3898 interpolation. Note the following examples:
3899
3900 Pattern PCRE matches Perl matches
3901
3902 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3903 \Qabc\$xyz\E abc\$xyz abc\$xyz
3904 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3905
3906 For compatibility with Perl, \Q...\E sequences are recognized inside character
3907 classes as well as outside them.
3908
3909 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3910 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3911 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3912 signed/unsigned warnings.
3913
3914 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3915 option for pcretest, so I've replaced it by a simple function that does just
3916 that job.
3917
3918 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3919 "pcregrep -".
3920
3921 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3922 Java package. This provides some syntactic sugar for simple cases of what my
3923 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3924 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3925 item, you can use this simplified notation. Note that only makes sense with
3926 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3927 greediness, whatever the setting of the PCRE_UNGREEDY option.
3928
3929 13. A change of greediness default within a pattern was not taking effect at
3930 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3931 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3932 was abstracted outside.
3933
3934 14. PCRE now supports the \G assertion. It is true when the current matching
3935 position is at the start point of the match. This differs from \A when the
3936 starting offset is non-zero. Used with the /g option of pcretest (or similar
3937 code), it works in the same way as it does for Perl's /g option. If all
3938 alternatives of a regex begin with \G, the expression is anchored to the start
3939 match position, and the "anchored" flag is set in the compiled expression.
3940
3941 15. Some bugs concerning the handling of certain option changes within patterns
3942 have been fixed. These applied to options other than (?ims). For example,
3943 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3944 been the other way round. Some of this was related to change 7 above.
3945
3946 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3947 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3948 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3949 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3950
3951 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3952 mode. Unlike ".", it always matches newline, whatever the setting of
3953 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3954 assertions. Perl allows it, but it doesn't (in general) work because it can't
3955 calculate the length of the lookbehind. At least, that's the case for Perl
3956 5.8.0 - I've been told they are going to document that it doesn't work in
3957 future.
3958
3959 18. Added an error diagnosis for escapes that PCRE does not support: these are
3960 \L, \l, \N, \P, \p, \U, \u, and \X.
3961
3962 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3963 reading past the end of the pattern in cases such as /[abcd/.
3964
3965 20. PCRE was getting more memory than necessary for patterns with classes that
3966 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3967
3968 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3969 compiling PCRE for use with Virtual Pascal.
3970
3971 22. Small fix to the Makefile to make it work properly if the build is done
3972 outside the source tree.
3973
3974 23. Added a new extension: a condition to go with recursion. If a conditional
3975 subpattern starts with (?(R) the "true" branch is used if recursion has
3976 happened, whereas the "false" branch is used only at the top level.
3977
3978 24. When there was a very long string of literal characters (over 255 bytes
3979 without UTF support, over 250 bytes with UTF support), the computation of how
3980 much memory was required could be incorrect, leading to segfaults or other
3981 strange effects.
3982
3983 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3984 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3985 there was a subsequent back reference to those brackets. This meant that, for
3986 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3987 possible to check for precisely this case. All we can do is abandon the
3988 optimization if .* occurs inside capturing brackets when there are any back
3989 references whatsoever. (See below for a better fix that came later.)
3990
3991 26. The handling of the optimization for finding the first character of a
3992 non-anchored pattern, and for finding a character that is required later in the
3993 match were failing in some cases. This didn't break the matching; it just
3994 failed to optimize when it could. The way this is done has been re-implemented.
3995
3996 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3997
3998 28. Added a new feature that provides some of the functionality that Perl
3999 provides with (?{...}). The facility is termed a "callout". The way it is done
4000 in PCRE is for the caller to provide an optional function, by setting
4001 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4002 global variable. By default it is unset, which disables all calling out. To get
4003 the function called, the regex must include (?C) at appropriate points. This
4004 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4005 This provides a means of identifying different callout points. When PCRE
4006 reaches such a point in the regex, if pcre_callout has been set, the external
4007 function is called. It is provided with data in a structure called
4008 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4009 matching continues; if it returns a non-zero value, the match at the current
4010 point fails. However, backtracking will occur if possible. [This was changed
4011 later and other features added - see item 49 below.]
4012
4013 29. pcretest is upgraded to test the callout functionality. It provides a
4014 callout function that displays information. By default, it shows the start of
4015 the match and the current position in the text. There are some new data escapes
4016 to vary what happens:
4017
4018 \C+ in addition, show current contents of captured substrings
4019 \C- do not supply a callout function
4020 \C!n return 1 when callout number n is reached
4021 \C!n!m return 1 when callout number n is reached for the mth time
4022
4023 30. If pcregrep was called with the -l option and just a single file name, it
4024 output "<stdin>" if a match was found, instead of the file name.
4025
4026 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4027 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4028 pcre_exec(). This saves a malloc/free per call. The default value of
4029 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4030 when configuring.
4031
4032 32. The default maximum size of a compiled pattern is 64K. There have been a
4033 few cases of people hitting this limit. The code now uses macros to handle the
4034 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4035 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4036 configuring. Tests 2 and 5 work only with 2-byte links because they output
4037 debugging information about compiled patterns.
4038
4039 33. Internal code re-arrangements:
4040
4041 (a) Moved the debugging function for printing out a compiled regex into
4042 its own source file (printint.c) and used #include to pull it into
4043 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4044 separate copies.
4045
4046 (b) Defined the list of op-code names for debugging as a macro in
4047 internal.h so that it is next to the definition of the opcodes.
4048
4049 (c) Defined a table of op-code lengths for simpler skipping along compiled
4050 code. This is again a macro in internal.h so that it is next to the
4051 definition of the opcodes.
4052
4053 34. Added support for recursive calls to individual subpatterns, along the
4054 lines of Robin Houston's patch (but implemented somewhat differently).
4055
4056 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4057 allow it to read and process whole directories in Win32. This code was
4058 contributed by Lionel Fourquaux; it has not been tested by me.
4059
4060 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4061 used to name a group. Names consist of alphanumerics and underscores, and must
4062 be unique. Back references use the syntax (?P=name) and recursive calls use
4063 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4064 numbers. The function pcre_fullinfo() can be used after compilation to extract
4065 a name/number map. There are three relevant calls:
4066
4067 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4068 PCRE_INFO_NAMECOUNT yields the number of entries
4069 PCRE_INFO_NAMETABLE yields a pointer to the map.
4070
4071 The map is a vector of fixed-size entries. The size of each entry depends on
4072 the length of the longest name used. The first two bytes of each entry are the
4073 group number, most significant byte first. There follows the corresponding
4074 name, zero terminated. The names are in alphabetical order.
4075
4076 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4077 case instead of 255. Making it the same both with and without UTF-8 support
4078 means that the same test output works with both.
4079
4080 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4081 calling malloc() with a zero argument.
4082
4083 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4084 optimization. I've improved things by keeping a bitmap of backreferences with
4085 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4086 fact referenced, the optimization can be applied. It is unlikely that a
4087 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4088 the match to follow \n) will appear inside brackets with a number greater than
4089 31, but if it does, any back reference > 31 suppresses the optimization.
4090
4091 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4092 of disabling numbered capturing parentheses. Any opening parenthesis that is
4093 not followed by ? behaves as if it were followed by ?: but named parentheses
4094 can still be used for capturing (and they will acquire numbers in the usual
4095 way).
4096
4097 41. Redesigned the return codes from the match() function into yes/no/error so
4098 that errors can be passed back from deep inside the nested calls. A malloc
4099 failure while inside a recursive subpattern call now causes the
4100 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4101
4102 42. It is now possible to set a limit on the number of times the match()
4103 function is called in a call to pcre_exec(). This facility makes it possible to
4104 limit the amount of recursion and backtracking, though not in a directly
4105 obvious way, because the match() function is used in a number of different
4106 circumstances. The count starts from zero for each position in the subject
4107 string (for non-anchored patterns). The default limit is, for compatibility, a
4108 large number, namely 10 000 000. You can change this in two ways:
4109
4110 (a) When configuring PCRE before making, you can use --with-match-limit=n
4111 to set a default value for the compiled library.
4112
4113 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4114 a different value is set. See 45 below.
4115
4116 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4117
4118 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4119 of things that can be changed at compile time. The first argument specifies
4120 what is wanted and the second points to where the information is to be placed.
4121 The current list of available information is:
4122
4123 PCRE_CONFIG_UTF8
4124
4125 The output is an integer that is set to one if UTF-8 support is available;
4126 otherwise it is set to zero.
4127
4128 PCRE_CONFIG_NEWLINE
4129
4130 The output is an integer that it set to the value of the code that is used for
4131 newline. It is either LF (10) or CR (13).
4132
4133 PCRE_CONFIG_LINK_SIZE
4134
4135 The output is an integer that contains the number of bytes used for internal
4136 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4137
4138 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4139
4140 The output is an integer that contains the threshold above which the POSIX
4141 interface uses malloc() for output vectors. See item 31 above.
4142
4143 PCRE_CONFIG_MATCH_LIMIT
4144
4145 The output is an unsigned integer that contains the default limit of the number
4146 of match() calls in a pcre_exec() execution. See 42 above.
4147
4148 44. pcretest has been upgraded by the addition of the -C option. This causes it
4149 to extract all the available output from the new pcre_config() function, and to
4150 output it. The program then exits immediately.
4151
4152 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4153 order to support additional features. One way would have been to define
4154 pcre_exec2() (for example) with extra arguments, but this would not have been
4155 extensible, and would also have required all calls to the original function to
4156 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4157 is used for passing in "extra" data from pcre_study().
4158
4159 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4160 contains the following fields:
4161
4162 flags a bitmap indicating which of the following fields are set
4163 study_data opaque data from pcre_study()
4164 match_limit a way of specifying a limit on match() calls for a specific
4165 call to pcre_exec()
4166 callout_data data for callouts (see 49 below)
4167
4168 The flag bits are also defined in pcre.h, and are
4169
4170 PCRE_EXTRA_STUDY_DATA
4171 PCRE_EXTRA_MATCH_LIMIT
4172 PCRE_EXTRA_CALLOUT_DATA
4173
4174 The pcre_study() function now returns one of these new pcre_extra blocks, with
4175 the actual study data pointed to by the study_data field, and the
4176 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4177 before. That is, this change is entirely upwards-compatible and requires no
4178 change to existing code.
4179
4180 If you want to pass in additional data to pcre_exec(), you can either place it
4181 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4182 block.
4183
4184 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4185 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4186 times with different match limits, until it finds the minimum value needed for
4187 pcre_exec() to complete. The value is then output. This can be instructive; for
4188 most simple matches the number is quite small, but for pathological cases it
4189 gets very large very quickly.
4190
4191 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4192 returns the size of the data block pointed to by the study_data field in a
4193 pcre_extra block, that is, the value that was passed as the argument to
4194 pcre_malloc() when PCRE was getting memory in which to place the information
4195 created by pcre_study(). The fourth argument should point to a size_t variable.
4196 pcretest has been extended so that this information is shown after a successful
4197 pcre_study() call when information about the compiled regex is being displayed.
4198
4199 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4200 because what follows is always an absolute path. (Later: it turns out that this
4201 is more than cosmetic for MinGW, because it doesn't like empty path
4202 components.)
4203
4204 49. Some changes have been made to the callout feature (see 28 above):
4205
4206 (i) A callout function now has three choices for what it returns:
4207
4208 0 => success, carry on matching
4209 > 0 => failure at this point, but backtrack if possible
4210 < 0 => serious error, return this value from pcre_exec()
4211
4212 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4213 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4214 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4215 use by callout functions. It will never be used by PCRE itself.
4216
4217 (ii) The pcre_extra structure (see 45 above) has a void * field called
4218 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4219 pcre_callout_block structure has a field of the same name. The contents of
4220 the field passed in the pcre_extra structure are passed to the callout
4221 function in the corresponding field in the callout block. This makes it
4222 easier to use the same callout-containing regex from multiple threads. For
4223 testing, the pcretest program has a new data escape
4224
4225 \C*n pass the number n (may be negative) as callout_data
4226
4227 If the callout function in pcretest receives a non-zero value as
4228 callout_data, it returns that value.
4229
4230 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4231 there were some redundant $(CFLAGS) in commands that are now specified as
4232 $(LINK), which already includes $(CFLAGS).
4233
4234 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4235 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4236 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4237 one-byte characters throughout. Note that case-insensitive matching applies
4238 only to characters whose values are less than 256. PCRE doesn't support the
4239 notion of cases for higher-valued characters.
4240
4241 (i) A character class whose characters are all within 0-255 is handled as
4242 a bit map, and the map is inverted for negative classes. Previously, a
4243 character > 255 always failed to match such a class; however it should
4244 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4245
4246 (ii) A negated character class with a single character < 255 is coded as
4247 "not this character" (OP_NOT). This wasn't working properly when the test
4248 character was multibyte, either singly or repeated.
4249
4250 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4251 mode, for example: \x{100}{2,3}.
4252
4253 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4254 singly or repeated) now correctly test multibyte characters. However,
4255 PCRE doesn't recognize any characters with values greater than 255 as
4256 digits, spaces, or word characters. Such characters always match \D, \S,
4257 and \W, and never match \d, \s, or \w.
4258
4259 (v) Classes may now contain characters and character ranges with values
4260 greater than 255. For example: [ab\x{100}-\x{400}].
4261
4262 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4263 PCRE in UTF-8 mode.
4264
4265 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4266 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4267 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4268 value.)
4269
4270 53. The single man page has become too large. I have therefore split it up into
4271 a number of separate man pages. These also give rise to individual HTML pages;
4272 these are now put in a separate directory, and there is an index.html page that
4273 lists them all. Some hyperlinking between the pages has been installed.
4274
4275 54. Added convenience functions for handling named capturing parentheses.
4276
4277 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4278 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4279 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4280 are faulted.
4281
4282 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4283 calling configure. These values are used when compiling the dftables.c program
4284 which is run to generate the source of the default character tables. They
4285 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4286 you will need to set these values.
4287
4288 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4289
4290
4291 Version 3.9 02-Jan-02
4292 ---------------------
4293
4294 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4295
4296 2. If --disable-static was given, the building process failed when trying to
4297 build pcretest and pcregrep. (For some reason it was using libtool to compile
4298 them, which is not right, as they aren't part of the library.)
4299
4300
4301 Version 3.8 18-Dec-01
4302 ---------------------
4303
4304 1. The experimental UTF-8 code was completely screwed up. It was packing the
4305 bytes in the wrong order. How dumb can you get?
4306
4307
4308 Version 3.7 29-Oct-01
4309 ---------------------
4310
4311 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4312 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4313 this didn't happen under Solaris 8, where I normally test things.
4314
4315 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4316 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4317 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4318 this fix an hour or so after the initial 3.7 release.)
4319
4320
4321 Version 3.6 23-Oct-01
4322 ---------------------
4323
4324 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4325 offsets passed as NULL with zero offset count.
4326
4327 2. The config.guess and config.sub files had not been updated when I moved to
4328 the latest autoconf.
4329
4330
4331 Version 3.5 15-Aug-01
4332 ---------------------
4333
4334 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4335 had been forgotten.
4336
4337 2. By using declared but undefined structures, we can avoid using "void"
4338 definitions in pcre.h while keeping the internal definitions of the structures
4339 private.
4340
4341 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4342 user point of view, this means that both static and shared libraries are built
4343 by default, but this can be individually controlled. More of the work of
4344 handling this static/shared cases is now inside libtool instead of PCRE's make
4345 file.
4346
4347 4. The pcretest utility is now installed along with pcregrep because it is
4348 useful for users (to test regexs) and by doing this, it automatically gets
4349 relinked by libtool. The documentation has been turned into a man page, so
4350 there are now .1, .txt, and .html versions in /doc.
4351
4352 5. Upgrades to pcregrep:
4353 (i) Added long-form option names like gnu grep.
4354 (ii) Added --help to list all options with an explanatory phrase.
4355 (iii) Added -r, --recursive to recurse into sub-directories.
4356 (iv) Added -f, --file to read patterns from a file.
4357
4358 6. pcre_exec() was referring to its "code" argument before testing that
4359 argument for NULL (and giving an error if it was NULL).
4360
4361 7. Upgraded Makefile.in to allow for compiling in a different directory from
4362 the source directory.
4363
4364 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4365 options bits, the pointer it was passed was to an int instead of to an unsigned
4366 long int. This mattered only on 64-bit systems.
4367
4368 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4369 generated) instead of pcre.in, which it its source. Also made the same change
4370 in several of the .c files.
4371
4372 10. A new release of gcc defines printf() as a macro, which broke pcretest
4373 because it had an ifdef in the middle of a string argument for printf(). Fixed
4374 by using separate calls to printf().
4375
4376 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4377 script, to force use of CR or LF instead of \n in the source. On non-Unix
4378 systems, the value can be set in config.h.
4379
4380 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4381 absolute limit. Changed the text of the error message to make this clear, and
4382 likewise updated the man page.
4383
4384 13. The limit of 99 on the number of capturing subpatterns has been removed.
4385 The new limit is 65535, which I hope will not be a "real" limit.
4386
4387
4388 Version 3.4 22-Aug-00
4389 ---------------------
4390
4391 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4392
4393 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4394
4395
4396 Version 3.3 01-Aug-00
4397 ---------------------
4398
4399 1. If an octal character was given, but the value was greater than \377, it
4400 was not getting masked to the least significant bits, as documented. This could
4401 lead to crashes in some systems.
4402
4403 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4404 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4405
4406 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4407 These just pass their arguments on to (pcre_free)(), but they are provided
4408 because some uses of PCRE bind it to non-C systems that can call its functions,
4409 but cannot call free() or pcre_free() directly.
4410
4411 4. Add "make test" as a synonym for "make check". Corrected some comments in
4412 the Makefile.
4413
4414 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4415 Makefile.
4416
4417 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4418 command called pgrep for grepping around the active processes.
4419
4420 7. Added the beginnings of support for UTF-8 character strings.
4421
4422 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4423 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4424 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4425 out for the ar command.)
4426
4427
4428 Version 3.2 12-May-00
4429 ---------------------
4430
4431 This is purely a bug fixing release.
4432
4433 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4434 of ZA. This was just one example of several cases that could provoke this bug,
4435 which was introduced by change 9 of version 2.00. The code for breaking
4436 infinite loops after an iteration that matches an empty string was't working
4437 correctly.
4438
4439 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4440 when matched against abbab (for example). After matching an empty string, it
4441 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4442 caused it to match further down the string than it should.
4443
4444 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4445 was there because it doesn't seem to be needed, and it causes trouble on some
4446 systems, as it is not a Standard C header. It has been removed.
4447
4448 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4449 were reported on the Macintosh. The changes were from
4450
4451 while ((c = *(++ptr)) != 0 && c != '\n');
4452 to
4453 while ((c = *(++ptr)) != 0 && c != '\n') ;
4454
4455 Totally extraordinary, but if that's what it takes...
4456
4457 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4458 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4459 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4460 assumes the way PCRE uses memmove() (always moving upwards).
4461
4462 6. PCRE is being used in one environment where strchr() is not available. There
4463 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4464 faster code anyway.
4465
4466
4467 Version 3.1 09-Feb-00
4468 ---------------------
4469
4470 The only change in this release is the fixing of some bugs in Makefile.in for
4471 the "install" target:
4472
4473 (1) It was failing to install pcreposix.h.
4474
4475 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4476
4477
4478 Version 3.0 01-Feb-00
4479 ---------------------
4480
4481 1. Add support for the /+ modifier to perltest (to output $` like it does in
4482 pcretest).
4483
4484 2. Add support for the /g modifier to perltest.
4485
4486 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4487 matches null strings.
4488
4489 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4490 pattern. Perl treats empty patterns specially - it reuses the most recent
4491 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4492 effect.
4493
4494 5. The POSIX interface was broken in that it was just handing over the POSIX
4495 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4496 required a bigger vector, with some working space on the end. This means that
4497 the POSIX wrapper now has to get and free some memory, and copy the results.
4498
4499 6. Added some simple autoconf support, placing the test data and the
4500 documentation in separate directories, re-organizing some of the
4501 information files, and making it build pcre-config (a GNU standard). Also added
4502 libtool support for building PCRE as a shared library, which is now the
4503 default.
4504
4505 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4506 09 are not valid octal constants. Single digits will be used for minor values
4507 less than 10.
4508
4509 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4510 existing programs that set these in the POSIX interface can use PCRE without
4511 modification.
4512
4513 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4514 return all that pcre_info() returns, plus additional data. The pcre_info()
4515 function is retained for compatibility, but is considered to be obsolete.
4516
4517 10. Added experimental recursion feature (?R) to handle one common case that
4518 Perl 5.6 will be able to do with (?p{...}).
4519
4520 11. Added support for POSIX character classes like [:alpha:], which Perl is
4521 adopting.
4522
4523
4524 Version 2.08 31-Aug-99
4525 ----------------------
4526
4527 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4528 trying to match at the startoffset position, but instead was moving forward to
4529 the next newline as if a previous match had failed.
4530
4531 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4532 and could get into a loop if a null string was matched other than at the start
4533 of the subject.
4534
4535 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4536 be distinguished at compile time, and for completeness also added PCRE_DATE.
4537
4538 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4539 in GnuWin32 environments.
4540
4541
4542 Version 2.07 29-Jul-99
4543 ----------------------
4544
4545 1. The documentation is now supplied in plain text form and HTML as well as in
4546 the form of man page sources.
4547
4548 2. C++ compilers don't like assigning (void *) values to other pointer types.
4549 In particular this affects malloc(). Although there is no problem in Standard
4550 C, I've put in casts to keep C++ compilers happy.
4551
4552 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4553 should be (const char *).
4554
4555 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4556 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4557 However, I haven't made this a standard facility. The documentation doesn't
4558 mention it, and the Makefile doesn't support it.
4559
4560 5. The Makefile now contains an "install" target, with editable destinations at
4561 the top of the file. The pcretest program is not installed.
4562
4563 6. pgrep -V now gives the PCRE version number and date.
4564
4565 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4566 causing the entire string to be ignored, instead of just the last character.
4567
4568 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4569 non-matching string, it can take a very, very long time, even for strings of
4570 quite modest length, because of the nested recursion. PCRE now does better in
4571 some of these cases. It does this by remembering the last required literal
4572 character in the pattern, and pre-searching the subject to ensure it is present
4573 before running the real match. In other words, it applies a heuristic to detect
4574 some types of certain failure quickly, and in the above example, if presented
4575 with a string that has no trailing " it gives "no match" very quickly.
4576
4577 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4578 other alternatives are tried instead.
4579
4580
4581 Version 2.06 09-Jun-99
4582 ----------------------
4583
4584 1. Change pcretest's output for amount of store used to show just the code
4585 space, because the remainder (the data block) varies in size between 32-bit and
4586 64-bit systems.
4587
4588 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4589 start matching at. This allows lookbehinds to work when searching for multiple
4590 occurrences in a string.
4591
4592 3. Added additional options to pcretest for testing multiple occurrences:
4593
4594 /+ outputs the rest of the string that follows a match
4595 /g loops for multiple occurrences, using the new startoffset argument
4596 /G loops for multiple occurrences by passing an incremented pointer
4597
4598 4. PCRE wasn't doing the "first character" optimization for patterns starting
4599 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4600 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4601 the letter 'x'. On long subject strings, this gives a significant speed-up.
4602
4603
4604 Version 2.05 21-Apr-99
4605 ----------------------
4606
4607 1. Changed the type of magic_number from int to long int so that it works
4608 properly on 16-bit systems.
4609
4610 2. Fixed a bug which caused patterns starting with .* not to work correctly
4611 when the subject string contained newline characters. PCRE was assuming
4612 anchoring for such patterns in all cases, which is not correct because .* will
4613 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4614 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4615 must be retried after every newline in the subject.
4616
4617
4618 Version 2.04 18-Feb-99
4619 ----------------------
4620
4621 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4622 computation of the store needed to hold the pattern was incorrect (too large).
4623 If such patterns were nested a few deep, this could multiply and become a real
4624 problem.
4625
4626 2. Added /M option to pcretest to show the memory requirement of a specific
4627 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4628
4629 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4630 compiled in such a way that the backtracking after subsequent failure was
4631 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4632 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4633
4634
4635 Version 2.03 02-Feb-99
4636 ----------------------
4637
4638 1. Fixed typo and small mistake in man page.
4639
4640 2. Added 4th condition (GPL supersedes if conflict) and created separate
4641 LICENCE file containing the conditions.
4642
4643 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4644 Perl, that is the internal \ allows the delimiter to be included in the
4645 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4646 the final delimiter, add \ to the end of the pattern (to test the error).
4647
4648 4. Added the convenience functions for extracting substrings after a successful
4649 match. Updated pcretest to make it able to test these functions.
4650
4651
4652 Version 2.02 14-Jan-99
4653 ----------------------
4654
4655 1. Initialized the working variables associated with each extraction so that
4656 their saving and restoring doesn't refer to uninitialized store.
4657
4658 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4659 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4660 fix the problem.
4661
4662 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4663 calls, and wasn't printing the correct value for compiling calls. Increased the
4664 default value of LOOPREPEAT, and the number of significant figures in the
4665 times.
4666
4667 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4668
4669 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4670 a building problem on Windows NT with a FAT file system.
4671
4672
4673 Version 2.01 21-Oct-98
4674 ----------------------
4675
4676 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4677 to character tables built by pcre_maketables() in the current locale. If NULL
4678 is passed, the default tables are used.
4679
4680
4681 Version 2.00 24-Sep-98
4682 ----------------------
4683
4684 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4685 it any more.
4686
4687 2. Allow quantification of (?>) groups, and make it work correctly.
4688
4689 3. The first character computation wasn't working for (?>) groups.
4690
4691 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4692 end of the subject) and add 5.005's \z, which really does match only at the
4693 very end of the subject.
4694
4695 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4696
4697 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4698 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4699 localized options. All options to pcre_study() were also removed.
4700
4701 7. Add other new features from 5.005:
4702
4703 $(?<= positive lookbehind
4704 $(?<! negative lookbehind
4705 (?imsx-imsx) added the unsetting capability
4706 such a setting is global if at outer level; local otherwise
4707 (?imsx-imsx:) non-capturing groups with option setting
4708 (?(cond)re|re) conditional pattern matching
4709
4710 A backreference to itself in a repeated group matches the previous
4711 captured string.
4712
4713 8. General tidying up of studying (both automatic and via "study")
4714 consequential on the addition of new assertions.
4715
4716 9. As in 5.005, unlimited repeated groups that could match an empty substring
4717 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4718 runtime if any iteration does actually match an empty substring.
4719
4720 10. Include the RunTest script in the distribution.
4721
4722 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4723 discrepancies, some of which were old and were also with respect to 5.004. They
4724 have now been fixed.
4725
4726
4727 Version 1.09 28-Apr-98
4728 ----------------------
4729
4730 1. A negated single character class followed by a quantifier with a minimum
4731 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4732 program crashes, or just wrong answers. This did not apply to negated classes
4733 containing more than one character, or to minima other than one.
4734
4735
4736 Version 1.08 27-Mar-98
4737 ----------------------
4738
4739 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4740
4741 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4742 latter must appear before anything that relies on it in the pattern.
4743
4744
4745 Version 1.07 16-Feb-98
4746 ----------------------
4747
4748 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4749 repeat of a potentially empty string).
4750
4751
4752 Version 1.06 23-Jan-98
4753 ----------------------
4754
4755 1. Added Markus Oberhumer's little patches for C++.
4756
4757 2. Literal strings longer than 255 characters were broken.
4758
4759
4760 Version 1.05 23-Dec-97
4761 ----------------------
4762
4763 1. Negated character classes containing more than one character were failing if
4764 PCRE_CASELESS was set at run time.
4765
4766
4767 Version 1.04 19-Dec-97
4768 ----------------------
4769
4770 1. Corrected the man page, where some "const" qualifiers had been omitted.
4771
4772 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4773 input syntax.
4774
4775 3. Fixed memory leak which occurred when a regex with back references was
4776 matched with an offsets vector that wasn't big enough. The temporary memory
4777 that is used in this case wasn't being freed if the match failed.
4778
4779 4. Tidied pcretest to ensure it frees memory that it gets.
4780
4781 5. Temporary memory was being obtained in the case where the passed offsets
4782 vector was exactly big enough.
4783
4784 6. Corrected definition of offsetof() from change 5 below.
4785
4786 7. I had screwed up change 6 below and broken the rules for the use of
4787 setjmp(). Now fixed.
4788
4789
4790 Version 1.03 18-Dec-97
4791 ----------------------
4792
4793 1. A erroneous regex with a missing opening parenthesis was correctly
4794 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4795 on some systems.
4796
4797 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4798 it was reported that one broken compiler failed on the former because "code" is
4799 also an independent variable.
4800
4801 3. The erroneous regex a[]b caused an array overrun reference.
4802
4803 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4804 fail on data ending with that character. (It was going on too far, and checking
4805 the next character, typically a binary zero.) This was specific to the
4806 optimized code for single-character negative classes.
4807
4808 5. Added a contributed patch from the TIN world which does the following:
4809
4810 + Add an undef for memmove, in case the the system defines a macro for it.
4811
4812 + Add a definition of offsetof(), in case there isn't one. (I don't know
4813 the reason behind this - offsetof() is part of the ANSI standard - but
4814 it does no harm).
4815
4816 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4817 most of the places where whitespace preceded '#'. I have given up and
4818 allowed the remaining 2 cases to be at the margin.
4819
4820 + Rename some variables in pcre to eliminate shadowing. This seems very
4821 pedantic, but does no harm, of course.
4822
4823 6. Moved the call to setjmp() into its own function, to get rid of warnings
4824 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4825
4826 7. Constructs such as \d{8,} were compiling into the equivalent of
4827 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4828 outcome, but in this particular case used more store than had been allocated,
4829 which caused the bug to be discovered because it threw up an internal error.
4830
4831 8. The debugging code in both pcre and pcretest for outputting the compiled
4832 form of a regex was going wrong in the case of back references followed by
4833 curly-bracketed repeats.
4834
4835
4836 Version 1.02 12-Dec-97
4837 ----------------------
4838
4839 1. Typos in pcre.3 and comments in the source fixed.
4840
4841 2. Applied a contributed patch to get rid of places where it used to remove
4842 'const' from variables, and fixed some signed/unsigned and uninitialized
4843 variable warnings.
4844
4845 3. Added the "runtest" target to Makefile.
4846
4847 4. Set default compiler flag to -O2 rather than just -O.
4848
4849
4850 Version 1.01 19-Nov-97
4851 ----------------------
4852
4853 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4854 like /([ab]*)*/, that is, for classes with more than one character in them.
4855
4856 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4857 as /((?>a*))*/ (a PCRE_EXTRA facility).
4858
4859
4860 Version 1.00 18-Nov-97
4861 ----------------------
4862
4863 1. Added compile-time macros to support systems such as SunOS4 which don't have
4864 memmove() or strerror() but have other things that can be used instead.
4865
4866 2. Arranged that "make clean" removes the executables.
4867
4868
4869 Version 0.99 27-Oct-97
4870 ----------------------
4871
4872 1. Fixed bug in code for optimizing classes with only one character. It was
4873 initializing a 32-byte map regardless, which could cause it to run off the end
4874 of the memory it had got.
4875
4876 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4877
4878
4879 Version 0.98 22-Oct-97
4880 ----------------------
4881
4882 1. Fixed bug in code for handling temporary memory usage when there are more
4883 back references than supplied space in the ovector. This could cause segfaults.
4884
4885
4886 Version 0.97 21-Oct-97
4887 ----------------------
4888
4889 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4890
4891 2. Optimized negated single characters not to use a bit map.
4892
4893 3. Brought error texts together as macro definitions; clarified some of them;
4894 fixed one that was wrong - it said "range out of order" when it meant "invalid
4895 escape sequence".
4896
4897 4. Changed some char * arguments to const char *.
4898
4899 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4900
4901 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4902 pcretest.
4903
4904
4905 Version 0.96 16-Oct-97
4906 ----------------------
4907
4908 1. Added a simple "pgrep" utility to the distribution.
4909
4910 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4911 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4912 where "ddd" means "one or more decimal digits".
4913
4914 3. Fixed serious bug. If a pattern had a back reference, but the call to
4915 pcre_exec() didn't supply a large enough ovector to record the related
4916 identifying subpattern, the match always failed. PCRE now remembers the number
4917 of the largest back reference, and gets some temporary memory in which to save
4918 the offsets during matching if necessary, in order to ensure that
4919 backreferences always work.
4920
4921 4. Increased the compatibility with Perl in a number of ways:
4922
4923 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4924 to request this handling. The option can be set at compile or exec time.
4925
4926 (b) $ matches before a terminating newline by default; an option
4927 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4928 mode). The option can be set at compile or exec time.
4929
4930 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4931 the same as Perl's. If the decimal number it represents is less than 10
4932 or there aren't that many previous left capturing parentheses, an octal
4933 escape is read. Inside a character class, it's always an octal escape,
4934 even if it is a single digit.
4935
4936 (d) An escaped but undefined alphabetic character is taken as a literal,
4937 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4938 escapes.
4939
4940 (e) {0} is now permitted. (The previous item is removed from the compiled
4941 pattern).
4942
4943 5. Changed all the names of code files so that the basic parts are no longer
4944 than 10 characters, and abolished the teeny "globals.c" file.
4945
4946 6. Changed the handling of character classes; they are now done with a 32-byte
4947 bit map always.
4948
4949 7. Added the -d and /D options to pcretest to make it possible to look at the
4950 internals of compilation without having to recompile pcre.
4951
4952
4953 Version 0.95 23-Sep-97
4954 ----------------------
4955
4956 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4957 \x20 at the start of a run of normal characters. These were being treated as
4958 real characters, instead of the source characters being re-checked.
4959
4960
4961 Version 0.94 18-Sep-97
4962 ----------------------
4963
4964 1. The functions are now thread-safe, with the caveat that the global variables
4965 containing pointers to malloc() and free() or alternative functions are the
4966 same for all threads.
4967
4968 2. Get pcre_study() to generate a bitmap of initial characters for non-
4969 anchored patterns when this is possible, and use it if passed to pcre_exec().
4970
4971
4972 Version 0.93 15-Sep-97
4973 ----------------------
4974
4975 1. /(b)|(:+)/ was computing an incorrect first character.
4976
4977 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4978 but not actually doing anything yet.
4979
4980 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4981 as Perl does (e.g. [-az] or [az-]).
4982
4983 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4984 all possible positions.
4985
4986 5. Split up into different modules to avoid including unneeded functions in a
4987 compiled binary. However, compile and exec are still in one module. The "study"
4988 function is split off.
4989
4990 6. The character tables are now in a separate module whose source is generated
4991 by an auxiliary program - but can then be edited by hand if required. There are
4992 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4993 toupper() in the code.
4994
4995 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4996 make them global. Abolish the function for setting them, as the caller can now
4997 set them directly.
4998
4999
5000 Version 0.92 11-Sep-97
5001 ----------------------
5002
5003 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5004 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5005
5006 2. Caseless matching was not working in character classes if the characters in
5007 the pattern were in upper case.
5008
5009 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5010
5011 4. Make PCRE_ANCHORED public and accept as a compile option.
5012
5013 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5014 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5015 pass them.
5016
5017 6. Give an error if bad option bits passed at compile or run time.
5018
5019 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5020 pcretest to cause it to pass that flag.
5021
5022 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5023 options, and the first character, if set.
5024
5025 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5026
5027
5028 Version 0.91 10-Sep-97
5029 ----------------------
5030
5031 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5032 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5033
5034 2. PCRE was looping on encountering an indefinitely repeated back reference to
5035 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5036 Perl does - treats the match as successful.
5037
5038 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5