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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1233 - (show annotations)
Sat Dec 8 12:43:37 2012 UTC (7 years ago) by ph10
File size: 240148 byte(s)
Error occurred while calculating annotation data.
Fix some Microsoft compiler warnings.
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
13 Version 8.32 30-November-2012
14 -----------------------------
15
16 1. Improved JIT compiler optimizations for first character search and single
17 character iterators.
18
19 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
20 Patch by Daniel Richard G.
21
22 3. Single character iterator optimizations in the JIT compiler.
23
24 4. Improved JIT compiler optimizations for character ranges.
25
26 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
27 Reported by Giuseppe D'Angelo.
28
29 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
30 of a line, was being set incorrectly in cases where .* appeared inside
31 atomic brackets at the start of a pattern, or where there was a subsequent
32 *PRUNE or *SKIP.
33
34 7. Improved instruction cache flush for POWER/PowerPC.
35 Patch by Daniel Richard G.
36
37 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
38 grep:
39
40 (a) There is now no limit to the number of patterns to be matched.
41
42 (b) An error is given if a pattern is too long.
43
44 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
45 are now supported.
46
47 (d) --exclude-from and --include-from (multiple use) have been added.
48
49 (e) Exclusions and inclusions now apply to all files and directories, not
50 just to those obtained from scanning a directory recursively.
51
52 (f) Multiple uses of -f and --file-list are now supported.
53
54 (g) In a Windows environment, the default for -d has been changed from
55 "read" (the GNU grep default) to "skip", because otherwise the presence
56 of a directory in the file list provokes an error.
57
58 (h) The documentation has been revised and clarified in places.
59
60 9. Improve the matching speed of capturing brackets.
61
62 10. Changed the meaning of \X so that it now matches a Unicode extended
63 grapheme cluster.
64
65 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
66 out POSIX threads when JIT support is configured.
67
68 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
69
70 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
71 structure could go wrong in environments where size_t is not the same size
72 as int.
73
74 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
75 set.
76
77 15. The EBCDIC support had decayed; later updates to the code had included
78 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
79 general tidy up of EBCDIC-related issues, and the documentation was also
80 not quite right. There is now a test that can be run on ASCII systems to
81 check some of the EBCDIC-related things (but is it not a full test).
82
83 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
84 in a small tidy to the code.
85
86 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
87
88 18. If the --only-matching (-o) option in pcregrep is specified multiple
89 times, each one causes appropriate output. For example, -o1 -o2 outputs the
90 substrings matched by the 1st and 2nd capturing parentheses. A separating
91 string can be specified by --om-separator (default empty).
92
93 19. Improving the first n character searches.
94
95 20. Turn case lists for horizontal and vertical white space into macros so that
96 they are defined only once.
97
98 21. This set of changes together give more compatible Unicode case-folding
99 behaviour for characters that have more than one other case when UCP
100 support is available.
101
102 (a) The Unicode property table now has offsets into a new table of sets of
103 three or more characters that are case-equivalent. The MultiStage2.py
104 script that generates these tables (the pcre_ucd.c file) now scans
105 CaseFolding.txt instead of UnicodeData.txt for character case
106 information.
107
108 (b) The code for adding characters or ranges of characters to a character
109 class has been abstracted into a generalized function that also handles
110 case-independence. In UTF-mode with UCP support, this uses the new data
111 to handle characters with more than one other case.
112
113 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
114 whose other case is greater than 256 are now correctly matched
115 caselessly. Previously, the high codepoint matched the low one, but not
116 vice versa.
117
118 (d) The processing of \h, \H, \v, and \ in character classes now makes use
119 of the new class addition function, using character lists defined as
120 macros alongside the case definitions of 20 above.
121
122 (e) Caseless back references now work with characters that have more than
123 one other case.
124
125 (f) General caseless matching of characters with more than one other case
126 is supported.
127
128 22. Unicode character properties were updated from Unicode 6.2.0
129
130 23. Improved CMake support under Windows. Patch by Daniel Richard G.
131
132 24. Add support for 32-bit character strings, and UTF-32
133
134 25. Major JIT compiler update (code refactoring and bugfixing).
135 Experimental Sparc 32 support is added.
136
137 26. Applied a modified version of Daniel Richard G's patch to create
138 pcre.h.generic and config.h.generic by "make" instead of in the
139 PrepareRelease script.
140
141 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
142 pcre_compile.c when checking for a zero character.
143
144 28. Introducing a native interface for JIT. Through this interface, the compiled
145 machine code can be directly executed. The purpose of this interface is to
146 provide fast pattern matching, so several sanity checks are not performed.
147 However, feature tests are still performed. The new interface provides
148 1.4x speedup compared to the old one.
149
150 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
151 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
152 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
153
154 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
155 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
156 the "old" RFC 2279). Instead, it ended up passing a negative length to
157 pcre_exec().
158
159 31. Add support for GCC's visibility feature to hide internal functions.
160
161 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
162 "unknown -C option" after outputting 0 or 1.
163
164 33. There is now support for generating a code coverage report for the test
165 suite in environments where gcc is the compiler and lcov is installed. This
166 is mainly for the benefit of the developers.
167
168 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
169 unaddressable using valgrind annotations, allowing valgrind to detect
170 invalid memory accesses. This is mainly for the benefit of the developers.
171
172 25. (*UTF) can now be used to start a pattern in any of the three libraries.
173
174 26. Give configure error if --enable-cpp but no C++ compiler found.
175
176
177 Version 8.31 06-July-2012
178 -------------------------
179
180 1. Fixing a wrong JIT test case and some compiler warnings.
181
182 2. Removed a bashism from the RunTest script.
183
184 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
185 to unsigned type, result still unsigned" that was given by an MS compiler
186 on encountering the code "-sizeof(xxx)".
187
188 4. Partial matching support is added to the JIT compiler.
189
190 5. Fixed several bugs concerned with partial matching of items that consist
191 of more than one character:
192
193 (a) /^(..)\1/ did not partially match "aba" because checking references was
194 done on an "all or nothing" basis. This also applied to repeated
195 references.
196
197 (b) \R did not give a hard partial match if \r was found at the end of the
198 subject.
199
200 (c) \X did not give a hard partial match after matching one or more
201 characters at the end of the subject.
202
203 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
204 a partial match for the string "\r".
205
206 (e) When newline was set to CRLF, the metacharacter "." did not recognize
207 a partial match for a CR character at the end of the subject string.
208
209 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
210 running pcretest, the text "(JIT)" added to the output whenever JIT is
211 actually used to run the match.
212
213 7. Individual JIT compile options can be set in pcretest by following -s+[+]
214 or /S+[+] with a digit between 1 and 7.
215
216 8. OP_NOT now supports any UTF character not just single-byte ones.
217
218 9. (*MARK) control verb is now supported by the JIT compiler.
219
220 10. The command "./RunTest list" lists the available tests without actually
221 running any of them. (Because I keep forgetting what they all are.)
222
223 11. Add PCRE_INFO_MAXLOOKBEHIND.
224
225 12. Applied a (slightly modified) user-supplied patch that improves performance
226 when the heap is used for recursion (compiled with --disable-stack-for-
227 recursion). Instead of malloc and free for each heap frame each time a
228 logical recursion happens, frames are retained on a chain and re-used where
229 possible. This sometimes gives as much as 30% improvement.
230
231 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
232 call.
233
234 14. As documented, (*COMMIT) is now confined to within a positive assertion.
235
236 15. It is now possible to link pcretest with libedit as an alternative to
237 libreadline.
238
239 16. (*COMMIT) control verb is now supported by the JIT compiler.
240
241 17. The Unicode data tables have been updated to Unicode 6.1.0.
242
243 18. Added --file-list option to pcregrep.
244
245 19. Added binary file support to pcregrep, including the -a, --binary-files,
246 -I, and --text options.
247
248 20. The madvise function is renamed for posix_madvise for QNX compatibility
249 reasons. Fixed by Giuseppe D'Angelo.
250
251 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
252
253 22. Changed the option for creating symbolic links for 16-bit man pages from
254 -s to -sf so that re-installing does not cause issues.
255
256 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
257
258 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
259 matches in certain environments (the workspace was not being correctly
260 retained). Also added to pcre_dfa_exec() a simple plausibility check on
261 some of the workspace data at the beginning of a restart.
262
263 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
264 was not doing so when it should - probably a typo introduced by SVN 528
265 (change 8.10/14).
266
267 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
268 \w+ when the character tables indicated that \x{c4} was a word character.
269 There were several related cases, all because the tests for doing a table
270 lookup were testing for characters less than 127 instead of 255.
271
272 27. If a pattern contains capturing parentheses that are not used in a match,
273 their slots in the ovector are set to -1. For those that are higher than
274 any matched groups, this happens at the end of processing. In the case when
275 there were back references that the ovector was too small to contain
276 (causing temporary malloc'd memory to be used during matching), and the
277 highest capturing number was not used, memory off the end of the ovector
278 was incorrectly being set to -1. (It was using the size of the temporary
279 memory instead of the true size.)
280
281 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
282 ovector size, it uses memory at the end of the block that it has got.
283
284 29. Check for an overlong MARK name and give an error at compile time. The
285 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
286
287 30. JIT compiler update.
288
289 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
290 Rill for the patch.
291
292 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
293
294 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
295
296 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
297 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
298 was enabled.
299
300 35. Improve JIT code generation for greedy plus quantifier.
301
302 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
303 1 to "aa" instead of to an empty string. The bug affected repeated groups
304 that could potentially match an empty string.
305
306 37. Optimizing single character iterators in JIT.
307
308 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
309 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
310 codepoints that are too big for the mode are faulted, and in a UTF mode,
311 disallowed codepoints are also faulted.
312
313 39. If PCRE was compiled with UTF support, in three places in the DFA
314 matcher there was code that should only have been obeyed in UTF mode, but
315 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
316 processing when bytes with values greater than 127 were present. In 16-bit
317 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
318 both cases the values are those that cannot be the first data item in a UTF
319 character. The three items that might have provoked this were recursions,
320 possessively repeated groups, and atomic groups.
321
322 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
323 and pcregrep, because some OS require shared objects to be explicitly
324 passed to ld, causing the link step to fail if they are not.
325
326 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
327 mode, patterns that started with \h* or \R* might be incorrectly matched.
328
329
330 Version 8.30 04-February-2012
331 -----------------------------
332
333 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
334 name is defined in ctype.h.
335
336 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
337 only in quite long subpatterns.
338
339 3. Removed the function pcre_info(), which has been obsolete and deprecated
340 since it was replaced by pcre_fullinfo() in February 2000.
341
342 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
343 match a (*MARK), and the match failed at the start of the subject, a
344 reference to memory before the start of the subject could occur. This bug
345 was introduced by fix 17 of release 8.21.
346
347 5. A reference to an unset group with zero minimum repetition was giving
348 totally wrong answers (in non-JavaScript-compatibility mode). For example,
349 /(another)?(\1?)test/ matched against "hello world test". This bug was
350 introduced in release 8.13.
351
352 6. Add support for 16-bit character strings (a large amount of work involving
353 many changes and refactorings).
354
355 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
356 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
357 from a file.
358
359 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
360 rounding is not applied in this particular case).
361
362 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
363 if they appear, or are escaped, in patterns.
364
365 10. Get rid of a number of -Wunused-but-set-variable warnings.
366
367 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
368 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
369 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
370 also returns the mark "x". This bug applied to capturing parentheses,
371 non-capturing parentheses, and atomic parentheses. It also applied to some
372 assertions.
373
374 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
375 information out of configure.ac instead of relying on pcre.h.generic, which
376 is not stored in the repository.
377
378 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
379 -lreadline.
380
381 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
382
383 15. Applied Graycode's patch to put the top-level frame on the stack rather
384 than the heap when not using the stack for recursion. This gives a
385 performance improvement in many cases when recursion is not deep.
386
387 16. Experimental code added to "pcretest -C" to output the stack frame size.
388
389
390 Version 8.21 12-Dec-2011
391 ------------------------
392
393 1. Updating the JIT compiler.
394
395 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
396 are added as well.
397
398 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
399 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
400 calling _pcre_jit_exec. Some extra comments are added.
401
402 4. (*MARK) settings inside atomic groups that do not contain any capturing
403 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
404 was introduced by change 18 for 8.20.
405
406 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
407 ECMA-262 standard.
408
409 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
410 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
411 This bug was probably introduced by change 9 of 8.13.
412
413 7. While fixing 6 above, I noticed that a number of other items were being
414 incorrectly rejected as "not fixed length". This arose partly because newer
415 opcodes had not been added to the fixed-length checking code. I have (a)
416 corrected the bug and added tests for these items, and (b) arranged for an
417 error to occur if an unknown opcode is encountered while checking for fixed
418 length instead of just assuming "not fixed length". The items that were
419 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
420 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
421 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
422
423 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
424 being incorrectly compiled and would have given unpredicatble results.
425
426 9. A possessively repeated subpattern with minimum repeat count greater than
427 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
428 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
429 the first (A) could occur when it should not.
430
431 10. Add a cast and remove a redundant test from the code.
432
433 11. JIT should use pcre_malloc/pcre_free for allocation.
434
435 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
436 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
437 is anything other than /usr, -L is still shown).
438
439 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
440
441 14. Perl does not support \N without a following name in a [] class; PCRE now
442 also gives an error.
443
444 15. If a forward reference was repeated with an upper limit of around 2000,
445 it caused the error "internal error: overran compiling workspace". The
446 maximum number of forward references (including repeats) was limited by the
447 internal workspace, and dependent on the LINK_SIZE. The code has been
448 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
449 the default depends on LINK_SIZE. There is a new upper limit (for safety)
450 of around 200,000 forward references. While doing this, I also speeded up
451 the filling in of repeated forward references.
452
453 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
454 incorrectly expecting the subject to contain another "a" after the start.
455
456 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
457 in the match, the SKIP should be ignored. This was not happening; instead
458 the SKIP was being treated as NOMATCH. For patterns such as
459 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
460 tested.
461
462 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
463 now much more compatible with Perl, in particular in cases where the result
464 is a non-match for a non-anchored pattern. For example, if
465 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
466 "m", where previously it did not return a name. A side effect of this
467 change is that for partial matches, the last encountered mark name is
468 returned, as for non matches. A number of tests that were previously not
469 Perl-compatible have been moved into the Perl-compatible test files. The
470 refactoring has had the pleasing side effect of removing one argument from
471 the match() function, thus reducing its stack requirements.
472
473 19. If the /S+ option was used in pcretest to study a pattern using JIT,
474 subsequent uses of /S (without +) incorrectly behaved like /S+.
475
476 21. Retrieve executable code size support for the JIT compiler and fixing
477 some warnings.
478
479 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
480 not work when the shorter character appeared right at the end of the
481 subject string.
482
483 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
484 systems.
485
486 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
487 output it when the /M option is used in pcretest.
488
489 25. The CheckMan script was not being included in the distribution. Also, added
490 an explicit "perl" to run Perl scripts from the PrepareRelease script
491 because this is reportedly needed in Windows.
492
493 26. If study data was being save in a file and studying had not found a set of
494 "starts with" bytes for the pattern, the data written to the file (though
495 never used) was taken from uninitialized memory and so caused valgrind to
496 complain.
497
498 27. Updated RunTest.bat as provided by Sheri Pierce.
499
500 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
501
502 29. Computation of memory usage for the table of capturing group names was
503 giving an unnecessarily large value.
504
505
506 Version 8.20 21-Oct-2011
507 ------------------------
508
509 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
510 a POSIX class. After further experiments with Perl, which convinced me that
511 Perl has bugs and confusions, a closing square bracket is no longer allowed
512 in a POSIX name. This bug also affected patterns with classes that started
513 with full stops.
514
515 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
516 captured substring, but while checking the failing first alternative,
517 substring 1 is temporarily captured. If the output vector supplied to
518 pcre_exec() was not big enough for this capture, the yield of the function
519 was still zero ("insufficient space for captured substrings"). This cannot
520 be totally fixed without adding another stack variable, which seems a lot
521 of expense for a edge case. However, I have improved the situation in cases
522 such as /(a)(b)x|abc/ matched against "abc", where the return code
523 indicates that fewer than the maximum number of slots in the ovector have
524 been set.
525
526 3. Related to (2) above: when there are more back references in a pattern than
527 slots in the output vector, pcre_exec() uses temporary memory during
528 matching, and copies in the captures as far as possible afterwards. It was
529 using the entire output vector, but this conflicts with the specification
530 that only 2/3 is used for passing back captured substrings. Now it uses
531 only the first 2/3, for compatibility. This is, of course, another edge
532 case.
533
534 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
535 main code base, and can be used by building with --enable-jit. When this is
536 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
537 runtime --no-jit option is given.
538
539 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
540 ovector, the return from the function was zero, implying that there were
541 other matches that did not fit. The correct "exactly full" value is now
542 returned.
543
544 6. If a subpattern that was called recursively or as a subroutine contained
545 (*PRUNE) or any other control that caused it to give a non-standard return,
546 invalid errors such as "Error -26 (nested recursion at the same subject
547 position)" or even infinite loops could occur.
548
549 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
550 computing the minimum length on reaching *ACCEPT, and so ended up with the
551 wrong value of 1 rather than 0. Further investigation indicates that
552 computing a minimum subject length in the presence of *ACCEPT is difficult
553 (think back references, subroutine calls), and so I have changed the code
554 so that no minimum is registered for a pattern that contains *ACCEPT.
555
556 8. If (*THEN) was present in the first (true) branch of a conditional group,
557 it was not handled as intended. [But see 16 below.]
558
559 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
560 Sheri Pierce.
561
562 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
563 the first byte in a match must be "a".
564
565 11. Change 17 for 8.13 increased the recursion depth for patterns like
566 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
567 pattern contains any instances of (*THEN). If it does not, the old
568 optimizations are restored. It would be nice to do this on a per-group
569 basis, but at the moment that is not feasible.
570
571 12. In some environments, the output of pcretest -C is CRLF terminated. This
572 broke RunTest's code that checks for the link size. A single white space
573 character after the value is now allowed for.
574
575 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
576 For "fr", it uses the Windows-specific input and output files.
577
578 14. If (*THEN) appeared in a group that was called recursively or as a
579 subroutine, it did not work as intended. [But see next item.]
580
581 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
582 pattern fragments (but not containing any | characters). If A and B are
583 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
584 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
585 D. In other words, Perl considers parentheses that do not contain any |
586 characters to be part of a surrounding alternative, whereas PCRE was
587 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
588 differently. PCRE now behaves in the same way as Perl, except in the case
589 of subroutine/recursion calls such as (?1) which have in any case always
590 been different (but PCRE had them first :-).
591
592 16. Related to 15 above: Perl does not treat the | in a conditional group as
593 creating alternatives. Such a group is treated in the same way as an
594 ordinary group without any | characters when processing (*THEN). PCRE has
595 been changed to match Perl's behaviour.
596
597 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
598 RunGrepTest script failed.
599
600 18. Change 22 for version 13 caused atomic groups to use more stack. This is
601 inevitable for groups that contain captures, but it can lead to a lot of
602 stack use in large patterns. The old behaviour has been restored for atomic
603 groups that do not contain any capturing parentheses.
604
605 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
606 suppress the check for a minimum subject length at run time. (If it was
607 given to pcre_exec() or pcre_dfa_exec() it did work.)
608
609 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
610 fail to work when decoding hex characters in data strings in EBCDIC
611 environments.
612
613 21. It appears that in at least one Mac OS environment, the isxdigit() function
614 is implemented as a macro that evaluates to its argument more than once,
615 contravening the C 90 Standard (I haven't checked a later standard). There
616 was an instance in pcretest which caused it to go wrong when processing
617 \x{...} escapes in subject strings. The has been rewritten to avoid using
618 things like p++ in the argument of isxdigit().
619
620
621 Version 8.13 16-Aug-2011
622 ------------------------
623
624 1. The Unicode data tables have been updated to Unicode 6.0.0.
625
626 2. Two minor typos in pcre_internal.h have been fixed.
627
628 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
629 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
630 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
631
632 4. There were a number of related bugs in the code for matching backrefences
633 caselessly in UTF-8 mode when codes for the characters concerned were
634 different numbers of bytes. For example, U+023A and U+2C65 are an upper
635 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
636 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
637 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
638 2-byte code at the end of the subject (it thought there wasn't enough data
639 left).
640
641 5. Comprehensive information about what went wrong is now returned by
642 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
643 as the output vector has at least 2 elements. The offset of the start of
644 the failing character and a reason code are placed in the vector.
645
646 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
647 now returned is for the first byte of the failing character, instead of the
648 last byte inspected. This is an incompatible change, but I hope it is small
649 enough not to be a problem. It makes the returned offset consistent with
650 pcre_exec() and pcre_dfa_exec().
651
652 7. pcretest now gives a text phrase as well as the error number when
653 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
654 failure, the offset and reason code are output.
655
656 8. When \R was used with a maximizing quantifier it failed to skip backwards
657 over a \r\n pair if the subsequent match failed. Instead, it just skipped
658 back over a single character (\n). This seems wrong (because it treated the
659 two characters as a single entity when going forwards), conflicts with the
660 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
661 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
662 has been changed.
663
664 9. Some internal refactoring has changed the processing so that the handling
665 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
666 time (the PCRE_DOTALL option was changed this way some time ago: version
667 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
668 which was always a bit of a fudge. It also means that there is one less
669 argument for the match() function, which reduces its stack requirements
670 slightly. This change also fixes an incompatibility with Perl: the pattern
671 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
672
673 10. More internal refactoring has drastically reduced the number of recursive
674 calls to match() for possessively repeated groups such as (abc)++ when
675 using pcre_exec().
676
677 11. While implementing 10, a number of bugs in the handling of groups were
678 discovered and fixed:
679
680 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
681 (a|)*(?1) gave a compile-time internal error.
682 ((a|)+)+ did not notice that the outer group could match an empty string.
683 (^a|^)+ was not marked as anchored.
684 (.*a|.*)+ was not marked as matching at start or after a newline.
685
686 12. Yet more internal refactoring has removed another argument from the match()
687 function. Special calls to this function are now indicated by setting a
688 value in a variable in the "match data" data block.
689
690 13. Be more explicit in pcre_study() instead of relying on "default" for
691 opcodes that mean there is no starting character; this means that when new
692 ones are added and accidentally left out of pcre_study(), testing should
693 pick them up.
694
695 14. The -s option of pcretest has been documented for ages as being an old
696 synonym of -m (show memory usage). I have changed it to mean "force study
697 for every regex", that is, assume /S for every regex. This is similar to -i
698 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
699 using it. It makes it easier to run collections of tests with and without
700 study enabled, and thereby test pcre_study() more easily. All the standard
701 tests are now run with and without -s (but some patterns can be marked as
702 "never study" - see 20 below).
703
704 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
705 restoration of the capturing data to the outer values was not happening
706 correctly.
707
708 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
709 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
710 pattern had matched an empty string, and so incorrectly returned a no
711 match.
712
713 17. There was optimizing code for the last branch of non-capturing parentheses,
714 and also for the obeyed branch of a conditional subexpression, which used
715 tail recursion to cut down on stack usage. Unfortunately, now that there is
716 the possibility of (*THEN) occurring in these branches, tail recursion is
717 no longer possible because the return has to be checked for (*THEN). These
718 two optimizations have therefore been removed. [But see 8.20/11 above.]
719
720 18. If a pattern containing \R was studied, it was assumed that \R always
721 matched two bytes, thus causing the minimum subject length to be
722 incorrectly computed because \R can also match just one byte.
723
724 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
725 was incorrectly computed.
726
727 20. If /S is present twice on a test pattern in pcretest input, it now
728 *disables* studying, thereby overriding the use of -s on the command line
729 (see 14 above). This is necessary for one or two tests to keep the output
730 identical in both cases.
731
732 21. When (*ACCEPT) was used in an assertion that matched an empty string and
733 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
734
735 22. When an atomic group that contained a capturing parenthesis was
736 successfully matched, but the branch in which it appeared failed, the
737 capturing was not being forgotten if a higher numbered group was later
738 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
739 group 1 to "a", when in fact it should be unset. This applied to multi-
740 branched capturing and non-capturing groups, repeated or not, and also to
741 positive assertions (capturing in negative assertions does not happen
742 in PCRE) and also to nested atomic groups.
743
744 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
745 subject after a captured substring, to make it easier to tell which of a
746 number of identical substrings has been captured.
747
748 24. The way atomic groups are processed by pcre_exec() has been changed so that
749 if they are repeated, backtracking one repetition now resets captured
750 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
751 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
752 "aaa". Previously, it would have been "a". As part of this code
753 refactoring, the way recursive calls are handled has also been changed.
754
755 25. If an assertion condition captured any substrings, they were not passed
756 back unless some other capturing happened later. For example, if
757 (?(?=(a))a) was matched against "a", no capturing was returned.
758
759 26. When studying a pattern that contained subroutine calls or assertions,
760 the code for finding the minimum length of a possible match was handling
761 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
762 group 1 called group 2 while simultaneously a separate group 2 called group
763 1). A stack overflow occurred in this case. I have fixed this by limiting
764 the recursion depth to 10.
765
766 27. Updated RunTest.bat in the distribution to the version supplied by Tom
767 Fortmann. This supports explicit test numbers on the command line, and has
768 argument validation and error reporting.
769
770 28. An instance of \X with an unlimited repeat could fail if at any point the
771 first character it looked at was a mark character.
772
773 29. Some minor code refactoring concerning Unicode properties and scripts
774 should reduce the stack requirement of match() slightly.
775
776 30. Added the '=' option to pcretest to check the setting of unused capturing
777 slots at the end of the pattern, which are documented as being -1, but are
778 not included in the return count.
779
780 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
781 compiled something random. Now it gives a compile-time error (as does
782 Perl).
783
784 32. A *MARK encountered during the processing of a positive assertion is now
785 recorded and passed back (compatible with Perl).
786
787 33. If --only-matching or --colour was set on a pcregrep call whose pattern
788 had alternative anchored branches, the search for a second match in a line
789 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
790 matched the line "0102" twice. The same bug affected patterns that started
791 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
792 twice.
793
794 34. Previously, PCRE did not allow quantification of assertions. However, Perl
795 does, and because of capturing effects, quantifying parenthesized
796 assertions may at times be useful. Quantifiers are now allowed for
797 parenthesized assertions.
798
799 35. A minor code tidy in pcre_compile() when checking options for \R usage.
800
801 36. \g was being checked for fancy things in a character class, when it should
802 just be a literal "g".
803
804 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
805 appearance of a nested POSIX class supersedes an apparent external class.
806 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
807 unescaped square brackets may also appear as part of class names. For
808 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
809 more like Perl. (But see 8.20/1 above.)
810
811 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
812 was because it thought it was \N{name}, which is not supported).
813
814 39. Add minix to OS list not supporting the -S option in pcretest.
815
816 40. PCRE tries to detect cases of infinite recursion at compile time, but it
817 cannot analyze patterns in sufficient detail to catch mutual recursions
818 such as ((?1))((?2)). There is now a runtime test that gives an error if a
819 subgroup is called recursively as a subpattern for a second time at the
820 same position in the subject string. In previous releases this might have
821 been caught by the recursion limit, or it might have run out of stack.
822
823 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
824 happen only once. PCRE was, however incorrectly giving a compile time error
825 "recursive call could loop indefinitely" because it cannot analyze the
826 pattern in sufficient detail. The compile time test no longer happens when
827 PCRE is compiling a conditional subpattern, but actual runaway loops are
828 now caught at runtime (see 40 above).
829
830 42. It seems that Perl allows any characters other than a closing parenthesis
831 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
832 has been changed to be the same.
833
834 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
835 as not to get warnings when autogen.sh is called. Also changed
836 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
837
838 44. To help people who use pcregrep to scan files containing exceedingly long
839 lines, the following changes have been made:
840
841 (a) The default value of the buffer size parameter has been increased from
842 8K to 20K. (The actual buffer used is three times this size.)
843
844 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
845 PCRE is built.
846
847 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
848 to be set at run time.
849
850 (d) Numerical values in pcregrep options can be followed by K or M, for
851 example --buffer-size=50K.
852
853 (e) If a line being scanned overflows pcregrep's buffer, an error is now
854 given and the return code is set to 2.
855
856 45. Add a pointer to the latest mark to the callout data block.
857
858 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
859 partial match of an empty string instead of no match. This was specific to
860 the use of ".".
861
862 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
863 complete match instead of a partial match. This bug was dependent on both
864 the PCRE_UTF8 and PCRE_DOTALL options being set.
865
866 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
867 starting byte set, because \b was not being ignored.
868
869
870 Version 8.12 15-Jan-2011
871 ------------------------
872
873 1. Fixed some typos in the markup of the man pages, and wrote a script that
874 checks for such things as part of the documentation building process.
875
876 2. On a big-endian 64-bit system, pcregrep did not correctly process the
877 --match-limit and --recursion-limit options (added for 8.11). In
878 particular, this made one of the standard tests fail. (The integer value
879 went into the wrong half of a long int.)
880
881 3. If the --colour option was given to pcregrep with -v (invert match), it
882 did strange things, either producing crazy output, or crashing. It should,
883 of course, ignore a request for colour when reporting lines that do not
884 match.
885
886 4. Another pcregrep bug caused similar problems if --colour was specified with
887 -M (multiline) and the pattern match finished with a line ending.
888
889 5. In pcregrep, when a pattern that ended with a literal newline sequence was
890 matched in multiline mode, the following line was shown as part of the
891 match. This seems wrong, so I have changed it.
892
893 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
894 the check for further matches in the same line (so they could be coloured)
895 to overrun the end of the current line. If another match was found, it was
896 incorrectly shown (and then shown again when found in the next line).
897
898 7. If pcregrep was compiled under Windows, there was a reference to the
899 function pcregrep_exit() before it was defined. I am assuming this was
900 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
901 reported by a user. I've moved the definition above the reference.
902
903
904 Version 8.11 10-Dec-2010
905 ------------------------
906
907 1. (*THEN) was not working properly if there were untried alternatives prior
908 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
909 backtracked to try for "b" instead of moving to the next alternative branch
910 at the same level (in this case, to look for "c"). The Perl documentation
911 is clear that when (*THEN) is backtracked onto, it goes to the "next
912 alternative in the innermost enclosing group".
913
914 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
915 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
916 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
917 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
918 (*THEN).
919
920 3. If \s appeared in a character class, it removed the VT character from
921 the class, even if it had been included by some previous item, for example
922 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
923 of \s, but is part of the POSIX "space" class.)
924
925 4. A partial match never returns an empty string (because you can always
926 match an empty string at the end of the subject); however the checking for
927 an empty string was starting at the "start of match" point. This has been
928 changed to the "earliest inspected character" point, because the returned
929 data for a partial match starts at this character. This means that, for
930 example, /(?<=abc)def/ gives a partial match for the subject "abc"
931 (previously it gave "no match").
932
933 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
934 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
935 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
936 has an implication that the given string is incomplete (because a partial
937 match is preferred over a full match). For this reason, these items now
938 give a partial match in this situation. [Aside: previously, the one case
939 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
940 match rather than a full match, which was wrong by the old rules, but is
941 now correct.]
942
943 6. There was a bug in the handling of #-introduced comments, recognized when
944 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
945 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
946 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
947 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
948 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
949 places in pcre_compile().
950
951 7. Related to (6) above, when pcre_compile() was skipping #-introduced
952 comments when looking ahead for named forward references to subpatterns,
953 the only newline sequence it recognized was NL. It now handles newlines
954 according to the set newline convention.
955
956 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
957 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
958 cater for a lack of strerror(). These oversights have been fixed.
959
960 9. Added --match-limit and --recursion-limit to pcregrep.
961
962 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
963
964 11. When the -o option was used, pcregrep was setting a return code of 1, even
965 when matches were found, and --line-buffered was not being honoured.
966
967 12. Added an optional parentheses number to the -o and --only-matching options
968 of pcregrep.
969
970 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
971 can match an empty string. The code to do it in pcretest and pcredemo
972 needed fixing:
973
974 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
975 only one byte after an empty string match just before CRLF (this case
976 just got forgotten; "any" and "anycrlf" were OK).
977
978 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
979 mode when an empty string match preceded an ASCII character followed by
980 a non-ASCII character. (The code for advancing by one character rather
981 than one byte was nonsense.)
982
983 (c) The pcredemo.c sample program did not have any code at all to handle
984 the cases when CRLF is a valid newline sequence.
985
986 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
987 as a starting offset was within the subject string. There is now a new
988 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
989 negative or greater than the length of the string. In order to test this,
990 pcretest is extended to allow the setting of negative starting offsets.
991
992 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
993 starting offset points to the beginning of a UTF-8 character was
994 unnecessarily clumsy. I tidied it up.
995
996 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
997 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
998
999 17. Nobody had reported that the --include_dir option, which was added in
1000 release 7.7 should have been called --include-dir (hyphen, not underscore)
1001 for compatibility with GNU grep. I have changed it to --include-dir, but
1002 left --include_dir as an undocumented synonym, and the same for
1003 --exclude-dir, though that is not available in GNU grep, at least as of
1004 release 2.5.4.
1005
1006 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1007 characters from a string of bytes) have been redefined so as not to use
1008 loops, in order to improve performance in some environments. At the same
1009 time, I abstracted some of the common code into auxiliary macros to save
1010 repetition (this should not affect the compiled code).
1011
1012 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1013 compile-time error is now given if \c is not followed by an ASCII
1014 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1015 different, and any byte value is allowed.)
1016
1017 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1018 START_OPTIMIZE option, which is now allowed at compile time - but just
1019 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1020 to pcregrep and other applications that have no direct access to PCRE
1021 options. The new /Y option in pcretest sets this option when calling
1022 pcre_compile().
1023
1024 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1025 back references. Groups containing recursive back references were forced to
1026 be atomic by that change, but in the case of named groups, the amount of
1027 memory required was incorrectly computed, leading to "Failed: internal
1028 error: code overflow". This has been fixed.
1029
1030 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1031 pcretest.c, to avoid build problems in some Borland environments.
1032
1033
1034 Version 8.10 25-Jun-2010
1035 ------------------------
1036
1037 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1038 THEN.
1039
1040 2. (*ACCEPT) was not working when inside an atomic group.
1041
1042 3. Inside a character class, \B is treated as a literal by default, but
1043 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1044 causes the error). The code is unchanged, but I tidied the documentation.
1045
1046 4. Inside a character class, PCRE always treated \R and \X as literals,
1047 whereas Perl faults them if its -w option is set. I have changed PCRE so
1048 that it faults them when PCRE_EXTRA is set.
1049
1050 5. Added support for \N, which always matches any character other than
1051 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1052
1053 6. When compiling pcregrep with newer versions of gcc which may have
1054 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1055 declared with attribute warn_unused_result" were given. Just casting the
1056 result to (void) does not stop the warnings; a more elaborate fudge is
1057 needed. I've used a macro to implement this.
1058
1059 7. Minor change to pcretest.c to avoid a compiler warning.
1060
1061 8. Added four artifical Unicode properties to help with an option to make
1062 \s etc use properties (see next item). The new properties are: Xan
1063 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1064
1065 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1066 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1067 this option. Modified pcretest to add /W to test this facility. Added
1068 REG_UCP to make it available via the POSIX interface.
1069
1070 10. Added --line-buffered to pcregrep.
1071
1072 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1073 studied, and the match started with a letter with a code point greater than
1074 127 whose first byte was different to the first byte of the other case of
1075 the letter, the other case of this starting letter was not recognized
1076 (#976).
1077
1078 12. If a pattern that was studied started with a repeated Unicode property
1079 test, for example, \p{Nd}+, there was the theoretical possibility of
1080 setting up an incorrect bitmap of starting bytes, but fortunately it could
1081 not have actually happened in practice until change 8 above was made (it
1082 added property types that matched character-matching opcodes).
1083
1084 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1085 possible starting bytes for non-anchored patterns.
1086
1087 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1088 \R, and also a number of cases that involve Unicode properties, both
1089 explicit and implicit when PCRE_UCP is set.
1090
1091 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1092 input, it could crash or give wrong results if characters with values
1093 greater than 0xc0 were present in the subject string. (Detail: it assumed
1094 UTF-8 input when processing these items.)
1095
1096 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1097 size_t is 64-bit (#991).
1098
1099 17. Added a check for running out of memory when PCRE is compiled with
1100 --disable-stack-for-recursion (#990).
1101
1102 18. If the last data line in a file for pcretest does not have a newline on
1103 the end, a newline was missing in the output.
1104
1105 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1106 less than 128) in its various bitmaps. However, there is a facility for
1107 generating tables according to the current locale when PCRE is compiled. It
1108 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1109 characters, are recognized by isspace() and therefore were getting set in
1110 these tables, and indeed these tables seem to approximate to ISO 8859. This
1111 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1112 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1113 which of course cannot start UTF-8 characters. I have changed the code so
1114 that only real ASCII characters (less than 128) and the correct starting
1115 bytes for UTF-8 encodings are set for characters greater than 127 when in
1116 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1117 altogether.)
1118
1119 20. Added the /T option to pcretest so as to be able to run tests with non-
1120 standard character tables, thus making it possible to include the tests
1121 used for 19 above in the standard set of tests.
1122
1123 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1124 reference to a subpattern the other side of a comment that contains an
1125 opening parenthesis caused either an internal compiling error, or a
1126 reference to the wrong subpattern.
1127
1128
1129 Version 8.02 19-Mar-2010
1130 ------------------------
1131
1132 1. The Unicode data tables have been updated to Unicode 5.2.0.
1133
1134 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1135 configured.
1136
1137 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1138 original author of that file, following a query about its status.
1139
1140 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1141 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1142
1143 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1144 quantifier applied to a forward-referencing subroutine call, could compile
1145 incorrect code or give the error "internal error: previously-checked
1146 referenced subpattern not found".
1147
1148 6. Both MS Visual Studio and Symbian OS have problems with initializing
1149 variables to point to external functions. For these systems, therefore,
1150 pcre_malloc etc. are now initialized to local functions that call the
1151 relevant global functions.
1152
1153 7. There were two entries missing in the vectors called coptable and poptable
1154 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1155 I've fixed the data, and added a kludgy way of testing at compile time that
1156 the lengths are correct (equal to the number of opcodes).
1157
1158 8. Following on from 7, I added a similar kludge to check the length of the
1159 eint vector in pcreposix.c.
1160
1161 9. Error texts for pcre_compile() are held as one long string to avoid too
1162 much relocation at load time. To find a text, the string is searched,
1163 counting zeros. There was no check for running off the end of the string,
1164 which could happen if a new error number was added without updating the
1165 string.
1166
1167 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1168
1169 11. \K was not working if it appeared in an atomic group or in a group that
1170 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1171 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1172 the assertion is positive, but not if it is negative.
1173
1174 12. Change 11 fortuitously reduced the size of the stack frame used in the
1175 "match()" function of pcre_exec.c by one pointer. Forthcoming
1176 implementation of support for (*MARK) will need an extra pointer on the
1177 stack; I have reserved it now, so that the stack frame size does not
1178 decrease.
1179
1180 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1181 item in branch that calls a recursion is a subroutine call - as in the
1182 second branch in the above example - was incorrectly given the compile-
1183 time error "recursive call could loop indefinitely" because pcre_compile()
1184 was not correctly checking the subroutine for matching a non-empty string.
1185
1186 14. The checks for overrunning compiling workspace could trigger after an
1187 overrun had occurred. This is a "should never occur" error, but it can be
1188 triggered by pathological patterns such as hundreds of nested parentheses.
1189 The checks now trigger 100 bytes before the end of the workspace.
1190
1191 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1192
1193
1194 Version 8.01 19-Jan-2010
1195 ------------------------
1196
1197 1. If a pattern contained a conditional subpattern with only one branch (in
1198 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1199 computed the wrong minimum data length (which is of course zero for such
1200 subpatterns). This could cause incorrect "no match" results.
1201
1202 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1203 the pattern is reset in the first branch, pcre_compile() failed with
1204 "internal error: code overflow at offset...". This happened only when
1205 the reset was to the original external option setting. (An optimization
1206 abstracts leading options settings into an external setting, which was the
1207 cause of this.)
1208
1209 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1210 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1211 assertion pattern did not match (meaning that the assertion was true), it
1212 was incorrectly treated as false if the SKIP had been reached during the
1213 matching. This also applied to assertions used as conditions.
1214
1215 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1216 assertion subpattern, including such a pattern used as a condition,
1217 unpredictable results occurred, instead of the error return
1218 PCRE_ERROR_DFA_UITEM.
1219
1220 5. The C++ GlobalReplace function was not working like Perl for the special
1221 situation when an empty string is matched. It now does the fancy magic
1222 stuff that is necessary.
1223
1224 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1225 removed. (These were left over from very, very early versions of PCRE.)
1226
1227 7. Some cosmetic changes to the code to make life easier when compiling it
1228 as part of something else:
1229
1230 (a) Change DEBUG to PCRE_DEBUG.
1231
1232 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1233 called "current" as "current_branch", to prevent a collision with the
1234 Linux macro when compiled as a kernel module.
1235
1236 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1237 prevent a collision with the Linux macro when compiled as a kernel
1238 module.
1239
1240 8. In pcre_compile() there are some checks for integer overflows that used to
1241 cast potentially large values to (double). This has been changed to that
1242 when building, a check for int64_t is made, and if it is found, it is used
1243 instead, thus avoiding the use of floating point arithmetic. (There is no
1244 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1245 double.
1246
1247 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1248 2005 (difference between two addresses compared to an unsigned value).
1249
1250 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1251 custom one, because of the following reported problem in Windows:
1252
1253 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1254 under Win32.
1255 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1256 therefore missing the function definition.
1257 - The compiler thus generates a "C" signature for the test function.
1258 - The linker fails to find the "C" function.
1259 - PCRE fails to configure if asked to do so against libbz2.
1260
1261 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1262 messages were output:
1263
1264 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1265 rerunning libtoolize, to keep the correct libtool macros in-tree.
1266 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1267
1268 I have done both of these things.
1269
1270 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1271 most of the time, it *can* run out if it is given a pattern that contains a
1272 runaway infinite recursion. I updated the discussion in the pcrestack man
1273 page.
1274
1275 13. Now that we have gone to the x.xx style of version numbers, the minor
1276 version may start with zero. Using 08 or 09 is a bad idea because users
1277 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1278 interpreted as invalid octal numbers. I've updated the previous comment in
1279 configure.ac, and also added a check that gives an error if 08 or 09 are
1280 used.
1281
1282 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1283 causing partial matching to fail when the end of the subject matched \W
1284 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1285
1286 15. There were some discrepancies between the declarations in pcre_internal.h
1287 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1288 their definitions. The declarations used "const uschar *" and the
1289 definitions used USPTR. Even though USPTR is normally defined as "const
1290 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1291 reported that: "This difference in casting confuses some C++ compilers, for
1292 example, SunCC recognizes above declarations as different functions and
1293 generates broken code for hbpcre." I have changed the declarations to use
1294 USPTR.
1295
1296 16. GNU libtool is named differently on some systems. The autogen.sh script now
1297 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1298 (FreeBSD).
1299
1300 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1301 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1302 comment: "Figure out how to create a longlong from a string: strtoll and
1303 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1304 instance, but it only takes 2 args instead of 3!"
1305
1306 18. A subtle bug concerned with back references has been fixed by a change of
1307 specification, with a corresponding code fix. A pattern such as
1308 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1309 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1310 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1311 same bug. Such groups have to be quantified to be useful, or contained
1312 inside another quantified group. (If there's no repetition, the reference
1313 can never match.) The problem arises because, having left the group and
1314 moved on to the rest of the pattern, a later failure that backtracks into
1315 the group uses the captured value from the final iteration of the group
1316 rather than the correct earlier one. I have fixed this in PCRE by forcing
1317 any group that contains a reference to itself to be an atomic group; that
1318 is, there cannot be any backtracking into it once it has completed. This is
1319 similar to recursive and subroutine calls.
1320
1321
1322 Version 8.00 19-Oct-09
1323 ----------------------
1324
1325 1. The table for translating pcre_compile() error codes into POSIX error codes
1326 was out-of-date, and there was no check on the pcre_compile() error code
1327 being within the table. This could lead to an OK return being given in
1328 error.
1329
1330 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1331 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1332 in a Windows environment.
1333
1334 3. The pcregrep --count option prints the count for each file even when it is
1335 zero, as does GNU grep. However, pcregrep was also printing all files when
1336 --files-with-matches was added. Now, when both options are given, it prints
1337 counts only for those files that have at least one match. (GNU grep just
1338 prints the file name in this circumstance, but including the count seems
1339 more useful - otherwise, why use --count?) Also ensured that the
1340 combination -clh just lists non-zero counts, with no names.
1341
1342 4. The long form of the pcregrep -F option was incorrectly implemented as
1343 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1344 but it seems right to fix it, and I didn't think it was worth preserving
1345 the old behaviour.
1346
1347 5. The command line items --regex=pattern and --regexp=pattern were not
1348 recognized by pcregrep, which required --regex pattern or --regexp pattern
1349 (with a space rather than an '='). The man page documented the '=' forms,
1350 which are compatible with GNU grep; these now work.
1351
1352 6. No libpcreposix.pc file was created for pkg-config; there was just
1353 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1354
1355 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1356 when UCP support is not needed, by modifying the Python script that
1357 generates it from Unicode data files. This should not matter if the module
1358 is correctly used as a library, but I received one complaint about 50K of
1359 unwanted data. My guess is that the person linked everything into his
1360 program rather than using a library. Anyway, it does no harm.
1361
1362 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1363 was a minimum greater than 1 for a wide character in a possessive
1364 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1365 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1366 character. Chaos in the form of incorrect output or a compiling loop could
1367 result.
1368
1369 9. The restrictions on what a pattern can contain when partial matching is
1370 requested for pcre_exec() have been removed. All patterns can now be
1371 partially matched by this function. In addition, if there are at least two
1372 slots in the offset vector, the offset of the earliest inspected character
1373 for the match and the offset of the end of the subject are set in them when
1374 PCRE_ERROR_PARTIAL is returned.
1375
1376 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1377 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1378 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1379 and may be more useful for multi-segment matching.
1380
1381 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1382 used to be given if ever the end of the subject was reached; now it is
1383 given only if matching could not proceed because another character was
1384 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1385 string "Z", which now yields "no match" instead of "partial match". In the
1386 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1387 final character ended with (*FAIL).
1388
1389 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1390 if the pattern had a "must contain" character that was already found in the
1391 earlier partial match, unless partial matching was again requested. For
1392 example, with the pattern /dog.(body)?/, the "must contain" character is
1393 "g". If the first part-match was for the string "dog", restarting with
1394 "sbody" failed. This bug has been fixed.
1395
1396 13. The string returned by pcre_dfa_exec() after a partial match has been
1397 changed so that it starts at the first inspected character rather than the
1398 first character of the match. This makes a difference only if the pattern
1399 starts with a lookbehind assertion or \b or \B (\K is not supported by
1400 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1401 matching functions compatible, and I think it's the right thing to do.
1402
1403 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1404 so that the demonstration program is easily available in environments where
1405 PCRE has not been installed from source.
1406
1407 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1408 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1409 library.
1410
1411 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1412 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1413 is not the first non-POSIX option to be added. Clearly some people find
1414 these options useful.
1415
1416 17. If a caller to the POSIX matching function regexec() passes a non-zero
1417 value for nmatch with a NULL value for pmatch, the value of
1418 nmatch is forced to zero.
1419
1420 18. RunGrepTest did not have a test for the availability of the -u option of
1421 the diff command, as RunTest does. It now checks in the same way as
1422 RunTest, and also checks for the -b option.
1423
1424 19. If an odd number of negated classes containing just a single character
1425 interposed, within parentheses, between a forward reference to a named
1426 subpattern and the definition of the subpattern, compilation crashed with
1427 an internal error, complaining that it could not find the referenced
1428 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1429 [The bug was that it was starting one character too far in when skipping
1430 over the character class, thus treating the ] as data rather than
1431 terminating the class. This meant it could skip too much.]
1432
1433 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1434 /g option in pcretest when the pattern contains \K, which makes it possible
1435 to have an empty string match not at the start, even when the pattern is
1436 anchored. Updated pcretest and pcredemo to use this option.
1437
1438 21. If the maximum number of capturing subpatterns in a recursion was greater
1439 than the maximum at the outer level, the higher number was returned, but
1440 with unset values at the outer level. The correct (outer level) value is
1441 now given.
1442
1443 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1444 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1445 make it do so. The string so far is captured, making this feature
1446 compatible with Perl.
1447
1448 23. The tests have been re-organized, adding tests 11 and 12, to make it
1449 possible to check the Perl 5.10 features against Perl 5.10.
1450
1451 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1452 pattern matches a fixed length string. PCRE did not allow this; now it
1453 does. Neither allows recursion.
1454
1455 25. I finally figured out how to implement a request to provide the minimum
1456 length of subject string that was needed in order to match a given pattern.
1457 (It was back references and recursion that I had previously got hung up
1458 on.) This code has now been added to pcre_study(); it finds a lower bound
1459 to the length of subject needed. It is not necessarily the greatest lower
1460 bound, but using it to avoid searching strings that are too short does give
1461 some useful speed-ups. The value is available to calling programs via
1462 pcre_fullinfo().
1463
1464 26. While implementing 25, I discovered to my embarrassment that pcretest had
1465 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1466 study optimizations had never been tested with that matching function.
1467 Oops. What is worse, even when it was passed study data, there was a bug in
1468 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1469 were also very few tests of studied patterns with pcre_dfa_exec().
1470
1471 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1472 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1473 on the other side of the coin, they are no longer allowed to have different
1474 names, because these cannot be distinguished in PCRE, and this has caused
1475 confusion. (This is a difference from Perl.)
1476
1477 28. When duplicate subpattern names are present (necessarily with different
1478 numbers, as required by 27 above), and a test is made by name in a
1479 conditional pattern, either for a subpattern having been matched, or for
1480 recursion in such a pattern, all the associated numbered subpatterns are
1481 tested, and the overall condition is true if the condition is true for any
1482 one of them. This is the way Perl works, and is also more like the way
1483 testing by number works.
1484
1485
1486 Version 7.9 11-Apr-09
1487 ---------------------
1488
1489 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1490 (pcretest), all targets were linked against these libraries. This included
1491 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1492 libraries. This caused unwanted dependencies to be created. This problem
1493 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1494 pcretest is linked with readline.
1495
1496 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1497 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1498 moved outside it again, because FALSE and TRUE are already defined in AIX,
1499 but BOOL is not.
1500
1501 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1502 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1503
1504 4. The pcregrep documentation said spaces were inserted as well as colons (or
1505 hyphens) following file names and line numbers when outputting matching
1506 lines. This is not true; no spaces are inserted. I have also clarified the
1507 wording for the --colour (or --color) option.
1508
1509 5. In pcregrep, when --colour was used with -o, the list of matching strings
1510 was not coloured; this is different to GNU grep, so I have changed it to be
1511 the same.
1512
1513 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1514 each matching line was coloured. Now it goes on to look for further matches
1515 of any of the test patterns, which is the same behaviour as GNU grep.
1516
1517 7. A pattern that could match an empty string could cause pcregrep to loop; it
1518 doesn't make sense to accept an empty string match in pcregrep, so I have
1519 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1520 seems to be how GNU grep behaves.
1521
1522 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1523 start or after a newline", because the conditional assertion was not being
1524 correctly handled. The rule now is that both the assertion and what follows
1525 in the first alternative must satisfy the test.
1526
1527 9. If auto-callout was enabled in a pattern with a conditional group whose
1528 condition was an assertion, PCRE could crash during matching, both with
1529 pcre_exec() and pcre_dfa_exec().
1530
1531 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1532 used for matching.
1533
1534 11. Unicode property support in character classes was not working for
1535 characters (bytes) greater than 127 when not in UTF-8 mode.
1536
1537 12. Added the -M command line option to pcretest.
1538
1539 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1540
1541 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1542
1543 16. Added comments and documentation about mis-use of no_arg in the C++
1544 wrapper.
1545
1546 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1547 from Martin Jerabek that uses macro names for all relevant character and
1548 string constants.
1549
1550 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1551 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1552 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1553 these, but not everybody uses configure.
1554
1555 19. A conditional group that had only one branch was not being correctly
1556 recognized as an item that could match an empty string. This meant that an
1557 enclosing group might also not be so recognized, causing infinite looping
1558 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1559 with the subject "ab", where knowledge that the repeated group can match
1560 nothing is needed in order to break the loop.
1561
1562 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1563 exec(), but without supplying a callout function, matching went wrong.
1564
1565 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1566 leak if the size of the offset vector was greater than 30. When the vector
1567 is smaller, the saved offsets during recursion go onto a local stack
1568 vector, but for larger vectors malloc() is used. It was failing to free
1569 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1570 error, in fact).
1571
1572 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1573 heapframe that is used only when UTF-8 support is enabled. This caused no
1574 problem, but was untidy.
1575
1576 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1577 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1578 included within another project.
1579
1580 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1581 slightly modified by me:
1582
1583 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1584 not building pcregrep.
1585
1586 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1587 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1588
1589 25. Forward references, both numeric and by name, in patterns that made use of
1590 duplicate group numbers, could behave incorrectly or give incorrect errors,
1591 because when scanning forward to find the reference group, PCRE was not
1592 taking into account the duplicate group numbers. A pattern such as
1593 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1594
1595 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1596 the feature of a custom pointer more persuasive (as requested by a user).
1597
1598 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1599 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1600 pre-defined.
1601
1602 28. Added support for (*UTF8) at the start of a pattern.
1603
1604 29. Arrange for flags added by the "release type" setting in CMake to be shown
1605 in the configuration summary.
1606
1607
1608 Version 7.8 05-Sep-08
1609 ---------------------
1610
1611 1. Replaced UCP searching code with optimized version as implemented for Ad
1612 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1613 stage table and inline lookup instead of a function, giving speed ups of 2
1614 to 5 times on some simple patterns that I tested. Permission was given to
1615 distribute the MultiStage2.py script that generates the tables (it's not in
1616 the tarball, but is in the Subversion repository).
1617
1618 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1619 scripts.
1620
1621 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1622 a group with a zero qualifier. The result of the study could be incorrect,
1623 or the function might crash, depending on the pattern.
1624
1625 4. Caseless matching was not working for non-ASCII characters in back
1626 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1627 It now works when Unicode Property Support is available.
1628
1629 5. In pcretest, an escape such as \x{de} in the data was always generating
1630 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1631 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1632 truncation.
1633
1634 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1635
1636 7. Added two (int) casts to pcregrep when printing the difference of two
1637 pointers, in case they are 64-bit values.
1638
1639 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1640 test 2 if it fails.
1641
1642 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1643 and a #define of that name to empty if it is not externally set. This is to
1644 allow users of MSVC to set it if necessary.
1645
1646 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1647 the convenience functions in the pcre_get.c source file.
1648
1649 11. An option change at the start of a pattern that had top-level alternatives
1650 could cause overwriting and/or a crash. This command provoked a crash in
1651 some environments:
1652
1653 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1654
1655 This potential security problem was recorded as CVE-2008-2371.
1656
1657 12. For a pattern where the match had to start at the beginning or immediately
1658 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1659 pcre_dfa_exec() could read past the end of the passed subject if there was
1660 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1661 pcretest so that it places the subject at the end of its malloc-ed buffer.
1662
1663 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1664 exec() might read past the end of the data buffer in UTF-8 mode.
1665
1666 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1667 the data contained the byte 0x85 as part of a UTF-8 character within its
1668 first line. This applied both to normal and DFA matching.
1669
1670 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1671 /^[^d]*?$/8 failed to match "abc".
1672
1673 16. Added a missing copyright notice to pcrecpp_internal.h.
1674
1675 17. Make it more clear in the documentation that values returned from
1676 pcre_exec() in ovector are byte offsets, not character counts.
1677
1678 18. Tidied a few places to stop certain compilers from issuing warnings.
1679
1680 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1681 supplied by Stefan Weber. I made a further small update for 7.8 because
1682 there is a change of source arrangements: the pcre_searchfuncs.c module is
1683 replaced by pcre_ucd.c.
1684
1685
1686 Version 7.7 07-May-08
1687 ---------------------
1688
1689 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1690 a string to a long long, pretend we don't even have a long long." This is
1691 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1692
1693 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1694 pre-7.6 versions, which defined a global no_arg variable instead of putting
1695 it in the RE class. (See also #8 below.)
1696
1697 3. Remove a line of dead code, identified by coverity and reported by Nuno
1698 Lopes.
1699
1700 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1701
1702 (1) The include/exclude patterns were being applied to the whole pathnames
1703 of files, instead of just to the final components.
1704
1705 (2) If there was more than one level of directory, the subdirectories were
1706 skipped unless they satisfied the include/exclude conditions. This is
1707 inconsistent with GNU grep (and could even be seen as contrary to the
1708 pcregrep specification - which I improved to make it absolutely clear).
1709 The action now is always to scan all levels of directory, and just
1710 apply the include/exclude patterns to regular files.
1711
1712 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1713 --exclude_dir in the tests to avoid scanning .svn directories.
1714
1715 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1716 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1717 doesn't support NULs in patterns.
1718
1719 7. Added some missing "const"s to declarations of static tables in
1720 pcre_compile.c and pcre_dfa_exec.c.
1721
1722 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1723 caused by fix #2 above. (Subsequently also a second patch to fix the
1724 first patch. And a third patch - this was a messy problem.)
1725
1726 9. Applied Craig's patch to remove the use of push_back().
1727
1728 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1729 matching function regexec().
1730
1731 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1732 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1733 references. PCRE supports relative numbers with this syntax (I don't think
1734 Oniguruma does).
1735
1736 12. Previously, a group with a zero repeat such as (...){0} was completely
1737 omitted from the compiled regex. However, this means that if the group
1738 was called as a subroutine from elsewhere in the pattern, things went wrong
1739 (an internal error was given). Such groups are now left in the compiled
1740 pattern, with a new opcode that causes them to be skipped at execution
1741 time.
1742
1743 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1744 to the way PCRE behaves:
1745
1746 (a) A lone ] character is dis-allowed (Perl treats it as data).
1747
1748 (b) A back reference to an unmatched subpattern matches an empty string
1749 (Perl fails the current match path).
1750
1751 (c) A data ] in a character class must be notated as \] because if the
1752 first data character in a class is ], it defines an empty class. (In
1753 Perl it is not possible to have an empty class.) The empty class []
1754 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1755 The negative empty class [^] matches any one character, independently
1756 of the DOTALL setting.
1757
1758 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1759 non-existent subpattern following a character class starting with ']' and
1760 containing () gave an internal compiling error instead of "reference to
1761 non-existent subpattern". Fortunately, when the pattern did exist, the
1762 compiled code was correct. (When scanning forwards to check for the
1763 existencd of the subpattern, it was treating the data ']' as terminating
1764 the class, so got the count wrong. When actually compiling, the reference
1765 was subsequently set up correctly.)
1766
1767 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1768 it was being rejected as not supported by pcre_dfa_exec(), even though
1769 other assertions are supported. I have made pcre_dfa_exec() support
1770 (*FAIL).
1771
1772 16. The implementation of 13c above involved the invention of a new opcode,
1773 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1774 cannot be changed at match time, I realized I could make a small
1775 improvement to matching performance by compiling OP_ALLANY instead of
1776 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1777 on the OP_ANY path.
1778
1779 17. Compiling pcretest on Windows with readline support failed without the
1780 following two fixes: (1) Make the unistd.h include conditional on
1781 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1782
1783 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1784 ncurses library to be included for pcretest when ReadLine support is
1785 requested, but also to allow for it to be overridden. This patch came from
1786 Daniel Bergström.
1787
1788 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1789 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1790 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1791 spotting this.
1792
1793
1794 Version 7.6 28-Jan-08
1795 ---------------------
1796
1797 1. A character class containing a very large number of characters with
1798 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1799 overflow.
1800
1801 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1802 HAVE_LONG_LONG is not defined.
1803
1804 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1805 bring them up to date and include new features. This patch includes:
1806
1807 - Fixed PH's badly added libz and libbz2 support.
1808 - Fixed a problem with static linking.
1809 - Added pcredemo. [But later removed - see 7 below.]
1810 - Fixed dftables problem and added an option.
1811 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1812 HAVE_LONG_LONG.
1813 - Added readline support for pcretest.
1814 - Added an listing of the option settings after cmake has run.
1815
1816 4. A user submitted a patch to Makefile that makes it easy to create
1817 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1818 Makefile.am that cause it to include this special target, without
1819 affecting anything else. Note that the same mingw target plus all
1820 the other distribution libraries and programs are now supported
1821 when configuring with CMake (see 6 below) instead of with
1822 Configure/Make.
1823
1824 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1825 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1826 exported in the Windows port". It has not yet been confirmed that the patch
1827 solves the problem, but it does no harm.
1828
1829 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1830 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1831 with CMake, and also correct the comment about stack recursion.
1832
1833 7. Remove the automatic building of pcredemo from the ./configure system and
1834 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1835 of a program that users should build themselves after PCRE is installed, so
1836 building it automatically is not really right. What is more, it gave
1837 trouble in some build environments.
1838
1839 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1840
1841
1842 Version 7.5 10-Jan-08
1843 ---------------------
1844
1845 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1846 values in parens when parsing an RE using the C++ wrapper."
1847
1848 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1849 Characters greater than 255 were excluded from the class instead of being
1850 included.
1851
1852 3. The same bug as (2) above applied to negated POSIX classes such as
1853 [:^space:].
1854
1855 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1856 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1857 I have changed it.
1858
1859 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1860 first named subpattern) and a construct such as (?&a) would reference the
1861 first named subpattern whose name started with "a" (in other words, the
1862 length check was missing). Both these problems are fixed. "Subpattern name
1863 expected" is now given for (?&) (a zero-length name), and this patch also
1864 makes it give the same error for \k'' (previously it complained that that
1865 was a reference to a non-existent subpattern).
1866
1867 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1868 this is right because (?- can be followed by option settings as well as by
1869 digits. I have, however, made the messages clearer.
1870
1871 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1872 than the number used in the conditional) now cause a compile-time error.
1873 This is actually not compatible with Perl, which accepts such patterns, but
1874 treats the conditional as always being FALSE (as PCRE used to), but it
1875 seems to me that giving a diagnostic is better.
1876
1877 8. Change "alphameric" to the more common word "alphanumeric" in comments
1878 and messages.
1879
1880 9. Fix two occurrences of "backslash" in comments that should have been
1881 "backspace".
1882
1883 10. Remove two redundant lines of code that can never be obeyed (their function
1884 was moved elsewhere).
1885
1886 11. The program that makes PCRE's Unicode character property table had a bug
1887 which caused it to generate incorrect table entries for sequences of
1888 characters that have the same character type, but are in different scripts.
1889 It amalgamated them into a single range, with the script of the first of
1890 them. In other words, some characters were in the wrong script. There were
1891 thirteen such cases, affecting characters in the following ranges:
1892
1893 U+002b0 - U+002c1
1894 U+0060c - U+0060d
1895 U+0061e - U+00612
1896 U+0064b - U+0065e
1897 U+0074d - U+0076d
1898 U+01800 - U+01805
1899 U+01d00 - U+01d77
1900 U+01d9b - U+01dbf
1901 U+0200b - U+0200f
1902 U+030fc - U+030fe
1903 U+03260 - U+0327f
1904 U+0fb46 - U+0fbb1
1905 U+10450 - U+1049d
1906
1907 12. The -o option (show only the matching part of a line) for pcregrep was not
1908 compatible with GNU grep in that, if there was more than one match in a
1909 line, it showed only the first of them. It now behaves in the same way as
1910 GNU grep.
1911
1912 13. If the -o and -v options were combined for pcregrep, it printed a blank
1913 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1914 does the same. The return code can be used to tell if there were any
1915 non-matching lines.
1916
1917 14. Added --file-offsets and --line-offsets to pcregrep.
1918
1919 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1920 infinitely looping recursion. The bug was that positive lookaheads were not
1921 being skipped when checking for a possible empty match (negative lookaheads
1922 and both kinds of lookbehind were skipped).
1923
1924 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1925 inclusion of <windows.h> to before rather than after the definition of
1926 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1927
1928 17. Specifying a possessive quantifier with a specific limit for a Unicode
1929 character property caused pcre_compile() to compile bad code, which led at
1930 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1931 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1932 caused the error; without that there was no problem.
1933
1934 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1935
1936 19. Added --enable-pcretest-libreadline.
1937
1938 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1939 RE::GlobalReplace(). As a result, the number of replacements returned was
1940 double what it should be. I removed one of the increments, but Craig sent a
1941 later patch that removed the other one (the right fix) and added unit tests
1942 that check the return values (which was not done before).
1943
1944 21. Several CMake things:
1945
1946 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1947 the names libpcre and libpcreposix, not just pcre and pcreposix.
1948
1949 (2) The above change means that pcretest and pcregrep are now correctly
1950 linked with the newly-built libraries, not previously installed ones.
1951
1952 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1953
1954 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1955 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1956 UTF-8 newline character). The key issue is that the pattern starts .*;
1957 this means that the match must be either at the beginning, or after a
1958 newline. The bug was in the code for advancing after a failed match and
1959 checking that the new position followed a newline. It was not taking
1960 account of UTF-8 characters correctly.
1961
1962 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1963 character classes. PCRE was not treating the sequence [:...:] as a
1964 character class unless the ... were all letters. Perl, however, seems to
1965 allow any characters between [: and :], though of course it rejects as
1966 unknown any "names" that contain non-letters, because all the known class
1967 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1968 for example, whereas PCRE did not - it did not recognize a POSIX character
1969 class. This seemed a bit dangerous, so the code has been changed to be
1970 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1971 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1972 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1973 Perl does, and where it didn't before.
1974
1975 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1976 Windows environments %n is disabled by default.
1977
1978
1979 Version 7.4 21-Sep-07
1980 ---------------------
1981
1982 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1983 means that a class such as [\s] counted as "explicit reference to CR or
1984 LF". That isn't really right - the whole point of the change was to try to
1985 help when there was an actual mention of one of the two characters. So now
1986 the change happens only if \r or \n (or a literal CR or LF) character is
1987 encountered.
1988
1989 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1990 of both had grown to the point where there were only 3 bits left.
1991 Fortunately, there was spare space in the data structure, and so I have
1992 moved the internal flags into a new 16-bit field to free up more option
1993 bits.
1994
1995 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1996 but did not set the internal JCHANGED flag - either of these is enough to
1997 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1998 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1999 start sets both bits.
2000
2001 4. Added options (at build time, compile time, exec time) to change \R from
2002 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2003
2004 5. doc/pcresyntax.html was missing from the distribution.
2005
2006 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2007 compatibility, even though it is no longer used.
2008
2009 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2010 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2011 windows.h file is present (where different names are used). [This was
2012 reversed later after testing - see 16 below.]
2013
2014 8. Changed all #include <config.h> to #include "config.h". There were also
2015 some further <pcre.h> cases that I changed to "pcre.h".
2016
2017 9. When pcregrep was used with the --colour option, it missed the line ending
2018 sequence off the lines that it output.
2019
2020 10. It was pointed out to me that arrays of string pointers cause lots of
2021 relocations when a shared library is dynamically loaded. A technique of
2022 using a single long string with a table of offsets can drastically reduce
2023 these. I have refactored PCRE in four places to do this. The result is
2024 dramatic:
2025
2026 Originally: 290
2027 After changing UCP table: 187
2028 After changing error message table: 43
2029 After changing table of "verbs" 36
2030 After changing table of Posix names 22
2031
2032 Thanks to the folks working on Gregex for glib for this insight.
2033
2034 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2035 unicode-properties was also set.
2036
2037 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2038
2039 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2040 checked only for CRLF.
2041
2042 14. Added casts to pcretest.c to avoid compiler warnings.
2043
2044 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2045
2046 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2047 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2048 entirely. This removes changes made in 7 above.
2049
2050 17. The CMake files have been updated, and there is now more information about
2051 building with CMake in the NON-UNIX-USE document.
2052
2053
2054 Version 7.3 28-Aug-07
2055 ---------------------
2056
2057 1. In the rejigging of the build system that eventually resulted in 7.1, the
2058 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2059 brackets there is not right, since it causes compilers to look for an
2060 installed pcre.h, not the version that is in the source that is being
2061 compiled (which of course may be different). I have changed it back to:
2062
2063 #include "pcre.h"
2064
2065 I have a vague recollection that the change was concerned with compiling in
2066 different directories, but in the new build system, that is taken care of
2067 by the VPATH setting the Makefile.
2068
2069 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2070 when the subject happened to end in the byte 0x85 (e.g. if the last
2071 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2072 characters but of course it shouldn't be taken as a newline when it is part
2073 of another character. The bug was that, for an unlimited repeat of . in
2074 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2075 characters when looking for a newline.
2076
2077 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2078
2079 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2080 in debug output.
2081
2082 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2083 long printing in the pcrecpp unittest when running under MinGW.
2084
2085 6. ESC_K was left out of the EBCDIC table.
2086
2087 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2088 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2089 limit also applies to "virtual nesting" when a pattern is recursive, and in
2090 this case 1000 isn't so big. I have been able to remove this limit at the
2091 expense of backing off one optimization in certain circumstances. Normally,
2092 when pcre_exec() would call its internal match() function recursively and
2093 immediately return the result unconditionally, it uses a "tail recursion"
2094 feature to save stack. However, when a subpattern that can match an empty
2095 string has an unlimited repetition quantifier, it no longer makes this
2096 optimization. That gives it a stack frame in which to save the data for
2097 checking that an empty string has been matched. Previously this was taken
2098 from the 1000-entry workspace that had been reserved. So now there is no
2099 explicit limit, but more stack is used.
2100
2101 8. Applied Daniel's patches to solve problems with the import/export magic
2102 syntax that is required for Windows, and which was going wrong for the
2103 pcreposix and pcrecpp parts of the library. These were overlooked when this
2104 problem was solved for the main library.
2105
2106 9. There were some crude static tests to avoid integer overflow when computing
2107 the size of patterns that contain repeated groups with explicit upper
2108 limits. As the maximum quantifier is 65535, the maximum group length was
2109 set at 30,000 so that the product of these two numbers did not overflow a
2110 32-bit integer. However, it turns out that people want to use groups that
2111 are longer than 30,000 bytes (though not repeat them that many times).
2112 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2113 made it possible to implement the integer overflow checks in a much more
2114 dynamic way, which I have now done. The artificial limitation on group
2115 length has been removed - we now have only the limit on the total length of
2116 the compiled pattern, which depends on the LINK_SIZE setting.
2117
2118 10. Fixed a bug in the documentation for get/copy named substring when
2119 duplicate names are permitted. If none of the named substrings are set, the
2120 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2121 empty string.
2122
2123 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2124 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2125 because the ] is interpreted as the first data character and the
2126 terminating ] is not found. PCRE has been made compatible with Perl in this
2127 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2128 cause memory overwriting.
2129
2130 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2131 string has been matched (to stop an infinite loop). It was not recognizing
2132 a conditional subpattern that could match an empty string if that
2133 subpattern was within another subpattern. For example, it looped when
2134 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2135 condition was not nested. This bug has been fixed.
2136
2137 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2138 past the start of the subject in the presence of bytes with the top bit
2139 set, for example "\x8aBCD".
2140
2141 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2142 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2143
2144 14. Optimized (?!) to (*FAIL).
2145
2146 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2147 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2148 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2149 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2150 does: it's just the validity check that is more restrictive.
2151
2152 16. Inserted checks for integer overflows during escape sequence (backslash)
2153 processing, and also fixed erroneous offset values for syntax errors during
2154 backslash processing.
2155
2156 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2157 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2158
2159 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2160 caused an overrun.
2161
2162 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2163 something other than just ASCII characters) inside a group that had an
2164 unlimited repeat caused a loop at compile time (while checking to see
2165 whether the group could match an empty string).
2166
2167 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2168 [\P{Any}] did so. (Error in the code for printing property names.)
2169
2170 21. An orphan \E inside a character class could cause a crash.
2171
2172 22. A repeated capturing bracket such as (A)? could cause a wild memory
2173 reference during compilation.
2174
2175 23. There are several functions in pcre_compile() that scan along a compiled
2176 expression for various reasons (e.g. to see if it's fixed length for look
2177 behind). There were bugs in these functions when a repeated \p or \P was
2178 present in the pattern. These operators have additional parameters compared
2179 with \d, etc, and these were not being taken into account when moving along
2180 the compiled data. Specifically:
2181
2182 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2183 length.
2184
2185 (b) An item such as \pL+ within a repeated group could cause crashes or
2186 loops.
2187
2188 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2189 "reference to non-existent subpattern" error.
2190
2191 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2192
2193 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2194 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2195
2196 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2197
2198 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2199 character were causing crashes (broken optimization).
2200
2201 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2202 \p or \P) caused a compile-time loop.
2203
2204 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2205 break. For example, the unstudied pattern [\r\n]A does not match the string
2206 "\r\nA" because change 7.0/46 below moves the current point on by two
2207 characters after failing to match at the start. However, the pattern \nA
2208 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2209 the same is true. There doesn't seem any very clean way out of this, but
2210 what I have chosen to do makes the common cases work: PCRE now takes note
2211 of whether there can be an explicit match for \r or \n anywhere in the
2212 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2213 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2214 pattern has explicit CR or LF references.
2215
2216 29. Added (*CR) etc for changing newline setting at start of pattern.
2217
2218
2219 Version 7.2 19-Jun-07
2220 ---------------------
2221
2222 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2223 which is apparently normally available under Windows.
2224
2225 2. Re-jig the pcregrep tests with different newline settings in an attempt
2226 to make them independent of the local environment's newline setting.
2227
2228 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2229
2230 4. Some of the "internals" tests were previously cut out when the link size
2231 was not 2, because the output contained actual offsets. The recent new
2232 "Z" feature of pcretest means that these can be cut out, making the tests
2233 usable with all link sizes.
2234
2235 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2236 stack recursion. This gives a massive performance boost under BSD, but just
2237 a small improvement under Linux. However, it saves one field in the frame
2238 in all cases.
2239
2240 6. Added more features from the forthcoming Perl 5.10:
2241
2242 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2243 recursion call. It refers to the nth most recently opened parentheses.
2244
2245 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2246 to be opened parentheses.
2247
2248 (c) Conditions that refer to capturing parentheses can be specified
2249 relatively, for example, (?(-2)... or (?(+3)...
2250
2251 (d) \K resets the start of the current match so that everything before
2252 is not part of it.
2253
2254 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2255
2256 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2257 reference syntax.
2258
2259 (g) (?| introduces a group in which the numbering of parentheses in each
2260 alternative starts with the same number.
2261
2262 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2263
2264 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2265 PCRE_INFO_JCHANGED.
2266
2267 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2268 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2269 for detecting groups that can match an empty string.
2270
2271 9. A pattern with a very large number of alternatives (more than several
2272 hundred) was running out of internal workspace during the pre-compile
2273 phase, where pcre_compile() figures out how much memory will be needed. A
2274 bit of new cunning has reduced the workspace needed for groups with
2275 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2276 workspace instead of running out of the 4096 that are available.
2277
2278 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2279
2280 11. Applied patch from Google to remove an optimization that didn't quite work.
2281 The report of the bug said:
2282
2283 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2284 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2285 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2286
2287 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2288 it matched the wrong number of bytes.
2289
2290
2291 Version 7.1 24-Apr-07
2292 ---------------------
2293
2294 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2295 that is more "standard", making use of automake and other Autotools. There
2296 is some re-arrangement of the files and adjustment of comments consequent
2297 on this.
2298
2299 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2300 for recursive directory scanning broke on some systems because the files
2301 are not scanned in any specific order and on different systems the order
2302 was different. A call to "sort" has been inserted into RunGrepTest for the
2303 approprate test as a short-term fix. In the longer term there may be an
2304 alternative.
2305
2306 3. I had an email from Eric Raymond about problems translating some of PCRE's
2307 man pages to HTML (despite the fact that I distribute HTML pages, some
2308 people do their own conversions for various reasons). The problems
2309 concerned the use of low-level troff macros .br and .in. I have therefore
2310 removed all such uses from the man pages (some were redundant, some could
2311 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2312 HTML has been updated to handle .nf/.fi and to complain if it encounters
2313 .br or .in.
2314
2315 4. Updated comments in configure.ac that get placed in config.h.in and also
2316 arranged for config.h to be included in the distribution, with the name
2317 config.h.generic, for the benefit of those who have to compile without
2318 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2319
2320 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2321 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2322 makevp.bat for the current PCRE, using the additional files
2323 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2324
2325 6. A Windows user reported a minor discrepancy with test 2, which turned out
2326 to be caused by a trailing space on an input line that had got lost in his
2327 copy. The trailing space was an accident, so I've just removed it.
2328
2329 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2330 that is needed.
2331
2332 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2333 as "const" (a) because they are and (b) because it helps the PHP
2334 maintainers who have recently made a script to detect big data structures
2335 in the php code that should be moved to the .rodata section. I remembered
2336 to update Builducptable as well, so it won't revert if ucptable.h is ever
2337 re-created.
2338
2339 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2340 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2341 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2342 support is not required. This saves 1.5-2K of code, which is important in
2343 some applications.
2344
2345 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2346 so as not to refer to the tables, even though these functions will never be
2347 called when UTF-8 support is disabled. Otherwise there are problems with a
2348 shared library.
2349
2350 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2351
2352 (a) It was defining its arguments as char * instead of void *.
2353
2354 (b) It was assuming that all moves were upwards in memory; this was true
2355 a long time ago when I wrote it, but is no longer the case.
2356
2357 The emulated memove() is provided for those environments that have neither
2358 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2359 is clearly not the case, as these two bugs were recently reported.
2360
2361 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2362 and Detrail to create the HTML documentation, the .txt form of the man
2363 pages, and it removes trailing spaces from listed files. It also creates
2364 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2365 case, it wraps all the #defines with #ifndefs. This script should be run
2366 before "make dist".
2367
2368 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2369 with Unicode property support.
2370
2371 (a) For a maximizing quantifier, if the two different cases of the
2372 character were of different lengths in their UTF-8 codings (there are
2373 some cases like this - I found 11), and the matching function had to
2374 back up over a mixture of the two cases, it incorrectly assumed they
2375 were both the same length.
2376
2377 (b) When PCRE was configured to use the heap rather than the stack for
2378 recursion during matching, it was not correctly preserving the data for
2379 the other case of a UTF-8 character when checking ahead for a match
2380 while processing a minimizing repeat. If the check also involved
2381 matching a wide character, but failed, corruption could cause an
2382 erroneous result when trying to check for a repeat of the original
2383 character.
2384
2385 13. Some tidying changes to the testing mechanism:
2386
2387 (a) The RunTest script now detects the internal link size and whether there
2388 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2389 values substituted by "configure". (The RunGrepTest script already did
2390 this for UTF-8.) The configure.ac script no longer substitutes the
2391 relevant variables.
2392
2393 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2394 with length and offset values. This means that the output is different
2395 for different internal link sizes. Test 2 is skipped for link sizes
2396 other than 2 because of this, bypassing the problem. Unfortunately,
2397 there was also a test in test 3 (the locale tests) that used /B and
2398 failed for link sizes other than 2. Rather than cut the whole test out,
2399 I have added a new /Z option to pcretest that replaces the length and
2400 offset values with spaces. This is now used to make test 3 independent
2401 of link size. (Test 2 will be tidied up later.)
2402
2403 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2404 segmentation fault instead of returning the appropriate error message.
2405
2406 15. In multiline mode when the newline sequence was set to "any", the pattern
2407 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2408 This doesn't seem right; it now treats the CRLF combination as the line
2409 ending, and so does not match in that case. It's only a pattern such as ^$
2410 that would hit this one: something like ^ABC$ would have failed after \r
2411 and then tried again after \r\n.
2412
2413 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2414 in an attempt to make files that differ only in their line terminators
2415 compare equal. This works on Linux.
2416
2417 17. Under certain error circumstances pcregrep might try to free random memory
2418 as it exited. This is now fixed, thanks to valgrind.
2419
2420 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2421 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2422 was because its rules for how to advance for /g after matching an empty
2423 string at the end of a line did not allow for this case. They now check for
2424 it specially.
2425
2426 20. pcretest is supposed to handle patterns and data of any length, by
2427 extending its buffers when necessary. It was getting this wrong when the
2428 buffer for a data line had to be extended.
2429
2430 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2431 CRLF as a newline sequence.
2432
2433 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2434 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2435 I have nevertheless tidied it up.
2436
2437 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2438
2439 24. Added a man page for pcre-config.
2440
2441
2442 Version 7.0 19-Dec-06
2443 ---------------------
2444
2445 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2446 moving to gcc 4.1.1.
2447
2448 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2449 sys/time.h, which is documented as needed for this function. It doesn't
2450 seem to matter on Linux, but it showed up on some releases of OS X.
2451
2452 3. It seems that there are systems where bytes whose values are greater than
2453 127 match isprint() in the "C" locale. The "C" locale should be the
2454 default when a C program starts up. In most systems, only ASCII printing
2455 characters match isprint(). This difference caused the output from pcretest
2456 to vary, making some of the tests fail. I have changed pcretest so that:
2457
2458 (a) When it is outputting text in the compiled version of a pattern, bytes
2459 other than 32-126 are always shown as hex escapes.
2460
2461 (b) When it is outputting text that is a matched part of a subject string,
2462 it does the same, unless a different locale has been set for the match
2463 (using the /L modifier). In this case, it uses isprint() to decide.
2464
2465 4. Fixed a major bug that caused incorrect computation of the amount of memory
2466 required for a compiled pattern when options that changed within the
2467 pattern affected the logic of the preliminary scan that determines the
2468 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2469 that the computed length was too small. The symptoms of this bug were
2470 either the PCRE error "internal error: code overflow" from pcre_compile(),
2471 or a glibc crash with a message such as "pcretest: free(): invalid next
2472 size (fast)". Examples of patterns that provoked this bug (shown in
2473 pcretest format) are:
2474
2475 /(?-x: )/x
2476 /(?x)(?-x: \s*#\s*)/
2477 /((?i)[\x{c0}])/8
2478 /(?i:[\x{c0}])/8
2479
2480 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2481 is now done differently.
2482
2483 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2484 wrapper classes; (b) implement a new function in the C++ scanner that is
2485 more efficient than the old way of doing things because it avoids levels of
2486 recursion in the regex matching; (c) add a paragraph to the documentation
2487 for the FullMatch() function.
2488
2489 6. The escape sequence \n was being treated as whatever was defined as
2490 "newline". Not only was this contrary to the documentation, which states
2491 that \n is character 10 (hex 0A), but it also went horribly wrong when
2492 "newline" was defined as CRLF. This has been fixed.
2493
2494 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2495 was being set to -1 for the "end of line" case (supposedly a value that no
2496 character can have). Though this value is never used (the check for end of
2497 line is "zero bytes in current character"), it caused compiler complaints.
2498 I've changed it to 0xffffffff.
2499
2500 8. In pcre_version.c, the version string was being built by a sequence of
2501 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2502 string (as it is for production releases) called a macro with an empty
2503 argument. The C standard says the result of this is undefined. The gcc
2504 compiler treats it as an empty string (which was what was wanted) but it is
2505 reported that Visual C gives an error. The source has been hacked around to
2506 avoid this problem.
2507
2508 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2509 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2510 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2511 of them did).
2512
2513 10. Originally, pcretest opened its input and output without "b"; then I was
2514 told that "b" was needed in some environments, so it was added for release
2515 5.0 to both the input and output. (It makes no difference on Unix-like
2516 systems.) Later I was told that it is wrong for the input on Windows. I've
2517 now abstracted the modes into two macros, to make it easier to fiddle with
2518 them, and removed "b" from the input mode under Windows.
2519
2520 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2521
2522 12. Added -help and --help to pcretest as an official way of being reminded
2523 of the options.
2524
2525 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2526 and pcrecpp.cc because they annoy compilers at high warning levels.
2527
2528 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2529
2530 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2531 scripts are not C programs :-) and which was not noticed because it works
2532 on Linux.
2533
2534 16. pcretest is supposed to handle any length of pattern and data line (as one
2535 line or as a continued sequence of lines) by extending its input buffer if
2536 necessary. This feature was broken for very long pattern lines, leading to
2537 a string of junk being passed to pcre_compile() if the pattern was longer
2538 than about 50K.
2539
2540 17. I have done a major re-factoring of the way pcre_compile() computes the
2541 amount of memory needed for a compiled pattern. Previously, there was code
2542 that made a preliminary scan of the pattern in order to do this. That was
2543 OK when PCRE was new, but as the facilities have expanded, it has become
2544 harder and harder to keep it in step with the real compile phase, and there
2545 have been a number of bugs (see for example, 4 above). I have now found a
2546 cunning way of running the real compile function in a "fake" mode that
2547 enables it to compute how much memory it would need, while actually only
2548 ever using a few hundred bytes of working memory and without too many
2549 tests of the mode. This should make future maintenance and development
2550 easier. A side effect of this work is that the limit of 200 on the nesting
2551 depth of parentheses has been removed (though this was never a serious
2552 limitation, I suspect). However, there is a downside: pcre_compile() now
2553 runs more slowly than before (30% or more, depending on the pattern). I
2554 hope this isn't a big issue. There is no effect on runtime performance.
2555
2556 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2557 newline (only possible for the last line of a file) and it was a
2558 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2559
2560 19. Added additional timing features to pcretest. (1) The -tm option now times
2561 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2562 separate command line item, by a number that specifies the number of
2563 repeats to use when timing. The default is 50000; this gives better
2564 precision, but takes uncomfortably long for very large patterns.
2565
2566 20. Extended pcre_study() to be more clever in cases where a branch of a
2567 subpattern has no definite first character. For example, (a*|b*)[cd] would
2568 previously give no result from pcre_study(). Now it recognizes that the
2569 first character must be a, b, c, or d.
2570
2571 21. There was an incorrect error "recursive call could loop indefinitely" if
2572 a subpattern (or the entire pattern) that was being tested for matching an
2573 empty string contained only one non-empty item after a nested subpattern.
2574 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2575 incorrectly, because the \d was being skipped in the check.
2576
2577 22. The pcretest program now has a new pattern option /B and a command line
2578 option -b, which is equivalent to adding /B to every pattern. This causes
2579 it to show the compiled bytecode, without the additional information that
2580 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2581 is the same as /B/I).
2582
2583 23. A new optimization is now able automatically to treat some sequences such
2584 as a*b as a*+b. More specifically, if something simple (such as a character
2585 or a simple class like \d) has an unlimited quantifier, and is followed by
2586 something that cannot possibly match the quantified thing, the quantifier
2587 is automatically "possessified".
2588
2589 24. A recursive reference to a subpattern whose number was greater than 39
2590 went wrong under certain circumstances in UTF-8 mode. This bug could also
2591 have affected the operation of pcre_study().
2592
2593 25. Realized that a little bit of performance could be had by replacing
2594 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2595
2596 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2597
2598 27. Possessive quantifiers such as a++ were previously implemented by turning
2599 them into atomic groups such as ($>a+). Now they have their own opcodes,
2600 which improves performance. This includes the automatically created ones
2601 from 23 above.
2602
2603 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2604 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2605 the first matched character to be a colon. This applied both to named and
2606 numbered groups.
2607
2608 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2609
2610 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2611 building PCRE on Windows easier, so I have included it in the distribution.
2612
2613 31. There is now a check in pcretest against a ridiculously large number being
2614 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2615 loop, the loop is abandoned.
2616
2617 32. Forward references to subpatterns in conditions such as (?(2)...) where
2618 subpattern 2 is defined later cause pcre_compile() to search forwards in
2619 the pattern for the relevant set of parentheses. This search went wrong
2620 when there were unescaped parentheses in a character class, parentheses
2621 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2622
2623 33. "Subroutine" calls and backreferences were previously restricted to
2624 referencing subpatterns earlier in the regex. This restriction has now
2625 been removed.
2626
2627 34. Added a number of extra features that are going to be in Perl 5.10. On the
2628 whole, these are just syntactic alternatives for features that PCRE had
2629 previously implemented using the Python syntax or my own invention. The
2630 other formats are all retained for compatibility.
2631
2632 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2633 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2634 also .NET compatible.
2635
2636 (b) A recursion or subroutine call to a named group can now be defined as
2637 (?&name) as well as (?P>name).
2638
2639 (c) A backreference to a named group can now be defined as \k<name> or
2640 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2641 5.10, are also .NET compatible.
2642
2643 (d) A conditional reference to a named group can now use the syntax
2644 (?(<name>) or (?('name') as well as (?(name).
2645
2646 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2647 groups (named and numbered) that are never evaluated inline, but can be
2648 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2649 is always false. There may be only one alternative in such a group.
2650
2651 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2652 as the simple (?(R). The condition is true only if the most recent
2653 recursion is that of the given number or name. It does not search out
2654 through the entire recursion stack.
2655
2656 (g) The escape \gN or \g{N} has been added, where N is a positive or
2657 negative number, specifying an absolute or relative reference.
2658
2659 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2660 some "unreachable code" warnings.
2661
2662 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2663 things, this adds five new scripts.
2664
2665 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2666 There were also incompatibilities regarding the handling of \Q..\E inside
2667 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2668 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2669
2670 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2671 matches an empty string, and forcibly breaks the loop. There were bugs in
2672 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2673 against aaaa the result was just "a" rather than "aaaa", for example. Two
2674 separate and independent bugs (that affected different cases) have been
2675 fixed.
2676
2677 39. Refactored the code to abolish the use of different opcodes for small
2678 capturing bracket numbers. This is a tidy that I avoided doing when I
2679 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2680 The new approach is not only tidier, it makes it possible to reduce the
2681 memory needed to fix the previous bug (38).
2682
2683 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2684 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2685 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2686 mode.
2687
2688 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2689 report.
2690
2691 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2692 copy construction and assignment in the C++ wrapper.
2693
2694 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2695 couple of bugs that could have given wrong results in the "--newline=crlf"
2696 case.
2697
2698 44. Added a number of casts and did some reorganization of signed/unsigned int
2699 variables following suggestions from Dair Grant. Also renamed the variable
2700 "this" as "item" because it is a C++ keyword.
2701
2702 45. Arranged for dftables to add
2703
2704 #include "pcre_internal.h"
2705
2706 to pcre_chartables.c because without it, gcc 4.x may remove the array
2707 definition from the final binary if PCRE is built into a static library and
2708 dead code stripping is activated.
2709
2710 46. For an unanchored pattern, if a match attempt fails at the start of a
2711 newline sequence, and the newline setting is CRLF or ANY, and the next two
2712 characters are CRLF, advance by two characters instead of one.
2713
2714
2715 Version 6.7 04-Jul-06
2716 ---------------------
2717
2718 1. In order to handle tests when input lines are enormously long, pcretest has
2719 been re-factored so that it automatically extends its buffers when
2720 necessary. The code is crude, but this _is_ just a test program. The
2721 default size has been increased from 32K to 50K.
2722
2723 2. The code in pcre_study() was using the value of the re argument before
2724 testing it for NULL. (Of course, in any sensible call of the function, it
2725 won't be NULL.)
2726
2727 3. The memmove() emulation function in pcre_internal.h, which is used on
2728 systems that lack both memmove() and bcopy() - that is, hardly ever -
2729 was missing a "static" storage class specifier.
2730
2731 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2732 containing an extended class (one that cannot be represented by a bitmap
2733 because it contains high-valued characters or Unicode property items, e.g.
2734 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2735 pattern, but PCRE should not loop if you do not (it no longer does).
2736 [Detail: two cases were found: (a) a repeated subpattern containing an
2737 extended class; (b) a recursive reference to a subpattern that followed a
2738 previous extended class. It wasn't skipping over the extended class
2739 correctly when UTF-8 mode was not set.]
2740
2741 5. A negated single-character class was not being recognized as fixed-length
2742 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2743 compile error "lookbehind assertion is not fixed length".
2744
2745 6. The RunPerlTest auxiliary script was showing an unexpected difference
2746 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2747 write a Perl script that can interpret lines of an input file either as
2748 byte characters or as UTF-8, which is what "perltest" was being required to
2749 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2750 can't do is switch easily at run time between having the "use utf8;" pragma
2751 or not. In the end, I fudged it by using the RunPerlTest script to insert
2752 "use utf8;" explicitly for the UTF-8 tests.
2753
2754 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2755 the end of the subject string, contrary to the documentation and to what
2756 Perl does. This was true of both matching functions. Now it matches only at
2757 the start of the subject and immediately after *internal* newlines.
2758
2759 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2760 a pointer to an int instead of a pointer to an unsigned long int. This
2761 caused problems on 64-bit systems.
2762
2763 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2764 instance of the 'standard' template library not being so standard".
2765
2766 10. There was no check on the number of named subpatterns nor the maximum
2767 length of a subpattern name. The product of these values is used to compute
2768 the size of the memory block for a compiled pattern. By supplying a very
2769 long subpattern name and a large number of named subpatterns, the size
2770 computation could be caused to overflow. This is now prevented by limiting
2771 the length of names to 32 characters, and the number of named subpatterns
2772 to 10,000.
2773
2774 11. Subpatterns that are repeated with specific counts have to be replicated in
2775 the compiled pattern. The size of memory for this was computed from the
2776 length of the subpattern and the repeat count. The latter is limited to
2777 65535, but there was no limit on the former, meaning that integer overflow
2778 could in principle occur. The compiled length of a repeated subpattern is
2779 now limited to 30,000 bytes in order to prevent this.
2780
2781 12. Added the optional facility to have named substrings with the same name.
2782
2783 13. Added the ability to use a named substring as a condition, using the
2784 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2785 are numbers (not recommended). Forward references are permitted.
2786
2787 14. Added forward references in named backreferences (if you see what I mean).
2788
2789 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2790 pattern could run off the end of the subject. For example, the pattern
2791 "(?s)(.{1,5})"8 did this with the subject "ab".
2792
2793 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2794 PCRE_CASELESS was set when matching characters that were quantified with ?
2795 or *.
2796
2797 17. A character class other than a single negated character that had a minimum
2798 but no maximum quantifier - for example [ab]{6,} - was not handled
2799 correctly by pce_dfa_exec(). It would match only one character.
2800
2801 18. A valid (though odd) pattern that looked like a POSIX character
2802 class but used an invalid character after [ (for example [[,abc,]]) caused
2803 pcre_compile() to give the error "Failed: internal error: code overflow" or
2804 in some cases to crash with a glibc free() error. This could even happen if
2805 the pattern terminated after [[ but there just happened to be a sequence of
2806 letters, a binary zero, and a closing ] in the memory that followed.
2807
2808 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2809 over the years. Originally (before any Unicode support), just the bottom 8
2810 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2811 output from "man perlunicode" includes this:
2812
2813 The regular expression compiler produces polymorphic opcodes. That
2814 is, the pattern adapts to the data and automatically switches to
2815 the Unicode character scheme when presented with Unicode data--or
2816 instead uses a traditional byte scheme when presented with byte
2817 data.
2818
2819 Sadly, a wide octal escape does not cause a switch, and in a string with
2820 no other multibyte characters, these octal escapes are treated as before.
2821 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2822 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2823 Unicode string.
2824
2825 I have not perpetrated such confusion in PCRE. Up till now, it took just
2826 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2827 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2828 translate to the appropriate multibyte character.
2829
2830 29. Applied some refactoring to reduce the number of warnings from Microsoft
2831 and Borland compilers. This has included removing the fudge introduced
2832 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2833 a warning about an unused variable.
2834
2835 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2836 characters since release 4.0, because Perl (from release 5.004) does not.
2837 [Or at least, is documented not to: some releases seem to be in conflict
2838 with the documentation.] However, when a pattern was studied with
2839 pcre_study() and all its branches started with \s, PCRE still included VT
2840 as a possible starting character. Of course, this did no harm; it just
2841 caused an unnecessary match attempt.
2842
2843 22. Removed a now-redundant internal flag bit that recorded the fact that case
2844 dependency changed within the pattern. This was once needed for "required
2845 byte" processing, but is no longer used. This recovers a now-scarce options
2846 bit. Also moved the least significant internal flag bit to the most-
2847 significant bit of the word, which was not previously used (hangover from
2848 the days when it was an int rather than a uint) to free up another bit for
2849 the future.
2850
2851 23. Added support for CRLF line endings as well as CR and LF. As well as the
2852 default being selectable at build time, it can now be changed at runtime
2853 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2854 specify that it is scanning data with non-default line endings.
2855
2856 24. Changed the definition of CXXLINK to make it agree with the definition of
2857 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2858
2859 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2860 recursions. This makes a big different to stack usage for some patterns.
2861
2862 26. If a subpattern containing a named recursion or subroutine reference such
2863 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2864 the space required for the compiled pattern went wrong and gave too small a
2865 value. Depending on the environment, this could lead to "Failed: internal
2866 error: code overflow at offset 49" or "glibc detected double free or
2867 corruption" errors.
2868
2869 27. Applied patches from Google (a) to support the new newline modes and (b) to
2870 advance over multibyte UTF-8 characters in GlobalReplace.
2871
2872 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2873 difference for some implementation of PCRE in some Windows version.
2874
2875 29. Added some extra testing facilities to pcretest:
2876
2877 \q<number> in a data line sets the "match limit" value
2878 \Q<number> in a data line sets the "match recursion limt" value
2879 -S <number> sets the stack size, where <number> is in megabytes
2880
2881 The -S option isn't available for Windows.
2882
2883
2884 Version 6.6 06-Feb-06
2885 ---------------------
2886
2887 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2888 in pcreposix.h. I have copied the definition from pcre.h.
2889
2890 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2891 because pcre.h is no longer a built file.
2892
2893 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2894 not normally included in the compiled code.
2895
2896
2897 Version 6.5 01-Feb-06
2898 ---------------------
2899
2900 1. When using the partial match feature with pcre_dfa_exec(), it was not
2901 anchoring the second and subsequent partial matches at the new starting
2902 point. This could lead to incorrect results. For example, with the pattern
2903 /1234/, partially matching against "123" and then "a4" gave a match.
2904
2905 2. Changes to pcregrep:
2906
2907 (a) All non-match returns from pcre_exec() were being treated as failures
2908 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2909 error message is output. Some extra information is given for the
2910 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2911 probably the only errors that are likely to be caused by users (by
2912 specifying a regex that has nested indefinite repeats, for instance).
2913 If there are more than 20 of these errors, pcregrep is abandoned.
2914
2915 (b) A binary zero was treated as data while matching, but terminated the
2916 output line if it was written out. This has been fixed: binary zeroes
2917 are now no different to any other data bytes.
2918
2919 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2920 used to set a locale for matching. The --locale=xxxx long option has
2921 been added (no short equivalent) to specify a locale explicitly on the
2922 pcregrep command, overriding the environment variables.
2923
2924 (d) When -B was used with -n, some line numbers in the output were one less
2925 than they should have been.
2926
2927 (e) Added the -o (--only-matching) option.
2928
2929 (f) If -A or -C was used with -c (count only), some lines of context were
2930 accidentally printed for the final match.
2931
2932 (g) Added the -H (--with-filename) option.
2933
2934 (h) The combination of options -rh failed to suppress file names for files
2935 that were found from directory arguments.
2936
2937 (i) Added the -D (--devices) and -d (--directories) options.
2938
2939 (j) Added the -F (--fixed-strings) option.
2940
2941 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2942
2943 (l) Added the --colo(u)r option.
2944
2945 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2946 is not present by default.
2947
2948 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2949 items such as (?R) or (?1), when the recursion could match a number of
2950 alternatives. If it matched one of the alternatives, but subsequently,
2951 outside the recursion, there was a failure, the code tried to back up into
2952 the recursion. However, because of the way PCRE is implemented, this is not
2953 possible, and the result was an incorrect result from the match.
2954
2955 In order to prevent this happening, the specification of recursion has
2956 been changed so that all such subpatterns are automatically treated as
2957 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2958
2959 4. I had overlooked the fact that, in some locales, there are characters for
2960 which isalpha() is true but neither isupper() nor islower() are true. In
2961 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2962 and ordfeminine) are like this. This affected the treatment of \w and \W
2963 when they appeared in character classes, but not when they appeared outside
2964 a character class. The bit map for "word" characters is now created
2965 separately from the results of isalnum() instead of just taking it from the
2966 upper, lower, and digit maps. (Plus the underscore character, of course.)
2967
2968 5. The above bug also affected the handling of POSIX character classes such as
2969 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2970 permanent tables. Instead, the bit maps for such a class were previously
2971 created as the appropriate unions of the upper, lower, and digit bitmaps.
2972 Now they are created by subtraction from the [[:word:]] class, which has
2973 its own bitmap.
2974
2975 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2976 It is created by subtracting the vertical space characters (\x09, \x0a,
2977 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2978 subtraction was done in the overall bitmap for a character class, meaning
2979 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2980 be recognized. This bug has been fixed.
2981
2982 7. Patches from the folks at Google:
2983
2984 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2985 real life, but is still worth protecting against".
2986
2987 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2988 regular expressions".
2989
2990 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2991 have it.
2992
2993 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2994 "configure" and the latter not, in order to fix a problem somebody had
2995 with compiling the Arg class on HP-UX.
2996
2997 (e) Improve the error-handling of the C++ wrapper a little bit.
2998
2999 (f) New tests for checking recursion limiting.
3000
3001 8. The pcre_memmove() function, which is used only if the environment does not
3002 have a standard memmove() function (and is therefore rarely compiled),
3003 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3004 returning a result (though PCRE never actually uses the result).
3005
3006 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3007 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3008 returned instead of calling malloc() with an overflowing number that would
3009 most likely cause subsequent chaos.
3010
3011 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3012
3013 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3014 with this option is matched, the nmatch and pmatch options of regexec() are
3015 ignored.
3016
3017 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3018 provided in case anyone wants to the the POSIX interface with UTF-8
3019 strings.
3020
3021 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3022 C++ linking (needed for some HP-UX environments).
3023
3024 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3025 (unused parameter) and in the pcre_printint() function (omitted "default"
3026 switch label when the default is to do nothing).
3027
3028 15. Added some code to make it possible, when PCRE is compiled as a C++
3029 library, to replace subject pointers for pcre_exec() with a smart pointer
3030 class, thus making it possible to process discontinuous strings.
3031
3032 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3033 much the same function. They were added by different people who were trying
3034 to make PCRE easy to compile on non-Unix systems. It has been suggested
3035 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3036 for compiling on Windows systems. I have therefore replaced it with
3037 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3038 defaults to "extern" for C or "extern C" for C++, which works fine on
3039 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3040 SCOPE with something explicit in config.h. In addition:
3041
3042 (a) pcreposix.h still had just "extern" instead of either of these macros;
3043 I have replaced it with PCRE_DATA_SCOPE.
3044
3045 (b) Functions such as _pcre_xclass(), which are internal to the library,
3046 but external in the C sense, all had PCRE_EXPORT in their definitions.
3047 This is apparently wrong for the Windows case, so I have removed it.
3048 (It makes no difference on Unix-like systems.)
3049
3050 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3051 of recursive calls to match(). This is different to MATCH_LIMIT because
3052 that limits the total number of calls to match(), not all of which increase
3053 the depth of recursion. Limiting the recursion depth limits the amount of
3054 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3055 when PCRE is compiled, and changed at run time. A patch from Google adds
3056 this functionality to the C++ interface.
3057
3058 18. Changes to the handling of Unicode character properties:
3059
3060 (a) Updated the table to Unicode 4.1.0.
3061
3062 (b) Recognize characters that are not in the table as "Cn" (undefined).
3063
3064 (c) I revised the way the table is implemented to a much improved format
3065 which includes recognition of ranges. It now supports the ranges that
3066 are defined in UnicodeData.txt, and it also amalgamates other
3067 characters into ranges. This has reduced the number of entries in the
3068 table from around 16,000 to around 3,000, thus reducing its size
3069 considerably. I realized I did not need to use a tree structure after
3070 all - a binary chop search is just as efficient. Having reduced the
3071 number of entries, I extended their size from 6 bytes to 8 bytes to
3072 allow for more data.
3073
3074 (d) Added support for Unicode script names via properties such as \p{Han}.
3075
3076 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3077 matching that character.
3078
3079 20. When matching a repeated Unicode property with a minimum greater than zero,
3080 (for example \pL{2,}), PCRE could look past the end of the subject if it
3081 reached it while seeking the minimum number of characters. This could
3082 happen only if some of the characters were more than one byte long, because
3083 there is a check for at least the minimum number of bytes.
3084
3085 21. Refactored the implementation of \p and \P so as to be more general, to
3086 allow for more different types of property in future. This has changed the
3087 compiled form incompatibly. Anybody with saved compiled patterns that use
3088 \p or \P will have to recompile them.
3089
3090 22. Added "Any" and "L&" to the supported property types.
3091
3092 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3093 but give a compile time error if the value is greater than 0xff.
3094
3095 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3096 accidentally not being installed or uninstalled.
3097
3098 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3099 made were to insert the current release number. This seemed silly, because
3100 it made things harder for people building PCRE on systems that don't run
3101 "configure". I have turned pcre.h into a distributed file, no longer built
3102 by "configure", with the version identification directly included. There is
3103 no longer a pcre.h.in file.
3104
3105 However, this change necessitated a change to the pcre-config script as
3106 well. It is built from pcre-config.in, and one of the substitutions was the
3107 release number. I have updated configure.ac so that ./configure now finds
3108 the release number by grepping pcre.h.
3109
3110 26. Added the ability to run the tests under valgrind.
3111
3112
3113 Version 6.4 05-Sep-05
3114 ---------------------
3115
3116 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3117 "--" to be printed when multiple files were scanned, even when none of the
3118 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3119 consider it to be a bug, and have restored the previous behaviour.
3120
3121 2. A couple of code tidies to get rid of compiler warnings.
3122
3123 3. The pcretest program used to cheat by referring to symbols in the library
3124 whose names begin with _pcre_. These are internal symbols that are not
3125 really supposed to be visible externally, and in some environments it is
3126 possible to suppress them. The cheating is now confined to including
3127 certain files from the library's source, which is a bit cleaner.
3128
3129 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3130 file's purpose clearer.
3131
3132 5. Reorganized pcre_ucp_findchar().
3133
3134
3135 Version 6.3 15-Aug-05
3136 ---------------------
3137
3138 1. The file libpcre.pc.in did not have general read permission in the tarball.
3139
3140 2. There were some problems when building without C++ support:
3141
3142 (a) If C++ support was not built, "make install" and "make test" still
3143 tried to test it.
3144
3145 (b) There were problems when the value of CXX was explicitly set. Some
3146 changes have been made to try to fix these, and ...
3147
3148 (c) --disable-cpp can now be used to explicitly disable C++ support.
3149
3150 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3151 backslash in a target when C++ was disabled. This confuses some
3152 versions of "make", apparently. Using an intermediate variable solves
3153 this. (Same for CPP_LOBJ.)
3154
3155 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3156 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3157 necessary on certain architectures.
3158
3159 4. Added a setting of -export-symbols-regex to the link command to remove
3160 those symbols that are exported in the C sense, but actually are local
3161 within the library, and not documented. Their names all begin with
3162 "_pcre_". This is not a perfect job, because (a) we have to except some
3163 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3164 available (and never for static libraries). I have made a note to try to
3165 find a way round (a) in the future.
3166
3167
3168 Version 6.2 01-Aug-05
3169 ---------------------
3170
3171 1. There was no test for integer overflow of quantifier values. A construction
3172 such as {1111111111111111} would give undefined results. What is worse, if
3173 a minimum quantifier for a parenthesized subpattern overflowed and became
3174 negative, the calculation of the memory size went wrong. This could have
3175 led to memory overwriting.
3176
3177 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3178
3179 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3180 operating environments where this matters.
3181
3182 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3183 PCRE options from within the C++ wrapper.
3184
3185 5. Named capturing subpatterns were not being correctly counted when a pattern
3186 was compiled. This caused two problems: (a) If there were more than 100
3187 such subpatterns, the calculation of the memory needed for the whole
3188 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3189 back references of the form \12, where the number was greater than 9, were
3190 not recognized as back references, even though there were sufficient
3191 previous subpatterns.
3192
3193 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3194 versions of gcc, e.g. 2.95.4.
3195
3196
3197 Version 6.1 21-Jun-05
3198 ---------------------
3199
3200 1. There was one reference to the variable "posix" in pcretest.c that was not
3201 surrounded by "#if !defined NOPOSIX".
3202
3203 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3204 the cross-check on the old pcre_info() function, for the benefit of the
3205 cut-down version of PCRE that is currently imported into Exim.
3206
3207 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3208 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3209 patterns that start (?i) but I don't think that matters. The use of (?i) is
3210 just an example; this all applies to the other options as well.
3211
3212 4. Since libtool seems to echo the compile commands it is issuing, the output
3213 from "make" can be reduced a bit by putting "@" in front of each libtool
3214 compile command.
3215
3216 5. Patch from the folks at Google for configure.in to be a bit more thorough
3217 in checking for a suitable C++ installation before trying to compile the
3218 C++ stuff. This should fix a reported problem when a compiler was present,
3219 but no suitable headers.
3220
3221 6. The man pages all had just "PCRE" as their title. I have changed them to
3222 be the relevant file name. I have also arranged that these names are
3223 retained in the file doc/pcre.txt, which is a concatenation in text format
3224 of all the man pages except the little individual ones for each function.
3225
3226 7. The NON-UNIX-USE file had not been updated for the different set of source
3227 files that come with release 6. I also added a few comments about the C++
3228 wrapper.
3229
3230
3231 Version 6.0 07-Jun-05
3232 ---------------------
3233
3234 1. Some minor internal re-organization to help with my DFA experiments.
3235
3236 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3237 didn't matter for the library itself when fully configured, but did matter
3238 when compiling without UCP support, or within Exim, where the ucp files are
3239 not imported.
3240
3241 3. Refactoring of the library code to split up the various functions into
3242 different source modules. The addition of the new DFA matching code (see
3243 below) to a single monolithic source would have made it really too
3244 unwieldy, quite apart from causing all the code to be include in a
3245 statically linked application, when only some functions are used. This is
3246 relevant even without the DFA addition now that patterns can be compiled in
3247 one application and matched in another.
3248
3249 The downside of splitting up is that there have to be some external
3250 functions and data tables that are used internally in different modules of
3251 the library but which are not part of the API. These have all had their
3252 names changed to start with "_pcre_" so that they are unlikely to clash
3253 with other external names.
3254
3255 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3256 a different (DFA) algorithm. Although it is slower than the original
3257 function, it does have some advantages for certain types of matching
3258 problem.
3259
3260 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3261 including restarting after a partial match.
3262
3263 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3264 defined when compiling for Windows was sent to me. I have put it into the
3265 code, though I have no means of testing or verifying it.
3266
3267 7. Added the pcre_refcount() auxiliary function.
3268
3269 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3270 match before or at the first newline in the subject string. In pcretest,
3271 the /f option on a pattern can be used to set this.
3272
3273 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3274 would behave wrongly. This has been present in PCRE since release 4.0.
3275
3276 10. A number of changes to the pcregrep command:
3277
3278 (a) Refactored how -x works; insert ^(...)$ instead of setting
3279 PCRE_ANCHORED and checking the length, in preparation for adding
3280 something similar for -w.
3281
3282 (b) Added the -w (match as a word) option.
3283
3284 (c) Refactored the way lines are read and buffered so as to have more
3285 than one at a time available.
3286
3287 (d) Implemented a pcregrep test script.
3288
3289 (e) Added the -M (multiline match) option. This allows patterns to match
3290 over several lines of the subject. The buffering ensures that at least
3291 8K, or the rest of the document (whichever is the shorter) is available
3292 for matching (and similarly the previous 8K for lookbehind assertions).
3293
3294 (f) Changed the --help output so that it now says
3295
3296 -w, --word-regex(p)
3297
3298 instead of two lines, one with "regex" and the other with "regexp"
3299 because that confused at least one person since the short forms are the
3300 same. (This required a bit of code, as the output is generated
3301 automatically from a table. It wasn't just a text change.)
3302
3303 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3304 option but starts with a hyphen. Could be a pattern or a path name
3305 starting with a hyphen, for instance.
3306
3307 (h) "-" can be given as a file name to represent stdin.
3308
3309 (i) When file names are being printed, "(standard input)" is used for
3310 the standard input, for compatibility with GNU grep. Previously
3311 "<stdin>" was used.
3312
3313 (j) The option --label=xxx can be used to supply a name to be used for
3314 stdin when file names are being printed. There is no short form.
3315
3316 (k) Re-factored the options decoding logic because we are going to add
3317 two more options that take data. Such options can now be given in four
3318 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3319
3320 (l) Added the -A, -B, and -C options for requesting that lines of context
3321 around matches be printed.
3322
3323 (m) Added the -L option to print the names of files that do not contain
3324 any matching lines, that is, the complement of -l.
3325
3326 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3327 continue to scan other files.
3328
3329 (o) The -s option was incorrectly implemented. For compatibility with other
3330 greps, it now suppresses the error message for a non-existent or non-
3331 accessible file (but not the return code). There is a new option called
3332 -q that suppresses the output of matching lines, which was what -s was
3333 previously doing.
3334
3335 (p) Added --include and --exclude options to specify files for inclusion
3336 and exclusion when recursing.
3337
3338 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3339 Hopefully, it now does.
3340
3341 12. Missing cast in pcre_study().
3342
3343 13. Added an "uninstall" target to the makefile.
3344
3345 14. Replaced "extern" in the function prototypes in Makefile.in with
3346 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3347 world, but is set differently for Windows.
3348
3349 15. Added a second compiling function called pcre_compile2(). The only
3350 difference is that it has an extra argument, which is a pointer to an
3351 integer error code. When there is a compile-time failure, this is set
3352 non-zero, in addition to the error test pointer being set to point to an
3353 error message. The new argument may be NULL if no error number is required
3354 (but then you may as well call pcre_compile(), which is now just a
3355 wrapper). This facility is provided because some applications need a
3356 numeric error indication, but it has also enabled me to tidy up the way
3357 compile-time errors are handled in the POSIX wrapper.
3358
3359 16. Added VPATH=.libs to the makefile; this should help when building with one
3360 prefix path and installing with another. (Or so I'm told by someone who
3361 knows more about this stuff than I do.)
3362
3363 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3364 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3365 match everything, including newlines. This is not POSIX-compatible, but
3366 somebody wanted the feature. From pcretest it can be activated by using
3367 both the P and the s flags.
3368
3369 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3370
3371 19. libpcre.pc was being incorrectly installed as executable.
3372
3373 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3374 it now also looks for '\r' so that it will work unmodified on Windows.
3375
3376 21. Added Google's contributed C++ wrapper to the distribution.
3377
3378 22. Added some untidy missing memory free() calls in pcretest, to keep
3379 Electric Fence happy when testing.
3380
3381
3382
3383 Version 5.0 13-Sep-04
3384 ---------------------
3385
3386 1. Internal change: literal characters are no longer packed up into items
3387 containing multiple characters in a single byte-string. Each character
3388 is now matched using a separate opcode. However, there may be more than one
3389 byte in the character in UTF-8 mode.
3390
3391 2. The pcre_callout_block structure has two new fields: pattern_position and
3392 next_item_length. These contain the offset in the pattern to the next match
3393 item, and its length, respectively.
3394
3395 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3396 insertion of callouts before each pattern item. Added the /C option to
3397 pcretest to make use of this.
3398
3399 4. On the advice of a Windows user, the lines
3400
3401 #if defined(_WIN32) || defined(WIN32)
3402 _setmode( _fileno( stdout ), 0x8000 );
3403 #endif /* defined(_WIN32) || defined(WIN32) */
3404
3405 have been added to the source of pcretest. This apparently does useful
3406 magic in relation to line terminators.
3407
3408 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3409 for the benefit of those environments where the "b" makes a difference.
3410
3411 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3412 to know about it. I have put a hack into configure.in that adds in code
3413 to set GCC=yes if CC=icc. This seems to end up at a point in the
3414 generated configure script that is early enough to affect the setting of
3415 compiler options, which is what is needed, but I have no means of testing
3416 whether it really works. (The user who reported this had patched the
3417 generated configure script, which of course I cannot do.)
3418
3419 LATER: After change 22 below (new libtool files), the configure script
3420 seems to know about icc (and also ecc). Therefore, I have commented out
3421 this hack in configure.in.
3422
3423 7. Added support for pkg-config (2 patches were sent in).
3424
3425 8. Negated POSIX character classes that used a combination of internal tables
3426 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3427 [[:^ascii]]. Typically, they would match almost any characters. The other
3428 POSIX classes were not broken in this way.
3429
3430 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3431 to find the match, as PCRE was deluded into thinking that the match had to
3432 start at the start point or following a newline. The same bug applied to
3433 patterns with negative forward assertions or any backward assertions
3434 preceding ".*" at the start, unless the pattern required a fixed first
3435 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3436
3437 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3438 starting at the last subject character, bytes beyond the end of the subject
3439 string were read.
3440
3441 11. Renamed the variable "class" as "classbits" to make life easier for C++
3442 users. (Previously there was a macro definition, but it apparently wasn't
3443 enough.)
3444
3445 12. Added the new field "tables" to the extra data so that tables can be passed
3446 in at exec time, or the internal tables can be re-selected. This allows
3447 a compiled regex to be saved and re-used at a later time by a different
3448 program that might have everything at different addresses.
3449
3450 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3451 -R library as well as a -L library.
3452
3453 14. The debugging options of pcretest (-d on the command line or D on a
3454 pattern) showed incorrect output for anything following an extended class
3455 that contained multibyte characters and which was followed by a quantifier.
3456
3457 15. Added optional support for general category Unicode character properties
3458 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3459 support. It adds about 90K to the size of the library. The meanings of the
3460 inbuilt class escapes such as \d and \s have NOT been changed.
3461
3462 16. Updated pcredemo.c to include calls to free() to release the memory for the
3463 compiled pattern.
3464
3465 17. The generated file chartables.c was being created in the source directory
3466 instead of in the building directory. This caused the build to fail if the
3467 source directory was different from the building directory, and was
3468 read-only.
3469
3470 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3471 file. No doubt somebody will tell me if they don't make sense... Also added
3472 Dan Mooney's comments about building on OpenVMS.
3473
3474 19. Added support for partial matching via the PCRE_PARTIAL option for
3475 pcre_exec() and the \P data escape in pcretest.
3476
3477 20. Extended pcretest with 3 new pattern features:
3478
3479 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3480 write the compiled pattern to the file whose name is "rest-of-line".
3481 This is a straight binary dump of the data, with the saved pointer to
3482 the character tables forced to be NULL. The study data, if any, is
3483 written too. After writing, pcretest reads a new pattern.
3484
3485 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3486 compiled pattern from the given file. There must not be any
3487 occurrences of "<" in the file name (pretty unlikely); if there are,
3488 pcretest will instead treat the initial "<" as a pattern delimiter.
3489 After reading in the pattern, pcretest goes on to read data lines as
3490 usual.
3491
3492 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3493 and 16-bit fields in a compiled pattern, to simulate a pattern that
3494 was compiled on a host of opposite endianness.
3495
3496 21. The pcre-exec() function can now cope with patterns that were compiled on
3497 hosts of opposite endianness, with this restriction:
3498
3499 As for any compiled expression that is saved and used later, the tables
3500 pointer field cannot be preserved; the extra_data field in the arguments
3501 to pcre_exec() should be used to pass in a tables address if a value
3502 other than the default internal tables were used at compile time.
3503
3504 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3505 now diagnosed as an error. Previously, most of the time, a negative number
3506 would have been treated as zero, but if in addition "ovector" was passed as
3507 NULL, a crash could occur.
3508
3509 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3510 new versions from the libtool 1.5 distribution (the last one is a copy of
3511 a file called libtool.m4). This seems to have fixed the need to patch
3512 "configure" to support Darwin 1.3 (which I used to do). However, I still
3513 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3514 workstation).
3515
3516 24. Changed the PCRE licence to be the more standard "BSD" licence.
3517
3518
3519 Version 4.5 01-Dec-03
3520 ---------------------
3521
3522 1. There has been some re-arrangement of the code for the match() function so
3523 that it can be compiled in a version that does not call itself recursively.
3524 Instead, it keeps those local variables that need separate instances for
3525 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3526 needs to "recurse". Keeping track of where control must go is done by means
3527 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3528 hide most of the details from the main code, and operates only if
3529 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3530 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3531 operating.
3532
3533 To make it easier for callers to provide specially tailored get/free
3534 functions for this usage, two new functions, pcre_stack_malloc, and
3535 pcre_stack_free, are used. They are always called in strict stacking order,
3536 and the size of block requested is always the same.
3537
3538 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3539 PCRE has been compiled to use the stack or the heap for recursion. The
3540 -C option of pcretest uses this to show which version is compiled.
3541
3542 A new data escape \S, is added to pcretest; it causes the amounts of store
3543 obtained and freed by both kinds of malloc/free at match time to be added
3544 to the output.
3545
3546 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3547 what's available on my current Linux desktop machine.
3548
3549 3. When matching a UTF-8 string, the test for a valid string at the start has
3550 been extended. If start_offset is not zero, PCRE now checks that it points
3551 to a byte that is the start of a UTF-8 character. If not, it returns
3552 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3553 this is necessary because there may be backward assertions in the pattern.
3554 When matching the same subject several times, it may save resources to use
3555 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3556
3557 4. The code for checking the validity of UTF-8 strings has been tightened so
3558 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3559 containing "overlong sequences".
3560
3561 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3562 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3563 should have been "&", but it just so happened that all the cases this let
3564 through by mistake were picked up later in the function.
3565
3566 6. I had used a variable called "isblank" - this is a C99 function, causing
3567 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3568
3569 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3570 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3571 so the version is shown; (c) stop "make test" echoing "./RunTest".
3572
3573 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3574
3575 9. The prototype for memmove() for systems that don't have it was using
3576 size_t, but the inclusion of the header that defines size_t was later. I've
3577 moved the #includes for the C headers earlier to avoid this.
3578
3579 10. Added some adjustments to the code to make it easier to compiler on certain
3580 special systems:
3581
3582 (a) Some "const" qualifiers were missing.
3583 (b) Added the macro EXPORT before all exported functions; by default this
3584 is defined to be empty.
3585 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3586 that it reads its output file name as an argument instead of writing
3587 to the standard output and assuming this can be redirected.
3588
3589 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3590 class containing characters with values greater than 255, PCRE compilation
3591 went into a loop.
3592
3593 12. A recursive reference to a subpattern that was within another subpattern
3594 that had a minimum quantifier of zero caused PCRE to crash. For example,
3595 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3596 recursion. If the recursively-called subpattern itself had a zero repeat,
3597 that was OK.
3598
3599 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3600 buffer into which it was copied (for escape processing) was still set at
3601 1024, so long lines caused crashes.
3602
3603 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3604 "internal error: code overflow...". This applied to any character class
3605 that was followed by a possessive quantifier.
3606
3607 15. Modified the Makefile to add libpcre.la as a prerequisite for
3608 libpcreposix.la because I was told this is needed for a parallel build to
3609 work.
3610
3611 16. If a pattern that contained .* following optional items at the start was
3612 studied, the wrong optimizing data was generated, leading to matching
3613 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3614 matching string must start with a or b or c. The correct conclusion for
3615 this pattern is that a match can start with any character.
3616
3617
3618 Version 4.4 13-Aug-03
3619 ---------------------
3620
3621 1. In UTF-8 mode, a character class containing characters with values between
3622 127 and 255 was not handled correctly if the compiled pattern was studied.
3623 In fixing this, I have also improved the studying algorithm for such
3624 classes (slightly).
3625
3626 2. Three internal functions had redundant arguments passed to them. Removal
3627 might give a very teeny performance improvement.
3628
3629 3. Documentation bug: the value of the capture_top field in a callout is *one
3630 more than* the number of the hightest numbered captured substring.
3631
3632 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3633 in incorrectly linking with a previously installed version. They now link
3634 explicitly with libpcre.la.
3635
3636 5. configure.in no longer needs to recognize Cygwin specially.
3637
3638 6. A problem in pcre.in for Windows platforms is fixed.
3639
3640 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3641 pcretest, it used to include the size of the study block as part of its
3642 output. Unfortunately, the structure contains a field that has a different
3643 size on different hardware architectures. This meant that the tests that
3644 showed this size failed. As the block is currently always of a fixed size,
3645 this information isn't actually particularly useful in pcretest output, so
3646 I have just removed it.
3647
3648 8. Three pre-processor statements accidentally did not start in column 1.
3649 Sadly, there are *still* compilers around that complain, even though
3650 standard C has not required this for well over a decade. Sigh.
3651
3652 9. In pcretest, the code for checking callouts passed small integers in the
3653 callout_data field, which is a void * field. However, some picky compilers
3654 complained about the casts involved for this on 64-bit systems. Now
3655 pcretest passes the address of the small integer instead, which should get
3656 rid of the warnings.
3657
3658 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3659 both compile and run time, and gives an error if an invalid UTF-8 sequence
3660 is found. There is a option for disabling this check in cases where the
3661 string is known to be correct and/or the maximum performance is wanted.
3662
3663 11. In response to a bug report, I changed one line in Makefile.in from
3664
3665 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3666 to
3667 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3668
3669 to look similar to other lines, but I have no way of telling whether this
3670 is the right thing to do, as I do not use Windows. No doubt I'll get told
3671 if it's wrong...
3672
3673
3674 Version 4.3 21-May-03
3675 ---------------------
3676
3677 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3678 Makefile.
3679
3680 2. Some refactoring to improve the quality of the code:
3681
3682 (i) The utf8_table... variables are now declared "const".
3683
3684 (ii) The code for \cx, which used the "case flipping" table to upper case
3685 lower case letters, now just substracts 32. This is ASCII-specific,
3686 but the whole concept of \cx is ASCII-specific, so it seems
3687 reasonable.
3688
3689 (iii) PCRE was using its character types table to recognize decimal and
3690 hexadecimal digits in the pattern. This is silly, because it handles
3691 only 0-9, a-f, and A-F, but the character types table is locale-
3692 specific, which means strange things might happen. A private
3693 table is now used for this - though it costs 256 bytes, a table is
3694 much faster than multiple explicit tests. Of course, the standard
3695 character types table is still used for matching digits in subject
3696 strings against \d.
3697
3698 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3699 ending in _t are). So I've renamed it as ESC_tee.
3700
3701 3. The first argument for regexec() in the POSIX wrapper should have been
3702 defined as "const".
3703
3704 4. Changed pcretest to use malloc() for its buffers so that they can be
3705 Electric Fenced for debugging.
3706
3707 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3708 to read one or more bytes before the start of the subject string. Often this
3709 had no effect on PCRE's behaviour, but in some circumstances it could
3710 provoke a segmentation fault.
3711
3712 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3713 to try to read one or more bytes before the start of the subject string.
3714
3715 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3716 UTF-8 support could misbehave in various ways if the subject string
3717 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3718 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3719 back over UTF-8 characters.)
3720
3721
3722 Version 4.2 14-Apr-03
3723 ---------------------
3724
3725 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3726
3727 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3728 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3729 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3730 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3731 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3732 and BUILD_EXEEXT
3733 Note: automatic setting of the BUILD variables is not yet working
3734 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3735 compile-time but not at link-time
3736 [LINK]: use for linking executables only
3737 make different versions for Windows and non-Windows
3738 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3739 libraries
3740 [LINK_FOR_BUILD]: new variable
3741 [OBJEXT]: use throughout
3742 [EXEEXT]: use throughout
3743 <winshared>: new target
3744 <wininstall>: new target
3745 <dftables.o>: use native compiler
3746 <dftables>: use native linker
3747 <install>: handle Windows platform correctly
3748 <clean>: ditto
3749 <check>: ditto
3750 copy DLL to top builddir before testing
3751
3752 As part of these changes, -no-undefined was removed again. This was reported
3753 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3754 in any case.
3755
3756 3. Some tidies to get rid of compiler warnings:
3757
3758 . In the match_data structure, match_limit was an unsigned long int, whereas
3759 match_call_count was an int. I've made them both unsigned long ints.
3760
3761 . In pcretest the fact that a const uschar * doesn't automatically cast to
3762 a void * provoked a warning.
3763
3764 . Turning on some more compiler warnings threw up some "shadow" variables
3765 and a few more missing casts.
3766
3767 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3768 option, a class that contained a single character with a value between 128
3769 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3770
3771 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3772 option, a class that contained several characters, but with at least one
3773 whose value was between 128 and 255 caused PCRE to crash.
3774
3775
3776 Version 4.1 12-Mar-03
3777 ---------------------
3778
3779 1. Compiling with gcc -pedantic found a couple of places where casts were
3780 needed, and a string in dftables.c that was longer than standard compilers are
3781 required to support.
3782
3783 2. Compiling with Sun's compiler found a few more places where the code could
3784 be tidied up in order to avoid warnings.
3785
3786 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3787 first of these names is deprecated in the latest Autoconf in favour of the name
3788 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3789 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3790 analogy I have changed it to CFLAGS_FOR_BUILD.
3791
3792 4. Added -no-undefined to the linking command in the Makefile, because this is
3793 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3794 linking step for the pcreposix library.
3795
3796 5. PCRE was failing to diagnose the case of two named groups with the same
3797 name.
3798
3799 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3800 literal character that is needed in the subject for a match, and scans along to
3801 ensure that it is present before embarking on the full matching process. This
3802 saves time in cases of nested unlimited repeats that are never going to match.
3803 Problem: the scan can take a lot of time if the subject is very long (e.g.
3804 megabytes), thus penalizing straightforward matches. It is now done only if the
3805 amount of subject to be scanned is less than 1000 bytes.
3806
3807 7. A lesser problem with the same optimization is that it was recording the
3808 first character of an anchored pattern as "needed", thus provoking a search
3809 right along the subject, even when the first match of the pattern was going to
3810 fail. The "needed" character is now not set for anchored patterns, unless it
3811 follows something in the pattern that is of non-fixed length. Thus, it still
3812 fulfils its original purpose of finding quick non-matches in cases of nested
3813 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3814
3815
3816 Version 4.0 17-Feb-03
3817 ---------------------
3818
3819 1. If a comment in an extended regex that started immediately after a meta-item
3820 extended to the end of string, PCRE compiled incorrect data. This could lead to
3821 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3822
3823 2. Moved to autoconf 2.53 and libtool 1.4.2.
3824
3825 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3826 the special perltest8 script is no longer needed - all the tests can be run
3827 from a single perltest script.
3828
3829 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3830 by \s. It has now been removed in PCRE. This means it isn't recognized as
3831 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3832 class [:space:] *does* include VT, thereby creating a mess.
3833
3834 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3835 space and tab.
3836
3837 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3838 its new features into the main test script, reducing the number of scripts.
3839
3840 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3841 were backward compatible, and made the (?i) apply to the whole pattern, as if
3842 /i were given. Now it behaves more logically, and applies the option setting
3843 only to what follows. PCRE has been changed to follow suit. However, if it
3844 finds options settings right at the start of the pattern, it extracts them into
3845 the global options, as before. Thus, they show up in the info data.
3846
3847 8. Added support for the \Q...\E escape sequence. Characters in between are
3848 treated as literals. This is slightly different from Perl in that $ and @ are
3849 also handled as literals inside the quotes. In Perl, they will cause variable
3850 interpolation. Note the following examples:
3851
3852 Pattern PCRE matches Perl matches
3853
3854 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3855 \Qabc\$xyz\E abc\$xyz abc\$xyz
3856 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3857
3858 For compatibility with Perl, \Q...\E sequences are recognized inside character
3859 classes as well as outside them.
3860
3861 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3862 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3863 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3864 signed/unsigned warnings.
3865
3866 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3867 option for pcretest, so I've replaced it by a simple function that does just
3868 that job.
3869
3870 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3871 "pcregrep -".
3872
3873 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3874 Java package. This provides some syntactic sugar for simple cases of what my
3875 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3876 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3877 item, you can use this simplified notation. Note that only makes sense with
3878 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3879 greediness, whatever the setting of the PCRE_UNGREEDY option.
3880
3881 13. A change of greediness default within a pattern was not taking effect at
3882 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3883 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3884 was abstracted outside.
3885
3886 14. PCRE now supports the \G assertion. It is true when the current matching
3887 position is at the start point of the match. This differs from \A when the
3888 starting offset is non-zero. Used with the /g option of pcretest (or similar
3889 code), it works in the same way as it does for Perl's /g option. If all
3890 alternatives of a regex begin with \G, the expression is anchored to the start
3891 match position, and the "anchored" flag is set in the compiled expression.
3892
3893 15. Some bugs concerning the handling of certain option changes within patterns
3894 have been fixed. These applied to options other than (?ims). For example,
3895 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3896 been the other way round. Some of this was related to change 7 above.
3897
3898 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3899 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3900 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3901 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3902
3903 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3904 mode. Unlike ".", it always matches newline, whatever the setting of
3905 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3906 assertions. Perl allows it, but it doesn't (in general) work because it can't
3907 calculate the length of the lookbehind. At least, that's the case for Perl
3908 5.8.0 - I've been told they are going to document that it doesn't work in
3909 future.
3910
3911 18. Added an error diagnosis for escapes that PCRE does not support: these are
3912 \L, \l, \N, \P, \p, \U, \u, and \X.
3913
3914 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3915 reading past the end of the pattern in cases such as /[abcd/.
3916
3917 20. PCRE was getting more memory than necessary for patterns with classes that
3918 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3919
3920 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3921 compiling PCRE for use with Virtual Pascal.
3922
3923 22. Small fix to the Makefile to make it work properly if the build is done
3924 outside the source tree.
3925
3926 23. Added a new extension: a condition to go with recursion. If a conditional
3927 subpattern starts with (?(R) the "true" branch is used if recursion has
3928 happened, whereas the "false" branch is used only at the top level.
3929
3930 24. When there was a very long string of literal characters (over 255 bytes
3931 without UTF support, over 250 bytes with UTF support), the computation of how
3932 much memory was required could be incorrect, leading to segfaults or other
3933 strange effects.
3934
3935 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3936 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3937 there was a subsequent back reference to those brackets. This meant that, for
3938 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3939 possible to check for precisely this case. All we can do is abandon the
3940 optimization if .* occurs inside capturing brackets when there are any back
3941 references whatsoever. (See below for a better fix that came later.)
3942
3943 26. The handling of the optimization for finding the first character of a
3944 non-anchored pattern, and for finding a character that is required later in the
3945 match were failing in some cases. This didn't break the matching; it just
3946 failed to optimize when it could. The way this is done has been re-implemented.
3947
3948 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3949
3950 28. Added a new feature that provides some of the functionality that Perl
3951 provides with (?{...}). The facility is termed a "callout". The way it is done
3952 in PCRE is for the caller to provide an optional function, by setting
3953 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3954 global variable. By default it is unset, which disables all calling out. To get
3955 the function called, the regex must include (?C) at appropriate points. This
3956 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3957 This provides a means of identifying different callout points. When PCRE
3958 reaches such a point in the regex, if pcre_callout has been set, the external
3959 function is called. It is provided with data in a structure called
3960 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3961 matching continues; if it returns a non-zero value, the match at the current
3962 point fails. However, backtracking will occur if possible. [This was changed
3963 later and other features added - see item 49 below.]
3964
3965 29. pcretest is upgraded to test the callout functionality. It provides a
3966 callout function that displays information. By default, it shows the start of
3967 the match and the current position in the text. There are some new data escapes
3968 to vary what happens:
3969
3970 \C+ in addition, show current contents of captured substrings
3971 \C- do not supply a callout function
3972 \C!n return 1 when callout number n is reached
3973 \C!n!m return 1 when callout number n is reached for the mth time
3974
3975 30. If pcregrep was called with the -l option and just a single file name, it
3976 output "<stdin>" if a match was found, instead of the file name.
3977
3978 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3979 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3980 pcre_exec(). This saves a malloc/free per call. The default value of
3981 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3982 when configuring.
3983
3984 32. The default maximum size of a compiled pattern is 64K. There have been a
3985 few cases of people hitting this limit. The code now uses macros to handle the
3986 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3987 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3988 configuring. Tests 2 and 5 work only with 2-byte links because they output
3989 debugging information about compiled patterns.
3990
3991 33. Internal code re-arrangements:
3992
3993 (a) Moved the debugging function for printing out a compiled regex into
3994 its own source file (printint.c) and used #include to pull it into
3995 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3996 separate copies.
3997
3998 (b) Defined the list of op-code names for debugging as a macro in
3999 internal.h so that it is next to the definition of the opcodes.
4000
4001 (c) Defined a table of op-code lengths for simpler skipping along compiled
4002 code. This is again a macro in internal.h so that it is next to the
4003 definition of the opcodes.
4004
4005 34. Added support for recursive calls to individual subpatterns, along the
4006 lines of Robin Houston's patch (but implemented somewhat differently).
4007
4008 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4009 allow it to read and process whole directories in Win32. This code was
4010 contributed by Lionel Fourquaux; it has not been tested by me.
4011
4012 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4013 used to name a group. Names consist of alphanumerics and underscores, and must
4014 be unique. Back references use the syntax (?P=name) and recursive calls use
4015 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4016 numbers. The function pcre_fullinfo() can be used after compilation to extract
4017 a name/number map. There are three relevant calls:
4018
4019 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4020 PCRE_INFO_NAMECOUNT yields the number of entries
4021 PCRE_INFO_NAMETABLE yields a pointer to the map.
4022
4023 The map is a vector of fixed-size entries. The size of each entry depends on
4024 the length of the longest name used. The first two bytes of each entry are the
4025 group number, most significant byte first. There follows the corresponding
4026 name, zero terminated. The names are in alphabetical order.
4027
4028 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4029 case instead of 255. Making it the same both with and without UTF-8 support
4030 means that the same test output works with both.
4031
4032 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4033 calling malloc() with a zero argument.
4034
4035 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4036 optimization. I've improved things by keeping a bitmap of backreferences with
4037 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4038 fact referenced, the optimization can be applied. It is unlikely that a
4039 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4040 the match to follow \n) will appear inside brackets with a number greater than
4041 31, but if it does, any back reference > 31 suppresses the optimization.
4042
4043 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4044 of disabling numbered capturing parentheses. Any opening parenthesis that is
4045 not followed by ? behaves as if it were followed by ?: but named parentheses
4046 can still be used for capturing (and they will acquire numbers in the usual
4047 way).
4048
4049 41. Redesigned the return codes from the match() function into yes/no/error so
4050 that errors can be passed back from deep inside the nested calls. A malloc
4051 failure while inside a recursive subpattern call now causes the
4052 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4053
4054 42. It is now possible to set a limit on the number of times the match()
4055 function is called in a call to pcre_exec(). This facility makes it possible to
4056 limit the amount of recursion and backtracking, though not in a directly
4057 obvious way, because the match() function is used in a number of different
4058 circumstances. The count starts from zero for each position in the subject
4059 string (for non-anchored patterns). The default limit is, for compatibility, a
4060 large number, namely 10 000 000. You can change this in two ways:
4061
4062 (a) When configuring PCRE before making, you can use --with-match-limit=n
4063 to set a default value for the compiled library.
4064
4065 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4066 a different value is set. See 45 below.
4067
4068 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4069
4070 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4071 of things that can be changed at compile time. The first argument specifies
4072 what is wanted and the second points to where the information is to be placed.
4073 The current list of available information is:
4074
4075 PCRE_CONFIG_UTF8
4076
4077 The output is an integer that is set to one if UTF-8 support is available;
4078 otherwise it is set to zero.
4079
4080 PCRE_CONFIG_NEWLINE
4081
4082 The output is an integer that it set to the value of the code that is used for
4083 newline. It is either LF (10) or CR (13).
4084
4085 PCRE_CONFIG_LINK_SIZE
4086
4087 The output is an integer that contains the number of bytes used for internal
4088 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4089
4090 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4091
4092 The output is an integer that contains the threshold above which the POSIX
4093 interface uses malloc() for output vectors. See item 31 above.
4094
4095 PCRE_CONFIG_MATCH_LIMIT
4096
4097 The output is an unsigned integer that contains the default limit of the number
4098 of match() calls in a pcre_exec() execution. See 42 above.
4099
4100 44. pcretest has been upgraded by the addition of the -C option. This causes it
4101 to extract all the available output from the new pcre_config() function, and to
4102 output it. The program then exits immediately.
4103
4104 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4105 order to support additional features. One way would have been to define
4106 pcre_exec2() (for example) with extra arguments, but this would not have been
4107 extensible, and would also have required all calls to the original function to
4108 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4109 is used for passing in "extra" data from pcre_study().
4110
4111 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4112 contains the following fields:
4113
4114 flags a bitmap indicating which of the following fields are set
4115 study_data opaque data from pcre_study()
4116 match_limit a way of specifying a limit on match() calls for a specific
4117 call to pcre_exec()
4118 callout_data data for callouts (see 49 below)
4119
4120 The flag bits are also defined in pcre.h, and are
4121
4122 PCRE_EXTRA_STUDY_DATA
4123 PCRE_EXTRA_MATCH_LIMIT
4124 PCRE_EXTRA_CALLOUT_DATA
4125
4126 The pcre_study() function now returns one of these new pcre_extra blocks, with
4127 the actual study data pointed to by the study_data field, and the
4128 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4129 before. That is, this change is entirely upwards-compatible and requires no
4130 change to existing code.
4131
4132 If you want to pass in additional data to pcre_exec(), you can either place it
4133 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4134 block.
4135
4136 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4137 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4138 times with different match limits, until it finds the minimum value needed for
4139 pcre_exec() to complete. The value is then output. This can be instructive; for
4140 most simple matches the number is quite small, but for pathological cases it
4141 gets very large very quickly.
4142
4143 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4144 returns the size of the data block pointed to by the study_data field in a
4145 pcre_extra block, that is, the value that was passed as the argument to
4146 pcre_malloc() when PCRE was getting memory in which to place the information
4147 created by pcre_study(). The fourth argument should point to a size_t variable.
4148 pcretest has been extended so that this information is shown after a successful
4149 pcre_study() call when information about the compiled regex is being displayed.
4150
4151 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4152 because what follows is always an absolute path. (Later: it turns out that this
4153 is more than cosmetic for MinGW, because it doesn't like empty path
4154 components.)
4155
4156 49. Some changes have been made to the callout feature (see 28 above):
4157
4158 (i) A callout function now has three choices for what it returns:
4159
4160 0 => success, carry on matching
4161 > 0 => failure at this point, but backtrack if possible
4162 < 0 => serious error, return this value from pcre_exec()
4163
4164 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4165 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4166 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4167 use by callout functions. It will never be used by PCRE itself.
4168
4169 (ii) The pcre_extra structure (see 45 above) has a void * field called
4170 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4171 pcre_callout_block structure has a field of the same name. The contents of
4172 the field passed in the pcre_extra structure are passed to the callout
4173 function in the corresponding field in the callout block. This makes it
4174 easier to use the same callout-containing regex from multiple threads. For
4175 testing, the pcretest program has a new data escape
4176
4177 \C*n pass the number n (may be negative) as callout_data
4178
4179 If the callout function in pcretest receives a non-zero value as
4180 callout_data, it returns that value.
4181
4182 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4183 there were some redundant $(CFLAGS) in commands that are now specified as
4184 $(LINK), which already includes $(CFLAGS).
4185
4186 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4187 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4188 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4189 one-byte characters throughout. Note that case-insensitive matching applies
4190 only to characters whose values are less than 256. PCRE doesn't support the
4191 notion of cases for higher-valued characters.
4192
4193 (i) A character class whose characters are all within 0-255 is handled as
4194 a bit map, and the map is inverted for negative classes. Previously, a
4195 character > 255 always failed to match such a class; however it should
4196 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4197
4198 (ii) A negated character class with a single character < 255 is coded as
4199 "not this character" (OP_NOT). This wasn't working properly when the test
4200 character was multibyte, either singly or repeated.
4201
4202 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4203 mode, for example: \x{100}{2,3}.
4204
4205 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4206 singly or repeated) now correctly test multibyte characters. However,
4207 PCRE doesn't recognize any characters with values greater than 255 as
4208 digits, spaces, or word characters. Such characters always match \D, \S,
4209 and \W, and never match \d, \s, or \w.
4210
4211 (v) Classes may now contain characters and character ranges with values
4212 greater than 255. For example: [ab\x{100}-\x{400}].
4213
4214 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4215 PCRE in UTF-8 mode.
4216
4217 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4218 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4219 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4220 value.)
4221
4222 53. The single man page has become too large. I have therefore split it up into
4223 a number of separate man pages. These also give rise to individual HTML pages;
4224 these are now put in a separate directory, and there is an index.html page that
4225 lists them all. Some hyperlinking between the pages has been installed.
4226
4227 54. Added convenience functions for handling named capturing parentheses.
4228
4229 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4230 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4231 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4232 are faulted.
4233
4234 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4235 calling configure. These values are used when compiling the dftables.c program
4236 which is run to generate the source of the default character tables. They
4237 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4238 you will need to set these values.
4239
4240 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4241
4242
4243 Version 3.9 02-Jan-02
4244 ---------------------
4245
4246 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4247
4248 2. If --disable-static was given, the building process failed when trying to
4249 build pcretest and pcregrep. (For some reason it was using libtool to compile
4250 them, which is not right, as they aren't part of the library.)
4251
4252
4253 Version 3.8 18-Dec-01
4254 ---------------------
4255
4256 1. The experimental UTF-8 code was completely screwed up. It was packing the
4257 bytes in the wrong order. How dumb can you get?
4258
4259
4260 Version 3.7 29-Oct-01
4261 ---------------------
4262
4263 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4264 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4265 this didn't happen under Solaris 8, where I normally test things.
4266
4267 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4268 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4269 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4270 this fix an hour or so after the initial 3.7 release.)
4271
4272
4273 Version 3.6 23-Oct-01
4274 ---------------------
4275
4276 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4277 offsets passed as NULL with zero offset count.
4278
4279 2. The config.guess and config.sub files had not been updated when I moved to
4280 the latest autoconf.
4281
4282
4283 Version 3.5 15-Aug-01
4284 ---------------------
4285
4286 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4287 had been forgotten.
4288
4289 2. By using declared but undefined structures, we can avoid using "void"
4290 definitions in pcre.h while keeping the internal definitions of the structures
4291 private.
4292
4293 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4294 user point of view, this means that both static and shared libraries are built
4295 by default, but this can be individually controlled. More of the work of
4296 handling this static/shared cases is now inside libtool instead of PCRE's make
4297 file.
4298
4299 4. The pcretest utility is now installed along with pcregrep because it is
4300 useful for users (to test regexs) and by doing this, it automatically gets
4301 relinked by libtool. The documentation has been turned into a man page, so
4302 there are now .1, .txt, and .html versions in /doc.
4303
4304 5. Upgrades to pcregrep:
4305 (i) Added long-form option names like gnu grep.
4306 (ii) Added --help to list all options with an explanatory phrase.
4307 (iii) Added -r, --recursive to recurse into sub-directories.
4308 (iv) Added -f, --file to read patterns from a file.
4309
4310 6. pcre_exec() was referring to its "code" argument before testing that
4311 argument for NULL (and giving an error if it was NULL).
4312
4313 7. Upgraded Makefile.in to allow for compiling in a different directory from
4314 the source directory.
4315
4316 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4317 options bits, the pointer it was passed was to an int instead of to an unsigned
4318 long int. This mattered only on 64-bit systems.
4319
4320 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4321 generated) instead of pcre.in, which it its source. Also made the same change
4322 in several of the .c files.
4323
4324 10. A new release of gcc defines printf() as a macro, which broke pcretest
4325 because it had an ifdef in the middle of a string argument for printf(). Fixed
4326 by using separate calls to printf().
4327
4328 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4329 script, to force use of CR or LF instead of \n in the source. On non-Unix
4330 systems, the value can be set in config.h.
4331
4332 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4333 absolute limit. Changed the text of the error message to make this clear, and
4334 likewise updated the man page.
4335
4336 13. The limit of 99 on the number of capturing subpatterns has been removed.
4337 The new limit is 65535, which I hope will not be a "real" limit.
4338
4339
4340 Version 3.4 22-Aug-00
4341 ---------------------
4342
4343 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4344
4345 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4346
4347
4348 Version 3.3 01-Aug-00
4349 ---------------------
4350
4351 1. If an octal character was given, but the value was greater than \377, it
4352 was not getting masked to the least significant bits, as documented. This could
4353 lead to crashes in some systems.
4354
4355 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4356 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4357
4358 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4359 These just pass their arguments on to (pcre_free)(), but they are provided
4360 because some uses of PCRE bind it to non-C systems that can call its functions,
4361 but cannot call free() or pcre_free() directly.
4362
4363 4. Add "make test" as a synonym for "make check". Corrected some comments in
4364 the Makefile.
4365
4366 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4367 Makefile.
4368
4369 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4370 command called pgrep for grepping around the active processes.
4371
4372 7. Added the beginnings of support for UTF-8 character strings.
4373
4374 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4375 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4376 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4377 out for the ar command.)
4378
4379
4380 Version 3.2 12-May-00
4381 ---------------------
4382
4383 This is purely a bug fixing release.
4384
4385 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4386 of ZA. This was just one example of several cases that could provoke this bug,
4387 which was introduced by change 9 of version 2.00. The code for breaking
4388 infinite loops after an iteration that matches an empty string was't working
4389 correctly.
4390
4391 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4392 when matched against abbab (for example). After matching an empty string, it
4393 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4394 caused it to match further down the string than it should.
4395
4396 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4397 was there because it doesn't seem to be needed, and it causes trouble on some
4398 systems, as it is not a Standard C header. It has been removed.
4399
4400 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4401 were reported on the Macintosh. The changes were from
4402
4403 while ((c = *(++ptr)) != 0 && c != '\n');
4404 to
4405 while ((c = *(++ptr)) != 0 && c != '\n') ;
4406
4407 Totally extraordinary, but if that's what it takes...
4408
4409 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4410 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4411 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4412 assumes the way PCRE uses memmove() (always moving upwards).
4413
4414 6. PCRE is being used in one environment where strchr() is not available. There
4415 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4416 faster code anyway.
4417
4418
4419 Version 3.1 09-Feb-00
4420 ---------------------
4421
4422 The only change in this release is the fixing of some bugs in Makefile.in for
4423 the "install" target:
4424
4425 (1) It was failing to install pcreposix.h.
4426
4427 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4428
4429
4430 Version 3.0 01-Feb-00
4431 ---------------------
4432
4433 1. Add support for the /+ modifier to perltest (to output $` like it does in
4434 pcretest).
4435
4436 2. Add support for the /g modifier to perltest.
4437
4438 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4439 matches null strings.
4440
4441 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4442 pattern. Perl treats empty patterns specially - it reuses the most recent
4443 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4444 effect.
4445
4446 5. The POSIX interface was broken in that it was just handing over the POSIX
4447 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4448 required a bigger vector, with some working space on the end. This means that
4449 the POSIX wrapper now has to get and free some memory, and copy the results.
4450
4451 6. Added some simple autoconf support, placing the test data and the
4452 documentation in separate directories, re-organizing some of the
4453 information files, and making it build pcre-config (a GNU standard). Also added
4454 libtool support for building PCRE as a shared library, which is now the
4455 default.
4456
4457 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4458 09 are not valid octal constants. Single digits will be used for minor values
4459 less than 10.
4460
4461 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4462 existing programs that set these in the POSIX interface can use PCRE without
4463 modification.
4464
4465 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4466 return all that pcre_info() returns, plus additional data. The pcre_info()
4467 function is retained for compatibility, but is considered to be obsolete.
4468
4469 10. Added experimental recursion feature (?R) to handle one common case that
4470 Perl 5.6 will be able to do with (?p{...}).
4471
4472 11. Added support for POSIX character classes like [:alpha:], which Perl is
4473 adopting.
4474
4475
4476 Version 2.08 31-Aug-99
4477 ----------------------
4478
4479 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4480 trying to match at the startoffset position, but instead was moving forward to
4481 the next newline as if a previous match had failed.
4482
4483 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4484 and could get into a loop if a null string was matched other than at the start
4485 of the subject.
4486
4487 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4488 be distinguished at compile time, and for completeness also added PCRE_DATE.
4489
4490 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4491 in GnuWin32 environments.
4492
4493
4494 Version 2.07 29-Jul-99
4495 ----------------------
4496
4497 1. The documentation is now supplied in plain text form and HTML as well as in
4498 the form of man page sources.
4499
4500 2. C++ compilers don't like assigning (void *) values to other pointer types.
4501 In particular this affects malloc(). Although there is no problem in Standard
4502 C, I've put in casts to keep C++ compilers happy.
4503
4504 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4505 should be (const char *).
4506
4507 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4508 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4509 However, I haven't made this a standard facility. The documentation doesn't
4510 mention it, and the Makefile doesn't support it.
4511
4512 5. The Makefile now contains an "install" target, with editable destinations at
4513 the top of the file. The pcretest program is not installed.
4514
4515 6. pgrep -V now gives the PCRE version number and date.
4516
4517 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4518 causing the entire string to be ignored, instead of just the last character.
4519
4520 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4521 non-matching string, it can take a very, very long time, even for strings of
4522 quite modest length, because of the nested recursion. PCRE now does better in
4523 some of these cases. It does this by remembering the last required literal
4524 character in the pattern, and pre-searching the subject to ensure it is present
4525 before running the real match. In other words, it applies a heuristic to detect
4526 some types of certain failure quickly, and in the above example, if presented
4527 with a string that has no trailing " it gives "no match" very quickly.
4528
4529 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4530 other alternatives are tried instead.
4531
4532
4533 Version 2.06 09-Jun-99
4534 ----------------------
4535
4536 1. Change pcretest's output for amount of store used to show just the code
4537 space, because the remainder (the data block) varies in size between 32-bit and
4538 64-bit systems.
4539
4540 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4541 start matching at. This allows lookbehinds to work when searching for multiple
4542 occurrences in a string.
4543
4544 3. Added additional options to pcretest for testing multiple occurrences:
4545
4546 /+ outputs the rest of the string that follows a match
4547 /g loops for multiple occurrences, using the new startoffset argument
4548 /G loops for multiple occurrences by passing an incremented pointer
4549
4550 4. PCRE wasn't doing the "first character" optimization for patterns starting
4551 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4552 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4553 the letter 'x'. On long subject strings, this gives a significant speed-up.
4554
4555
4556 Version 2.05 21-Apr-99
4557 ----------------------
4558
4559 1. Changed the type of magic_number from int to long int so that it works
4560 properly on 16-bit systems.
4561
4562 2. Fixed a bug which caused patterns starting with .* not to work correctly
4563 when the subject string contained newline characters. PCRE was assuming
4564 anchoring for such patterns in all cases, which is not correct because .* will
4565 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4566 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4567 must be retried after every newline in the subject.
4568
4569
4570 Version 2.04 18-Feb-99
4571 ----------------------
4572
4573 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4574 computation of the store needed to hold the pattern was incorrect (too large).
4575 If such patterns were nested a few deep, this could multiply and become a real
4576 problem.
4577
4578 2. Added /M option to pcretest to show the memory requirement of a specific
4579 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4580
4581 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4582 compiled in such a way that the backtracking after subsequent failure was
4583 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4584 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4585
4586
4587 Version 2.03 02-Feb-99
4588 ----------------------
4589
4590 1. Fixed typo and small mistake in man page.
4591
4592 2. Added 4th condition (GPL supersedes if conflict) and created separate
4593 LICENCE file containing the conditions.
4594
4595 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4596 Perl, that is the internal \ allows the delimiter to be included in the
4597 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4598 the final delimiter, add \ to the end of the pattern (to test the error).
4599
4600 4. Added the convenience functions for extracting substrings after a successful
4601 match. Updated pcretest to make it able to test these functions.
4602
4603
4604 Version 2.02 14-Jan-99
4605 ----------------------
4606
4607 1. Initialized the working variables associated with each extraction so that
4608 their saving and restoring doesn't refer to uninitialized store.
4609
4610 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4611 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4612 fix the problem.
4613
4614 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4615 calls, and wasn't printing the correct value for compiling calls. Increased the
4616 default value of LOOPREPEAT, and the number of significant figures in the
4617 times.
4618
4619 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4620
4621 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4622 a building problem on Windows NT with a FAT file system.
4623
4624
4625 Version 2.01 21-Oct-98
4626 ----------------------
4627
4628 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4629 to character tables built by pcre_maketables() in the current locale. If NULL
4630 is passed, the default tables are used.
4631
4632
4633 Version 2.00 24-Sep-98
4634 ----------------------
4635
4636 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4637 it any more.
4638
4639 2. Allow quantification of (?>) groups, and make it work correctly.
4640
4641 3. The first character computation wasn't working for (?>) groups.
4642
4643 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4644 end of the subject) and add 5.005's \z, which really does match only at the
4645 very end of the subject.
4646
4647 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4648
4649 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4650 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4651 localized options. All options to pcre_study() were also removed.
4652
4653 7. Add other new features from 5.005:
4654
4655 $(?<= positive lookbehind
4656 $(?<! negative lookbehind
4657 (?imsx-imsx) added the unsetting capability
4658 such a setting is global if at outer level; local otherwise
4659 (?imsx-imsx:) non-capturing groups with option setting
4660 (?(cond)re|re) conditional pattern matching
4661
4662 A backreference to itself in a repeated group matches the previous
4663 captured string.
4664
4665 8. General tidying up of studying (both automatic and via "study")
4666 consequential on the addition of new assertions.
4667
4668 9. As in 5.005, unlimited repeated groups that could match an empty substring
4669 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4670 runtime if any iteration does actually match an empty substring.
4671
4672 10. Include the RunTest script in the distribution.
4673
4674 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4675 discrepancies, some of which were old and were also with respect to 5.004. They
4676 have now been fixed.
4677
4678
4679 Version 1.09 28-Apr-98
4680 ----------------------
4681
4682 1. A negated single character class followed by a quantifier with a minimum
4683 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4684 program crashes, or just wrong answers. This did not apply to negated classes
4685 containing more than one character, or to minima other than one.
4686
4687
4688 Version 1.08 27-Mar-98
4689 ----------------------
4690
4691 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4692
4693 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4694 latter must appear before anything that relies on it in the pattern.
4695
4696
4697 Version 1.07 16-Feb-98
4698 ----------------------
4699
4700 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4701 repeat of a potentially empty string).
4702
4703
4704 Version 1.06 23-Jan-98
4705 ----------------------
4706
4707 1. Added Markus Oberhumer's little patches for C++.
4708
4709 2. Literal strings longer than 255 characters were broken.
4710
4711
4712 Version 1.05 23-Dec-97
4713 ----------------------
4714
4715 1. Negated character classes containing more than one character were failing if
4716 PCRE_CASELESS was set at run time.
4717
4718
4719 Version 1.04 19-Dec-97
4720 ----------------------
4721
4722 1. Corrected the man page, where some "const" qualifiers had been omitted.
4723
4724 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4725 input syntax.
4726
4727 3. Fixed memory leak which occurred when a regex with back references was
4728 matched with an offsets vector that wasn't big enough. The temporary memory
4729 that is used in this case wasn't being freed if the match failed.
4730
4731 4. Tidied pcretest to ensure it frees memory that it gets.
4732
4733 5. Temporary memory was being obtained in the case where the passed offsets
4734 vector was exactly big enough.
4735
4736 6. Corrected definition of offsetof() from change 5 below.
4737
4738 7. I had screwed up change 6 below and broken the rules for the use of
4739 setjmp(). Now fixed.
4740
4741
4742 Version 1.03 18-Dec-97
4743 ----------------------
4744
4745 1. A erroneous regex with a missing opening parenthesis was correctly
4746 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4747 on some systems.
4748
4749 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4750 it was reported that one broken compiler failed on the former because "code" is
4751 also an independent variable.
4752
4753 3. The erroneous regex a[]b caused an array overrun reference.
4754
4755 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4756 fail on data ending with that character. (It was going on too far, and checking
4757 the next character, typically a binary zero.) This was specific to the
4758 optimized code for single-character negative classes.
4759
4760 5. Added a contributed patch from the TIN world which does the following:
4761
4762 + Add an undef for memmove, in case the the system defines a macro for it.
4763
4764 + Add a definition of offsetof(), in case there isn't one. (I don't know
4765 the reason behind this - offsetof() is part of the ANSI standard - but
4766 it does no harm).
4767
4768 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4769 most of the places where whitespace preceded '#'. I have given up and
4770 allowed the remaining 2 cases to be at the margin.
4771
4772 + Rename some variables in pcre to eliminate shadowing. This seems very
4773 pedantic, but does no harm, of course.
4774
4775 6. Moved the call to setjmp() into its own function, to get rid of warnings
4776 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4777
4778 7. Constructs such as \d{8,} were compiling into the equivalent of
4779 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4780 outcome, but in this particular case used more store than had been allocated,
4781 which caused the bug to be discovered because it threw up an internal error.
4782
4783 8. The debugging code in both pcre and pcretest for outputting the compiled
4784 form of a regex was going wrong in the case of back references followed by
4785 curly-bracketed repeats.
4786
4787
4788 Version 1.02 12-Dec-97
4789 ----------------------
4790
4791 1. Typos in pcre.3 and comments in the source fixed.
4792
4793 2. Applied a contributed patch to get rid of places where it used to remove
4794 'const' from variables, and fixed some signed/unsigned and uninitialized
4795 variable warnings.
4796
4797 3. Added the "runtest" target to Makefile.
4798
4799 4. Set default compiler flag to -O2 rather than just -O.
4800
4801
4802 Version 1.01 19-Nov-97
4803 ----------------------
4804
4805 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4806 like /([ab]*)*/, that is, for classes with more than one character in them.
4807
4808 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4809 as /((?>a*))*/ (a PCRE_EXTRA facility).
4810
4811
4812 Version 1.00 18-Nov-97
4813 ----------------------
4814
4815 1. Added compile-time macros to support systems such as SunOS4 which don't have
4816 memmove() or strerror() but have other things that can be used instead.
4817
4818 2. Arranged that "make clean" removes the executables.
4819
4820
4821 Version 0.99 27-Oct-97
4822 ----------------------
4823
4824 1. Fixed bug in code for optimizing classes with only one character. It was
4825 initializing a 32-byte map regardless, which could cause it to run off the end
4826 of the memory it had got.
4827
4828 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4829
4830
4831 Version 0.98 22-Oct-97
4832 ----------------------
4833
4834 1. Fixed bug in code for handling temporary memory usage when there are more
4835 back references than supplied space in the ovector. This could cause segfaults.
4836
4837
4838 Version 0.97 21-Oct-97
4839 ----------------------
4840
4841 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4842
4843 2. Optimized negated single characters not to use a bit map.
4844
4845 3. Brought error texts together as macro definitions; clarified some of them;
4846 fixed one that was wrong - it said "range out of order" when it meant "invalid
4847 escape sequence".
4848
4849 4. Changed some char * arguments to const char *.
4850
4851 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4852
4853 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4854 pcretest.
4855
4856
4857 Version 0.96 16-Oct-97
4858 ----------------------
4859
4860 1. Added a simple "pgrep" utility to the distribution.
4861
4862 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4863 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4864 where "ddd" means "one or more decimal digits".
4865
4866 3. Fixed serious bug. If a pattern had a back reference, but the call to
4867 pcre_exec() didn't supply a large enough ovector to record the related
4868 identifying subpattern, the match always failed. PCRE now remembers the number
4869 of the largest back reference, and gets some temporary memory in which to save
4870 the offsets during matching if necessary, in order to ensure that
4871 backreferences always work.
4872
4873 4. Increased the compatibility with Perl in a number of ways:
4874
4875 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4876 to request this handling. The option can be set at compile or exec time.
4877
4878 (b) $ matches before a terminating newline by default; an option
4879 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4880 mode). The option can be set at compile or exec time.
4881
4882 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4883 the same as Perl's. If the decimal number it represents is less than 10
4884 or there aren't that many previous left capturing parentheses, an octal
4885 escape is read. Inside a character class, it's always an octal escape,
4886 even if it is a single digit.
4887
4888 (d) An escaped but undefined alphabetic character is taken as a literal,
4889 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4890 escapes.
4891
4892 (e) {0} is now permitted. (The previous item is removed from the compiled
4893 pattern).
4894
4895 5. Changed all the names of code files so that the basic parts are no longer
4896 than 10 characters, and abolished the teeny "globals.c" file.
4897
4898 6. Changed the handling of character classes; they are now done with a 32-byte
4899 bit map always.
4900
4901 7. Added the -d and /D options to pcretest to make it possible to look at the
4902 internals of compilation without having to recompile pcre.
4903
4904
4905 Version 0.95 23-Sep-97
4906 ----------------------
4907
4908 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4909 \x20 at the start of a run of normal characters. These were being treated as
4910 real characters, instead of the source characters being re-checked.
4911
4912
4913 Version 0.94 18-Sep-97
4914 ----------------------
4915
4916 1. The functions are now thread-safe, with the caveat that the global variables
4917 containing pointers to malloc() and free() or alternative functions are the
4918 same for all threads.
4919
4920 2. Get pcre_study() to generate a bitmap of initial characters for non-
4921 anchored patterns when this is possible, and use it if passed to pcre_exec().
4922
4923
4924 Version 0.93 15-Sep-97
4925 ----------------------
4926
4927 1. /(b)|(:+)/ was computing an incorrect first character.
4928
4929 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4930 but not actually doing anything yet.
4931
4932 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4933 as Perl does (e.g. [-az] or [az-]).
4934
4935 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4936 all possible positions.
4937
4938 5. Split up into different modules to avoid including unneeded functions in a
4939 compiled binary. However, compile and exec are still in one module. The "study"
4940 function is split off.
4941
4942 6. The character tables are now in a separate module whose source is generated
4943 by an auxiliary program - but can then be edited by hand if required. There are
4944 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4945 toupper() in the code.
4946
4947 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4948 make them global. Abolish the function for setting them, as the caller can now
4949 set them directly.
4950
4951
4952 Version 0.92 11-Sep-97
4953 ----------------------
4954
4955 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4956 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4957
4958 2. Caseless matching was not working in character classes if the characters in
4959 the pattern were in upper case.
4960
4961 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4962
4963 4. Make PCRE_ANCHORED public and accept as a compile option.
4964
4965 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4966 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4967 pass them.
4968
4969 6. Give an error if bad option bits passed at compile or run time.
4970
4971 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4972 pcretest to cause it to pass that flag.
4973
4974 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4975 options, and the first character, if set.
4976
4977 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4978
4979
4980 Version 0.91 10-Sep-97
4981 ----------------------
4982
4983 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4984 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4985
4986 2. PCRE was looping on encountering an indefinitely repeated back reference to
4987 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4988 Perl does - treats the match as successful.
4989
4990 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5