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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1219 - (show annotations)
Sun Nov 11 18:04:37 2012 UTC (7 years ago) by ph10
File size: 239962 byte(s)
Error occurred while calculating annotation data.
Support (*UTF) in all libraries.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32 11-November-2012
5 -----------------------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51 9. Improve the matching speed of capturing brackets.
52
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
55
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
58
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60
61 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62 structure could go wrong in environments where size_t is not the same size
63 as int.
64
65 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66 set.
67
68 15. The EBCDIC support had decayed; later updates to the code had included
69 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70 general tidy up of EBCDIC-related issues, and the documentation was also
71 not quite right. There is now a test that can be run on ASCII systems to
72 check some of the EBCDIC-related things (but is it not a full test).
73
74 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75 in a small tidy to the code.
76
77 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
78
79 18. If the --only-matching (-o) option in pcregrep is specified multiple
80 times, each one causes appropriate output. For example, -o1 -o2 outputs the
81 substrings matched by the 1st and 2nd capturing parentheses. A separating
82 string can be specified by --om-separator (default empty).
83
84 19. Improving the first n character searches.
85
86 20. Turn case lists for horizontal and vertical white space into macros so that
87 they are defined only once.
88
89 21. This set of changes together give more compatible Unicode case-folding
90 behaviour for characters that have more than one other case when UCP
91 support is available.
92
93 (a) The Unicode property table now has offsets into a new table of sets of
94 three or more characters that are case-equivalent. The MultiStage2.py
95 script that generates these tables (the pcre_ucd.c file) now scans
96 CaseFolding.txt instead of UnicodeData.txt for character case
97 information.
98
99 (b) The code for adding characters or ranges of characters to a character
100 class has been abstracted into a generalized function that also handles
101 case-independence. In UTF-mode with UCP support, this uses the new data
102 to handle characters with more than one other case.
103
104 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105 whose other case is greater than 256 are now correctly matched
106 caselessly. Previously, the high codepoint matched the low one, but not
107 vice versa.
108
109 (d) The processing of \h, \H, \v, and \ in character classes now makes use
110 of the new class addition function, using character lists defined as
111 macros alongside the case definitions of 20 above.
112
113 (e) Caseless back references now work with characters that have more than
114 one other case.
115
116 (f) General caseless matching of characters with more than one other case
117 is supported.
118
119 22. Unicode character properties were updated from Unicode 6.2.0
120
121 23. Improved CMake support under Windows. Patch by Daniel Richard G.
122
123 24. Add support for 32-bit character strings, and UTF-32
124
125 25. Major JIT compiler update (code refactoring and bugfixing).
126 Experimental Sparc 32 support is added.
127
128 26. Applied a modified version of Daniel Richard G's patch to create
129 pcre.h.generic and config.h.generic by "make" instead of in the
130 PrepareRelease script.
131
132 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
133 pcre_compile.c when checking for a zero character.
134
135 28. Introducing a native interface for JIT. Through this interface, the compiled
136 machine code can be directly executed. The purpose of this interface is to
137 provide fast pattern matching, so several sanity checks are not performed.
138 However, feature tests are still performed. The new interface provides
139 1.4x speedup compared to the old one.
140
141 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
142 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
143 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
144
145 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
146 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
147 the "old" RFC 2279). Instead, it ended up passing a negative length to
148 pcre_exec().
149
150 31. Add support for GCC's visibility feature to hide internal functions.
151
152 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
153 "unknown -C option" after outputting 0 or 1.
154
155 33. There is now support for generating a code coverage report for the test
156 suite in environments where gcc is the compiler and lcov is installed. This
157 is mainly for the benefit of the developers.
158
159 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
160 unaddressable using valgrind annotations, allowing valgrind to detect
161 invalid memory accesses. This is mainly for the benefit of the developers.
162
163 25. (*UTF) can now be used to start a pattern in any of the three libraries.
164
165
166 Version 8.31 06-July-2012
167 -------------------------
168
169 1. Fixing a wrong JIT test case and some compiler warnings.
170
171 2. Removed a bashism from the RunTest script.
172
173 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
174 to unsigned type, result still unsigned" that was given by an MS compiler
175 on encountering the code "-sizeof(xxx)".
176
177 4. Partial matching support is added to the JIT compiler.
178
179 5. Fixed several bugs concerned with partial matching of items that consist
180 of more than one character:
181
182 (a) /^(..)\1/ did not partially match "aba" because checking references was
183 done on an "all or nothing" basis. This also applied to repeated
184 references.
185
186 (b) \R did not give a hard partial match if \r was found at the end of the
187 subject.
188
189 (c) \X did not give a hard partial match after matching one or more
190 characters at the end of the subject.
191
192 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
193 a partial match for the string "\r".
194
195 (e) When newline was set to CRLF, the metacharacter "." did not recognize
196 a partial match for a CR character at the end of the subject string.
197
198 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
199 running pcretest, the text "(JIT)" added to the output whenever JIT is
200 actually used to run the match.
201
202 7. Individual JIT compile options can be set in pcretest by following -s+[+]
203 or /S+[+] with a digit between 1 and 7.
204
205 8. OP_NOT now supports any UTF character not just single-byte ones.
206
207 9. (*MARK) control verb is now supported by the JIT compiler.
208
209 10. The command "./RunTest list" lists the available tests without actually
210 running any of them. (Because I keep forgetting what they all are.)
211
212 11. Add PCRE_INFO_MAXLOOKBEHIND.
213
214 12. Applied a (slightly modified) user-supplied patch that improves performance
215 when the heap is used for recursion (compiled with --disable-stack-for-
216 recursion). Instead of malloc and free for each heap frame each time a
217 logical recursion happens, frames are retained on a chain and re-used where
218 possible. This sometimes gives as much as 30% improvement.
219
220 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
221 call.
222
223 14. As documented, (*COMMIT) is now confined to within a positive assertion.
224
225 15. It is now possible to link pcretest with libedit as an alternative to
226 libreadline.
227
228 16. (*COMMIT) control verb is now supported by the JIT compiler.
229
230 17. The Unicode data tables have been updated to Unicode 6.1.0.
231
232 18. Added --file-list option to pcregrep.
233
234 19. Added binary file support to pcregrep, including the -a, --binary-files,
235 -I, and --text options.
236
237 20. The madvise function is renamed for posix_madvise for QNX compatibility
238 reasons. Fixed by Giuseppe D'Angelo.
239
240 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
241
242 22. Changed the option for creating symbolic links for 16-bit man pages from
243 -s to -sf so that re-installing does not cause issues.
244
245 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
246
247 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
248 matches in certain environments (the workspace was not being correctly
249 retained). Also added to pcre_dfa_exec() a simple plausibility check on
250 some of the workspace data at the beginning of a restart.
251
252 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
253 was not doing so when it should - probably a typo introduced by SVN 528
254 (change 8.10/14).
255
256 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
257 \w+ when the character tables indicated that \x{c4} was a word character.
258 There were several related cases, all because the tests for doing a table
259 lookup were testing for characters less than 127 instead of 255.
260
261 27. If a pattern contains capturing parentheses that are not used in a match,
262 their slots in the ovector are set to -1. For those that are higher than
263 any matched groups, this happens at the end of processing. In the case when
264 there were back references that the ovector was too small to contain
265 (causing temporary malloc'd memory to be used during matching), and the
266 highest capturing number was not used, memory off the end of the ovector
267 was incorrectly being set to -1. (It was using the size of the temporary
268 memory instead of the true size.)
269
270 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
271 ovector size, it uses memory at the end of the block that it has got.
272
273 29. Check for an overlong MARK name and give an error at compile time. The
274 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
275
276 30. JIT compiler update.
277
278 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
279 Rill for the patch.
280
281 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
282
283 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
284
285 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
286 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
287 was enabled.
288
289 35. Improve JIT code generation for greedy plus quantifier.
290
291 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
292 1 to "aa" instead of to an empty string. The bug affected repeated groups
293 that could potentially match an empty string.
294
295 37. Optimizing single character iterators in JIT.
296
297 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
298 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
299 codepoints that are too big for the mode are faulted, and in a UTF mode,
300 disallowed codepoints are also faulted.
301
302 39. If PCRE was compiled with UTF support, in three places in the DFA
303 matcher there was code that should only have been obeyed in UTF mode, but
304 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
305 processing when bytes with values greater than 127 were present. In 16-bit
306 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
307 both cases the values are those that cannot be the first data item in a UTF
308 character. The three items that might have provoked this were recursions,
309 possessively repeated groups, and atomic groups.
310
311 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
312 and pcregrep, because some OS require shared objects to be explicitly
313 passed to ld, causing the link step to fail if they are not.
314
315 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
316 mode, patterns that started with \h* or \R* might be incorrectly matched.
317
318
319 Version 8.30 04-February-2012
320 -----------------------------
321
322 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
323 name is defined in ctype.h.
324
325 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
326 only in quite long subpatterns.
327
328 3. Removed the function pcre_info(), which has been obsolete and deprecated
329 since it was replaced by pcre_fullinfo() in February 2000.
330
331 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
332 match a (*MARK), and the match failed at the start of the subject, a
333 reference to memory before the start of the subject could occur. This bug
334 was introduced by fix 17 of release 8.21.
335
336 5. A reference to an unset group with zero minimum repetition was giving
337 totally wrong answers (in non-JavaScript-compatibility mode). For example,
338 /(another)?(\1?)test/ matched against "hello world test". This bug was
339 introduced in release 8.13.
340
341 6. Add support for 16-bit character strings (a large amount of work involving
342 many changes and refactorings).
343
344 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
345 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
346 from a file.
347
348 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
349 rounding is not applied in this particular case).
350
351 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
352 if they appear, or are escaped, in patterns.
353
354 10. Get rid of a number of -Wunused-but-set-variable warnings.
355
356 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
357 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
358 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
359 also returns the mark "x". This bug applied to capturing parentheses,
360 non-capturing parentheses, and atomic parentheses. It also applied to some
361 assertions.
362
363 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
364 information out of configure.ac instead of relying on pcre.h.generic, which
365 is not stored in the repository.
366
367 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
368 -lreadline.
369
370 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
371
372 15. Applied Graycode's patch to put the top-level frame on the stack rather
373 than the heap when not using the stack for recursion. This gives a
374 performance improvement in many cases when recursion is not deep.
375
376 16. Experimental code added to "pcretest -C" to output the stack frame size.
377
378
379 Version 8.21 12-Dec-2011
380 ------------------------
381
382 1. Updating the JIT compiler.
383
384 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
385 are added as well.
386
387 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
388 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
389 calling _pcre_jit_exec. Some extra comments are added.
390
391 4. (*MARK) settings inside atomic groups that do not contain any capturing
392 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
393 was introduced by change 18 for 8.20.
394
395 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
396 ECMA-262 standard.
397
398 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
399 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
400 This bug was probably introduced by change 9 of 8.13.
401
402 7. While fixing 6 above, I noticed that a number of other items were being
403 incorrectly rejected as "not fixed length". This arose partly because newer
404 opcodes had not been added to the fixed-length checking code. I have (a)
405 corrected the bug and added tests for these items, and (b) arranged for an
406 error to occur if an unknown opcode is encountered while checking for fixed
407 length instead of just assuming "not fixed length". The items that were
408 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
409 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
410 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
411
412 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
413 being incorrectly compiled and would have given unpredicatble results.
414
415 9. A possessively repeated subpattern with minimum repeat count greater than
416 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
417 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
418 the first (A) could occur when it should not.
419
420 10. Add a cast and remove a redundant test from the code.
421
422 11. JIT should use pcre_malloc/pcre_free for allocation.
423
424 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
425 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
426 is anything other than /usr, -L is still shown).
427
428 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
429
430 14. Perl does not support \N without a following name in a [] class; PCRE now
431 also gives an error.
432
433 15. If a forward reference was repeated with an upper limit of around 2000,
434 it caused the error "internal error: overran compiling workspace". The
435 maximum number of forward references (including repeats) was limited by the
436 internal workspace, and dependent on the LINK_SIZE. The code has been
437 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
438 the default depends on LINK_SIZE. There is a new upper limit (for safety)
439 of around 200,000 forward references. While doing this, I also speeded up
440 the filling in of repeated forward references.
441
442 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
443 incorrectly expecting the subject to contain another "a" after the start.
444
445 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
446 in the match, the SKIP should be ignored. This was not happening; instead
447 the SKIP was being treated as NOMATCH. For patterns such as
448 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
449 tested.
450
451 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
452 now much more compatible with Perl, in particular in cases where the result
453 is a non-match for a non-anchored pattern. For example, if
454 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
455 "m", where previously it did not return a name. A side effect of this
456 change is that for partial matches, the last encountered mark name is
457 returned, as for non matches. A number of tests that were previously not
458 Perl-compatible have been moved into the Perl-compatible test files. The
459 refactoring has had the pleasing side effect of removing one argument from
460 the match() function, thus reducing its stack requirements.
461
462 19. If the /S+ option was used in pcretest to study a pattern using JIT,
463 subsequent uses of /S (without +) incorrectly behaved like /S+.
464
465 21. Retrieve executable code size support for the JIT compiler and fixing
466 some warnings.
467
468 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
469 not work when the shorter character appeared right at the end of the
470 subject string.
471
472 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
473 systems.
474
475 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
476 output it when the /M option is used in pcretest.
477
478 25. The CheckMan script was not being included in the distribution. Also, added
479 an explicit "perl" to run Perl scripts from the PrepareRelease script
480 because this is reportedly needed in Windows.
481
482 26. If study data was being save in a file and studying had not found a set of
483 "starts with" bytes for the pattern, the data written to the file (though
484 never used) was taken from uninitialized memory and so caused valgrind to
485 complain.
486
487 27. Updated RunTest.bat as provided by Sheri Pierce.
488
489 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
490
491 29. Computation of memory usage for the table of capturing group names was
492 giving an unnecessarily large value.
493
494
495 Version 8.20 21-Oct-2011
496 ------------------------
497
498 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
499 a POSIX class. After further experiments with Perl, which convinced me that
500 Perl has bugs and confusions, a closing square bracket is no longer allowed
501 in a POSIX name. This bug also affected patterns with classes that started
502 with full stops.
503
504 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
505 captured substring, but while checking the failing first alternative,
506 substring 1 is temporarily captured. If the output vector supplied to
507 pcre_exec() was not big enough for this capture, the yield of the function
508 was still zero ("insufficient space for captured substrings"). This cannot
509 be totally fixed without adding another stack variable, which seems a lot
510 of expense for a edge case. However, I have improved the situation in cases
511 such as /(a)(b)x|abc/ matched against "abc", where the return code
512 indicates that fewer than the maximum number of slots in the ovector have
513 been set.
514
515 3. Related to (2) above: when there are more back references in a pattern than
516 slots in the output vector, pcre_exec() uses temporary memory during
517 matching, and copies in the captures as far as possible afterwards. It was
518 using the entire output vector, but this conflicts with the specification
519 that only 2/3 is used for passing back captured substrings. Now it uses
520 only the first 2/3, for compatibility. This is, of course, another edge
521 case.
522
523 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
524 main code base, and can be used by building with --enable-jit. When this is
525 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
526 runtime --no-jit option is given.
527
528 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
529 ovector, the return from the function was zero, implying that there were
530 other matches that did not fit. The correct "exactly full" value is now
531 returned.
532
533 6. If a subpattern that was called recursively or as a subroutine contained
534 (*PRUNE) or any other control that caused it to give a non-standard return,
535 invalid errors such as "Error -26 (nested recursion at the same subject
536 position)" or even infinite loops could occur.
537
538 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
539 computing the minimum length on reaching *ACCEPT, and so ended up with the
540 wrong value of 1 rather than 0. Further investigation indicates that
541 computing a minimum subject length in the presence of *ACCEPT is difficult
542 (think back references, subroutine calls), and so I have changed the code
543 so that no minimum is registered for a pattern that contains *ACCEPT.
544
545 8. If (*THEN) was present in the first (true) branch of a conditional group,
546 it was not handled as intended. [But see 16 below.]
547
548 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
549 Sheri Pierce.
550
551 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
552 the first byte in a match must be "a".
553
554 11. Change 17 for 8.13 increased the recursion depth for patterns like
555 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
556 pattern contains any instances of (*THEN). If it does not, the old
557 optimizations are restored. It would be nice to do this on a per-group
558 basis, but at the moment that is not feasible.
559
560 12. In some environments, the output of pcretest -C is CRLF terminated. This
561 broke RunTest's code that checks for the link size. A single white space
562 character after the value is now allowed for.
563
564 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
565 For "fr", it uses the Windows-specific input and output files.
566
567 14. If (*THEN) appeared in a group that was called recursively or as a
568 subroutine, it did not work as intended. [But see next item.]
569
570 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
571 pattern fragments (but not containing any | characters). If A and B are
572 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
573 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
574 D. In other words, Perl considers parentheses that do not contain any |
575 characters to be part of a surrounding alternative, whereas PCRE was
576 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
577 differently. PCRE now behaves in the same way as Perl, except in the case
578 of subroutine/recursion calls such as (?1) which have in any case always
579 been different (but PCRE had them first :-).
580
581 16. Related to 15 above: Perl does not treat the | in a conditional group as
582 creating alternatives. Such a group is treated in the same way as an
583 ordinary group without any | characters when processing (*THEN). PCRE has
584 been changed to match Perl's behaviour.
585
586 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
587 RunGrepTest script failed.
588
589 18. Change 22 for version 13 caused atomic groups to use more stack. This is
590 inevitable for groups that contain captures, but it can lead to a lot of
591 stack use in large patterns. The old behaviour has been restored for atomic
592 groups that do not contain any capturing parentheses.
593
594 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
595 suppress the check for a minimum subject length at run time. (If it was
596 given to pcre_exec() or pcre_dfa_exec() it did work.)
597
598 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
599 fail to work when decoding hex characters in data strings in EBCDIC
600 environments.
601
602 21. It appears that in at least one Mac OS environment, the isxdigit() function
603 is implemented as a macro that evaluates to its argument more than once,
604 contravening the C 90 Standard (I haven't checked a later standard). There
605 was an instance in pcretest which caused it to go wrong when processing
606 \x{...} escapes in subject strings. The has been rewritten to avoid using
607 things like p++ in the argument of isxdigit().
608
609
610 Version 8.13 16-Aug-2011
611 ------------------------
612
613 1. The Unicode data tables have been updated to Unicode 6.0.0.
614
615 2. Two minor typos in pcre_internal.h have been fixed.
616
617 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
618 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
619 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
620
621 4. There were a number of related bugs in the code for matching backrefences
622 caselessly in UTF-8 mode when codes for the characters concerned were
623 different numbers of bytes. For example, U+023A and U+2C65 are an upper
624 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
625 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
626 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
627 2-byte code at the end of the subject (it thought there wasn't enough data
628 left).
629
630 5. Comprehensive information about what went wrong is now returned by
631 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
632 as the output vector has at least 2 elements. The offset of the start of
633 the failing character and a reason code are placed in the vector.
634
635 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
636 now returned is for the first byte of the failing character, instead of the
637 last byte inspected. This is an incompatible change, but I hope it is small
638 enough not to be a problem. It makes the returned offset consistent with
639 pcre_exec() and pcre_dfa_exec().
640
641 7. pcretest now gives a text phrase as well as the error number when
642 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
643 failure, the offset and reason code are output.
644
645 8. When \R was used with a maximizing quantifier it failed to skip backwards
646 over a \r\n pair if the subsequent match failed. Instead, it just skipped
647 back over a single character (\n). This seems wrong (because it treated the
648 two characters as a single entity when going forwards), conflicts with the
649 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
650 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
651 has been changed.
652
653 9. Some internal refactoring has changed the processing so that the handling
654 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
655 time (the PCRE_DOTALL option was changed this way some time ago: version
656 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
657 which was always a bit of a fudge. It also means that there is one less
658 argument for the match() function, which reduces its stack requirements
659 slightly. This change also fixes an incompatibility with Perl: the pattern
660 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
661
662 10. More internal refactoring has drastically reduced the number of recursive
663 calls to match() for possessively repeated groups such as (abc)++ when
664 using pcre_exec().
665
666 11. While implementing 10, a number of bugs in the handling of groups were
667 discovered and fixed:
668
669 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
670 (a|)*(?1) gave a compile-time internal error.
671 ((a|)+)+ did not notice that the outer group could match an empty string.
672 (^a|^)+ was not marked as anchored.
673 (.*a|.*)+ was not marked as matching at start or after a newline.
674
675 12. Yet more internal refactoring has removed another argument from the match()
676 function. Special calls to this function are now indicated by setting a
677 value in a variable in the "match data" data block.
678
679 13. Be more explicit in pcre_study() instead of relying on "default" for
680 opcodes that mean there is no starting character; this means that when new
681 ones are added and accidentally left out of pcre_study(), testing should
682 pick them up.
683
684 14. The -s option of pcretest has been documented for ages as being an old
685 synonym of -m (show memory usage). I have changed it to mean "force study
686 for every regex", that is, assume /S for every regex. This is similar to -i
687 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
688 using it. It makes it easier to run collections of tests with and without
689 study enabled, and thereby test pcre_study() more easily. All the standard
690 tests are now run with and without -s (but some patterns can be marked as
691 "never study" - see 20 below).
692
693 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
694 restoration of the capturing data to the outer values was not happening
695 correctly.
696
697 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
698 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
699 pattern had matched an empty string, and so incorrectly returned a no
700 match.
701
702 17. There was optimizing code for the last branch of non-capturing parentheses,
703 and also for the obeyed branch of a conditional subexpression, which used
704 tail recursion to cut down on stack usage. Unfortunately, now that there is
705 the possibility of (*THEN) occurring in these branches, tail recursion is
706 no longer possible because the return has to be checked for (*THEN). These
707 two optimizations have therefore been removed. [But see 8.20/11 above.]
708
709 18. If a pattern containing \R was studied, it was assumed that \R always
710 matched two bytes, thus causing the minimum subject length to be
711 incorrectly computed because \R can also match just one byte.
712
713 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
714 was incorrectly computed.
715
716 20. If /S is present twice on a test pattern in pcretest input, it now
717 *disables* studying, thereby overriding the use of -s on the command line
718 (see 14 above). This is necessary for one or two tests to keep the output
719 identical in both cases.
720
721 21. When (*ACCEPT) was used in an assertion that matched an empty string and
722 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
723
724 22. When an atomic group that contained a capturing parenthesis was
725 successfully matched, but the branch in which it appeared failed, the
726 capturing was not being forgotten if a higher numbered group was later
727 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
728 group 1 to "a", when in fact it should be unset. This applied to multi-
729 branched capturing and non-capturing groups, repeated or not, and also to
730 positive assertions (capturing in negative assertions does not happen
731 in PCRE) and also to nested atomic groups.
732
733 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
734 subject after a captured substring, to make it easier to tell which of a
735 number of identical substrings has been captured.
736
737 24. The way atomic groups are processed by pcre_exec() has been changed so that
738 if they are repeated, backtracking one repetition now resets captured
739 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
740 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
741 "aaa". Previously, it would have been "a". As part of this code
742 refactoring, the way recursive calls are handled has also been changed.
743
744 25. If an assertion condition captured any substrings, they were not passed
745 back unless some other capturing happened later. For example, if
746 (?(?=(a))a) was matched against "a", no capturing was returned.
747
748 26. When studying a pattern that contained subroutine calls or assertions,
749 the code for finding the minimum length of a possible match was handling
750 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
751 group 1 called group 2 while simultaneously a separate group 2 called group
752 1). A stack overflow occurred in this case. I have fixed this by limiting
753 the recursion depth to 10.
754
755 27. Updated RunTest.bat in the distribution to the version supplied by Tom
756 Fortmann. This supports explicit test numbers on the command line, and has
757 argument validation and error reporting.
758
759 28. An instance of \X with an unlimited repeat could fail if at any point the
760 first character it looked at was a mark character.
761
762 29. Some minor code refactoring concerning Unicode properties and scripts
763 should reduce the stack requirement of match() slightly.
764
765 30. Added the '=' option to pcretest to check the setting of unused capturing
766 slots at the end of the pattern, which are documented as being -1, but are
767 not included in the return count.
768
769 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
770 compiled something random. Now it gives a compile-time error (as does
771 Perl).
772
773 32. A *MARK encountered during the processing of a positive assertion is now
774 recorded and passed back (compatible with Perl).
775
776 33. If --only-matching or --colour was set on a pcregrep call whose pattern
777 had alternative anchored branches, the search for a second match in a line
778 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
779 matched the line "0102" twice. The same bug affected patterns that started
780 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
781 twice.
782
783 34. Previously, PCRE did not allow quantification of assertions. However, Perl
784 does, and because of capturing effects, quantifying parenthesized
785 assertions may at times be useful. Quantifiers are now allowed for
786 parenthesized assertions.
787
788 35. A minor code tidy in pcre_compile() when checking options for \R usage.
789
790 36. \g was being checked for fancy things in a character class, when it should
791 just be a literal "g".
792
793 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
794 appearance of a nested POSIX class supersedes an apparent external class.
795 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
796 unescaped square brackets may also appear as part of class names. For
797 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
798 more like Perl. (But see 8.20/1 above.)
799
800 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
801 was because it thought it was \N{name}, which is not supported).
802
803 39. Add minix to OS list not supporting the -S option in pcretest.
804
805 40. PCRE tries to detect cases of infinite recursion at compile time, but it
806 cannot analyze patterns in sufficient detail to catch mutual recursions
807 such as ((?1))((?2)). There is now a runtime test that gives an error if a
808 subgroup is called recursively as a subpattern for a second time at the
809 same position in the subject string. In previous releases this might have
810 been caught by the recursion limit, or it might have run out of stack.
811
812 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
813 happen only once. PCRE was, however incorrectly giving a compile time error
814 "recursive call could loop indefinitely" because it cannot analyze the
815 pattern in sufficient detail. The compile time test no longer happens when
816 PCRE is compiling a conditional subpattern, but actual runaway loops are
817 now caught at runtime (see 40 above).
818
819 42. It seems that Perl allows any characters other than a closing parenthesis
820 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
821 has been changed to be the same.
822
823 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
824 as not to get warnings when autogen.sh is called. Also changed
825 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
826
827 44. To help people who use pcregrep to scan files containing exceedingly long
828 lines, the following changes have been made:
829
830 (a) The default value of the buffer size parameter has been increased from
831 8K to 20K. (The actual buffer used is three times this size.)
832
833 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
834 PCRE is built.
835
836 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
837 to be set at run time.
838
839 (d) Numerical values in pcregrep options can be followed by K or M, for
840 example --buffer-size=50K.
841
842 (e) If a line being scanned overflows pcregrep's buffer, an error is now
843 given and the return code is set to 2.
844
845 45. Add a pointer to the latest mark to the callout data block.
846
847 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
848 partial match of an empty string instead of no match. This was specific to
849 the use of ".".
850
851 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
852 complete match instead of a partial match. This bug was dependent on both
853 the PCRE_UTF8 and PCRE_DOTALL options being set.
854
855 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
856 starting byte set, because \b was not being ignored.
857
858
859 Version 8.12 15-Jan-2011
860 ------------------------
861
862 1. Fixed some typos in the markup of the man pages, and wrote a script that
863 checks for such things as part of the documentation building process.
864
865 2. On a big-endian 64-bit system, pcregrep did not correctly process the
866 --match-limit and --recursion-limit options (added for 8.11). In
867 particular, this made one of the standard tests fail. (The integer value
868 went into the wrong half of a long int.)
869
870 3. If the --colour option was given to pcregrep with -v (invert match), it
871 did strange things, either producing crazy output, or crashing. It should,
872 of course, ignore a request for colour when reporting lines that do not
873 match.
874
875 4. Another pcregrep bug caused similar problems if --colour was specified with
876 -M (multiline) and the pattern match finished with a line ending.
877
878 5. In pcregrep, when a pattern that ended with a literal newline sequence was
879 matched in multiline mode, the following line was shown as part of the
880 match. This seems wrong, so I have changed it.
881
882 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
883 the check for further matches in the same line (so they could be coloured)
884 to overrun the end of the current line. If another match was found, it was
885 incorrectly shown (and then shown again when found in the next line).
886
887 7. If pcregrep was compiled under Windows, there was a reference to the
888 function pcregrep_exit() before it was defined. I am assuming this was
889 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
890 reported by a user. I've moved the definition above the reference.
891
892
893 Version 8.11 10-Dec-2010
894 ------------------------
895
896 1. (*THEN) was not working properly if there were untried alternatives prior
897 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
898 backtracked to try for "b" instead of moving to the next alternative branch
899 at the same level (in this case, to look for "c"). The Perl documentation
900 is clear that when (*THEN) is backtracked onto, it goes to the "next
901 alternative in the innermost enclosing group".
902
903 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
904 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
905 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
906 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
907 (*THEN).
908
909 3. If \s appeared in a character class, it removed the VT character from
910 the class, even if it had been included by some previous item, for example
911 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
912 of \s, but is part of the POSIX "space" class.)
913
914 4. A partial match never returns an empty string (because you can always
915 match an empty string at the end of the subject); however the checking for
916 an empty string was starting at the "start of match" point. This has been
917 changed to the "earliest inspected character" point, because the returned
918 data for a partial match starts at this character. This means that, for
919 example, /(?<=abc)def/ gives a partial match for the subject "abc"
920 (previously it gave "no match").
921
922 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
923 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
924 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
925 has an implication that the given string is incomplete (because a partial
926 match is preferred over a full match). For this reason, these items now
927 give a partial match in this situation. [Aside: previously, the one case
928 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
929 match rather than a full match, which was wrong by the old rules, but is
930 now correct.]
931
932 6. There was a bug in the handling of #-introduced comments, recognized when
933 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
934 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
935 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
936 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
937 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
938 places in pcre_compile().
939
940 7. Related to (6) above, when pcre_compile() was skipping #-introduced
941 comments when looking ahead for named forward references to subpatterns,
942 the only newline sequence it recognized was NL. It now handles newlines
943 according to the set newline convention.
944
945 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
946 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
947 cater for a lack of strerror(). These oversights have been fixed.
948
949 9. Added --match-limit and --recursion-limit to pcregrep.
950
951 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
952
953 11. When the -o option was used, pcregrep was setting a return code of 1, even
954 when matches were found, and --line-buffered was not being honoured.
955
956 12. Added an optional parentheses number to the -o and --only-matching options
957 of pcregrep.
958
959 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
960 can match an empty string. The code to do it in pcretest and pcredemo
961 needed fixing:
962
963 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
964 only one byte after an empty string match just before CRLF (this case
965 just got forgotten; "any" and "anycrlf" were OK).
966
967 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
968 mode when an empty string match preceded an ASCII character followed by
969 a non-ASCII character. (The code for advancing by one character rather
970 than one byte was nonsense.)
971
972 (c) The pcredemo.c sample program did not have any code at all to handle
973 the cases when CRLF is a valid newline sequence.
974
975 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
976 as a starting offset was within the subject string. There is now a new
977 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
978 negative or greater than the length of the string. In order to test this,
979 pcretest is extended to allow the setting of negative starting offsets.
980
981 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
982 starting offset points to the beginning of a UTF-8 character was
983 unnecessarily clumsy. I tidied it up.
984
985 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
986 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
987
988 17. Nobody had reported that the --include_dir option, which was added in
989 release 7.7 should have been called --include-dir (hyphen, not underscore)
990 for compatibility with GNU grep. I have changed it to --include-dir, but
991 left --include_dir as an undocumented synonym, and the same for
992 --exclude-dir, though that is not available in GNU grep, at least as of
993 release 2.5.4.
994
995 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
996 characters from a string of bytes) have been redefined so as not to use
997 loops, in order to improve performance in some environments. At the same
998 time, I abstracted some of the common code into auxiliary macros to save
999 repetition (this should not affect the compiled code).
1000
1001 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1002 compile-time error is now given if \c is not followed by an ASCII
1003 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1004 different, and any byte value is allowed.)
1005
1006 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1007 START_OPTIMIZE option, which is now allowed at compile time - but just
1008 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1009 to pcregrep and other applications that have no direct access to PCRE
1010 options. The new /Y option in pcretest sets this option when calling
1011 pcre_compile().
1012
1013 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1014 back references. Groups containing recursive back references were forced to
1015 be atomic by that change, but in the case of named groups, the amount of
1016 memory required was incorrectly computed, leading to "Failed: internal
1017 error: code overflow". This has been fixed.
1018
1019 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1020 pcretest.c, to avoid build problems in some Borland environments.
1021
1022
1023 Version 8.10 25-Jun-2010
1024 ------------------------
1025
1026 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1027 THEN.
1028
1029 2. (*ACCEPT) was not working when inside an atomic group.
1030
1031 3. Inside a character class, \B is treated as a literal by default, but
1032 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1033 causes the error). The code is unchanged, but I tidied the documentation.
1034
1035 4. Inside a character class, PCRE always treated \R and \X as literals,
1036 whereas Perl faults them if its -w option is set. I have changed PCRE so
1037 that it faults them when PCRE_EXTRA is set.
1038
1039 5. Added support for \N, which always matches any character other than
1040 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1041
1042 6. When compiling pcregrep with newer versions of gcc which may have
1043 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1044 declared with attribute warn_unused_result" were given. Just casting the
1045 result to (void) does not stop the warnings; a more elaborate fudge is
1046 needed. I've used a macro to implement this.
1047
1048 7. Minor change to pcretest.c to avoid a compiler warning.
1049
1050 8. Added four artifical Unicode properties to help with an option to make
1051 \s etc use properties (see next item). The new properties are: Xan
1052 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1053
1054 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1055 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1056 this option. Modified pcretest to add /W to test this facility. Added
1057 REG_UCP to make it available via the POSIX interface.
1058
1059 10. Added --line-buffered to pcregrep.
1060
1061 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1062 studied, and the match started with a letter with a code point greater than
1063 127 whose first byte was different to the first byte of the other case of
1064 the letter, the other case of this starting letter was not recognized
1065 (#976).
1066
1067 12. If a pattern that was studied started with a repeated Unicode property
1068 test, for example, \p{Nd}+, there was the theoretical possibility of
1069 setting up an incorrect bitmap of starting bytes, but fortunately it could
1070 not have actually happened in practice until change 8 above was made (it
1071 added property types that matched character-matching opcodes).
1072
1073 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1074 possible starting bytes for non-anchored patterns.
1075
1076 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1077 \R, and also a number of cases that involve Unicode properties, both
1078 explicit and implicit when PCRE_UCP is set.
1079
1080 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1081 input, it could crash or give wrong results if characters with values
1082 greater than 0xc0 were present in the subject string. (Detail: it assumed
1083 UTF-8 input when processing these items.)
1084
1085 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1086 size_t is 64-bit (#991).
1087
1088 17. Added a check for running out of memory when PCRE is compiled with
1089 --disable-stack-for-recursion (#990).
1090
1091 18. If the last data line in a file for pcretest does not have a newline on
1092 the end, a newline was missing in the output.
1093
1094 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1095 less than 128) in its various bitmaps. However, there is a facility for
1096 generating tables according to the current locale when PCRE is compiled. It
1097 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1098 characters, are recognized by isspace() and therefore were getting set in
1099 these tables, and indeed these tables seem to approximate to ISO 8859. This
1100 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1101 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1102 which of course cannot start UTF-8 characters. I have changed the code so
1103 that only real ASCII characters (less than 128) and the correct starting
1104 bytes for UTF-8 encodings are set for characters greater than 127 when in
1105 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1106 altogether.)
1107
1108 20. Added the /T option to pcretest so as to be able to run tests with non-
1109 standard character tables, thus making it possible to include the tests
1110 used for 19 above in the standard set of tests.
1111
1112 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1113 reference to a subpattern the other side of a comment that contains an
1114 opening parenthesis caused either an internal compiling error, or a
1115 reference to the wrong subpattern.
1116
1117
1118 Version 8.02 19-Mar-2010
1119 ------------------------
1120
1121 1. The Unicode data tables have been updated to Unicode 5.2.0.
1122
1123 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1124 configured.
1125
1126 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1127 original author of that file, following a query about its status.
1128
1129 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1130 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1131
1132 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1133 quantifier applied to a forward-referencing subroutine call, could compile
1134 incorrect code or give the error "internal error: previously-checked
1135 referenced subpattern not found".
1136
1137 6. Both MS Visual Studio and Symbian OS have problems with initializing
1138 variables to point to external functions. For these systems, therefore,
1139 pcre_malloc etc. are now initialized to local functions that call the
1140 relevant global functions.
1141
1142 7. There were two entries missing in the vectors called coptable and poptable
1143 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1144 I've fixed the data, and added a kludgy way of testing at compile time that
1145 the lengths are correct (equal to the number of opcodes).
1146
1147 8. Following on from 7, I added a similar kludge to check the length of the
1148 eint vector in pcreposix.c.
1149
1150 9. Error texts for pcre_compile() are held as one long string to avoid too
1151 much relocation at load time. To find a text, the string is searched,
1152 counting zeros. There was no check for running off the end of the string,
1153 which could happen if a new error number was added without updating the
1154 string.
1155
1156 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1157
1158 11. \K was not working if it appeared in an atomic group or in a group that
1159 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1160 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1161 the assertion is positive, but not if it is negative.
1162
1163 12. Change 11 fortuitously reduced the size of the stack frame used in the
1164 "match()" function of pcre_exec.c by one pointer. Forthcoming
1165 implementation of support for (*MARK) will need an extra pointer on the
1166 stack; I have reserved it now, so that the stack frame size does not
1167 decrease.
1168
1169 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1170 item in branch that calls a recursion is a subroutine call - as in the
1171 second branch in the above example - was incorrectly given the compile-
1172 time error "recursive call could loop indefinitely" because pcre_compile()
1173 was not correctly checking the subroutine for matching a non-empty string.
1174
1175 14. The checks for overrunning compiling workspace could trigger after an
1176 overrun had occurred. This is a "should never occur" error, but it can be
1177 triggered by pathological patterns such as hundreds of nested parentheses.
1178 The checks now trigger 100 bytes before the end of the workspace.
1179
1180 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1181
1182
1183 Version 8.01 19-Jan-2010
1184 ------------------------
1185
1186 1. If a pattern contained a conditional subpattern with only one branch (in
1187 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1188 computed the wrong minimum data length (which is of course zero for such
1189 subpatterns). This could cause incorrect "no match" results.
1190
1191 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1192 the pattern is reset in the first branch, pcre_compile() failed with
1193 "internal error: code overflow at offset...". This happened only when
1194 the reset was to the original external option setting. (An optimization
1195 abstracts leading options settings into an external setting, which was the
1196 cause of this.)
1197
1198 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1199 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1200 assertion pattern did not match (meaning that the assertion was true), it
1201 was incorrectly treated as false if the SKIP had been reached during the
1202 matching. This also applied to assertions used as conditions.
1203
1204 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1205 assertion subpattern, including such a pattern used as a condition,
1206 unpredictable results occurred, instead of the error return
1207 PCRE_ERROR_DFA_UITEM.
1208
1209 5. The C++ GlobalReplace function was not working like Perl for the special
1210 situation when an empty string is matched. It now does the fancy magic
1211 stuff that is necessary.
1212
1213 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1214 removed. (These were left over from very, very early versions of PCRE.)
1215
1216 7. Some cosmetic changes to the code to make life easier when compiling it
1217 as part of something else:
1218
1219 (a) Change DEBUG to PCRE_DEBUG.
1220
1221 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1222 called "current" as "current_branch", to prevent a collision with the
1223 Linux macro when compiled as a kernel module.
1224
1225 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1226 prevent a collision with the Linux macro when compiled as a kernel
1227 module.
1228
1229 8. In pcre_compile() there are some checks for integer overflows that used to
1230 cast potentially large values to (double). This has been changed to that
1231 when building, a check for int64_t is made, and if it is found, it is used
1232 instead, thus avoiding the use of floating point arithmetic. (There is no
1233 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1234 double.
1235
1236 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1237 2005 (difference between two addresses compared to an unsigned value).
1238
1239 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1240 custom one, because of the following reported problem in Windows:
1241
1242 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1243 under Win32.
1244 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1245 therefore missing the function definition.
1246 - The compiler thus generates a "C" signature for the test function.
1247 - The linker fails to find the "C" function.
1248 - PCRE fails to configure if asked to do so against libbz2.
1249
1250 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1251 messages were output:
1252
1253 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1254 rerunning libtoolize, to keep the correct libtool macros in-tree.
1255 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1256
1257 I have done both of these things.
1258
1259 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1260 most of the time, it *can* run out if it is given a pattern that contains a
1261 runaway infinite recursion. I updated the discussion in the pcrestack man
1262 page.
1263
1264 13. Now that we have gone to the x.xx style of version numbers, the minor
1265 version may start with zero. Using 08 or 09 is a bad idea because users
1266 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1267 interpreted as invalid octal numbers. I've updated the previous comment in
1268 configure.ac, and also added a check that gives an error if 08 or 09 are
1269 used.
1270
1271 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1272 causing partial matching to fail when the end of the subject matched \W
1273 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1274
1275 15. There were some discrepancies between the declarations in pcre_internal.h
1276 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1277 their definitions. The declarations used "const uschar *" and the
1278 definitions used USPTR. Even though USPTR is normally defined as "const
1279 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1280 reported that: "This difference in casting confuses some C++ compilers, for
1281 example, SunCC recognizes above declarations as different functions and
1282 generates broken code for hbpcre." I have changed the declarations to use
1283 USPTR.
1284
1285 16. GNU libtool is named differently on some systems. The autogen.sh script now
1286 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1287 (FreeBSD).
1288
1289 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1290 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1291 comment: "Figure out how to create a longlong from a string: strtoll and
1292 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1293 instance, but it only takes 2 args instead of 3!"
1294
1295 18. A subtle bug concerned with back references has been fixed by a change of
1296 specification, with a corresponding code fix. A pattern such as
1297 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1298 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1299 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1300 same bug. Such groups have to be quantified to be useful, or contained
1301 inside another quantified group. (If there's no repetition, the reference
1302 can never match.) The problem arises because, having left the group and
1303 moved on to the rest of the pattern, a later failure that backtracks into
1304 the group uses the captured value from the final iteration of the group
1305 rather than the correct earlier one. I have fixed this in PCRE by forcing
1306 any group that contains a reference to itself to be an atomic group; that
1307 is, there cannot be any backtracking into it once it has completed. This is
1308 similar to recursive and subroutine calls.
1309
1310
1311 Version 8.00 19-Oct-09
1312 ----------------------
1313
1314 1. The table for translating pcre_compile() error codes into POSIX error codes
1315 was out-of-date, and there was no check on the pcre_compile() error code
1316 being within the table. This could lead to an OK return being given in
1317 error.
1318
1319 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1320 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1321 in a Windows environment.
1322
1323 3. The pcregrep --count option prints the count for each file even when it is
1324 zero, as does GNU grep. However, pcregrep was also printing all files when
1325 --files-with-matches was added. Now, when both options are given, it prints
1326 counts only for those files that have at least one match. (GNU grep just
1327 prints the file name in this circumstance, but including the count seems
1328 more useful - otherwise, why use --count?) Also ensured that the
1329 combination -clh just lists non-zero counts, with no names.
1330
1331 4. The long form of the pcregrep -F option was incorrectly implemented as
1332 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1333 but it seems right to fix it, and I didn't think it was worth preserving
1334 the old behaviour.
1335
1336 5. The command line items --regex=pattern and --regexp=pattern were not
1337 recognized by pcregrep, which required --regex pattern or --regexp pattern
1338 (with a space rather than an '='). The man page documented the '=' forms,
1339 which are compatible with GNU grep; these now work.
1340
1341 6. No libpcreposix.pc file was created for pkg-config; there was just
1342 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1343
1344 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1345 when UCP support is not needed, by modifying the Python script that
1346 generates it from Unicode data files. This should not matter if the module
1347 is correctly used as a library, but I received one complaint about 50K of
1348 unwanted data. My guess is that the person linked everything into his
1349 program rather than using a library. Anyway, it does no harm.
1350
1351 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1352 was a minimum greater than 1 for a wide character in a possessive
1353 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1354 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1355 character. Chaos in the form of incorrect output or a compiling loop could
1356 result.
1357
1358 9. The restrictions on what a pattern can contain when partial matching is
1359 requested for pcre_exec() have been removed. All patterns can now be
1360 partially matched by this function. In addition, if there are at least two
1361 slots in the offset vector, the offset of the earliest inspected character
1362 for the match and the offset of the end of the subject are set in them when
1363 PCRE_ERROR_PARTIAL is returned.
1364
1365 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1366 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1367 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1368 and may be more useful for multi-segment matching.
1369
1370 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1371 used to be given if ever the end of the subject was reached; now it is
1372 given only if matching could not proceed because another character was
1373 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1374 string "Z", which now yields "no match" instead of "partial match". In the
1375 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1376 final character ended with (*FAIL).
1377
1378 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1379 if the pattern had a "must contain" character that was already found in the
1380 earlier partial match, unless partial matching was again requested. For
1381 example, with the pattern /dog.(body)?/, the "must contain" character is
1382 "g". If the first part-match was for the string "dog", restarting with
1383 "sbody" failed. This bug has been fixed.
1384
1385 13. The string returned by pcre_dfa_exec() after a partial match has been
1386 changed so that it starts at the first inspected character rather than the
1387 first character of the match. This makes a difference only if the pattern
1388 starts with a lookbehind assertion or \b or \B (\K is not supported by
1389 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1390 matching functions compatible, and I think it's the right thing to do.
1391
1392 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1393 so that the demonstration program is easily available in environments where
1394 PCRE has not been installed from source.
1395
1396 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1397 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1398 library.
1399
1400 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1401 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1402 is not the first non-POSIX option to be added. Clearly some people find
1403 these options useful.
1404
1405 17. If a caller to the POSIX matching function regexec() passes a non-zero
1406 value for nmatch with a NULL value for pmatch, the value of
1407 nmatch is forced to zero.
1408
1409 18. RunGrepTest did not have a test for the availability of the -u option of
1410 the diff command, as RunTest does. It now checks in the same way as
1411 RunTest, and also checks for the -b option.
1412
1413 19. If an odd number of negated classes containing just a single character
1414 interposed, within parentheses, between a forward reference to a named
1415 subpattern and the definition of the subpattern, compilation crashed with
1416 an internal error, complaining that it could not find the referenced
1417 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1418 [The bug was that it was starting one character too far in when skipping
1419 over the character class, thus treating the ] as data rather than
1420 terminating the class. This meant it could skip too much.]
1421
1422 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1423 /g option in pcretest when the pattern contains \K, which makes it possible
1424 to have an empty string match not at the start, even when the pattern is
1425 anchored. Updated pcretest and pcredemo to use this option.
1426
1427 21. If the maximum number of capturing subpatterns in a recursion was greater
1428 than the maximum at the outer level, the higher number was returned, but
1429 with unset values at the outer level. The correct (outer level) value is
1430 now given.
1431
1432 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1433 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1434 make it do so. The string so far is captured, making this feature
1435 compatible with Perl.
1436
1437 23. The tests have been re-organized, adding tests 11 and 12, to make it
1438 possible to check the Perl 5.10 features against Perl 5.10.
1439
1440 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1441 pattern matches a fixed length string. PCRE did not allow this; now it
1442 does. Neither allows recursion.
1443
1444 25. I finally figured out how to implement a request to provide the minimum
1445 length of subject string that was needed in order to match a given pattern.
1446 (It was back references and recursion that I had previously got hung up
1447 on.) This code has now been added to pcre_study(); it finds a lower bound
1448 to the length of subject needed. It is not necessarily the greatest lower
1449 bound, but using it to avoid searching strings that are too short does give
1450 some useful speed-ups. The value is available to calling programs via
1451 pcre_fullinfo().
1452
1453 26. While implementing 25, I discovered to my embarrassment that pcretest had
1454 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1455 study optimizations had never been tested with that matching function.
1456 Oops. What is worse, even when it was passed study data, there was a bug in
1457 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1458 were also very few tests of studied patterns with pcre_dfa_exec().
1459
1460 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1461 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1462 on the other side of the coin, they are no longer allowed to have different
1463 names, because these cannot be distinguished in PCRE, and this has caused
1464 confusion. (This is a difference from Perl.)
1465
1466 28. When duplicate subpattern names are present (necessarily with different
1467 numbers, as required by 27 above), and a test is made by name in a
1468 conditional pattern, either for a subpattern having been matched, or for
1469 recursion in such a pattern, all the associated numbered subpatterns are
1470 tested, and the overall condition is true if the condition is true for any
1471 one of them. This is the way Perl works, and is also more like the way
1472 testing by number works.
1473
1474
1475 Version 7.9 11-Apr-09
1476 ---------------------
1477
1478 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1479 (pcretest), all targets were linked against these libraries. This included
1480 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1481 libraries. This caused unwanted dependencies to be created. This problem
1482 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1483 pcretest is linked with readline.
1484
1485 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1486 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1487 moved outside it again, because FALSE and TRUE are already defined in AIX,
1488 but BOOL is not.
1489
1490 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1491 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1492
1493 4. The pcregrep documentation said spaces were inserted as well as colons (or
1494 hyphens) following file names and line numbers when outputting matching
1495 lines. This is not true; no spaces are inserted. I have also clarified the
1496 wording for the --colour (or --color) option.
1497
1498 5. In pcregrep, when --colour was used with -o, the list of matching strings
1499 was not coloured; this is different to GNU grep, so I have changed it to be
1500 the same.
1501
1502 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1503 each matching line was coloured. Now it goes on to look for further matches
1504 of any of the test patterns, which is the same behaviour as GNU grep.
1505
1506 7. A pattern that could match an empty string could cause pcregrep to loop; it
1507 doesn't make sense to accept an empty string match in pcregrep, so I have
1508 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1509 seems to be how GNU grep behaves.
1510
1511 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1512 start or after a newline", because the conditional assertion was not being
1513 correctly handled. The rule now is that both the assertion and what follows
1514 in the first alternative must satisfy the test.
1515
1516 9. If auto-callout was enabled in a pattern with a conditional group whose
1517 condition was an assertion, PCRE could crash during matching, both with
1518 pcre_exec() and pcre_dfa_exec().
1519
1520 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1521 used for matching.
1522
1523 11. Unicode property support in character classes was not working for
1524 characters (bytes) greater than 127 when not in UTF-8 mode.
1525
1526 12. Added the -M command line option to pcretest.
1527
1528 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1529
1530 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1531
1532 16. Added comments and documentation about mis-use of no_arg in the C++
1533 wrapper.
1534
1535 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1536 from Martin Jerabek that uses macro names for all relevant character and
1537 string constants.
1538
1539 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1540 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1541 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1542 these, but not everybody uses configure.
1543
1544 19. A conditional group that had only one branch was not being correctly
1545 recognized as an item that could match an empty string. This meant that an
1546 enclosing group might also not be so recognized, causing infinite looping
1547 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1548 with the subject "ab", where knowledge that the repeated group can match
1549 nothing is needed in order to break the loop.
1550
1551 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1552 exec(), but without supplying a callout function, matching went wrong.
1553
1554 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1555 leak if the size of the offset vector was greater than 30. When the vector
1556 is smaller, the saved offsets during recursion go onto a local stack
1557 vector, but for larger vectors malloc() is used. It was failing to free
1558 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1559 error, in fact).
1560
1561 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1562 heapframe that is used only when UTF-8 support is enabled. This caused no
1563 problem, but was untidy.
1564
1565 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1566 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1567 included within another project.
1568
1569 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1570 slightly modified by me:
1571
1572 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1573 not building pcregrep.
1574
1575 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1576 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1577
1578 25. Forward references, both numeric and by name, in patterns that made use of
1579 duplicate group numbers, could behave incorrectly or give incorrect errors,
1580 because when scanning forward to find the reference group, PCRE was not
1581 taking into account the duplicate group numbers. A pattern such as
1582 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1583
1584 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1585 the feature of a custom pointer more persuasive (as requested by a user).
1586
1587 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1588 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1589 pre-defined.
1590
1591 28. Added support for (*UTF8) at the start of a pattern.
1592
1593 29. Arrange for flags added by the "release type" setting in CMake to be shown
1594 in the configuration summary.
1595
1596
1597 Version 7.8 05-Sep-08
1598 ---------------------
1599
1600 1. Replaced UCP searching code with optimized version as implemented for Ad
1601 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1602 stage table and inline lookup instead of a function, giving speed ups of 2
1603 to 5 times on some simple patterns that I tested. Permission was given to
1604 distribute the MultiStage2.py script that generates the tables (it's not in
1605 the tarball, but is in the Subversion repository).
1606
1607 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1608 scripts.
1609
1610 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1611 a group with a zero qualifier. The result of the study could be incorrect,
1612 or the function might crash, depending on the pattern.
1613
1614 4. Caseless matching was not working for non-ASCII characters in back
1615 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1616 It now works when Unicode Property Support is available.
1617
1618 5. In pcretest, an escape such as \x{de} in the data was always generating
1619 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1620 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1621 truncation.
1622
1623 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1624
1625 7. Added two (int) casts to pcregrep when printing the difference of two
1626 pointers, in case they are 64-bit values.
1627
1628 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1629 test 2 if it fails.
1630
1631 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1632 and a #define of that name to empty if it is not externally set. This is to
1633 allow users of MSVC to set it if necessary.
1634
1635 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1636 the convenience functions in the pcre_get.c source file.
1637
1638 11. An option change at the start of a pattern that had top-level alternatives
1639 could cause overwriting and/or a crash. This command provoked a crash in
1640 some environments:
1641
1642 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1643
1644 This potential security problem was recorded as CVE-2008-2371.
1645
1646 12. For a pattern where the match had to start at the beginning or immediately
1647 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1648 pcre_dfa_exec() could read past the end of the passed subject if there was
1649 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1650 pcretest so that it places the subject at the end of its malloc-ed buffer.
1651
1652 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1653 exec() might read past the end of the data buffer in UTF-8 mode.
1654
1655 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1656 the data contained the byte 0x85 as part of a UTF-8 character within its
1657 first line. This applied both to normal and DFA matching.
1658
1659 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1660 /^[^d]*?$/8 failed to match "abc".
1661
1662 16. Added a missing copyright notice to pcrecpp_internal.h.
1663
1664 17. Make it more clear in the documentation that values returned from
1665 pcre_exec() in ovector are byte offsets, not character counts.
1666
1667 18. Tidied a few places to stop certain compilers from issuing warnings.
1668
1669 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1670 supplied by Stefan Weber. I made a further small update for 7.8 because
1671 there is a change of source arrangements: the pcre_searchfuncs.c module is
1672 replaced by pcre_ucd.c.
1673
1674
1675 Version 7.7 07-May-08
1676 ---------------------
1677
1678 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1679 a string to a long long, pretend we don't even have a long long." This is
1680 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1681
1682 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1683 pre-7.6 versions, which defined a global no_arg variable instead of putting
1684 it in the RE class. (See also #8 below.)
1685
1686 3. Remove a line of dead code, identified by coverity and reported by Nuno
1687 Lopes.
1688
1689 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1690
1691 (1) The include/exclude patterns were being applied to the whole pathnames
1692 of files, instead of just to the final components.
1693
1694 (2) If there was more than one level of directory, the subdirectories were
1695 skipped unless they satisfied the include/exclude conditions. This is
1696 inconsistent with GNU grep (and could even be seen as contrary to the
1697 pcregrep specification - which I improved to make it absolutely clear).
1698 The action now is always to scan all levels of directory, and just
1699 apply the include/exclude patterns to regular files.
1700
1701 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1702 --exclude_dir in the tests to avoid scanning .svn directories.
1703
1704 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1705 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1706 doesn't support NULs in patterns.
1707
1708 7. Added some missing "const"s to declarations of static tables in
1709 pcre_compile.c and pcre_dfa_exec.c.
1710
1711 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1712 caused by fix #2 above. (Subsequently also a second patch to fix the
1713 first patch. And a third patch - this was a messy problem.)
1714
1715 9. Applied Craig's patch to remove the use of push_back().
1716
1717 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1718 matching function regexec().
1719
1720 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1721 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1722 references. PCRE supports relative numbers with this syntax (I don't think
1723 Oniguruma does).
1724
1725 12. Previously, a group with a zero repeat such as (...){0} was completely
1726 omitted from the compiled regex. However, this means that if the group
1727 was called as a subroutine from elsewhere in the pattern, things went wrong
1728 (an internal error was given). Such groups are now left in the compiled
1729 pattern, with a new opcode that causes them to be skipped at execution
1730 time.
1731
1732 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1733 to the way PCRE behaves:
1734
1735 (a) A lone ] character is dis-allowed (Perl treats it as data).
1736
1737 (b) A back reference to an unmatched subpattern matches an empty string
1738 (Perl fails the current match path).
1739
1740 (c) A data ] in a character class must be notated as \] because if the
1741 first data character in a class is ], it defines an empty class. (In
1742 Perl it is not possible to have an empty class.) The empty class []
1743 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1744 The negative empty class [^] matches any one character, independently
1745 of the DOTALL setting.
1746
1747 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1748 non-existent subpattern following a character class starting with ']' and
1749 containing () gave an internal compiling error instead of "reference to
1750 non-existent subpattern". Fortunately, when the pattern did exist, the
1751 compiled code was correct. (When scanning forwards to check for the
1752 existencd of the subpattern, it was treating the data ']' as terminating
1753 the class, so got the count wrong. When actually compiling, the reference
1754 was subsequently set up correctly.)
1755
1756 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1757 it was being rejected as not supported by pcre_dfa_exec(), even though
1758 other assertions are supported. I have made pcre_dfa_exec() support
1759 (*FAIL).
1760
1761 16. The implementation of 13c above involved the invention of a new opcode,
1762 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1763 cannot be changed at match time, I realized I could make a small
1764 improvement to matching performance by compiling OP_ALLANY instead of
1765 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1766 on the OP_ANY path.
1767
1768 17. Compiling pcretest on Windows with readline support failed without the
1769 following two fixes: (1) Make the unistd.h include conditional on
1770 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1771
1772 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1773 ncurses library to be included for pcretest when ReadLine support is
1774 requested, but also to allow for it to be overridden. This patch came from
1775 Daniel Bergström.
1776
1777 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1778 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1779 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1780 spotting this.
1781
1782
1783 Version 7.6 28-Jan-08
1784 ---------------------
1785
1786 1. A character class containing a very large number of characters with
1787 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1788 overflow.
1789
1790 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1791 HAVE_LONG_LONG is not defined.
1792
1793 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1794 bring them up to date and include new features. This patch includes:
1795
1796 - Fixed PH's badly added libz and libbz2 support.
1797 - Fixed a problem with static linking.
1798 - Added pcredemo. [But later removed - see 7 below.]
1799 - Fixed dftables problem and added an option.
1800 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1801 HAVE_LONG_LONG.
1802 - Added readline support for pcretest.
1803 - Added an listing of the option settings after cmake has run.
1804
1805 4. A user submitted a patch to Makefile that makes it easy to create
1806 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1807 Makefile.am that cause it to include this special target, without
1808 affecting anything else. Note that the same mingw target plus all
1809 the other distribution libraries and programs are now supported
1810 when configuring with CMake (see 6 below) instead of with
1811 Configure/Make.
1812
1813 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1814 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1815 exported in the Windows port". It has not yet been confirmed that the patch
1816 solves the problem, but it does no harm.
1817
1818 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1819 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1820 with CMake, and also correct the comment about stack recursion.
1821
1822 7. Remove the automatic building of pcredemo from the ./configure system and
1823 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1824 of a program that users should build themselves after PCRE is installed, so
1825 building it automatically is not really right. What is more, it gave
1826 trouble in some build environments.
1827
1828 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1829
1830
1831 Version 7.5 10-Jan-08
1832 ---------------------
1833
1834 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1835 values in parens when parsing an RE using the C++ wrapper."
1836
1837 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1838 Characters greater than 255 were excluded from the class instead of being
1839 included.
1840
1841 3. The same bug as (2) above applied to negated POSIX classes such as
1842 [:^space:].
1843
1844 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1845 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1846 I have changed it.
1847
1848 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1849 first named subpattern) and a construct such as (?&a) would reference the
1850 first named subpattern whose name started with "a" (in other words, the
1851 length check was missing). Both these problems are fixed. "Subpattern name
1852 expected" is now given for (?&) (a zero-length name), and this patch also
1853 makes it give the same error for \k'' (previously it complained that that
1854 was a reference to a non-existent subpattern).
1855
1856 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1857 this is right because (?- can be followed by option settings as well as by
1858 digits. I have, however, made the messages clearer.
1859
1860 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1861 than the number used in the conditional) now cause a compile-time error.
1862 This is actually not compatible with Perl, which accepts such patterns, but
1863 treats the conditional as always being FALSE (as PCRE used to), but it
1864 seems to me that giving a diagnostic is better.
1865
1866 8. Change "alphameric" to the more common word "alphanumeric" in comments
1867 and messages.
1868
1869 9. Fix two occurrences of "backslash" in comments that should have been
1870 "backspace".
1871
1872 10. Remove two redundant lines of code that can never be obeyed (their function
1873 was moved elsewhere).
1874
1875 11. The program that makes PCRE's Unicode character property table had a bug
1876 which caused it to generate incorrect table entries for sequences of
1877 characters that have the same character type, but are in different scripts.
1878 It amalgamated them into a single range, with the script of the first of
1879 them. In other words, some characters were in the wrong script. There were
1880 thirteen such cases, affecting characters in the following ranges:
1881
1882 U+002b0 - U+002c1
1883 U+0060c - U+0060d
1884 U+0061e - U+00612
1885 U+0064b - U+0065e
1886 U+0074d - U+0076d
1887 U+01800 - U+01805
1888 U+01d00 - U+01d77
1889 U+01d9b - U+01dbf
1890 U+0200b - U+0200f
1891 U+030fc - U+030fe
1892 U+03260 - U+0327f
1893 U+0fb46 - U+0fbb1
1894 U+10450 - U+1049d
1895
1896 12. The -o option (show only the matching part of a line) for pcregrep was not
1897 compatible with GNU grep in that, if there was more than one match in a
1898 line, it showed only the first of them. It now behaves in the same way as
1899 GNU grep.
1900
1901 13. If the -o and -v options were combined for pcregrep, it printed a blank
1902 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1903 does the same. The return code can be used to tell if there were any
1904 non-matching lines.
1905
1906 14. Added --file-offsets and --line-offsets to pcregrep.
1907
1908 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1909 infinitely looping recursion. The bug was that positive lookaheads were not
1910 being skipped when checking for a possible empty match (negative lookaheads
1911 and both kinds of lookbehind were skipped).
1912
1913 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1914 inclusion of <windows.h> to before rather than after the definition of
1915 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1916
1917 17. Specifying a possessive quantifier with a specific limit for a Unicode
1918 character property caused pcre_compile() to compile bad code, which led at
1919 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1920 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1921 caused the error; without that there was no problem.
1922
1923 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1924
1925 19. Added --enable-pcretest-libreadline.
1926
1927 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1928 RE::GlobalReplace(). As a result, the number of replacements returned was
1929 double what it should be. I removed one of the increments, but Craig sent a
1930 later patch that removed the other one (the right fix) and added unit tests
1931 that check the return values (which was not done before).
1932
1933 21. Several CMake things:
1934
1935 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1936 the names libpcre and libpcreposix, not just pcre and pcreposix.
1937
1938 (2) The above change means that pcretest and pcregrep are now correctly
1939 linked with the newly-built libraries, not previously installed ones.
1940
1941 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1942
1943 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1944 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1945 UTF-8 newline character). The key issue is that the pattern starts .*;
1946 this means that the match must be either at the beginning, or after a
1947 newline. The bug was in the code for advancing after a failed match and
1948 checking that the new position followed a newline. It was not taking
1949 account of UTF-8 characters correctly.
1950
1951 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1952 character classes. PCRE was not treating the sequence [:...:] as a
1953 character class unless the ... were all letters. Perl, however, seems to
1954 allow any characters between [: and :], though of course it rejects as
1955 unknown any "names" that contain non-letters, because all the known class
1956 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1957 for example, whereas PCRE did not - it did not recognize a POSIX character
1958 class. This seemed a bit dangerous, so the code has been changed to be
1959 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1960 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1961 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1962 Perl does, and where it didn't before.
1963
1964 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1965 Windows environments %n is disabled by default.
1966
1967
1968 Version 7.4 21-Sep-07
1969 ---------------------
1970
1971 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1972 means that a class such as [\s] counted as "explicit reference to CR or
1973 LF". That isn't really right - the whole point of the change was to try to
1974 help when there was an actual mention of one of the two characters. So now
1975 the change happens only if \r or \n (or a literal CR or LF) character is
1976 encountered.
1977
1978 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1979 of both had grown to the point where there were only 3 bits left.
1980 Fortunately, there was spare space in the data structure, and so I have
1981 moved the internal flags into a new 16-bit field to free up more option
1982 bits.
1983
1984 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1985 but did not set the internal JCHANGED flag - either of these is enough to
1986 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1987 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1988 start sets both bits.
1989
1990 4. Added options (at build time, compile time, exec time) to change \R from
1991 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1992
1993 5. doc/pcresyntax.html was missing from the distribution.
1994
1995 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1996 compatibility, even though it is no longer used.
1997
1998 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1999 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2000 windows.h file is present (where different names are used). [This was
2001 reversed later after testing - see 16 below.]
2002
2003 8. Changed all #include <config.h> to #include "config.h". There were also
2004 some further <pcre.h> cases that I changed to "pcre.h".
2005
2006 9. When pcregrep was used with the --colour option, it missed the line ending
2007 sequence off the lines that it output.
2008
2009 10. It was pointed out to me that arrays of string pointers cause lots of
2010 relocations when a shared library is dynamically loaded. A technique of
2011 using a single long string with a table of offsets can drastically reduce
2012 these. I have refactored PCRE in four places to do this. The result is
2013 dramatic:
2014
2015 Originally: 290
2016 After changing UCP table: 187
2017 After changing error message table: 43
2018 After changing table of "verbs" 36
2019 After changing table of Posix names 22
2020
2021 Thanks to the folks working on Gregex for glib for this insight.
2022
2023 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2024 unicode-properties was also set.
2025
2026 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2027
2028 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2029 checked only for CRLF.
2030
2031 14. Added casts to pcretest.c to avoid compiler warnings.
2032
2033 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2034
2035 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2036 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2037 entirely. This removes changes made in 7 above.
2038
2039 17. The CMake files have been updated, and there is now more information about
2040 building with CMake in the NON-UNIX-USE document.
2041
2042
2043 Version 7.3 28-Aug-07
2044 ---------------------
2045
2046 1. In the rejigging of the build system that eventually resulted in 7.1, the
2047 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2048 brackets there is not right, since it causes compilers to look for an
2049 installed pcre.h, not the version that is in the source that is being
2050 compiled (which of course may be different). I have changed it back to:
2051
2052 #include "pcre.h"
2053
2054 I have a vague recollection that the change was concerned with compiling in
2055 different directories, but in the new build system, that is taken care of
2056 by the VPATH setting the Makefile.
2057
2058 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2059 when the subject happened to end in the byte 0x85 (e.g. if the last
2060 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2061 characters but of course it shouldn't be taken as a newline when it is part
2062 of another character. The bug was that, for an unlimited repeat of . in
2063 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2064 characters when looking for a newline.
2065
2066 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2067
2068 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2069 in debug output.
2070
2071 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2072 long printing in the pcrecpp unittest when running under MinGW.
2073
2074 6. ESC_K was left out of the EBCDIC table.
2075
2076 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2077 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2078 limit also applies to "virtual nesting" when a pattern is recursive, and in
2079 this case 1000 isn't so big. I have been able to remove this limit at the
2080 expense of backing off one optimization in certain circumstances. Normally,
2081 when pcre_exec() would call its internal match() function recursively and
2082 immediately return the result unconditionally, it uses a "tail recursion"
2083 feature to save stack. However, when a subpattern that can match an empty
2084 string has an unlimited repetition quantifier, it no longer makes this
2085 optimization. That gives it a stack frame in which to save the data for
2086 checking that an empty string has been matched. Previously this was taken
2087 from the 1000-entry workspace that had been reserved. So now there is no
2088 explicit limit, but more stack is used.
2089
2090 8. Applied Daniel's patches to solve problems with the import/export magic
2091 syntax that is required for Windows, and which was going wrong for the
2092 pcreposix and pcrecpp parts of the library. These were overlooked when this
2093 problem was solved for the main library.
2094
2095 9. There were some crude static tests to avoid integer overflow when computing
2096 the size of patterns that contain repeated groups with explicit upper
2097 limits. As the maximum quantifier is 65535, the maximum group length was
2098 set at 30,000 so that the product of these two numbers did not overflow a
2099 32-bit integer. However, it turns out that people want to use groups that
2100 are longer than 30,000 bytes (though not repeat them that many times).
2101 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2102 made it possible to implement the integer overflow checks in a much more
2103 dynamic way, which I have now done. The artificial limitation on group
2104 length has been removed - we now have only the limit on the total length of
2105 the compiled pattern, which depends on the LINK_SIZE setting.
2106
2107 10. Fixed a bug in the documentation for get/copy named substring when
2108 duplicate names are permitted. If none of the named substrings are set, the
2109 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2110 empty string.
2111
2112 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2113 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2114 because the ] is interpreted as the first data character and the
2115 terminating ] is not found. PCRE has been made compatible with Perl in this
2116 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2117 cause memory overwriting.
2118
2119 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2120 string has been matched (to stop an infinite loop). It was not recognizing
2121 a conditional subpattern that could match an empty string if that
2122 subpattern was within another subpattern. For example, it looped when
2123 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2124 condition was not nested. This bug has been fixed.
2125
2126 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2127 past the start of the subject in the presence of bytes with the top bit
2128 set, for example "\x8aBCD".
2129
2130 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2131 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2132
2133 14. Optimized (?!) to (*FAIL).
2134
2135 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2136 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2137 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2138 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2139 does: it's just the validity check that is more restrictive.
2140
2141 16. Inserted checks for integer overflows during escape sequence (backslash)
2142 processing, and also fixed erroneous offset values for syntax errors during
2143 backslash processing.
2144
2145 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2146 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2147
2148 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2149 caused an overrun.
2150
2151 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2152 something other than just ASCII characters) inside a group that had an
2153 unlimited repeat caused a loop at compile time (while checking to see
2154 whether the group could match an empty string).
2155
2156 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2157 [\P{Any}] did so. (Error in the code for printing property names.)
2158
2159 21. An orphan \E inside a character class could cause a crash.
2160
2161 22. A repeated capturing bracket such as (A)? could cause a wild memory
2162 reference during compilation.
2163
2164 23. There are several functions in pcre_compile() that scan along a compiled
2165 expression for various reasons (e.g. to see if it's fixed length for look
2166 behind). There were bugs in these functions when a repeated \p or \P was
2167 present in the pattern. These operators have additional parameters compared
2168 with \d, etc, and these were not being taken into account when moving along
2169 the compiled data. Specifically:
2170
2171 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2172 length.
2173
2174 (b) An item such as \pL+ within a repeated group could cause crashes or
2175 loops.
2176
2177 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2178 "reference to non-existent subpattern" error.
2179
2180 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2181
2182 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2183 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2184
2185 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2186
2187 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2188 character were causing crashes (broken optimization).
2189
2190 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2191 \p or \P) caused a compile-time loop.
2192
2193 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2194 break. For example, the unstudied pattern [\r\n]A does not match the string
2195 "\r\nA" because change 7.0/46 below moves the current point on by two
2196 characters after failing to match at the start. However, the pattern \nA
2197 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2198 the same is true. There doesn't seem any very clean way out of this, but
2199 what I have chosen to do makes the common cases work: PCRE now takes note
2200 of whether there can be an explicit match for \r or \n anywhere in the
2201 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2202 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2203 pattern has explicit CR or LF references.
2204
2205 29. Added (*CR) etc for changing newline setting at start of pattern.
2206
2207
2208 Version 7.2 19-Jun-07
2209 ---------------------
2210
2211 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2212 which is apparently normally available under Windows.
2213
2214 2. Re-jig the pcregrep tests with different newline settings in an attempt
2215 to make them independent of the local environment's newline setting.
2216
2217 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2218
2219 4. Some of the "internals" tests were previously cut out when the link size
2220 was not 2, because the output contained actual offsets. The recent new
2221 "Z" feature of pcretest means that these can be cut out, making the tests
2222 usable with all link sizes.
2223
2224 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2225 stack recursion. This gives a massive performance boost under BSD, but just
2226 a small improvement under Linux. However, it saves one field in the frame
2227 in all cases.
2228
2229 6. Added more features from the forthcoming Perl 5.10:
2230
2231 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2232 recursion call. It refers to the nth most recently opened parentheses.
2233
2234 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2235 to be opened parentheses.
2236
2237 (c) Conditions that refer to capturing parentheses can be specified
2238 relatively, for example, (?(-2)... or (?(+3)...
2239
2240 (d) \K resets the start of the current match so that everything before
2241 is not part of it.
2242
2243 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2244
2245 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2246 reference syntax.
2247
2248 (g) (?| introduces a group in which the numbering of parentheses in each
2249 alternative starts with the same number.
2250
2251 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2252
2253 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2254 PCRE_INFO_JCHANGED.
2255
2256 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2257 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2258 for detecting groups that can match an empty string.
2259
2260 9. A pattern with a very large number of alternatives (more than several
2261 hundred) was running out of internal workspace during the pre-compile
2262 phase, where pcre_compile() figures out how much memory will be needed. A
2263 bit of new cunning has reduced the workspace needed for groups with
2264 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2265 workspace instead of running out of the 4096 that are available.
2266
2267 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2268
2269 11. Applied patch from Google to remove an optimization that didn't quite work.
2270 The report of the bug said:
2271
2272 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2273 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2274 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2275
2276 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2277 it matched the wrong number of bytes.
2278
2279
2280 Version 7.1 24-Apr-07
2281 ---------------------
2282
2283 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2284 that is more "standard", making use of automake and other Autotools. There
2285 is some re-arrangement of the files and adjustment of comments consequent
2286 on this.
2287
2288 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2289 for recursive directory scanning broke on some systems because the files
2290 are not scanned in any specific order and on different systems the order
2291 was different. A call to "sort" has been inserted into RunGrepTest for the
2292 approprate test as a short-term fix. In the longer term there may be an
2293 alternative.
2294
2295 3. I had an email from Eric Raymond about problems translating some of PCRE's
2296 man pages to HTML (despite the fact that I distribute HTML pages, some
2297 people do their own conversions for various reasons). The problems
2298 concerned the use of low-level troff macros .br and .in. I have therefore
2299 removed all such uses from the man pages (some were redundant, some could
2300 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2301 HTML has been updated to handle .nf/.fi and to complain if it encounters
2302 .br or .in.
2303
2304 4. Updated comments in configure.ac that get placed in config.h.in and also
2305 arranged for config.h to be included in the distribution, with the name
2306 config.h.generic, for the benefit of those who have to compile without
2307 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2308
2309 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2310 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2311 makevp.bat for the current PCRE, using the additional files
2312 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2313
2314 6. A Windows user reported a minor discrepancy with test 2, which turned out
2315 to be caused by a trailing space on an input line that had got lost in his
2316 copy. The trailing space was an accident, so I've just removed it.
2317
2318 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2319 that is needed.
2320
2321 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2322 as "const" (a) because they are and (b) because it helps the PHP
2323 maintainers who have recently made a script to detect big data structures
2324 in the php code that should be moved to the .rodata section. I remembered
2325 to update Builducptable as well, so it won't revert if ucptable.h is ever
2326 re-created.
2327
2328 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2329 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2330 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2331 support is not required. This saves 1.5-2K of code, which is important in
2332 some applications.
2333
2334 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2335 so as not to refer to the tables, even though these functions will never be
2336 called when UTF-8 support is disabled. Otherwise there are problems with a
2337 shared library.
2338
2339 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2340
2341 (a) It was defining its arguments as char * instead of void *.
2342
2343 (b) It was assuming that all moves were upwards in memory; this was true
2344 a long time ago when I wrote it, but is no longer the case.
2345
2346 The emulated memove() is provided for those environments that have neither
2347 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2348 is clearly not the case, as these two bugs were recently reported.
2349
2350 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2351 and Detrail to create the HTML documentation, the .txt form of the man
2352 pages, and it removes trailing spaces from listed files. It also creates
2353 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2354 case, it wraps all the #defines with #ifndefs. This script should be run
2355 before "make dist".
2356
2357 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2358 with Unicode property support.
2359
2360 (a) For a maximizing quantifier, if the two different cases of the
2361 character were of different lengths in their UTF-8 codings (there are
2362 some cases like this - I found 11), and the matching function had to
2363 back up over a mixture of the two cases, it incorrectly assumed they
2364 were both the same length.
2365
2366 (b) When PCRE was configured to use the heap rather than the stack for
2367 recursion during matching, it was not correctly preserving the data for
2368 the other case of a UTF-8 character when checking ahead for a match
2369 while processing a minimizing repeat. If the check also involved
2370 matching a wide character, but failed, corruption could cause an
2371 erroneous result when trying to check for a repeat of the original
2372 character.
2373
2374 13. Some tidying changes to the testing mechanism:
2375
2376 (a) The RunTest script now detects the internal link size and whether there
2377 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2378 values substituted by "configure". (The RunGrepTest script already did
2379 this for UTF-8.) The configure.ac script no longer substitutes the
2380 relevant variables.
2381
2382 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2383 with length and offset values. This means that the output is different
2384 for different internal link sizes. Test 2 is skipped for link sizes
2385 other than 2 because of this, bypassing the problem. Unfortunately,
2386 there was also a test in test 3 (the locale tests) that used /B and
2387 failed for link sizes other than 2. Rather than cut the whole test out,
2388 I have added a new /Z option to pcretest that replaces the length and
2389 offset values with spaces. This is now used to make test 3 independent
2390 of link size. (Test 2 will be tidied up later.)
2391
2392 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2393 segmentation fault instead of returning the appropriate error message.
2394
2395 15. In multiline mode when the newline sequence was set to "any", the pattern
2396 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2397 This doesn't seem right; it now treats the CRLF combination as the line
2398 ending, and so does not match in that case. It's only a pattern such as ^$
2399 that would hit this one: something like ^ABC$ would have failed after \r
2400 and then tried again after \r\n.
2401
2402 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2403 in an attempt to make files that differ only in their line terminators
2404 compare equal. This works on Linux.
2405
2406 17. Under certain error circumstances pcregrep might try to free random memory
2407 as it exited. This is now fixed, thanks to valgrind.
2408
2409 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2410 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2411 was because its rules for how to advance for /g after matching an empty
2412 string at the end of a line did not allow for this case. They now check for
2413 it specially.
2414
2415 20. pcretest is supposed to handle patterns and data of any length, by
2416 extending its buffers when necessary. It was getting this wrong when the
2417 buffer for a data line had to be extended.
2418
2419 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2420 CRLF as a newline sequence.
2421
2422 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2423 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2424 I have nevertheless tidied it up.
2425
2426 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2427
2428 24. Added a man page for pcre-config.
2429
2430
2431 Version 7.0 19-Dec-06
2432 ---------------------
2433
2434 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2435 moving to gcc 4.1.1.
2436
2437 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2438 sys/time.h, which is documented as needed for this function. It doesn't
2439 seem to matter on Linux, but it showed up on some releases of OS X.
2440
2441 3. It seems that there are systems where bytes whose values are greater than
2442 127 match isprint() in the "C" locale. The "C" locale should be the
2443 default when a C program starts up. In most systems, only ASCII printing
2444 characters match isprint(). This difference caused the output from pcretest
2445 to vary, making some of the tests fail. I have changed pcretest so that:
2446
2447 (a) When it is outputting text in the compiled version of a pattern, bytes
2448 other than 32-126 are always shown as hex escapes.
2449
2450 (b) When it is outputting text that is a matched part of a subject string,
2451 it does the same, unless a different locale has been set for the match
2452 (using the /L modifier). In this case, it uses isprint() to decide.
2453
2454 4. Fixed a major bug that caused incorrect computation of the amount of memory
2455 required for a compiled pattern when options that changed within the
2456 pattern affected the logic of the preliminary scan that determines the
2457 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2458 that the computed length was too small. The symptoms of this bug were
2459 either the PCRE error "internal error: code overflow" from pcre_compile(),
2460 or a glibc crash with a message such as "pcretest: free(): invalid next
2461 size (fast)". Examples of patterns that provoked this bug (shown in
2462 pcretest format) are:
2463
2464 /(?-x: )/x
2465 /(?x)(?-x: \s*#\s*)/
2466 /((?i)[\x{c0}])/8
2467 /(?i:[\x{c0}])/8
2468
2469 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2470 is now done differently.
2471
2472 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2473 wrapper classes; (b) implement a new function in the C++ scanner that is
2474 more efficient than the old way of doing things because it avoids levels of
2475 recursion in the regex matching; (c) add a paragraph to the documentation
2476 for the FullMatch() function.
2477
2478 6. The escape sequence \n was being treated as whatever was defined as
2479 "newline". Not only was this contrary to the documentation, which states
2480 that \n is character 10 (hex 0A), but it also went horribly wrong when
2481 "newline" was defined as CRLF. This has been fixed.
2482
2483 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2484 was being set to -1 for the "end of line" case (supposedly a value that no
2485 character can have). Though this value is never used (the check for end of
2486 line is "zero bytes in current character"), it caused compiler complaints.
2487 I've changed it to 0xffffffff.
2488
2489 8. In pcre_version.c, the version string was being built by a sequence of
2490 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2491 string (as it is for production releases) called a macro with an empty
2492 argument. The C standard says the result of this is undefined. The gcc
2493 compiler treats it as an empty string (which was what was wanted) but it is
2494 reported that Visual C gives an error. The source has been hacked around to
2495 avoid this problem.
2496
2497 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2498 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2499 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2500 of them did).
2501
2502 10. Originally, pcretest opened its input and output without "b"; then I was
2503 told that "b" was needed in some environments, so it was added for release
2504 5.0 to both the input and output. (It makes no difference on Unix-like
2505 systems.) Later I was told that it is wrong for the input on Windows. I've
2506 now abstracted the modes into two macros, to make it easier to fiddle with
2507 them, and removed "b" from the input mode under Windows.
2508
2509 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2510
2511 12. Added -help and --help to pcretest as an official way of being reminded
2512 of the options.
2513
2514 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2515 and pcrecpp.cc because they annoy compilers at high warning levels.
2516
2517 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2518
2519 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2520 scripts are not C programs :-) and which was not noticed because it works
2521 on Linux.
2522
2523 16. pcretest is supposed to handle any length of pattern and data line (as one
2524 line or as a continued sequence of lines) by extending its input buffer if
2525 necessary. This feature was broken for very long pattern lines, leading to
2526 a string of junk being passed to pcre_compile() if the pattern was longer
2527 than about 50K.
2528
2529 17. I have done a major re-factoring of the way pcre_compile() computes the
2530 amount of memory needed for a compiled pattern. Previously, there was code
2531 that made a preliminary scan of the pattern in order to do this. That was
2532 OK when PCRE was new, but as the facilities have expanded, it has become
2533 harder and harder to keep it in step with the real compile phase, and there
2534 have been a number of bugs (see for example, 4 above). I have now found a
2535 cunning way of running the real compile function in a "fake" mode that
2536 enables it to compute how much memory it would need, while actually only
2537 ever using a few hundred bytes of working memory and without too many
2538 tests of the mode. This should make future maintenance and development
2539 easier. A side effect of this work is that the limit of 200 on the nesting
2540 depth of parentheses has been removed (though this was never a serious
2541 limitation, I suspect). However, there is a downside: pcre_compile() now
2542 runs more slowly than before (30% or more, depending on the pattern). I
2543 hope this isn't a big issue. There is no effect on runtime performance.
2544
2545 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2546 newline (only possible for the last line of a file) and it was a
2547 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2548
2549 19. Added additional timing features to pcretest. (1) The -tm option now times
2550 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2551 separate command line item, by a number that specifies the number of
2552 repeats to use when timing. The default is 50000; this gives better
2553 precision, but takes uncomfortably long for very large patterns.
2554
2555 20. Extended pcre_study() to be more clever in cases where a branch of a
2556 subpattern has no definite first character. For example, (a*|b*)[cd] would
2557 previously give no result from pcre_study(). Now it recognizes that the
2558 first character must be a, b, c, or d.
2559
2560 21. There was an incorrect error "recursive call could loop indefinitely" if
2561 a subpattern (or the entire pattern) that was being tested for matching an
2562 empty string contained only one non-empty item after a nested subpattern.
2563 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2564 incorrectly, because the \d was being skipped in the check.
2565
2566 22. The pcretest program now has a new pattern option /B and a command line
2567 option -b, which is equivalent to adding /B to every pattern. This causes
2568 it to show the compiled bytecode, without the additional information that
2569 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2570 is the same as /B/I).
2571
2572 23. A new optimization is now able automatically to treat some sequences such
2573 as a*b as a*+b. More specifically, if something simple (such as a character
2574 or a simple class like \d) has an unlimited quantifier, and is followed by
2575 something that cannot possibly match the quantified thing, the quantifier
2576 is automatically "possessified".
2577
2578 24. A recursive reference to a subpattern whose number was greater than 39
2579 went wrong under certain circumstances in UTF-8 mode. This bug could also
2580 have affected the operation of pcre_study().
2581
2582 25. Realized that a little bit of performance could be had by replacing
2583 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2584
2585 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2586
2587 27. Possessive quantifiers such as a++ were previously implemented by turning
2588 them into atomic groups such as ($>a+). Now they have their own opcodes,
2589 which improves performance. This includes the automatically created ones
2590 from 23 above.
2591
2592 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2593 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2594 the first matched character to be a colon. This applied both to named and
2595 numbered groups.
2596
2597 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2598
2599 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2600 building PCRE on Windows easier, so I have included it in the distribution.
2601
2602 31. There is now a check in pcretest against a ridiculously large number being
2603 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2604 loop, the loop is abandoned.
2605
2606 32. Forward references to subpatterns in conditions such as (?(2)...) where
2607 subpattern 2 is defined later cause pcre_compile() to search forwards in
2608 the pattern for the relevant set of parentheses. This search went wrong
2609 when there were unescaped parentheses in a character class, parentheses
2610 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2611
2612 33. "Subroutine" calls and backreferences were previously restricted to
2613 referencing subpatterns earlier in the regex. This restriction has now
2614 been removed.
2615
2616 34. Added a number of extra features that are going to be in Perl 5.10. On the
2617 whole, these are just syntactic alternatives for features that PCRE had
2618 previously implemented using the Python syntax or my own invention. The
2619 other formats are all retained for compatibility.
2620
2621 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2622 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2623 also .NET compatible.
2624
2625 (b) A recursion or subroutine call to a named group can now be defined as
2626 (?&name) as well as (?P>name).
2627
2628 (c) A backreference to a named group can now be defined as \k<name> or
2629 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2630 5.10, are also .NET compatible.
2631
2632 (d) A conditional reference to a named group can now use the syntax
2633 (?(<name>) or (?('name') as well as (?(name).
2634
2635 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2636 groups (named and numbered) that are never evaluated inline, but can be
2637 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2638 is always false. There may be only one alternative in such a group.
2639
2640 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2641 as the simple (?(R). The condition is true only if the most recent
2642 recursion is that of the given number or name. It does not search out
2643 through the entire recursion stack.
2644
2645 (g) The escape \gN or \g{N} has been added, where N is a positive or
2646 negative number, specifying an absolute or relative reference.
2647
2648 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2649 some "unreachable code" warnings.
2650
2651 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2652 things, this adds five new scripts.
2653
2654 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2655 There were also incompatibilities regarding the handling of \Q..\E inside
2656 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2657 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2658
2659 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2660 matches an empty string, and forcibly breaks the loop. There were bugs in
2661 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2662 against aaaa the result was just "a" rather than "aaaa", for example. Two
2663 separate and independent bugs (that affected different cases) have been
2664 fixed.
2665
2666 39. Refactored the code to abolish the use of different opcodes for small
2667 capturing bracket numbers. This is a tidy that I avoided doing when I
2668 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2669 The new approach is not only tidier, it makes it possible to reduce the
2670 memory needed to fix the previous bug (38).
2671
2672 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2673 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2674 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2675 mode.
2676
2677 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2678 report.
2679
2680 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2681 copy construction and assignment in the C++ wrapper.
2682
2683 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2684 couple of bugs that could have given wrong results in the "--newline=crlf"
2685 case.
2686
2687 44. Added a number of casts and did some reorganization of signed/unsigned int
2688 variables following suggestions from Dair Grant. Also renamed the variable
2689 "this" as "item" because it is a C++ keyword.
2690
2691 45. Arranged for dftables to add
2692
2693 #include "pcre_internal.h"
2694
2695 to pcre_chartables.c because without it, gcc 4.x may remove the array
2696 definition from the final binary if PCRE is built into a static library and
2697 dead code stripping is activated.
2698
2699 46. For an unanchored pattern, if a match attempt fails at the start of a
2700 newline sequence, and the newline setting is CRLF or ANY, and the next two
2701 characters are CRLF, advance by two characters instead of one.
2702
2703
2704 Version 6.7 04-Jul-06
2705 ---------------------
2706
2707 1. In order to handle tests when input lines are enormously long, pcretest has
2708 been re-factored so that it automatically extends its buffers when
2709 necessary. The code is crude, but this _is_ just a test program. The
2710 default size has been increased from 32K to 50K.
2711
2712 2. The code in pcre_study() was using the value of the re argument before
2713 testing it for NULL. (Of course, in any sensible call of the function, it
2714 won't be NULL.)
2715
2716 3. The memmove() emulation function in pcre_internal.h, which is used on
2717 systems that lack both memmove() and bcopy() - that is, hardly ever -
2718 was missing a "static" storage class specifier.
2719
2720 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2721 containing an extended class (one that cannot be represented by a bitmap
2722 because it contains high-valued characters or Unicode property items, e.g.
2723 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2724 pattern, but PCRE should not loop if you do not (it no longer does).
2725 [Detail: two cases were found: (a) a repeated subpattern containing an
2726 extended class; (b) a recursive reference to a subpattern that followed a
2727 previous extended class. It wasn't skipping over the extended class
2728 correctly when UTF-8 mode was not set.]
2729
2730 5. A negated single-character class was not being recognized as fixed-length
2731 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2732 compile error "lookbehind assertion is not fixed length".
2733
2734 6. The RunPerlTest auxiliary script was showing an unexpected difference
2735 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2736 write a Perl script that can interpret lines of an input file either as
2737 byte characters or as UTF-8, which is what "perltest" was being required to
2738 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2739 can't do is switch easily at run time between having the "use utf8;" pragma
2740 or not. In the end, I fudged it by using the RunPerlTest script to insert
2741 "use utf8;" explicitly for the UTF-8 tests.
2742
2743 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2744 the end of the subject string, contrary to the documentation and to what
2745 Perl does. This was true of both matching functions. Now it matches only at
2746 the start of the subject and immediately after *internal* newlines.
2747
2748 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2749 a pointer to an int instead of a pointer to an unsigned long int. This
2750 caused problems on 64-bit systems.
2751
2752 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2753 instance of the 'standard' template library not being so standard".
2754
2755 10. There was no check on the number of named subpatterns nor the maximum
2756 length of a subpattern name. The product of these values is used to compute
2757 the size of the memory block for a compiled pattern. By supplying a very
2758 long subpattern name and a large number of named subpatterns, the size
2759 computation could be caused to overflow. This is now prevented by limiting
2760 the length of names to 32 characters, and the number of named subpatterns
2761 to 10,000.
2762
2763 11. Subpatterns that are repeated with specific counts have to be replicated in
2764 the compiled pattern. The size of memory for this was computed from the
2765 length of the subpattern and the repeat count. The latter is limited to
2766 65535, but there was no limit on the former, meaning that integer overflow
2767 could in principle occur. The compiled length of a repeated subpattern is
2768 now limited to 30,000 bytes in order to prevent this.
2769
2770 12. Added the optional facility to have named substrings with the same name.
2771
2772 13. Added the ability to use a named substring as a condition, using the
2773 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2774 are numbers (not recommended). Forward references are permitted.
2775
2776 14. Added forward references in named backreferences (if you see what I mean).
2777
2778 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2779 pattern could run off the end of the subject. For example, the pattern
2780 "(?s)(.{1,5})"8 did this with the subject "ab".
2781
2782 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2783 PCRE_CASELESS was set when matching characters that were quantified with ?
2784 or *.
2785
2786 17. A character class other than a single negated character that had a minimum
2787 but no maximum quantifier - for example [ab]{6,} - was not handled
2788 correctly by pce_dfa_exec(). It would match only one character.
2789
2790 18. A valid (though odd) pattern that looked like a POSIX character
2791 class but used an invalid character after [ (for example [[,abc,]]) caused
2792 pcre_compile() to give the error "Failed: internal error: code overflow" or
2793 in some cases to crash with a glibc free() error. This could even happen if
2794 the pattern terminated after [[ but there just happened to be a sequence of
2795 letters, a binary zero, and a closing ] in the memory that followed.
2796
2797 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2798 over the years. Originally (before any Unicode support), just the bottom 8
2799 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2800 output from "man perlunicode" includes this:
2801
2802 The regular expression compiler produces polymorphic opcodes. That
2803 is, the pattern adapts to the data and automatically switches to
2804 the Unicode character scheme when presented with Unicode data--or
2805 instead uses a traditional byte scheme when presented with byte
2806 data.
2807
2808 Sadly, a wide octal escape does not cause a switch, and in a string with
2809 no other multibyte characters, these octal escapes are treated as before.
2810 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2811 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2812 Unicode string.
2813
2814 I have not perpetrated such confusion in PCRE. Up till now, it took just
2815 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2816 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2817 translate to the appropriate multibyte character.
2818
2819 29. Applied some refactoring to reduce the number of warnings from Microsoft
2820 and Borland compilers. This has included removing the fudge introduced
2821 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2822 a warning about an unused variable.
2823
2824 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2825 characters since release 4.0, because Perl (from release 5.004) does not.
2826 [Or at least, is documented not to: some releases seem to be in conflict
2827 with the documentation.] However, when a pattern was studied with
2828 pcre_study() and all its branches started with \s, PCRE still included VT
2829 as a possible starting character. Of course, this did no harm; it just
2830 caused an unnecessary match attempt.
2831
2832 22. Removed a now-redundant internal flag bit that recorded the fact that case
2833 dependency changed within the pattern. This was once needed for "required
2834 byte" processing, but is no longer used. This recovers a now-scarce options
2835 bit. Also moved the least significant internal flag bit to the most-
2836 significant bit of the word, which was not previously used (hangover from
2837 the days when it was an int rather than a uint) to free up another bit for
2838 the future.
2839
2840 23. Added support for CRLF line endings as well as CR and LF. As well as the
2841 default being selectable at build time, it can now be changed at runtime
2842 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2843 specify that it is scanning data with non-default line endings.
2844
2845 24. Changed the definition of CXXLINK to make it agree with the definition of
2846 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2847
2848 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2849 recursions. This makes a big different to stack usage for some patterns.
2850
2851 26. If a subpattern containing a named recursion or subroutine reference such
2852 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2853 the space required for the compiled pattern went wrong and gave too small a
2854 value. Depending on the environment, this could lead to "Failed: internal
2855 error: code overflow at offset 49" or "glibc detected double free or
2856 corruption" errors.
2857
2858 27. Applied patches from Google (a) to support the new newline modes and (b) to
2859 advance over multibyte UTF-8 characters in GlobalReplace.
2860
2861 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2862 difference for some implementation of PCRE in some Windows version.
2863
2864 29. Added some extra testing facilities to pcretest:
2865
2866 \q<number> in a data line sets the "match limit" value
2867 \Q<number> in a data line sets the "match recursion limt" value
2868 -S <number> sets the stack size, where <number> is in megabytes
2869
2870 The -S option isn't available for Windows.
2871
2872
2873 Version 6.6 06-Feb-06
2874 ---------------------
2875
2876 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2877 in pcreposix.h. I have copied the definition from pcre.h.
2878
2879 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2880 because pcre.h is no longer a built file.
2881
2882 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2883 not normally included in the compiled code.
2884
2885
2886 Version 6.5 01-Feb-06
2887 ---------------------
2888
2889 1. When using the partial match feature with pcre_dfa_exec(), it was not
2890 anchoring the second and subsequent partial matches at the new starting
2891 point. This could lead to incorrect results. For example, with the pattern
2892 /1234/, partially matching against "123" and then "a4" gave a match.
2893
2894 2. Changes to pcregrep:
2895
2896 (a) All non-match returns from pcre_exec() were being treated as failures
2897 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2898 error message is output. Some extra information is given for the
2899 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2900 probably the only errors that are likely to be caused by users (by
2901 specifying a regex that has nested indefinite repeats, for instance).
2902 If there are more than 20 of these errors, pcregrep is abandoned.
2903
2904 (b) A binary zero was treated as data while matching, but terminated the
2905 output line if it was written out. This has been fixed: binary zeroes
2906 are now no different to any other data bytes.
2907
2908 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2909 used to set a locale for matching. The --locale=xxxx long option has
2910 been added (no short equivalent) to specify a locale explicitly on the
2911 pcregrep command, overriding the environment variables.
2912
2913 (d) When -B was used with -n, some line numbers in the output were one less
2914 than they should have been.
2915
2916 (e) Added the -o (--only-matching) option.
2917
2918 (f) If -A or -C was used with -c (count only), some lines of context were
2919 accidentally printed for the final match.
2920
2921 (g) Added the -H (--with-filename) option.
2922
2923 (h) The combination of options -rh failed to suppress file names for files
2924 that were found from directory arguments.
2925
2926 (i) Added the -D (--devices) and -d (--directories) options.
2927
2928 (j) Added the -F (--fixed-strings) option.
2929
2930 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2931
2932 (l) Added the --colo(u)r option.
2933
2934 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2935 is not present by default.
2936
2937 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2938 items such as (?R) or (?1), when the recursion could match a number of
2939 alternatives. If it matched one of the alternatives, but subsequently,
2940 outside the recursion, there was a failure, the code tried to back up into
2941 the recursion. However, because of the way PCRE is implemented, this is not
2942 possible, and the result was an incorrect result from the match.
2943
2944 In order to prevent this happening, the specification of recursion has
2945 been changed so that all such subpatterns are automatically treated as
2946 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2947
2948 4. I had overlooked the fact that, in some locales, there are characters for
2949 which isalpha() is true but neither isupper() nor islower() are true. In
2950 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2951 and ordfeminine) are like this. This affected the treatment of \w and \W
2952 when they appeared in character classes, but not when they appeared outside
2953 a character class. The bit map for "word" characters is now created
2954 separately from the results of isalnum() instead of just taking it from the
2955 upper, lower, and digit maps. (Plus the underscore character, of course.)
2956
2957 5. The above bug also affected the handling of POSIX character classes such as
2958 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2959 permanent tables. Instead, the bit maps for such a class were previously
2960 created as the appropriate unions of the upper, lower, and digit bitmaps.
2961 Now they are created by subtraction from the [[:word:]] class, which has
2962 its own bitmap.
2963
2964 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2965 It is created by subtracting the vertical space characters (\x09, \x0a,
2966 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2967 subtraction was done in the overall bitmap for a character class, meaning
2968 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2969 be recognized. This bug has been fixed.
2970
2971 7. Patches from the folks at Google:
2972
2973 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2974 real life, but is still worth protecting against".
2975
2976 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2977 regular expressions".
2978
2979 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2980 have it.
2981
2982 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2983 "configure" and the latter not, in order to fix a problem somebody had
2984 with compiling the Arg class on HP-UX.
2985
2986 (e) Improve the error-handling of the C++ wrapper a little bit.
2987
2988 (f) New tests for checking recursion limiting.
2989
2990 8. The pcre_memmove() function, which is used only if the environment does not
2991 have a standard memmove() function (and is therefore rarely compiled),
2992 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2993 returning a result (though PCRE never actually uses the result).
2994
2995 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2996 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2997 returned instead of calling malloc() with an overflowing number that would
2998 most likely cause subsequent chaos.
2999
3000 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3001
3002 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3003 with this option is matched, the nmatch and pmatch options of regexec() are
3004 ignored.
3005
3006 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3007 provided in case anyone wants to the the POSIX interface with UTF-8
3008 strings.
3009
3010 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3011 C++ linking (needed for some HP-UX environments).
3012
3013 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3014 (unused parameter) and in the pcre_printint() function (omitted "default"
3015 switch label when the default is to do nothing).
3016
3017 15. Added some code to make it possible, when PCRE is compiled as a C++
3018 library, to replace subject pointers for pcre_exec() with a smart pointer
3019 class, thus making it possible to process discontinuous strings.
3020
3021 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3022 much the same function. They were added by different people who were trying
3023 to make PCRE easy to compile on non-Unix systems. It has been suggested
3024 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3025 for compiling on Windows systems. I have therefore replaced it with
3026 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3027 defaults to "extern" for C or "extern C" for C++, which works fine on
3028 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3029 SCOPE with something explicit in config.h. In addition:
3030
3031 (a) pcreposix.h still had just "extern" instead of either of these macros;
3032 I have replaced it with PCRE_DATA_SCOPE.
3033
3034 (b) Functions such as _pcre_xclass(), which are internal to the library,
3035 but external in the C sense, all had PCRE_EXPORT in their definitions.
3036 This is apparently wrong for the Windows case, so I have removed it.
3037 (It makes no difference on Unix-like systems.)
3038
3039 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3040 of recursive calls to match(). This is different to MATCH_LIMIT because
3041 that limits the total number of calls to match(), not all of which increase
3042 the depth of recursion. Limiting the recursion depth limits the amount of
3043 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3044 when PCRE is compiled, and changed at run time. A patch from Google adds
3045 this functionality to the C++ interface.
3046
3047 18. Changes to the handling of Unicode character properties:
3048
3049 (a) Updated the table to Unicode 4.1.0.
3050
3051 (b) Recognize characters that are not in the table as "Cn" (undefined).
3052
3053 (c) I revised the way the table is implemented to a much improved format
3054 which includes recognition of ranges. It now supports the ranges that
3055 are defined in UnicodeData.txt, and it also amalgamates other
3056 characters into ranges. This has reduced the number of entries in the
3057 table from around 16,000 to around 3,000, thus reducing its size
3058 considerably. I realized I did not need to use a tree structure after
3059 all - a binary chop search is just as efficient. Having reduced the
3060 number of entries, I extended their size from 6 bytes to 8 bytes to
3061 allow for more data.
3062
3063 (d) Added support for Unicode script names via properties such as \p{Han}.
3064
3065 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3066 matching that character.
3067
3068 20. When matching a repeated Unicode property with a minimum greater than zero,
3069 (for example \pL{2,}), PCRE could look past the end of the subject if it
3070 reached it while seeking the minimum number of characters. This could
3071 happen only if some of the characters were more than one byte long, because
3072 there is a check for at least the minimum number of bytes.
3073
3074 21. Refactored the implementation of \p and \P so as to be more general, to
3075 allow for more different types of property in future. This has changed the
3076 compiled form incompatibly. Anybody with saved compiled patterns that use
3077 \p or \P will have to recompile them.
3078
3079 22. Added "Any" and "L&" to the supported property types.
3080
3081 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3082 but give a compile time error if the value is greater than 0xff.
3083
3084 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3085 accidentally not being installed or uninstalled.
3086
3087 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3088 made were to insert the current release number. This seemed silly, because
3089 it made things harder for people building PCRE on systems that don't run
3090 "configure". I have turned pcre.h into a distributed file, no longer built
3091 by "configure", with the version identification directly included. There is
3092 no longer a pcre.h.in file.
3093
3094 However, this change necessitated a change to the pcre-config script as
3095 well. It is built from pcre-config.in, and one of the substitutions was the
3096 release number. I have updated configure.ac so that ./configure now finds
3097 the release number by grepping pcre.h.
3098
3099 26. Added the ability to run the tests under valgrind.
3100
3101
3102 Version 6.4 05-Sep-05
3103 ---------------------
3104
3105 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3106 "--" to be printed when multiple files were scanned, even when none of the
3107 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3108 consider it to be a bug, and have restored the previous behaviour.
3109
3110 2. A couple of code tidies to get rid of compiler warnings.
3111
3112 3. The pcretest program used to cheat by referring to symbols in the library
3113 whose names begin with _pcre_. These are internal symbols that are not
3114 really supposed to be visible externally, and in some environments it is
3115 possible to suppress them. The cheating is now confined to including
3116 certain files from the library's source, which is a bit cleaner.
3117
3118 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3119 file's purpose clearer.
3120
3121 5. Reorganized pcre_ucp_findchar().
3122
3123
3124 Version 6.3 15-Aug-05
3125 ---------------------
3126
3127 1. The file libpcre.pc.in did not have general read permission in the tarball.
3128
3129 2. There were some problems when building without C++ support:
3130
3131 (a) If C++ support was not built, "make install" and "make test" still
3132 tried to test it.
3133
3134 (b) There were problems when the value of CXX was explicitly set. Some
3135 changes have been made to try to fix these, and ...
3136
3137 (c) --disable-cpp can now be used to explicitly disable C++ support.
3138
3139 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3140 backslash in a target when C++ was disabled. This confuses some
3141 versions of "make", apparently. Using an intermediate variable solves
3142 this. (Same for CPP_LOBJ.)
3143
3144 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3145 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3146 necessary on certain architectures.
3147
3148 4. Added a setting of -export-symbols-regex to the link command to remove
3149 those symbols that are exported in the C sense, but actually are local
3150 within the library, and not documented. Their names all begin with
3151 "_pcre_". This is not a perfect job, because (a) we have to except some
3152 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3153 available (and never for static libraries). I have made a note to try to
3154 find a way round (a) in the future.
3155
3156
3157 Version 6.2 01-Aug-05
3158 ---------------------
3159
3160 1. There was no test for integer overflow of quantifier values. A construction
3161 such as {1111111111111111} would give undefined results. What is worse, if
3162 a minimum quantifier for a parenthesized subpattern overflowed and became
3163 negative, the calculation of the memory size went wrong. This could have
3164 led to memory overwriting.
3165
3166 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3167
3168 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3169 operating environments where this matters.
3170
3171 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3172 PCRE options from within the C++ wrapper.
3173
3174 5. Named capturing subpatterns were not being correctly counted when a pattern
3175 was compiled. This caused two problems: (a) If there were more than 100
3176 such subpatterns, the calculation of the memory needed for the whole
3177 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3178 back references of the form \12, where the number was greater than 9, were
3179 not recognized as back references, even though there were sufficient
3180 previous subpatterns.
3181
3182 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3183 versions of gcc, e.g. 2.95.4.
3184
3185
3186 Version 6.1 21-Jun-05
3187 ---------------------
3188
3189 1. There was one reference to the variable "posix" in pcretest.c that was not
3190 surrounded by "#if !defined NOPOSIX".
3191
3192 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3193 the cross-check on the old pcre_info() function, for the benefit of the
3194 cut-down version of PCRE that is currently imported into Exim.
3195
3196 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3197 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3198 patterns that start (?i) but I don't think that matters. The use of (?i) is
3199 just an example; this all applies to the other options as well.
3200
3201 4. Since libtool seems to echo the compile commands it is issuing, the output
3202 from "make" can be reduced a bit by putting "@" in front of each libtool
3203 compile command.
3204
3205 5. Patch from the folks at Google for configure.in to be a bit more thorough
3206 in checking for a suitable C++ installation before trying to compile the
3207 C++ stuff. This should fix a reported problem when a compiler was present,
3208 but no suitable headers.
3209
3210 6. The man pages all had just "PCRE" as their title. I have changed them to
3211 be the relevant file name. I have also arranged that these names are
3212 retained in the file doc/pcre.txt, which is a concatenation in text format
3213 of all the man pages except the little individual ones for each function.
3214
3215 7. The NON-UNIX-USE file had not been updated for the different set of source
3216 files that come with release 6. I also added a few comments about the C++
3217 wrapper.
3218
3219
3220 Version 6.0 07-Jun-05
3221 ---------------------
3222
3223 1. Some minor internal re-organization to help with my DFA experiments.
3224
3225 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3226 didn't matter for the library itself when fully configured, but did matter
3227 when compiling without UCP support, or within Exim, where the ucp files are
3228 not imported.
3229
3230 3. Refactoring of the library code to split up the various functions into
3231 different source modules. The addition of the new DFA matching code (see
3232 below) to a single monolithic source would have made it really too
3233 unwieldy, quite apart from causing all the code to be include in a
3234 statically linked application, when only some functions are used. This is
3235 relevant even without the DFA addition now that patterns can be compiled in
3236 one application and matched in another.
3237
3238 The downside of splitting up is that there have to be some external
3239 functions and data tables that are used internally in different modules of
3240 the library but which are not part of the API. These have all had their
3241 names changed to start with "_pcre_" so that they are unlikely to clash
3242 with other external names.
3243
3244 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3245 a different (DFA) algorithm. Although it is slower than the original
3246 function, it does have some advantages for certain types of matching
3247 problem.
3248
3249 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3250 including restarting after a partial match.
3251
3252 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3253 defined when compiling for Windows was sent to me. I have put it into the
3254 code, though I have no means of testing or verifying it.
3255
3256 7. Added the pcre_refcount() auxiliary function.
3257
3258 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3259 match before or at the first newline in the subject string. In pcretest,
3260 the /f option on a pattern can be used to set this.
3261
3262 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3263 would behave wrongly. This has been present in PCRE since release 4.0.
3264
3265 10. A number of changes to the pcregrep command:
3266
3267 (a) Refactored how -x works; insert ^(...)$ instead of setting
3268 PCRE_ANCHORED and checking the length, in preparation for adding
3269 something similar for -w.
3270
3271 (b) Added the -w (match as a word) option.
3272
3273 (c) Refactored the way lines are read and buffered so as to have more
3274 than one at a time available.
3275
3276 (d) Implemented a pcregrep test script.
3277
3278 (e) Added the -M (multiline match) option. This allows patterns to match
3279 over several lines of the subject. The buffering ensures that at least
3280 8K, or the rest of the document (whichever is the shorter) is available
3281 for matching (and similarly the previous 8K for lookbehind assertions).
3282
3283 (f) Changed the --help output so that it now says
3284
3285 -w, --word-regex(p)
3286
3287 instead of two lines, one with "regex" and the other with "regexp"
3288 because that confused at least one person since the short forms are the
3289 same. (This required a bit of code, as the output is generated
3290 automatically from a table. It wasn't just a text change.)
3291
3292 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3293 option but starts with a hyphen. Could be a pattern or a path name
3294 starting with a hyphen, for instance.
3295
3296 (h) "-" can be given as a file name to represent stdin.
3297
3298 (i) When file names are being printed, "(standard input)" is used for
3299 the standard input, for compatibility with GNU grep. Previously
3300 "<stdin>" was used.
3301
3302 (j) The option --label=xxx can be used to supply a name to be used for
3303 stdin when file names are being printed. There is no short form.
3304
3305 (k) Re-factored the options decoding logic because we are going to add
3306 two more options that take data. Such options can now be given in four
3307 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3308
3309 (l) Added the -A, -B, and -C options for requesting that lines of context
3310 around matches be printed.
3311
3312 (m) Added the -L option to print the names of files that do not contain
3313 any matching lines, that is, the complement of -l.
3314
3315 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3316 continue to scan other files.
3317
3318 (o) The -s option was incorrectly implemented. For compatibility with other
3319 greps, it now suppresses the error message for a non-existent or non-
3320 accessible file (but not the return code). There is a new option called
3321 -q that suppresses the output of matching lines, which was what -s was
3322 previously doing.
3323
3324 (p) Added --include and --exclude options to specify files for inclusion
3325 and exclusion when recursing.
3326
3327 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3328 Hopefully, it now does.
3329
3330 12. Missing cast in pcre_study().
3331
3332 13. Added an "uninstall" target to the makefile.
3333
3334 14. Replaced "extern" in the function prototypes in Makefile.in with
3335 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3336 world, but is set differently for Windows.
3337
3338 15. Added a second compiling function called pcre_compile2(). The only
3339 difference is that it has an extra argument, which is a pointer to an
3340 integer error code. When there is a compile-time failure, this is set
3341 non-zero, in addition to the error test pointer being set to point to an
3342 error message. The new argument may be NULL if no error number is required
3343 (but then you may as well call pcre_compile(), which is now just a
3344 wrapper). This facility is provided because some applications need a
3345 numeric error indication, but it has also enabled me to tidy up the way
3346 compile-time errors are handled in the POSIX wrapper.
3347
3348 16. Added VPATH=.libs to the makefile; this should help when building with one
3349 prefix path and installing with another. (Or so I'm told by someone who
3350 knows more about this stuff than I do.)
3351
3352 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3353 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3354 match everything, including newlines. This is not POSIX-compatible, but
3355 somebody wanted the feature. From pcretest it can be activated by using
3356 both the P and the s flags.
3357
3358 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3359
3360 19. libpcre.pc was being incorrectly installed as executable.
3361
3362 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3363 it now also looks for '\r' so that it will work unmodified on Windows.
3364
3365 21. Added Google's contributed C++ wrapper to the distribution.
3366
3367 22. Added some untidy missing memory free() calls in pcretest, to keep
3368 Electric Fence happy when testing.
3369
3370
3371
3372 Version 5.0 13-Sep-04
3373 ---------------------
3374
3375 1. Internal change: literal characters are no longer packed up into items
3376 containing multiple characters in a single byte-string. Each character
3377 is now matched using a separate opcode. However, there may be more than one
3378 byte in the character in UTF-8 mode.
3379
3380 2. The pcre_callout_block structure has two new fields: pattern_position and
3381 next_item_length. These contain the offset in the pattern to the next match
3382 item, and its length, respectively.
3383
3384 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3385 insertion of callouts before each pattern item. Added the /C option to
3386 pcretest to make use of this.
3387
3388 4. On the advice of a Windows user, the lines
3389
3390 #if defined(_WIN32) || defined(WIN32)
3391 _setmode( _fileno( stdout ), 0x8000 );
3392 #endif /* defined(_WIN32) || defined(WIN32) */
3393
3394 have been added to the source of pcretest. This apparently does useful
3395 magic in relation to line terminators.
3396
3397 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3398 for the benefit of those environments where the "b" makes a difference.
3399
3400 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3401 to know about it. I have put a hack into configure.in that adds in code
3402 to set GCC=yes if CC=icc. This seems to end up at a point in the
3403 generated configure script that is early enough to affect the setting of
3404 compiler options, which is what is needed, but I have no means of testing
3405 whether it really works. (The user who reported this had patched the
3406 generated configure script, which of course I cannot do.)
3407
3408 LATER: After change 22 below (new libtool files), the configure script
3409 seems to know about icc (and also ecc). Therefore, I have commented out
3410 this hack in configure.in.
3411
3412 7. Added support for pkg-config (2 patches were sent in).
3413
3414 8. Negated POSIX character classes that used a combination of internal tables
3415 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3416 [[:^ascii]]. Typically, they would match almost any characters. The other
3417 POSIX classes were not broken in this way.
3418
3419 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3420 to find the match, as PCRE was deluded into thinking that the match had to
3421 start at the start point or following a newline. The same bug applied to
3422 patterns with negative forward assertions or any backward assertions
3423 preceding ".*" at the start, unless the pattern required a fixed first
3424 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3425
3426 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3427 starting at the last subject character, bytes beyond the end of the subject
3428 string were read.
3429
3430 11. Renamed the variable "class" as "classbits" to make life easier for C++
3431 users. (Previously there was a macro definition, but it apparently wasn't
3432 enough.)
3433
3434 12. Added the new field "tables" to the extra data so that tables can be passed
3435 in at exec time, or the internal tables can be re-selected. This allows
3436 a compiled regex to be saved and re-used at a later time by a different
3437 program that might have everything at different addresses.
3438
3439 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3440 -R library as well as a -L library.
3441
3442 14. The debugging options of pcretest (-d on the command line or D on a
3443 pattern) showed incorrect output for anything following an extended class
3444 that contained multibyte characters and which was followed by a quantifier.
3445
3446 15. Added optional support for general category Unicode character properties
3447 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3448 support. It adds about 90K to the size of the library. The meanings of the
3449 inbuilt class escapes such as \d and \s have NOT been changed.
3450
3451 16. Updated pcredemo.c to include calls to free() to release the memory for the
3452 compiled pattern.
3453
3454 17. The generated file chartables.c was being created in the source directory
3455 instead of in the building directory. This caused the build to fail if the
3456 source directory was different from the building directory, and was
3457 read-only.
3458
3459 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3460 file. No doubt somebody will tell me if they don't make sense... Also added
3461 Dan Mooney's comments about building on OpenVMS.
3462
3463 19. Added support for partial matching via the PCRE_PARTIAL option for
3464 pcre_exec() and the \P data escape in pcretest.
3465
3466 20. Extended pcretest with 3 new pattern features:
3467
3468 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3469 write the compiled pattern to the file whose name is "rest-of-line".
3470 This is a straight binary dump of the data, with the saved pointer to
3471 the character tables forced to be NULL. The study data, if any, is
3472 written too. After writing, pcretest reads a new pattern.
3473
3474 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3475 compiled pattern from the given file. There must not be any
3476 occurrences of "<" in the file name (pretty unlikely); if there are,
3477 pcretest will instead treat the initial "<" as a pattern delimiter.
3478 After reading in the pattern, pcretest goes on to read data lines as
3479 usual.
3480
3481 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3482 and 16-bit fields in a compiled pattern, to simulate a pattern that
3483 was compiled on a host of opposite endianness.
3484
3485 21. The pcre-exec() function can now cope with patterns that were compiled on
3486 hosts of opposite endianness, with this restriction:
3487
3488 As for any compiled expression that is saved and used later, the tables
3489 pointer field cannot be preserved; the extra_data field in the arguments
3490 to pcre_exec() should be used to pass in a tables address if a value
3491 other than the default internal tables were used at compile time.
3492
3493 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3494 now diagnosed as an error. Previously, most of the time, a negative number
3495 would have been treated as zero, but if in addition "ovector" was passed as
3496 NULL, a crash could occur.
3497
3498 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3499 new versions from the libtool 1.5 distribution (the last one is a copy of
3500 a file called libtool.m4). This seems to have fixed the need to patch
3501 "configure" to support Darwin 1.3 (which I used to do). However, I still
3502 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3503 workstation).
3504
3505 24. Changed the PCRE licence to be the more standard "BSD" licence.
3506
3507
3508 Version 4.5 01-Dec-03
3509 ---------------------
3510
3511 1. There has been some re-arrangement of the code for the match() function so
3512 that it can be compiled in a version that does not call itself recursively.
3513 Instead, it keeps those local variables that need separate instances for
3514 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3515 needs to "recurse". Keeping track of where control must go is done by means
3516 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3517 hide most of the details from the main code, and operates only if
3518 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3519 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3520 operating.
3521
3522 To make it easier for callers to provide specially tailored get/free
3523 functions for this usage, two new functions, pcre_stack_malloc, and
3524 pcre_stack_free, are used. They are always called in strict stacking order,
3525 and the size of block requested is always the same.
3526
3527 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3528 PCRE has been compiled to use the stack or the heap for recursion. The
3529 -C option of pcretest uses this to show which version is compiled.
3530
3531 A new data escape \S, is added to pcretest; it causes the amounts of store
3532 obtained and freed by both kinds of malloc/free at match time to be added
3533 to the output.
3534
3535 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3536 what's available on my current Linux desktop machine.
3537
3538 3. When matching a UTF-8 string, the test for a valid string at the start has
3539 been extended. If start_offset is not zero, PCRE now checks that it points
3540 to a byte that is the start of a UTF-8 character. If not, it returns
3541 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3542 this is necessary because there may be backward assertions in the pattern.
3543 When matching the same subject several times, it may save resources to use
3544 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3545
3546 4. The code for checking the validity of UTF-8 strings has been tightened so
3547 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3548 containing "overlong sequences".
3549
3550 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3551 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3552 should have been "&", but it just so happened that all the cases this let
3553 through by mistake were picked up later in the function.
3554
3555 6. I had used a variable called "isblank" - this is a C99 function, causing
3556 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3557
3558 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3559 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3560 so the version is shown; (c) stop "make test" echoing "./RunTest".
3561
3562 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3563
3564 9. The prototype for memmove() for systems that don't have it was using
3565 size_t, but the inclusion of the header that defines size_t was later. I've
3566 moved the #includes for the C headers earlier to avoid this.
3567
3568 10. Added some adjustments to the code to make it easier to compiler on certain
3569 special systems:
3570
3571 (a) Some "const" qualifiers were missing.
3572 (b) Added the macro EXPORT before all exported functions; by default this
3573 is defined to be empty.
3574 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3575 that it reads its output file name as an argument instead of writing
3576 to the standard output and assuming this can be redirected.
3577
3578 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3579 class containing characters with values greater than 255, PCRE compilation
3580 went into a loop.
3581
3582 12. A recursive reference to a subpattern that was within another subpattern
3583 that had a minimum quantifier of zero caused PCRE to crash. For example,
3584 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3585 recursion. If the recursively-called subpattern itself had a zero repeat,
3586 that was OK.
3587
3588 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3589 buffer into which it was copied (for escape processing) was still set at
3590 1024, so long lines caused crashes.
3591
3592 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3593 "internal error: code overflow...". This applied to any character class
3594 that was followed by a possessive quantifier.
3595
3596 15. Modified the Makefile to add libpcre.la as a prerequisite for
3597 libpcreposix.la because I was told this is needed for a parallel build to
3598 work.
3599
3600 16. If a pattern that contained .* following optional items at the start was
3601 studied, the wrong optimizing data was generated, leading to matching
3602 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3603 matching string must start with a or b or c. The correct conclusion for
3604 this pattern is that a match can start with any character.
3605
3606
3607 Version 4.4 13-Aug-03
3608 ---------------------
3609
3610 1. In UTF-8 mode, a character class containing characters with values between
3611 127 and 255 was not handled correctly if the compiled pattern was studied.
3612 In fixing this, I have also improved the studying algorithm for such
3613 classes (slightly).
3614
3615 2. Three internal functions had redundant arguments passed to them. Removal
3616 might give a very teeny performance improvement.
3617
3618 3. Documentation bug: the value of the capture_top field in a callout is *one
3619 more than* the number of the hightest numbered captured substring.
3620
3621 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3622 in incorrectly linking with a previously installed version. They now link
3623 explicitly with libpcre.la.
3624
3625 5. configure.in no longer needs to recognize Cygwin specially.
3626
3627 6. A problem in pcre.in for Windows platforms is fixed.
3628
3629 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3630 pcretest, it used to include the size of the study block as part of its
3631 output. Unfortunately, the structure contains a field that has a different
3632 size on different hardware architectures. This meant that the tests that
3633 showed this size failed. As the block is currently always of a fixed size,
3634 this information isn't actually particularly useful in pcretest output, so
3635 I have just removed it.
3636
3637 8. Three pre-processor statements accidentally did not start in column 1.
3638 Sadly, there are *still* compilers around that complain, even though
3639 standard C has not required this for well over a decade. Sigh.
3640
3641 9. In pcretest, the code for checking callouts passed small integers in the
3642 callout_data field, which is a void * field. However, some picky compilers
3643 complained about the casts involved for this on 64-bit systems. Now
3644 pcretest passes the address of the small integer instead, which should get
3645 rid of the warnings.
3646
3647 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3648 both compile and run time, and gives an error if an invalid UTF-8 sequence
3649 is found. There is a option for disabling this check in cases where the
3650 string is known to be correct and/or the maximum performance is wanted.
3651
3652 11. In response to a bug report, I changed one line in Makefile.in from
3653
3654 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3655 to
3656 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3657
3658 to look similar to other lines, but I have no way of telling whether this
3659 is the right thing to do, as I do not use Windows. No doubt I'll get told
3660 if it's wrong...
3661
3662
3663 Version 4.3 21-May-03
3664 ---------------------
3665
3666 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3667 Makefile.
3668
3669 2. Some refactoring to improve the quality of the code:
3670
3671 (i) The utf8_table... variables are now declared "const".
3672
3673 (ii) The code for \cx, which used the "case flipping" table to upper case
3674 lower case letters, now just substracts 32. This is ASCII-specific,
3675 but the whole concept of \cx is ASCII-specific, so it seems
3676 reasonable.
3677
3678 (iii) PCRE was using its character types table to recognize decimal and
3679 hexadecimal digits in the pattern. This is silly, because it handles
3680 only 0-9, a-f, and A-F, but the character types table is locale-
3681 specific, which means strange things might happen. A private
3682 table is now used for this - though it costs 256 bytes, a table is
3683 much faster than multiple explicit tests. Of course, the standard
3684 character types table is still used for matching digits in subject
3685 strings against \d.
3686
3687 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3688 ending in _t are). So I've renamed it as ESC_tee.
3689
3690 3. The first argument for regexec() in the POSIX wrapper should have been
3691 defined as "const".
3692
3693 4. Changed pcretest to use malloc() for its buffers so that they can be
3694 Electric Fenced for debugging.
3695
3696 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3697 to read one or more bytes before the start of the subject string. Often this
3698 had no effect on PCRE's behaviour, but in some circumstances it could
3699 provoke a segmentation fault.
3700
3701 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3702 to try to read one or more bytes before the start of the subject string.
3703
3704 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3705 UTF-8 support could misbehave in various ways if the subject string
3706 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3707 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3708 back over UTF-8 characters.)
3709
3710
3711 Version 4.2 14-Apr-03
3712 ---------------------
3713
3714 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3715
3716 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3717 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3718 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3719 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3720 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3721 and BUILD_EXEEXT
3722 Note: automatic setting of the BUILD variables is not yet working
3723 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3724 compile-time but not at link-time
3725 [LINK]: use for linking executables only
3726 make different versions for Windows and non-Windows
3727 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3728 libraries
3729 [LINK_FOR_BUILD]: new variable
3730 [OBJEXT]: use throughout
3731 [EXEEXT]: use throughout
3732 <winshared>: new target
3733 <wininstall>: new target
3734 <dftables.o>: use native compiler
3735 <dftables>: use native linker
3736 <install>: handle Windows platform correctly
3737 <clean>: ditto
3738 <check>: ditto
3739 copy DLL to top builddir before testing
3740
3741 As part of these changes, -no-undefined was removed again. This was reported
3742 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3743 in any case.
3744
3745 3. Some tidies to get rid of compiler warnings:
3746
3747 . In the match_data structure, match_limit was an unsigned long int, whereas
3748 match_call_count was an int. I've made them both unsigned long ints.
3749
3750 . In pcretest the fact that a const uschar * doesn't automatically cast to
3751 a void * provoked a warning.
3752
3753 . Turning on some more compiler warnings threw up some "shadow" variables
3754 and a few more missing casts.
3755
3756 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3757 option, a class that contained a single character with a value between 128
3758 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3759
3760 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3761 option, a class that contained several characters, but with at least one
3762 whose value was between 128 and 255 caused PCRE to crash.
3763
3764
3765 Version 4.1 12-Mar-03
3766 ---------------------
3767
3768 1. Compiling with gcc -pedantic found a couple of places where casts were
3769 needed, and a string in dftables.c that was longer than standard compilers are
3770 required to support.
3771
3772 2. Compiling with Sun's compiler found a few more places where the code could
3773 be tidied up in order to avoid warnings.
3774
3775 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3776 first of these names is deprecated in the latest Autoconf in favour of the name
3777 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3778 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3779 analogy I have changed it to CFLAGS_FOR_BUILD.
3780
3781 4. Added -no-undefined to the linking command in the Makefile, because this is
3782 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3783 linking step for the pcreposix library.
3784
3785 5. PCRE was failing to diagnose the case of two named groups with the same
3786 name.
3787
3788 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3789 literal character that is needed in the subject for a match, and scans along to
3790 ensure that it is present before embarking on the full matching process. This
3791 saves time in cases of nested unlimited repeats that are never going to match.
3792 Problem: the scan can take a lot of time if the subject is very long (e.g.
3793 megabytes), thus penalizing straightforward matches. It is now done only if the
3794 amount of subject to be scanned is less than 1000 bytes.
3795
3796 7. A lesser problem with the same optimization is that it was recording the
3797 first character of an anchored pattern as "needed", thus provoking a search
3798 right along the subject, even when the first match of the pattern was going to
3799 fail. The "needed" character is now not set for anchored patterns, unless it
3800 follows something in the pattern that is of non-fixed length. Thus, it still
3801 fulfils its original purpose of finding quick non-matches in cases of nested
3802 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3803
3804
3805 Version 4.0 17-Feb-03
3806 ---------------------
3807
3808 1. If a comment in an extended regex that started immediately after a meta-item
3809 extended to the end of string, PCRE compiled incorrect data. This could lead to
3810 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3811
3812 2. Moved to autoconf 2.53 and libtool 1.4.2.
3813
3814 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3815 the special perltest8 script is no longer needed - all the tests can be run
3816 from a single perltest script.
3817
3818 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3819 by \s. It has now been removed in PCRE. This means it isn't recognized as
3820 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3821 class [:space:] *does* include VT, thereby creating a mess.
3822
3823 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3824 space and tab.
3825
3826 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3827 its new features into the main test script, reducing the number of scripts.
3828
3829 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3830 were backward compatible, and made the (?i) apply to the whole pattern, as if
3831 /i were given. Now it behaves more logically, and applies the option setting
3832 only to what follows. PCRE has been changed to follow suit. However, if it
3833 finds options settings right at the start of the pattern, it extracts them into
3834 the global options, as before. Thus, they show up in the info data.
3835
3836 8. Added support for the \Q...\E escape sequence. Characters in between are
3837 treated as literals. This is slightly different from Perl in that $ and @ are
3838 also handled as literals inside the quotes. In Perl, they will cause variable
3839 interpolation. Note the following examples:
3840
3841 Pattern PCRE matches Perl matches
3842
3843 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3844 \Qabc\$xyz\E abc\$xyz abc\$xyz
3845 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3846
3847 For compatibility with Perl, \Q...\E sequences are recognized inside character
3848 classes as well as outside them.
3849
3850 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3851 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3852 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3853 signed/unsigned warnings.
3854
3855 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3856 option for pcretest, so I've replaced it by a simple function that does just
3857 that job.
3858
3859 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3860 "pcregrep -".
3861
3862 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3863 Java package. This provides some syntactic sugar for simple cases of what my
3864 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3865 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3866 item, you can use this simplified notation. Note that only makes sense with
3867 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3868 greediness, whatever the setting of the PCRE_UNGREEDY option.
3869
3870 13. A change of greediness default within a pattern was not taking effect at
3871 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3872 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3873 was abstracted outside.
3874
3875 14. PCRE now supports the \G assertion. It is true when the current matching
3876 position is at the start point of the match. This differs from \A when the
3877 starting offset is non-zero. Used with the /g option of pcretest (or similar
3878 code), it works in the same way as it does for Perl's /g option. If all
3879 alternatives of a regex begin with \G, the expression is anchored to the start
3880 match position, and the "anchored" flag is set in the compiled expression.
3881
3882 15. Some bugs concerning the handling of certain option changes within patterns
3883 have been fixed. These applied to options other than (?ims). For example,
3884 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3885 been the other way round. Some of this was related to change 7 above.
3886
3887 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3888 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3889 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3890 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3891
3892 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3893 mode. Unlike ".", it always matches newline, whatever the setting of
3894 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3895 assertions. Perl allows it, but it doesn't (in general) work because it can't
3896 calculate the length of the lookbehind. At least, that's the case for Perl
3897 5.8.0 - I've been told they are going to document that it doesn't work in
3898 future.
3899
3900 18. Added an error diagnosis for escapes that PCRE does not support: these are
3901 \L, \l, \N, \P, \p, \U, \u, and \X.
3902
3903 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3904 reading past the end of the pattern in cases such as /[abcd/.
3905
3906 20. PCRE was getting more memory than necessary for patterns with classes that
3907 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3908
3909 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3910 compiling PCRE for use with Virtual Pascal.
3911
3912 22. Small fix to the Makefile to make it work properly if the build is done
3913 outside the source tree.
3914
3915 23. Added a new extension: a condition to go with recursion. If a conditional
3916 subpattern starts with (?(R) the "true" branch is used if recursion has
3917 happened, whereas the "false" branch is used only at the top level.
3918
3919 24. When there was a very long string of literal characters (over 255 bytes
3920 without UTF support, over 250 bytes with UTF support), the computation of how
3921 much memory was required could be incorrect, leading to segfaults or other
3922 strange effects.
3923
3924 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3925 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3926 there was a subsequent back reference to those brackets. This meant that, for
3927 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3928 possible to check for precisely this case. All we can do is abandon the
3929 optimization if .* occurs inside capturing brackets when there are any back
3930 references whatsoever. (See below for a better fix that came later.)
3931
3932 26. The handling of the optimization for finding the first character of a
3933 non-anchored pattern, and for finding a character that is required later in the
3934 match were failing in some cases. This didn't break the matching; it just
3935 failed to optimize when it could. The way this is done has been re-implemented.
3936
3937 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3938
3939 28. Added a new feature that provides some of the functionality that Perl
3940 provides with (?{...}). The facility is termed a "callout". The way it is done
3941 in PCRE is for the caller to provide an optional function, by setting
3942 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3943 global variable. By default it is unset, which disables all calling out. To get
3944 the function called, the regex must include (?C) at appropriate points. This
3945 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3946 This provides a means of identifying different callout points. When PCRE
3947 reaches such a point in the regex, if pcre_callout has been set, the external
3948 function is called. It is provided with data in a structure called
3949 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3950 matching continues; if it returns a non-zero value, the match at the current
3951 point fails. However, backtracking will occur if possible. [This was changed
3952 later and other features added - see item 49 below.]
3953
3954 29. pcretest is upgraded to test the callout functionality. It provides a
3955 callout function that displays information. By default, it shows the start of
3956 the match and the current position in the text. There are some new data escapes
3957 to vary what happens:
3958
3959 \C+ in addition, show current contents of captured substrings
3960 \C- do not supply a callout function
3961 \C!n return 1 when callout number n is reached
3962 \C!n!m return 1 when callout number n is reached for the mth time
3963
3964 30. If pcregrep was called with the -l option and just a single file name, it
3965 output "<stdin>" if a match was found, instead of the file name.
3966
3967 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3968 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3969 pcre_exec(). This saves a malloc/free per call. The default value of
3970 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3971 when configuring.
3972
3973 32. The default maximum size of a compiled pattern is 64K. There have been a
3974 few cases of people hitting this limit. The code now uses macros to handle the
3975 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3976 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3977 configuring. Tests 2 and 5 work only with 2-byte links because they output
3978 debugging information about compiled patterns.
3979
3980 33. Internal code re-arrangements:
3981
3982 (a) Moved the debugging function for printing out a compiled regex into
3983 its own source file (printint.c) and used #include to pull it into
3984 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3985 separate copies.
3986
3987 (b) Defined the list of op-code names for debugging as a macro in
3988 internal.h so that it is next to the definition of the opcodes.
3989
3990 (c) Defined a table of op-code lengths for simpler skipping along compiled
3991 code. This is again a macro in internal.h so that it is next to the
3992 definition of the opcodes.
3993
3994 34. Added support for recursive calls to individual subpatterns, along the
3995 lines of Robin Houston's patch (but implemented somewhat differently).
3996
3997 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3998 allow it to read and process whole directories in Win32. This code was
3999 contributed by Lionel Fourquaux; it has not been tested by me.
4000
4001 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4002 used to name a group. Names consist of alphanumerics and underscores, and must
4003 be unique. Back references use the syntax (?P=name) and recursive calls use
4004 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4005 numbers. The function pcre_fullinfo() can be used after compilation to extract
4006 a name/number map. There are three relevant calls:
4007
4008 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4009 PCRE_INFO_NAMECOUNT yields the number of entries
4010 PCRE_INFO_NAMETABLE yields a pointer to the map.
4011
4012 The map is a vector of fixed-size entries. The size of each entry depends on
4013 the length of the longest name used. The first two bytes of each entry are the
4014 group number, most significant byte first. There follows the corresponding
4015 name, zero terminated. The names are in alphabetical order.
4016
4017 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4018 case instead of 255. Making it the same both with and without UTF-8 support
4019 means that the same test output works with both.
4020
4021 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4022 calling malloc() with a zero argument.
4023
4024 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4025 optimization. I've improved things by keeping a bitmap of backreferences with
4026 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4027 fact referenced, the optimization can be applied. It is unlikely that a
4028 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4029 the match to follow \n) will appear inside brackets with a number greater than
4030 31, but if it does, any back reference > 31 suppresses the optimization.
4031
4032 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4033 of disabling numbered capturing parentheses. Any opening parenthesis that is
4034 not followed by ? behaves as if it were followed by ?: but named parentheses
4035 can still be used for capturing (and they will acquire numbers in the usual
4036 way).
4037
4038 41. Redesigned the return codes from the match() function into yes/no/error so
4039 that errors can be passed back from deep inside the nested calls. A malloc
4040 failure while inside a recursive subpattern call now causes the
4041 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4042
4043 42. It is now possible to set a limit on the number of times the match()
4044 function is called in a call to pcre_exec(). This facility makes it possible to
4045 limit the amount of recursion and backtracking, though not in a directly
4046 obvious way, because the match() function is used in a number of different
4047 circumstances. The count starts from zero for each position in the subject
4048 string (for non-anchored patterns). The default limit is, for compatibility, a
4049 large number, namely 10 000 000. You can change this in two ways:
4050
4051 (a) When configuring PCRE before making, you can use --with-match-limit=n
4052 to set a default value for the compiled library.
4053
4054 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4055 a different value is set. See 45 below.
4056
4057 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4058
4059 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4060 of things that can be changed at compile time. The first argument specifies
4061 what is wanted and the second points to where the information is to be placed.
4062 The current list of available information is:
4063
4064 PCRE_CONFIG_UTF8
4065
4066 The output is an integer that is set to one if UTF-8 support is available;
4067 otherwise it is set to zero.
4068
4069 PCRE_CONFIG_NEWLINE
4070
4071 The output is an integer that it set to the value of the code that is used for
4072 newline. It is either LF (10) or CR (13).
4073
4074 PCRE_CONFIG_LINK_SIZE
4075
4076 The output is an integer that contains the number of bytes used for internal
4077 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4078
4079 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4080
4081 The output is an integer that contains the threshold above which the POSIX
4082 interface uses malloc() for output vectors. See item 31 above.
4083
4084 PCRE_CONFIG_MATCH_LIMIT
4085
4086 The output is an unsigned integer that contains the default limit of the number
4087 of match() calls in a pcre_exec() execution. See 42 above.
4088
4089 44. pcretest has been upgraded by the addition of the -C option. This causes it
4090 to extract all the available output from the new pcre_config() function, and to
4091 output it. The program then exits immediately.
4092
4093 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4094 order to support additional features. One way would have been to define
4095 pcre_exec2() (for example) with extra arguments, but this would not have been
4096 extensible, and would also have required all calls to the original function to
4097 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4098 is used for passing in "extra" data from pcre_study().
4099
4100 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4101 contains the following fields:
4102
4103 flags a bitmap indicating which of the following fields are set
4104 study_data opaque data from pcre_study()
4105 match_limit a way of specifying a limit on match() calls for a specific
4106 call to pcre_exec()
4107 callout_data data for callouts (see 49 below)
4108
4109 The flag bits are also defined in pcre.h, and are
4110
4111 PCRE_EXTRA_STUDY_DATA
4112 PCRE_EXTRA_MATCH_LIMIT
4113 PCRE_EXTRA_CALLOUT_DATA
4114
4115 The pcre_study() function now returns one of these new pcre_extra blocks, with
4116 the actual study data pointed to by the study_data field, and the
4117 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4118 before. That is, this change is entirely upwards-compatible and requires no
4119 change to existing code.
4120
4121 If you want to pass in additional data to pcre_exec(), you can either place it
4122 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4123 block.
4124
4125 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4126 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4127 times with different match limits, until it finds the minimum value needed for
4128 pcre_exec() to complete. The value is then output. This can be instructive; for
4129 most simple matches the number is quite small, but for pathological cases it
4130 gets very large very quickly.
4131
4132 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4133 returns the size of the data block pointed to by the study_data field in a
4134 pcre_extra block, that is, the value that was passed as the argument to
4135 pcre_malloc() when PCRE was getting memory in which to place the information
4136 created by pcre_study(). The fourth argument should point to a size_t variable.
4137 pcretest has been extended so that this information is shown after a successful
4138 pcre_study() call when information about the compiled regex is being displayed.
4139
4140 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4141 because what follows is always an absolute path. (Later: it turns out that this
4142 is more than cosmetic for MinGW, because it doesn't like empty path
4143 components.)
4144
4145 49. Some changes have been made to the callout feature (see 28 above):
4146
4147 (i) A callout function now has three choices for what it returns:
4148
4149 0 => success, carry on matching
4150 > 0 => failure at this point, but backtrack if possible
4151 < 0 => serious error, return this value from pcre_exec()
4152
4153 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4154 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4155 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4156 use by callout functions. It will never be used by PCRE itself.
4157
4158 (ii) The pcre_extra structure (see 45 above) has a void * field called
4159 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4160 pcre_callout_block structure has a field of the same name. The contents of
4161 the field passed in the pcre_extra structure are passed to the callout
4162 function in the corresponding field in the callout block. This makes it
4163 easier to use the same callout-containing regex from multiple threads. For
4164 testing, the pcretest program has a new data escape
4165
4166 \C*n pass the number n (may be negative) as callout_data
4167
4168 If the callout function in pcretest receives a non-zero value as
4169 callout_data, it returns that value.
4170
4171 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4172 there were some redundant $(CFLAGS) in commands that are now specified as
4173 $(LINK), which already includes $(CFLAGS).
4174
4175 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4176 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4177 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4178 one-byte characters throughout. Note that case-insensitive matching applies
4179 only to characters whose values are less than 256. PCRE doesn't support the
4180 notion of cases for higher-valued characters.
4181
4182 (i) A character class whose characters are all within 0-255 is handled as
4183 a bit map, and the map is inverted for negative classes. Previously, a
4184 character > 255 always failed to match such a class; however it should
4185 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4186
4187 (ii) A negated character class with a single character < 255 is coded as
4188 "not this character" (OP_NOT). This wasn't working properly when the test
4189 character was multibyte, either singly or repeated.
4190
4191 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4192 mode, for example: \x{100}{2,3}.
4193
4194 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4195 singly or repeated) now correctly test multibyte characters. However,
4196 PCRE doesn't recognize any characters with values greater than 255 as
4197 digits, spaces, or word characters. Such characters always match \D, \S,
4198 and \W, and never match \d, \s, or \w.
4199
4200 (v) Classes may now contain characters and character ranges with values
4201 greater than 255. For example: [ab\x{100}-\x{400}].
4202
4203 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4204 PCRE in UTF-8 mode.
4205
4206 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4207 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4208 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4209 value.)
4210
4211 53. The single man page has become too large. I have therefore split it up into
4212 a number of separate man pages. These also give rise to individual HTML pages;
4213 these are now put in a separate directory, and there is an index.html page that
4214 lists them all. Some hyperlinking between the pages has been installed.
4215
4216 54. Added convenience functions for handling named capturing parentheses.
4217
4218 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4219 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4220 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4221 are faulted.
4222
4223 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4224 calling configure. These values are used when compiling the dftables.c program
4225 which is run to generate the source of the default character tables. They
4226 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4227 you will need to set these values.
4228
4229 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4230
4231
4232 Version 3.9 02-Jan-02
4233 ---------------------
4234
4235 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4236
4237 2. If --disable-static was given, the building process failed when trying to
4238 build pcretest and pcregrep. (For some reason it was using libtool to compile
4239 them, which is not right, as they aren't part of the library.)
4240
4241
4242 Version 3.8 18-Dec-01
4243 ---------------------
4244
4245 1. The experimental UTF-8 code was completely screwed up. It was packing the
4246 bytes in the wrong order. How dumb can you get?
4247
4248
4249 Version 3.7 29-Oct-01
4250 ---------------------
4251
4252 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4253 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4254 this didn't happen under Solaris 8, where I normally test things.
4255
4256 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4257 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4258 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4259 this fix an hour or so after the initial 3.7 release.)
4260
4261
4262 Version 3.6 23-Oct-01
4263 ---------------------
4264
4265 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4266 offsets passed as NULL with zero offset count.
4267
4268 2. The config.guess and config.sub files had not been updated when I moved to
4269 the latest autoconf.
4270
4271
4272 Version 3.5 15-Aug-01
4273 ---------------------
4274
4275 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4276 had been forgotten.
4277
4278 2. By using declared but undefined structures, we can avoid using "void"
4279 definitions in pcre.h while keeping the internal definitions of the structures
4280 private.
4281
4282 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4283 user point of view, this means that both static and shared libraries are built
4284 by default, but this can be individually controlled. More of the work of
4285 handling this static/shared cases is now inside libtool instead of PCRE's make
4286 file.
4287
4288 4. The pcretest utility is now installed along with pcregrep because it is
4289 useful for users (to test regexs) and by doing this, it automatically gets
4290 relinked by libtool. The documentation has been turned into a man page, so
4291 there are now .1, .txt, and .html versions in /doc.
4292
4293 5. Upgrades to pcregrep:
4294 (i) Added long-form option names like gnu grep.
4295 (ii) Added --help to list all options with an explanatory phrase.
4296 (iii) Added -r, --recursive to recurse into sub-directories.
4297 (iv) Added -f, --file to read patterns from a file.
4298
4299 6. pcre_exec() was referring to its "code" argument before testing that
4300 argument for NULL (and giving an error if it was NULL).
4301
4302 7. Upgraded Makefile.in to allow for compiling in a different directory from
4303 the source directory.
4304
4305 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4306 options bits, the pointer it was passed was to an int instead of to an unsigned
4307 long int. This mattered only on 64-bit systems.
4308
4309 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4310 generated) instead of pcre.in, which it its source. Also made the same change
4311 in several of the .c files.
4312
4313 10. A new release of gcc defines printf() as a macro, which broke pcretest
4314 because it had an ifdef in the middle of a string argument for printf(). Fixed
4315 by using separate calls to printf().
4316
4317 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4318 script, to force use of CR or LF instead of \n in the source. On non-Unix
4319 systems, the value can be set in config.h.
4320
4321 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4322 absolute limit. Changed the text of the error message to make this clear, and
4323 likewise updated the man page.
4324
4325 13. The limit of 99 on the number of capturing subpatterns has been removed.
4326 The new limit is 65535, which I hope will not be a "real" limit.
4327
4328
4329 Version 3.4 22-Aug-00
4330 ---------------------
4331
4332 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4333
4334 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4335
4336
4337 Version 3.3 01-Aug-00
4338 ---------------------
4339
4340 1. If an octal character was given, but the value was greater than \377, it
4341 was not getting masked to the least significant bits, as documented. This could
4342 lead to crashes in some systems.
4343
4344 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4345 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4346
4347 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4348 These just pass their arguments on to (pcre_free)(), but they are provided
4349 because some uses of PCRE bind it to non-C systems that can call its functions,
4350 but cannot call free() or pcre_free() directly.
4351
4352 4. Add "make test" as a synonym for "make check". Corrected some comments in
4353 the Makefile.
4354
4355 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4356 Makefile.
4357
4358 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4359 command called pgrep for grepping around the active processes.
4360
4361 7. Added the beginnings of support for UTF-8 character strings.
4362
4363 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4364 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4365 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4366 out for the ar command.)
4367
4368
4369 Version 3.2 12-May-00
4370 ---------------------
4371
4372 This is purely a bug fixing release.
4373
4374 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4375 of ZA. This was just one example of several cases that could provoke this bug,
4376 which was introduced by change 9 of version 2.00. The code for breaking
4377 infinite loops after an iteration that matches an empty string was't working
4378 correctly.
4379
4380 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4381 when matched against abbab (for example). After matching an empty string, it
4382 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4383 caused it to match further down the string than it should.
4384
4385 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4386 was there because it doesn't seem to be needed, and it causes trouble on some
4387 systems, as it is not a Standard C header. It has been removed.
4388
4389 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4390 were reported on the Macintosh. The changes were from
4391
4392 while ((c = *(++ptr)) != 0 && c != '\n');
4393 to
4394 while ((c = *(++ptr)) != 0 && c != '\n') ;
4395
4396 Totally extraordinary, but if that's what it takes...
4397
4398 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4399 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4400 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4401 assumes the way PCRE uses memmove() (always moving upwards).
4402
4403 6. PCRE is being used in one environment where strchr() is not available. There
4404 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4405 faster code anyway.
4406
4407
4408 Version 3.1 09-Feb-00
4409 ---------------------
4410
4411 The only change in this release is the fixing of some bugs in Makefile.in for
4412 the "install" target:
4413
4414 (1) It was failing to install pcreposix.h.
4415
4416 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4417
4418
4419 Version 3.0 01-Feb-00
4420 ---------------------
4421
4422 1. Add support for the /+ modifier to perltest (to output $` like it does in
4423 pcretest).
4424
4425 2. Add support for the /g modifier to perltest.
4426
4427 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4428 matches null strings.
4429
4430 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4431 pattern. Perl treats empty patterns specially - it reuses the most recent
4432 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4433 effect.
4434
4435 5. The POSIX interface was broken in that it was just handing over the POSIX
4436 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4437 required a bigger vector, with some working space on the end. This means that
4438 the POSIX wrapper now has to get and free some memory, and copy the results.
4439
4440 6. Added some simple autoconf support, placing the test data and the
4441 documentation in separate directories, re-organizing some of the
4442 information files, and making it build pcre-config (a GNU standard). Also added
4443 libtool support for building PCRE as a shared library, which is now the
4444 default.
4445
4446 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4447 09 are not valid octal constants. Single digits will be used for minor values
4448 less than 10.
4449
4450 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4451 existing programs that set these in the POSIX interface can use PCRE without
4452 modification.
4453
4454 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4455 return all that pcre_info() returns, plus additional data. The pcre_info()
4456 function is retained for compatibility, but is considered to be obsolete.
4457
4458 10. Added experimental recursion feature (?R) to handle one common case that
4459 Perl 5.6 will be able to do with (?p{...}).
4460
4461 11. Added support for POSIX character classes like [:alpha:], which Perl is
4462 adopting.
4463
4464
4465 Version 2.08 31-Aug-99
4466 ----------------------
4467
4468 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4469 trying to match at the startoffset position, but instead was moving forward to
4470 the next newline as if a previous match had failed.
4471
4472 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4473 and could get into a loop if a null string was matched other than at the start
4474 of the subject.
4475
4476 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4477 be distinguished at compile time, and for completeness also added PCRE_DATE.
4478
4479 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4480 in GnuWin32 environments.
4481
4482
4483 Version 2.07 29-Jul-99
4484 ----------------------
4485
4486 1. The documentation is now supplied in plain text form and HTML as well as in
4487 the form of man page sources.
4488
4489 2. C++ compilers don't like assigning (void *) values to other pointer types.
4490 In particular this affects malloc(). Although there is no problem in Standard
4491 C, I've put in casts to keep C++ compilers happy.
4492
4493 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4494 should be (const char *).
4495
4496 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4497 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4498 However, I haven't made this a standard facility. The documentation doesn't
4499 mention it, and the Makefile doesn't support it.
4500
4501 5. The Makefile now contains an "install" target, with editable destinations at
4502 the top of the file. The pcretest program is not installed.
4503
4504 6. pgrep -V now gives the PCRE version number and date.
4505
4506 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4507 causing the entire string to be ignored, instead of just the last character.
4508
4509 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4510 non-matching string, it can take a very, very long time, even for strings of
4511 quite modest length, because of the nested recursion. PCRE now does better in
4512 some of these cases. It does this by remembering the last required literal
4513 character in the pattern, and pre-searching the subject to ensure it is present
4514 before running the real match. In other words, it applies a heuristic to detect
4515 some types of certain failure quickly, and in the above example, if presented
4516 with a string that has no trailing " it gives "no match" very quickly.
4517
4518 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4519 other alternatives are tried instead.
4520
4521
4522 Version 2.06 09-Jun-99
4523 ----------------------
4524
4525 1. Change pcretest's output for amount of store used to show just the code
4526 space, because the remainder (the data block) varies in size between 32-bit and
4527 64-bit systems.
4528
4529 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4530 start matching at. This allows lookbehinds to work when searching for multiple
4531 occurrences in a string.
4532
4533 3. Added additional options to pcretest for testing multiple occurrences:
4534
4535 /+ outputs the rest of the string that follows a match
4536 /g loops for multiple occurrences, using the new startoffset argument
4537 /G loops for multiple occurrences by passing an incremented pointer
4538
4539 4. PCRE wasn't doing the "first character" optimization for patterns starting
4540 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4541 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4542 the letter 'x'. On long subject strings, this gives a significant speed-up.
4543
4544
4545 Version 2.05 21-Apr-99
4546 ----------------------
4547
4548 1. Changed the type of magic_number from int to long int so that it works
4549 properly on 16-bit systems.
4550
4551 2. Fixed a bug which caused patterns starting with .* not to work correctly
4552 when the subject string contained newline characters. PCRE was assuming
4553 anchoring for such patterns in all cases, which is not correct because .* will
4554 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4555 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4556 must be retried after every newline in the subject.
4557
4558
4559 Version 2.04 18-Feb-99
4560 ----------------------
4561
4562 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4563 computation of the store needed to hold the pattern was incorrect (too large).
4564 If such patterns were nested a few deep, this could multiply and become a real
4565 problem.
4566
4567 2. Added /M option to pcretest to show the memory requirement of a specific
4568 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4569
4570 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4571 compiled in such a way that the backtracking after subsequent failure was
4572 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4573 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4574
4575
4576 Version 2.03 02-Feb-99
4577 ----------------------
4578
4579 1. Fixed typo and small mistake in man page.
4580
4581 2. Added 4th condition (GPL supersedes if conflict) and created separate
4582 LICENCE file containing the conditions.
4583
4584 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4585 Perl, that is the internal \ allows the delimiter to be included in the
4586 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4587 the final delimiter, add \ to the end of the pattern (to test the error).
4588
4589 4. Added the convenience functions for extracting substrings after a successful
4590 match. Updated pcretest to make it able to test these functions.
4591
4592
4593 Version 2.02 14-Jan-99
4594 ----------------------
4595
4596 1. Initialized the working variables associated with each extraction so that
4597 their saving and restoring doesn't refer to uninitialized store.
4598
4599 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4600 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4601 fix the problem.
4602
4603 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4604 calls, and wasn't printing the correct value for compiling calls. Increased the
4605 default value of LOOPREPEAT, and the number of significant figures in the
4606 times.
4607
4608 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4609
4610 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4611 a building problem on Windows NT with a FAT file system.
4612
4613
4614 Version 2.01 21-Oct-98
4615 ----------------------
4616
4617 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4618 to character tables built by pcre_maketables() in the current locale. If NULL
4619 is passed, the default tables are used.
4620
4621
4622 Version 2.00 24-Sep-98
4623 ----------------------
4624
4625 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4626 it any more.
4627
4628 2. Allow quantification of (?>) groups, and make it work correctly.
4629
4630 3. The first character computation wasn't working for (?>) groups.
4631
4632 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4633 end of the subject) and add 5.005's \z, which really does match only at the
4634 very end of the subject.
4635
4636 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4637
4638 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4639 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4640 localized options. All options to pcre_study() were also removed.
4641
4642 7. Add other new features from 5.005:
4643
4644 $(?<= positive lookbehind
4645 $(?<! negative lookbehind
4646 (?imsx-imsx) added the unsetting capability
4647 such a setting is global if at outer level; local otherwise
4648 (?imsx-imsx:) non-capturing groups with option setting
4649 (?(cond)re|re) conditional pattern matching
4650
4651 A backreference to itself in a repeated group matches the previous
4652 captured string.
4653
4654 8. General tidying up of studying (both automatic and via "study")
4655 consequential on the addition of new assertions.
4656
4657 9. As in 5.005, unlimited repeated groups that could match an empty substring
4658 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4659 runtime if any iteration does actually match an empty substring.
4660
4661 10. Include the RunTest script in the distribution.
4662
4663 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4664 discrepancies, some of which were old and were also with respect to 5.004. They
4665 have now been fixed.
4666
4667
4668 Version 1.09 28-Apr-98
4669 ----------------------
4670
4671 1. A negated single character class followed by a quantifier with a minimum
4672 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4673 program crashes, or just wrong answers. This did not apply to negated classes
4674 containing more than one character, or to minima other than one.
4675
4676
4677 Version 1.08 27-Mar-98
4678 ----------------------
4679
4680 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4681
4682 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4683 latter must appear before anything that relies on it in the pattern.
4684
4685
4686 Version 1.07 16-Feb-98
4687 ----------------------
4688
4689 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4690 repeat of a potentially empty string).
4691
4692
4693 Version 1.06 23-Jan-98
4694 ----------------------
4695
4696 1. Added Markus Oberhumer's little patches for C++.
4697
4698 2. Literal strings longer than 255 characters were broken.
4699
4700
4701 Version 1.05 23-Dec-97
4702 ----------------------
4703
4704 1. Negated character classes containing more than one character were failing if
4705 PCRE_CASELESS was set at run time.
4706
4707
4708 Version 1.04 19-Dec-97
4709 ----------------------
4710
4711 1. Corrected the man page, where some "const" qualifiers had been omitted.
4712
4713 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4714 input syntax.
4715
4716 3. Fixed memory leak which occurred when a regex with back references was
4717 matched with an offsets vector that wasn't big enough. The temporary memory
4718 that is used in this case wasn't being freed if the match failed.
4719
4720 4. Tidied pcretest to ensure it frees memory that it gets.
4721
4722 5. Temporary memory was being obtained in the case where the passed offsets
4723 vector was exactly big enough.
4724
4725 6. Corrected definition of offsetof() from change 5 below.
4726
4727 7. I had screwed up change 6 below and broken the rules for the use of
4728 setjmp(). Now fixed.
4729
4730
4731 Version 1.03 18-Dec-97
4732 ----------------------
4733
4734 1. A erroneous regex with a missing opening parenthesis was correctly
4735 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4736 on some systems.
4737
4738 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4739 it was reported that one broken compiler failed on the former because "code" is
4740 also an independent variable.
4741
4742 3. The erroneous regex a[]b caused an array overrun reference.
4743
4744 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4745 fail on data ending with that character. (It was going on too far, and checking
4746 the next character, typically a binary zero.) This was specific to the
4747 optimized code for single-character negative classes.
4748
4749 5. Added a contributed patch from the TIN world which does the following:
4750
4751 + Add an undef for memmove, in case the the system defines a macro for it.
4752
4753 + Add a definition of offsetof(), in case there isn't one. (I don't know
4754 the reason behind this - offsetof() is part of the ANSI standard - but
4755 it does no harm).
4756
4757 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4758 most of the places where whitespace preceded '#'. I have given up and
4759 allowed the remaining 2 cases to be at the margin.
4760
4761 + Rename some variables in pcre to eliminate shadowing. This seems very
4762 pedantic, but does no harm, of course.
4763
4764 6. Moved the call to setjmp() into its own function, to get rid of warnings
4765 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4766
4767 7. Constructs such as \d{8,} were compiling into the equivalent of
4768 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4769 outcome, but in this particular case used more store than had been allocated,
4770 which caused the bug to be discovered because it threw up an internal error.
4771
4772 8. The debugging code in both pcre and pcretest for outputting the compiled
4773 form of a regex was going wrong in the case of back references followed by
4774 curly-bracketed repeats.
4775
4776
4777 Version 1.02 12-Dec-97
4778 ----------------------
4779
4780 1. Typos in pcre.3 and comments in the source fixed.
4781
4782 2. Applied a contributed patch to get rid of places where it used to remove
4783 'const' from variables, and fixed some signed/unsigned and uninitialized
4784 variable warnings.
4785
4786 3. Added the "runtest" target to Makefile.
4787
4788 4. Set default compiler flag to -O2 rather than just -O.
4789
4790
4791 Version 1.01 19-Nov-97
4792 ----------------------
4793
4794 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4795 like /([ab]*)*/, that is, for classes with more than one character in them.
4796
4797 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4798 as /((?>a*))*/ (a PCRE_EXTRA facility).
4799
4800
4801 Version 1.00 18-Nov-97
4802 ----------------------
4803
4804 1. Added compile-time macros to support systems such as SunOS4 which don't have
4805 memmove() or strerror() but have other things that can be used instead.
4806
4807 2. Arranged that "make clean" removes the executables.
4808
4809
4810 Version 0.99 27-Oct-97
4811 ----------------------
4812
4813 1. Fixed bug in code for optimizing classes with only one character. It was
4814 initializing a 32-byte map regardless, which could cause it to run off the end
4815 of the memory it had got.
4816
4817 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4818
4819
4820 Version 0.98 22-Oct-97
4821 ----------------------
4822
4823 1. Fixed bug in code for handling temporary memory usage when there are more
4824 back references than supplied space in the ovector. This could cause segfaults.
4825
4826
4827 Version 0.97 21-Oct-97
4828 ----------------------
4829
4830 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4831
4832 2. Optimized negated single characters not to use a bit map.
4833
4834 3. Brought error texts together as macro definitions; clarified some of them;
4835 fixed one that was wrong - it said "range out of order" when it meant "invalid
4836 escape sequence".
4837
4838 4. Changed some char * arguments to const char *.
4839
4840 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4841
4842 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4843 pcretest.
4844
4845
4846 Version 0.96 16-Oct-97
4847 ----------------------
4848
4849 1. Added a simple "pgrep" utility to the distribution.
4850
4851 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4852 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4853 where "ddd" means "one or more decimal digits".
4854
4855 3. Fixed serious bug. If a pattern had a back reference, but the call to
4856 pcre_exec() didn't supply a large enough ovector to record the related
4857 identifying subpattern, the match always failed. PCRE now remembers the number
4858 of the largest back reference, and gets some temporary memory in which to save
4859 the offsets during matching if necessary, in order to ensure that
4860 backreferences always work.
4861
4862 4. Increased the compatibility with Perl in a number of ways:
4863
4864 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4865 to request this handling. The option can be set at compile or exec time.
4866
4867 (b) $ matches before a terminating newline by default; an option
4868 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4869 mode). The option can be set at compile or exec time.
4870
4871 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4872 the same as Perl's. If the decimal number it represents is less than 10
4873 or there aren't that many previous left capturing parentheses, an octal
4874 escape is read. Inside a character class, it's always an octal escape,
4875 even if it is a single digit.
4876
4877 (d) An escaped but undefined alphabetic character is taken as a literal,
4878 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4879 escapes.
4880
4881 (e) {0} is now permitted. (The previous item is removed from the compiled
4882 pattern).
4883
4884 5. Changed all the names of code files so that the basic parts are no longer
4885 than 10 characters, and abolished the teeny "globals.c" file.
4886
4887 6. Changed the handling of character classes; they are now done with a 32-byte
4888 bit map always.
4889
4890 7. Added the -d and /D options to pcretest to make it possible to look at the
4891 internals of compilation without having to recompile pcre.
4892
4893
4894 Version 0.95 23-Sep-97
4895 ----------------------
4896
4897 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4898 \x20 at the start of a run of normal characters. These were being treated as
4899 real characters, instead of the source characters being re-checked.
4900
4901
4902 Version 0.94 18-Sep-97
4903 ----------------------
4904
4905 1. The functions are now thread-safe, with the caveat that the global variables
4906 containing pointers to malloc() and free() or alternative functions are the
4907 same for all threads.
4908
4909 2. Get pcre_study() to generate a bitmap of initial characters for non-
4910 anchored patterns when this is possible, and use it if passed to pcre_exec().
4911
4912
4913 Version 0.93 15-Sep-97
4914 ----------------------
4915
4916 1. /(b)|(:+)/ was computing an incorrect first character.
4917
4918 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4919 but not actually doing anything yet.
4920
4921 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4922 as Perl does (e.g. [-az] or [az-]).
4923
4924 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4925 all possible positions.
4926
4927 5. Split up into different modules to avoid including unneeded functions in a
4928 compiled binary. However, compile and exec are still in one module. The "study"
4929 function is split off.
4930
4931 6. The character tables are now in a separate module whose source is generated
4932 by an auxiliary program - but can then be edited by hand if required. There are
4933 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4934 toupper() in the code.
4935
4936 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4937 make them global. Abolish the function for setting them, as the caller can now
4938 set them directly.
4939
4940
4941 Version 0.92 11-Sep-97
4942 ----------------------
4943
4944 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4945 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4946
4947 2. Caseless matching was not working in character classes if the characters in
4948 the pattern were in upper case.
4949
4950 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4951
4952 4. Make PCRE_ANCHORED public and accept as a compile option.
4953
4954 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4955 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4956 pass them.
4957
4958 6. Give an error if bad option bits passed at compile or run time.
4959
4960 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4961 pcretest to cause it to pass that flag.
4962
4963 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4964 options, and the first character, if set.
4965
4966 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4967
4968
4969 Version 0.91 10-Sep-97
4970 ----------------------
4971
4972 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4973 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4974
4975 2. PCRE was looping on encountering an indefinitely repeated back reference to
4976 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4977 Perl does - treats the match as successful.
4978
4979 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5