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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1242 - (show annotations)
Sat Jan 26 17:51:43 2013 UTC (6 years, 8 months ago) by zherczeg
File size: 241256 byte(s)
Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37
38 Version 8.32 30-November-2012
39 -----------------------------
40
41 1. Improved JIT compiler optimizations for first character search and single
42 character iterators.
43
44 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
45 Patch by Daniel Richard G.
46
47 3. Single character iterator optimizations in the JIT compiler.
48
49 4. Improved JIT compiler optimizations for character ranges.
50
51 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
52 Reported by Giuseppe D'Angelo.
53
54 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
55 of a line, was being set incorrectly in cases where .* appeared inside
56 atomic brackets at the start of a pattern, or where there was a subsequent
57 *PRUNE or *SKIP.
58
59 7. Improved instruction cache flush for POWER/PowerPC.
60 Patch by Daniel Richard G.
61
62 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
63 grep:
64
65 (a) There is now no limit to the number of patterns to be matched.
66
67 (b) An error is given if a pattern is too long.
68
69 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
70 are now supported.
71
72 (d) --exclude-from and --include-from (multiple use) have been added.
73
74 (e) Exclusions and inclusions now apply to all files and directories, not
75 just to those obtained from scanning a directory recursively.
76
77 (f) Multiple uses of -f and --file-list are now supported.
78
79 (g) In a Windows environment, the default for -d has been changed from
80 "read" (the GNU grep default) to "skip", because otherwise the presence
81 of a directory in the file list provokes an error.
82
83 (h) The documentation has been revised and clarified in places.
84
85 9. Improve the matching speed of capturing brackets.
86
87 10. Changed the meaning of \X so that it now matches a Unicode extended
88 grapheme cluster.
89
90 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
91 out POSIX threads when JIT support is configured.
92
93 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
94
95 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
96 structure could go wrong in environments where size_t is not the same size
97 as int.
98
99 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
100 set.
101
102 15. The EBCDIC support had decayed; later updates to the code had included
103 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
104 general tidy up of EBCDIC-related issues, and the documentation was also
105 not quite right. There is now a test that can be run on ASCII systems to
106 check some of the EBCDIC-related things (but is it not a full test).
107
108 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
109 in a small tidy to the code.
110
111 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
112
113 18. If the --only-matching (-o) option in pcregrep is specified multiple
114 times, each one causes appropriate output. For example, -o1 -o2 outputs the
115 substrings matched by the 1st and 2nd capturing parentheses. A separating
116 string can be specified by --om-separator (default empty).
117
118 19. Improving the first n character searches.
119
120 20. Turn case lists for horizontal and vertical white space into macros so that
121 they are defined only once.
122
123 21. This set of changes together give more compatible Unicode case-folding
124 behaviour for characters that have more than one other case when UCP
125 support is available.
126
127 (a) The Unicode property table now has offsets into a new table of sets of
128 three or more characters that are case-equivalent. The MultiStage2.py
129 script that generates these tables (the pcre_ucd.c file) now scans
130 CaseFolding.txt instead of UnicodeData.txt for character case
131 information.
132
133 (b) The code for adding characters or ranges of characters to a character
134 class has been abstracted into a generalized function that also handles
135 case-independence. In UTF-mode with UCP support, this uses the new data
136 to handle characters with more than one other case.
137
138 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
139 whose other case is greater than 256 are now correctly matched
140 caselessly. Previously, the high codepoint matched the low one, but not
141 vice versa.
142
143 (d) The processing of \h, \H, \v, and \ in character classes now makes use
144 of the new class addition function, using character lists defined as
145 macros alongside the case definitions of 20 above.
146
147 (e) Caseless back references now work with characters that have more than
148 one other case.
149
150 (f) General caseless matching of characters with more than one other case
151 is supported.
152
153 22. Unicode character properties were updated from Unicode 6.2.0
154
155 23. Improved CMake support under Windows. Patch by Daniel Richard G.
156
157 24. Add support for 32-bit character strings, and UTF-32
158
159 25. Major JIT compiler update (code refactoring and bugfixing).
160 Experimental Sparc 32 support is added.
161
162 26. Applied a modified version of Daniel Richard G's patch to create
163 pcre.h.generic and config.h.generic by "make" instead of in the
164 PrepareRelease script.
165
166 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
167 pcre_compile.c when checking for a zero character.
168
169 28. Introducing a native interface for JIT. Through this interface, the compiled
170 machine code can be directly executed. The purpose of this interface is to
171 provide fast pattern matching, so several sanity checks are not performed.
172 However, feature tests are still performed. The new interface provides
173 1.4x speedup compared to the old one.
174
175 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
176 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
177 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
178
179 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
180 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
181 the "old" RFC 2279). Instead, it ended up passing a negative length to
182 pcre_exec().
183
184 31. Add support for GCC's visibility feature to hide internal functions.
185
186 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
187 "unknown -C option" after outputting 0 or 1.
188
189 33. There is now support for generating a code coverage report for the test
190 suite in environments where gcc is the compiler and lcov is installed. This
191 is mainly for the benefit of the developers.
192
193 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
194 unaddressable using valgrind annotations, allowing valgrind to detect
195 invalid memory accesses. This is mainly for the benefit of the developers.
196
197 25. (*UTF) can now be used to start a pattern in any of the three libraries.
198
199 26. Give configure error if --enable-cpp but no C++ compiler found.
200
201
202 Version 8.31 06-July-2012
203 -------------------------
204
205 1. Fixing a wrong JIT test case and some compiler warnings.
206
207 2. Removed a bashism from the RunTest script.
208
209 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
210 to unsigned type, result still unsigned" that was given by an MS compiler
211 on encountering the code "-sizeof(xxx)".
212
213 4. Partial matching support is added to the JIT compiler.
214
215 5. Fixed several bugs concerned with partial matching of items that consist
216 of more than one character:
217
218 (a) /^(..)\1/ did not partially match "aba" because checking references was
219 done on an "all or nothing" basis. This also applied to repeated
220 references.
221
222 (b) \R did not give a hard partial match if \r was found at the end of the
223 subject.
224
225 (c) \X did not give a hard partial match after matching one or more
226 characters at the end of the subject.
227
228 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
229 a partial match for the string "\r".
230
231 (e) When newline was set to CRLF, the metacharacter "." did not recognize
232 a partial match for a CR character at the end of the subject string.
233
234 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
235 running pcretest, the text "(JIT)" added to the output whenever JIT is
236 actually used to run the match.
237
238 7. Individual JIT compile options can be set in pcretest by following -s+[+]
239 or /S+[+] with a digit between 1 and 7.
240
241 8. OP_NOT now supports any UTF character not just single-byte ones.
242
243 9. (*MARK) control verb is now supported by the JIT compiler.
244
245 10. The command "./RunTest list" lists the available tests without actually
246 running any of them. (Because I keep forgetting what they all are.)
247
248 11. Add PCRE_INFO_MAXLOOKBEHIND.
249
250 12. Applied a (slightly modified) user-supplied patch that improves performance
251 when the heap is used for recursion (compiled with --disable-stack-for-
252 recursion). Instead of malloc and free for each heap frame each time a
253 logical recursion happens, frames are retained on a chain and re-used where
254 possible. This sometimes gives as much as 30% improvement.
255
256 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
257 call.
258
259 14. As documented, (*COMMIT) is now confined to within a positive assertion.
260
261 15. It is now possible to link pcretest with libedit as an alternative to
262 libreadline.
263
264 16. (*COMMIT) control verb is now supported by the JIT compiler.
265
266 17. The Unicode data tables have been updated to Unicode 6.1.0.
267
268 18. Added --file-list option to pcregrep.
269
270 19. Added binary file support to pcregrep, including the -a, --binary-files,
271 -I, and --text options.
272
273 20. The madvise function is renamed for posix_madvise for QNX compatibility
274 reasons. Fixed by Giuseppe D'Angelo.
275
276 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
277
278 22. Changed the option for creating symbolic links for 16-bit man pages from
279 -s to -sf so that re-installing does not cause issues.
280
281 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
282
283 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
284 matches in certain environments (the workspace was not being correctly
285 retained). Also added to pcre_dfa_exec() a simple plausibility check on
286 some of the workspace data at the beginning of a restart.
287
288 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
289 was not doing so when it should - probably a typo introduced by SVN 528
290 (change 8.10/14).
291
292 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
293 \w+ when the character tables indicated that \x{c4} was a word character.
294 There were several related cases, all because the tests for doing a table
295 lookup were testing for characters less than 127 instead of 255.
296
297 27. If a pattern contains capturing parentheses that are not used in a match,
298 their slots in the ovector are set to -1. For those that are higher than
299 any matched groups, this happens at the end of processing. In the case when
300 there were back references that the ovector was too small to contain
301 (causing temporary malloc'd memory to be used during matching), and the
302 highest capturing number was not used, memory off the end of the ovector
303 was incorrectly being set to -1. (It was using the size of the temporary
304 memory instead of the true size.)
305
306 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
307 ovector size, it uses memory at the end of the block that it has got.
308
309 29. Check for an overlong MARK name and give an error at compile time. The
310 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
311
312 30. JIT compiler update.
313
314 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
315 Rill for the patch.
316
317 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
318
319 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
320
321 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
322 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
323 was enabled.
324
325 35. Improve JIT code generation for greedy plus quantifier.
326
327 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
328 1 to "aa" instead of to an empty string. The bug affected repeated groups
329 that could potentially match an empty string.
330
331 37. Optimizing single character iterators in JIT.
332
333 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
334 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
335 codepoints that are too big for the mode are faulted, and in a UTF mode,
336 disallowed codepoints are also faulted.
337
338 39. If PCRE was compiled with UTF support, in three places in the DFA
339 matcher there was code that should only have been obeyed in UTF mode, but
340 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
341 processing when bytes with values greater than 127 were present. In 16-bit
342 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
343 both cases the values are those that cannot be the first data item in a UTF
344 character. The three items that might have provoked this were recursions,
345 possessively repeated groups, and atomic groups.
346
347 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
348 and pcregrep, because some OS require shared objects to be explicitly
349 passed to ld, causing the link step to fail if they are not.
350
351 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
352 mode, patterns that started with \h* or \R* might be incorrectly matched.
353
354
355 Version 8.30 04-February-2012
356 -----------------------------
357
358 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
359 name is defined in ctype.h.
360
361 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
362 only in quite long subpatterns.
363
364 3. Removed the function pcre_info(), which has been obsolete and deprecated
365 since it was replaced by pcre_fullinfo() in February 2000.
366
367 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
368 match a (*MARK), and the match failed at the start of the subject, a
369 reference to memory before the start of the subject could occur. This bug
370 was introduced by fix 17 of release 8.21.
371
372 5. A reference to an unset group with zero minimum repetition was giving
373 totally wrong answers (in non-JavaScript-compatibility mode). For example,
374 /(another)?(\1?)test/ matched against "hello world test". This bug was
375 introduced in release 8.13.
376
377 6. Add support for 16-bit character strings (a large amount of work involving
378 many changes and refactorings).
379
380 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
381 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
382 from a file.
383
384 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
385 rounding is not applied in this particular case).
386
387 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
388 if they appear, or are escaped, in patterns.
389
390 10. Get rid of a number of -Wunused-but-set-variable warnings.
391
392 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
393 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
394 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
395 also returns the mark "x". This bug applied to capturing parentheses,
396 non-capturing parentheses, and atomic parentheses. It also applied to some
397 assertions.
398
399 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
400 information out of configure.ac instead of relying on pcre.h.generic, which
401 is not stored in the repository.
402
403 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
404 -lreadline.
405
406 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
407
408 15. Applied Graycode's patch to put the top-level frame on the stack rather
409 than the heap when not using the stack for recursion. This gives a
410 performance improvement in many cases when recursion is not deep.
411
412 16. Experimental code added to "pcretest -C" to output the stack frame size.
413
414
415 Version 8.21 12-Dec-2011
416 ------------------------
417
418 1. Updating the JIT compiler.
419
420 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
421 are added as well.
422
423 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
424 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
425 calling _pcre_jit_exec. Some extra comments are added.
426
427 4. (*MARK) settings inside atomic groups that do not contain any capturing
428 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
429 was introduced by change 18 for 8.20.
430
431 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
432 ECMA-262 standard.
433
434 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
435 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
436 This bug was probably introduced by change 9 of 8.13.
437
438 7. While fixing 6 above, I noticed that a number of other items were being
439 incorrectly rejected as "not fixed length". This arose partly because newer
440 opcodes had not been added to the fixed-length checking code. I have (a)
441 corrected the bug and added tests for these items, and (b) arranged for an
442 error to occur if an unknown opcode is encountered while checking for fixed
443 length instead of just assuming "not fixed length". The items that were
444 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
445 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
446 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
447
448 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
449 being incorrectly compiled and would have given unpredicatble results.
450
451 9. A possessively repeated subpattern with minimum repeat count greater than
452 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
453 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
454 the first (A) could occur when it should not.
455
456 10. Add a cast and remove a redundant test from the code.
457
458 11. JIT should use pcre_malloc/pcre_free for allocation.
459
460 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
461 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
462 is anything other than /usr, -L is still shown).
463
464 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
465
466 14. Perl does not support \N without a following name in a [] class; PCRE now
467 also gives an error.
468
469 15. If a forward reference was repeated with an upper limit of around 2000,
470 it caused the error "internal error: overran compiling workspace". The
471 maximum number of forward references (including repeats) was limited by the
472 internal workspace, and dependent on the LINK_SIZE. The code has been
473 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
474 the default depends on LINK_SIZE. There is a new upper limit (for safety)
475 of around 200,000 forward references. While doing this, I also speeded up
476 the filling in of repeated forward references.
477
478 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
479 incorrectly expecting the subject to contain another "a" after the start.
480
481 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
482 in the match, the SKIP should be ignored. This was not happening; instead
483 the SKIP was being treated as NOMATCH. For patterns such as
484 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
485 tested.
486
487 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
488 now much more compatible with Perl, in particular in cases where the result
489 is a non-match for a non-anchored pattern. For example, if
490 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
491 "m", where previously it did not return a name. A side effect of this
492 change is that for partial matches, the last encountered mark name is
493 returned, as for non matches. A number of tests that were previously not
494 Perl-compatible have been moved into the Perl-compatible test files. The
495 refactoring has had the pleasing side effect of removing one argument from
496 the match() function, thus reducing its stack requirements.
497
498 19. If the /S+ option was used in pcretest to study a pattern using JIT,
499 subsequent uses of /S (without +) incorrectly behaved like /S+.
500
501 21. Retrieve executable code size support for the JIT compiler and fixing
502 some warnings.
503
504 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
505 not work when the shorter character appeared right at the end of the
506 subject string.
507
508 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
509 systems.
510
511 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
512 output it when the /M option is used in pcretest.
513
514 25. The CheckMan script was not being included in the distribution. Also, added
515 an explicit "perl" to run Perl scripts from the PrepareRelease script
516 because this is reportedly needed in Windows.
517
518 26. If study data was being save in a file and studying had not found a set of
519 "starts with" bytes for the pattern, the data written to the file (though
520 never used) was taken from uninitialized memory and so caused valgrind to
521 complain.
522
523 27. Updated RunTest.bat as provided by Sheri Pierce.
524
525 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
526
527 29. Computation of memory usage for the table of capturing group names was
528 giving an unnecessarily large value.
529
530
531 Version 8.20 21-Oct-2011
532 ------------------------
533
534 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
535 a POSIX class. After further experiments with Perl, which convinced me that
536 Perl has bugs and confusions, a closing square bracket is no longer allowed
537 in a POSIX name. This bug also affected patterns with classes that started
538 with full stops.
539
540 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
541 captured substring, but while checking the failing first alternative,
542 substring 1 is temporarily captured. If the output vector supplied to
543 pcre_exec() was not big enough for this capture, the yield of the function
544 was still zero ("insufficient space for captured substrings"). This cannot
545 be totally fixed without adding another stack variable, which seems a lot
546 of expense for a edge case. However, I have improved the situation in cases
547 such as /(a)(b)x|abc/ matched against "abc", where the return code
548 indicates that fewer than the maximum number of slots in the ovector have
549 been set.
550
551 3. Related to (2) above: when there are more back references in a pattern than
552 slots in the output vector, pcre_exec() uses temporary memory during
553 matching, and copies in the captures as far as possible afterwards. It was
554 using the entire output vector, but this conflicts with the specification
555 that only 2/3 is used for passing back captured substrings. Now it uses
556 only the first 2/3, for compatibility. This is, of course, another edge
557 case.
558
559 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
560 main code base, and can be used by building with --enable-jit. When this is
561 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
562 runtime --no-jit option is given.
563
564 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
565 ovector, the return from the function was zero, implying that there were
566 other matches that did not fit. The correct "exactly full" value is now
567 returned.
568
569 6. If a subpattern that was called recursively or as a subroutine contained
570 (*PRUNE) or any other control that caused it to give a non-standard return,
571 invalid errors such as "Error -26 (nested recursion at the same subject
572 position)" or even infinite loops could occur.
573
574 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
575 computing the minimum length on reaching *ACCEPT, and so ended up with the
576 wrong value of 1 rather than 0. Further investigation indicates that
577 computing a minimum subject length in the presence of *ACCEPT is difficult
578 (think back references, subroutine calls), and so I have changed the code
579 so that no minimum is registered for a pattern that contains *ACCEPT.
580
581 8. If (*THEN) was present in the first (true) branch of a conditional group,
582 it was not handled as intended. [But see 16 below.]
583
584 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
585 Sheri Pierce.
586
587 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
588 the first byte in a match must be "a".
589
590 11. Change 17 for 8.13 increased the recursion depth for patterns like
591 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
592 pattern contains any instances of (*THEN). If it does not, the old
593 optimizations are restored. It would be nice to do this on a per-group
594 basis, but at the moment that is not feasible.
595
596 12. In some environments, the output of pcretest -C is CRLF terminated. This
597 broke RunTest's code that checks for the link size. A single white space
598 character after the value is now allowed for.
599
600 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
601 For "fr", it uses the Windows-specific input and output files.
602
603 14. If (*THEN) appeared in a group that was called recursively or as a
604 subroutine, it did not work as intended. [But see next item.]
605
606 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
607 pattern fragments (but not containing any | characters). If A and B are
608 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
609 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
610 D. In other words, Perl considers parentheses that do not contain any |
611 characters to be part of a surrounding alternative, whereas PCRE was
612 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
613 differently. PCRE now behaves in the same way as Perl, except in the case
614 of subroutine/recursion calls such as (?1) which have in any case always
615 been different (but PCRE had them first :-).
616
617 16. Related to 15 above: Perl does not treat the | in a conditional group as
618 creating alternatives. Such a group is treated in the same way as an
619 ordinary group without any | characters when processing (*THEN). PCRE has
620 been changed to match Perl's behaviour.
621
622 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
623 RunGrepTest script failed.
624
625 18. Change 22 for version 13 caused atomic groups to use more stack. This is
626 inevitable for groups that contain captures, but it can lead to a lot of
627 stack use in large patterns. The old behaviour has been restored for atomic
628 groups that do not contain any capturing parentheses.
629
630 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
631 suppress the check for a minimum subject length at run time. (If it was
632 given to pcre_exec() or pcre_dfa_exec() it did work.)
633
634 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
635 fail to work when decoding hex characters in data strings in EBCDIC
636 environments.
637
638 21. It appears that in at least one Mac OS environment, the isxdigit() function
639 is implemented as a macro that evaluates to its argument more than once,
640 contravening the C 90 Standard (I haven't checked a later standard). There
641 was an instance in pcretest which caused it to go wrong when processing
642 \x{...} escapes in subject strings. The has been rewritten to avoid using
643 things like p++ in the argument of isxdigit().
644
645
646 Version 8.13 16-Aug-2011
647 ------------------------
648
649 1. The Unicode data tables have been updated to Unicode 6.0.0.
650
651 2. Two minor typos in pcre_internal.h have been fixed.
652
653 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
654 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
655 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
656
657 4. There were a number of related bugs in the code for matching backrefences
658 caselessly in UTF-8 mode when codes for the characters concerned were
659 different numbers of bytes. For example, U+023A and U+2C65 are an upper
660 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
661 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
662 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
663 2-byte code at the end of the subject (it thought there wasn't enough data
664 left).
665
666 5. Comprehensive information about what went wrong is now returned by
667 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
668 as the output vector has at least 2 elements. The offset of the start of
669 the failing character and a reason code are placed in the vector.
670
671 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
672 now returned is for the first byte of the failing character, instead of the
673 last byte inspected. This is an incompatible change, but I hope it is small
674 enough not to be a problem. It makes the returned offset consistent with
675 pcre_exec() and pcre_dfa_exec().
676
677 7. pcretest now gives a text phrase as well as the error number when
678 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
679 failure, the offset and reason code are output.
680
681 8. When \R was used with a maximizing quantifier it failed to skip backwards
682 over a \r\n pair if the subsequent match failed. Instead, it just skipped
683 back over a single character (\n). This seems wrong (because it treated the
684 two characters as a single entity when going forwards), conflicts with the
685 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
686 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
687 has been changed.
688
689 9. Some internal refactoring has changed the processing so that the handling
690 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
691 time (the PCRE_DOTALL option was changed this way some time ago: version
692 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
693 which was always a bit of a fudge. It also means that there is one less
694 argument for the match() function, which reduces its stack requirements
695 slightly. This change also fixes an incompatibility with Perl: the pattern
696 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
697
698 10. More internal refactoring has drastically reduced the number of recursive
699 calls to match() for possessively repeated groups such as (abc)++ when
700 using pcre_exec().
701
702 11. While implementing 10, a number of bugs in the handling of groups were
703 discovered and fixed:
704
705 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
706 (a|)*(?1) gave a compile-time internal error.
707 ((a|)+)+ did not notice that the outer group could match an empty string.
708 (^a|^)+ was not marked as anchored.
709 (.*a|.*)+ was not marked as matching at start or after a newline.
710
711 12. Yet more internal refactoring has removed another argument from the match()
712 function. Special calls to this function are now indicated by setting a
713 value in a variable in the "match data" data block.
714
715 13. Be more explicit in pcre_study() instead of relying on "default" for
716 opcodes that mean there is no starting character; this means that when new
717 ones are added and accidentally left out of pcre_study(), testing should
718 pick them up.
719
720 14. The -s option of pcretest has been documented for ages as being an old
721 synonym of -m (show memory usage). I have changed it to mean "force study
722 for every regex", that is, assume /S for every regex. This is similar to -i
723 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
724 using it. It makes it easier to run collections of tests with and without
725 study enabled, and thereby test pcre_study() more easily. All the standard
726 tests are now run with and without -s (but some patterns can be marked as
727 "never study" - see 20 below).
728
729 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
730 restoration of the capturing data to the outer values was not happening
731 correctly.
732
733 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
734 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
735 pattern had matched an empty string, and so incorrectly returned a no
736 match.
737
738 17. There was optimizing code for the last branch of non-capturing parentheses,
739 and also for the obeyed branch of a conditional subexpression, which used
740 tail recursion to cut down on stack usage. Unfortunately, now that there is
741 the possibility of (*THEN) occurring in these branches, tail recursion is
742 no longer possible because the return has to be checked for (*THEN). These
743 two optimizations have therefore been removed. [But see 8.20/11 above.]
744
745 18. If a pattern containing \R was studied, it was assumed that \R always
746 matched two bytes, thus causing the minimum subject length to be
747 incorrectly computed because \R can also match just one byte.
748
749 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
750 was incorrectly computed.
751
752 20. If /S is present twice on a test pattern in pcretest input, it now
753 *disables* studying, thereby overriding the use of -s on the command line
754 (see 14 above). This is necessary for one or two tests to keep the output
755 identical in both cases.
756
757 21. When (*ACCEPT) was used in an assertion that matched an empty string and
758 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
759
760 22. When an atomic group that contained a capturing parenthesis was
761 successfully matched, but the branch in which it appeared failed, the
762 capturing was not being forgotten if a higher numbered group was later
763 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
764 group 1 to "a", when in fact it should be unset. This applied to multi-
765 branched capturing and non-capturing groups, repeated or not, and also to
766 positive assertions (capturing in negative assertions does not happen
767 in PCRE) and also to nested atomic groups.
768
769 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
770 subject after a captured substring, to make it easier to tell which of a
771 number of identical substrings has been captured.
772
773 24. The way atomic groups are processed by pcre_exec() has been changed so that
774 if they are repeated, backtracking one repetition now resets captured
775 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
776 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
777 "aaa". Previously, it would have been "a". As part of this code
778 refactoring, the way recursive calls are handled has also been changed.
779
780 25. If an assertion condition captured any substrings, they were not passed
781 back unless some other capturing happened later. For example, if
782 (?(?=(a))a) was matched against "a", no capturing was returned.
783
784 26. When studying a pattern that contained subroutine calls or assertions,
785 the code for finding the minimum length of a possible match was handling
786 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
787 group 1 called group 2 while simultaneously a separate group 2 called group
788 1). A stack overflow occurred in this case. I have fixed this by limiting
789 the recursion depth to 10.
790
791 27. Updated RunTest.bat in the distribution to the version supplied by Tom
792 Fortmann. This supports explicit test numbers on the command line, and has
793 argument validation and error reporting.
794
795 28. An instance of \X with an unlimited repeat could fail if at any point the
796 first character it looked at was a mark character.
797
798 29. Some minor code refactoring concerning Unicode properties and scripts
799 should reduce the stack requirement of match() slightly.
800
801 30. Added the '=' option to pcretest to check the setting of unused capturing
802 slots at the end of the pattern, which are documented as being -1, but are
803 not included in the return count.
804
805 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
806 compiled something random. Now it gives a compile-time error (as does
807 Perl).
808
809 32. A *MARK encountered during the processing of a positive assertion is now
810 recorded and passed back (compatible with Perl).
811
812 33. If --only-matching or --colour was set on a pcregrep call whose pattern
813 had alternative anchored branches, the search for a second match in a line
814 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
815 matched the line "0102" twice. The same bug affected patterns that started
816 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
817 twice.
818
819 34. Previously, PCRE did not allow quantification of assertions. However, Perl
820 does, and because of capturing effects, quantifying parenthesized
821 assertions may at times be useful. Quantifiers are now allowed for
822 parenthesized assertions.
823
824 35. A minor code tidy in pcre_compile() when checking options for \R usage.
825
826 36. \g was being checked for fancy things in a character class, when it should
827 just be a literal "g".
828
829 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
830 appearance of a nested POSIX class supersedes an apparent external class.
831 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
832 unescaped square brackets may also appear as part of class names. For
833 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
834 more like Perl. (But see 8.20/1 above.)
835
836 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
837 was because it thought it was \N{name}, which is not supported).
838
839 39. Add minix to OS list not supporting the -S option in pcretest.
840
841 40. PCRE tries to detect cases of infinite recursion at compile time, but it
842 cannot analyze patterns in sufficient detail to catch mutual recursions
843 such as ((?1))((?2)). There is now a runtime test that gives an error if a
844 subgroup is called recursively as a subpattern for a second time at the
845 same position in the subject string. In previous releases this might have
846 been caught by the recursion limit, or it might have run out of stack.
847
848 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
849 happen only once. PCRE was, however incorrectly giving a compile time error
850 "recursive call could loop indefinitely" because it cannot analyze the
851 pattern in sufficient detail. The compile time test no longer happens when
852 PCRE is compiling a conditional subpattern, but actual runaway loops are
853 now caught at runtime (see 40 above).
854
855 42. It seems that Perl allows any characters other than a closing parenthesis
856 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
857 has been changed to be the same.
858
859 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
860 as not to get warnings when autogen.sh is called. Also changed
861 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
862
863 44. To help people who use pcregrep to scan files containing exceedingly long
864 lines, the following changes have been made:
865
866 (a) The default value of the buffer size parameter has been increased from
867 8K to 20K. (The actual buffer used is three times this size.)
868
869 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
870 PCRE is built.
871
872 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
873 to be set at run time.
874
875 (d) Numerical values in pcregrep options can be followed by K or M, for
876 example --buffer-size=50K.
877
878 (e) If a line being scanned overflows pcregrep's buffer, an error is now
879 given and the return code is set to 2.
880
881 45. Add a pointer to the latest mark to the callout data block.
882
883 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
884 partial match of an empty string instead of no match. This was specific to
885 the use of ".".
886
887 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
888 complete match instead of a partial match. This bug was dependent on both
889 the PCRE_UTF8 and PCRE_DOTALL options being set.
890
891 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
892 starting byte set, because \b was not being ignored.
893
894
895 Version 8.12 15-Jan-2011
896 ------------------------
897
898 1. Fixed some typos in the markup of the man pages, and wrote a script that
899 checks for such things as part of the documentation building process.
900
901 2. On a big-endian 64-bit system, pcregrep did not correctly process the
902 --match-limit and --recursion-limit options (added for 8.11). In
903 particular, this made one of the standard tests fail. (The integer value
904 went into the wrong half of a long int.)
905
906 3. If the --colour option was given to pcregrep with -v (invert match), it
907 did strange things, either producing crazy output, or crashing. It should,
908 of course, ignore a request for colour when reporting lines that do not
909 match.
910
911 4. Another pcregrep bug caused similar problems if --colour was specified with
912 -M (multiline) and the pattern match finished with a line ending.
913
914 5. In pcregrep, when a pattern that ended with a literal newline sequence was
915 matched in multiline mode, the following line was shown as part of the
916 match. This seems wrong, so I have changed it.
917
918 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
919 the check for further matches in the same line (so they could be coloured)
920 to overrun the end of the current line. If another match was found, it was
921 incorrectly shown (and then shown again when found in the next line).
922
923 7. If pcregrep was compiled under Windows, there was a reference to the
924 function pcregrep_exit() before it was defined. I am assuming this was
925 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
926 reported by a user. I've moved the definition above the reference.
927
928
929 Version 8.11 10-Dec-2010
930 ------------------------
931
932 1. (*THEN) was not working properly if there were untried alternatives prior
933 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
934 backtracked to try for "b" instead of moving to the next alternative branch
935 at the same level (in this case, to look for "c"). The Perl documentation
936 is clear that when (*THEN) is backtracked onto, it goes to the "next
937 alternative in the innermost enclosing group".
938
939 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
940 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
941 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
942 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
943 (*THEN).
944
945 3. If \s appeared in a character class, it removed the VT character from
946 the class, even if it had been included by some previous item, for example
947 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
948 of \s, but is part of the POSIX "space" class.)
949
950 4. A partial match never returns an empty string (because you can always
951 match an empty string at the end of the subject); however the checking for
952 an empty string was starting at the "start of match" point. This has been
953 changed to the "earliest inspected character" point, because the returned
954 data for a partial match starts at this character. This means that, for
955 example, /(?<=abc)def/ gives a partial match for the subject "abc"
956 (previously it gave "no match").
957
958 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
959 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
960 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
961 has an implication that the given string is incomplete (because a partial
962 match is preferred over a full match). For this reason, these items now
963 give a partial match in this situation. [Aside: previously, the one case
964 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
965 match rather than a full match, which was wrong by the old rules, but is
966 now correct.]
967
968 6. There was a bug in the handling of #-introduced comments, recognized when
969 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
970 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
971 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
972 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
973 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
974 places in pcre_compile().
975
976 7. Related to (6) above, when pcre_compile() was skipping #-introduced
977 comments when looking ahead for named forward references to subpatterns,
978 the only newline sequence it recognized was NL. It now handles newlines
979 according to the set newline convention.
980
981 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
982 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
983 cater for a lack of strerror(). These oversights have been fixed.
984
985 9. Added --match-limit and --recursion-limit to pcregrep.
986
987 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
988
989 11. When the -o option was used, pcregrep was setting a return code of 1, even
990 when matches were found, and --line-buffered was not being honoured.
991
992 12. Added an optional parentheses number to the -o and --only-matching options
993 of pcregrep.
994
995 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
996 can match an empty string. The code to do it in pcretest and pcredemo
997 needed fixing:
998
999 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1000 only one byte after an empty string match just before CRLF (this case
1001 just got forgotten; "any" and "anycrlf" were OK).
1002
1003 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1004 mode when an empty string match preceded an ASCII character followed by
1005 a non-ASCII character. (The code for advancing by one character rather
1006 than one byte was nonsense.)
1007
1008 (c) The pcredemo.c sample program did not have any code at all to handle
1009 the cases when CRLF is a valid newline sequence.
1010
1011 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1012 as a starting offset was within the subject string. There is now a new
1013 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1014 negative or greater than the length of the string. In order to test this,
1015 pcretest is extended to allow the setting of negative starting offsets.
1016
1017 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1018 starting offset points to the beginning of a UTF-8 character was
1019 unnecessarily clumsy. I tidied it up.
1020
1021 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1022 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1023
1024 17. Nobody had reported that the --include_dir option, which was added in
1025 release 7.7 should have been called --include-dir (hyphen, not underscore)
1026 for compatibility with GNU grep. I have changed it to --include-dir, but
1027 left --include_dir as an undocumented synonym, and the same for
1028 --exclude-dir, though that is not available in GNU grep, at least as of
1029 release 2.5.4.
1030
1031 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1032 characters from a string of bytes) have been redefined so as not to use
1033 loops, in order to improve performance in some environments. At the same
1034 time, I abstracted some of the common code into auxiliary macros to save
1035 repetition (this should not affect the compiled code).
1036
1037 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1038 compile-time error is now given if \c is not followed by an ASCII
1039 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1040 different, and any byte value is allowed.)
1041
1042 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1043 START_OPTIMIZE option, which is now allowed at compile time - but just
1044 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1045 to pcregrep and other applications that have no direct access to PCRE
1046 options. The new /Y option in pcretest sets this option when calling
1047 pcre_compile().
1048
1049 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1050 back references. Groups containing recursive back references were forced to
1051 be atomic by that change, but in the case of named groups, the amount of
1052 memory required was incorrectly computed, leading to "Failed: internal
1053 error: code overflow". This has been fixed.
1054
1055 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1056 pcretest.c, to avoid build problems in some Borland environments.
1057
1058
1059 Version 8.10 25-Jun-2010
1060 ------------------------
1061
1062 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1063 THEN.
1064
1065 2. (*ACCEPT) was not working when inside an atomic group.
1066
1067 3. Inside a character class, \B is treated as a literal by default, but
1068 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1069 causes the error). The code is unchanged, but I tidied the documentation.
1070
1071 4. Inside a character class, PCRE always treated \R and \X as literals,
1072 whereas Perl faults them if its -w option is set. I have changed PCRE so
1073 that it faults them when PCRE_EXTRA is set.
1074
1075 5. Added support for \N, which always matches any character other than
1076 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1077
1078 6. When compiling pcregrep with newer versions of gcc which may have
1079 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1080 declared with attribute warn_unused_result" were given. Just casting the
1081 result to (void) does not stop the warnings; a more elaborate fudge is
1082 needed. I've used a macro to implement this.
1083
1084 7. Minor change to pcretest.c to avoid a compiler warning.
1085
1086 8. Added four artifical Unicode properties to help with an option to make
1087 \s etc use properties (see next item). The new properties are: Xan
1088 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1089
1090 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1091 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1092 this option. Modified pcretest to add /W to test this facility. Added
1093 REG_UCP to make it available via the POSIX interface.
1094
1095 10. Added --line-buffered to pcregrep.
1096
1097 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1098 studied, and the match started with a letter with a code point greater than
1099 127 whose first byte was different to the first byte of the other case of
1100 the letter, the other case of this starting letter was not recognized
1101 (#976).
1102
1103 12. If a pattern that was studied started with a repeated Unicode property
1104 test, for example, \p{Nd}+, there was the theoretical possibility of
1105 setting up an incorrect bitmap of starting bytes, but fortunately it could
1106 not have actually happened in practice until change 8 above was made (it
1107 added property types that matched character-matching opcodes).
1108
1109 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1110 possible starting bytes for non-anchored patterns.
1111
1112 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1113 \R, and also a number of cases that involve Unicode properties, both
1114 explicit and implicit when PCRE_UCP is set.
1115
1116 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1117 input, it could crash or give wrong results if characters with values
1118 greater than 0xc0 were present in the subject string. (Detail: it assumed
1119 UTF-8 input when processing these items.)
1120
1121 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1122 size_t is 64-bit (#991).
1123
1124 17. Added a check for running out of memory when PCRE is compiled with
1125 --disable-stack-for-recursion (#990).
1126
1127 18. If the last data line in a file for pcretest does not have a newline on
1128 the end, a newline was missing in the output.
1129
1130 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1131 less than 128) in its various bitmaps. However, there is a facility for
1132 generating tables according to the current locale when PCRE is compiled. It
1133 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1134 characters, are recognized by isspace() and therefore were getting set in
1135 these tables, and indeed these tables seem to approximate to ISO 8859. This
1136 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1137 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1138 which of course cannot start UTF-8 characters. I have changed the code so
1139 that only real ASCII characters (less than 128) and the correct starting
1140 bytes for UTF-8 encodings are set for characters greater than 127 when in
1141 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1142 altogether.)
1143
1144 20. Added the /T option to pcretest so as to be able to run tests with non-
1145 standard character tables, thus making it possible to include the tests
1146 used for 19 above in the standard set of tests.
1147
1148 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1149 reference to a subpattern the other side of a comment that contains an
1150 opening parenthesis caused either an internal compiling error, or a
1151 reference to the wrong subpattern.
1152
1153
1154 Version 8.02 19-Mar-2010
1155 ------------------------
1156
1157 1. The Unicode data tables have been updated to Unicode 5.2.0.
1158
1159 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1160 configured.
1161
1162 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1163 original author of that file, following a query about its status.
1164
1165 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1166 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1167
1168 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1169 quantifier applied to a forward-referencing subroutine call, could compile
1170 incorrect code or give the error "internal error: previously-checked
1171 referenced subpattern not found".
1172
1173 6. Both MS Visual Studio and Symbian OS have problems with initializing
1174 variables to point to external functions. For these systems, therefore,
1175 pcre_malloc etc. are now initialized to local functions that call the
1176 relevant global functions.
1177
1178 7. There were two entries missing in the vectors called coptable and poptable
1179 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1180 I've fixed the data, and added a kludgy way of testing at compile time that
1181 the lengths are correct (equal to the number of opcodes).
1182
1183 8. Following on from 7, I added a similar kludge to check the length of the
1184 eint vector in pcreposix.c.
1185
1186 9. Error texts for pcre_compile() are held as one long string to avoid too
1187 much relocation at load time. To find a text, the string is searched,
1188 counting zeros. There was no check for running off the end of the string,
1189 which could happen if a new error number was added without updating the
1190 string.
1191
1192 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1193
1194 11. \K was not working if it appeared in an atomic group or in a group that
1195 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1196 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1197 the assertion is positive, but not if it is negative.
1198
1199 12. Change 11 fortuitously reduced the size of the stack frame used in the
1200 "match()" function of pcre_exec.c by one pointer. Forthcoming
1201 implementation of support for (*MARK) will need an extra pointer on the
1202 stack; I have reserved it now, so that the stack frame size does not
1203 decrease.
1204
1205 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1206 item in branch that calls a recursion is a subroutine call - as in the
1207 second branch in the above example - was incorrectly given the compile-
1208 time error "recursive call could loop indefinitely" because pcre_compile()
1209 was not correctly checking the subroutine for matching a non-empty string.
1210
1211 14. The checks for overrunning compiling workspace could trigger after an
1212 overrun had occurred. This is a "should never occur" error, but it can be
1213 triggered by pathological patterns such as hundreds of nested parentheses.
1214 The checks now trigger 100 bytes before the end of the workspace.
1215
1216 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1217
1218
1219 Version 8.01 19-Jan-2010
1220 ------------------------
1221
1222 1. If a pattern contained a conditional subpattern with only one branch (in
1223 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1224 computed the wrong minimum data length (which is of course zero for such
1225 subpatterns). This could cause incorrect "no match" results.
1226
1227 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1228 the pattern is reset in the first branch, pcre_compile() failed with
1229 "internal error: code overflow at offset...". This happened only when
1230 the reset was to the original external option setting. (An optimization
1231 abstracts leading options settings into an external setting, which was the
1232 cause of this.)
1233
1234 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1235 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1236 assertion pattern did not match (meaning that the assertion was true), it
1237 was incorrectly treated as false if the SKIP had been reached during the
1238 matching. This also applied to assertions used as conditions.
1239
1240 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1241 assertion subpattern, including such a pattern used as a condition,
1242 unpredictable results occurred, instead of the error return
1243 PCRE_ERROR_DFA_UITEM.
1244
1245 5. The C++ GlobalReplace function was not working like Perl for the special
1246 situation when an empty string is matched. It now does the fancy magic
1247 stuff that is necessary.
1248
1249 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1250 removed. (These were left over from very, very early versions of PCRE.)
1251
1252 7. Some cosmetic changes to the code to make life easier when compiling it
1253 as part of something else:
1254
1255 (a) Change DEBUG to PCRE_DEBUG.
1256
1257 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1258 called "current" as "current_branch", to prevent a collision with the
1259 Linux macro when compiled as a kernel module.
1260
1261 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1262 prevent a collision with the Linux macro when compiled as a kernel
1263 module.
1264
1265 8. In pcre_compile() there are some checks for integer overflows that used to
1266 cast potentially large values to (double). This has been changed to that
1267 when building, a check for int64_t is made, and if it is found, it is used
1268 instead, thus avoiding the use of floating point arithmetic. (There is no
1269 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1270 double.
1271
1272 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1273 2005 (difference between two addresses compared to an unsigned value).
1274
1275 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1276 custom one, because of the following reported problem in Windows:
1277
1278 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1279 under Win32.
1280 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1281 therefore missing the function definition.
1282 - The compiler thus generates a "C" signature for the test function.
1283 - The linker fails to find the "C" function.
1284 - PCRE fails to configure if asked to do so against libbz2.
1285
1286 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1287 messages were output:
1288
1289 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1290 rerunning libtoolize, to keep the correct libtool macros in-tree.
1291 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1292
1293 I have done both of these things.
1294
1295 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1296 most of the time, it *can* run out if it is given a pattern that contains a
1297 runaway infinite recursion. I updated the discussion in the pcrestack man
1298 page.
1299
1300 13. Now that we have gone to the x.xx style of version numbers, the minor
1301 version may start with zero. Using 08 or 09 is a bad idea because users
1302 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1303 interpreted as invalid octal numbers. I've updated the previous comment in
1304 configure.ac, and also added a check that gives an error if 08 or 09 are
1305 used.
1306
1307 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1308 causing partial matching to fail when the end of the subject matched \W
1309 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1310
1311 15. There were some discrepancies between the declarations in pcre_internal.h
1312 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1313 their definitions. The declarations used "const uschar *" and the
1314 definitions used USPTR. Even though USPTR is normally defined as "const
1315 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1316 reported that: "This difference in casting confuses some C++ compilers, for
1317 example, SunCC recognizes above declarations as different functions and
1318 generates broken code for hbpcre." I have changed the declarations to use
1319 USPTR.
1320
1321 16. GNU libtool is named differently on some systems. The autogen.sh script now
1322 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1323 (FreeBSD).
1324
1325 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1326 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1327 comment: "Figure out how to create a longlong from a string: strtoll and
1328 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1329 instance, but it only takes 2 args instead of 3!"
1330
1331 18. A subtle bug concerned with back references has been fixed by a change of
1332 specification, with a corresponding code fix. A pattern such as
1333 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1334 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1335 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1336 same bug. Such groups have to be quantified to be useful, or contained
1337 inside another quantified group. (If there's no repetition, the reference
1338 can never match.) The problem arises because, having left the group and
1339 moved on to the rest of the pattern, a later failure that backtracks into
1340 the group uses the captured value from the final iteration of the group
1341 rather than the correct earlier one. I have fixed this in PCRE by forcing
1342 any group that contains a reference to itself to be an atomic group; that
1343 is, there cannot be any backtracking into it once it has completed. This is
1344 similar to recursive and subroutine calls.
1345
1346
1347 Version 8.00 19-Oct-09
1348 ----------------------
1349
1350 1. The table for translating pcre_compile() error codes into POSIX error codes
1351 was out-of-date, and there was no check on the pcre_compile() error code
1352 being within the table. This could lead to an OK return being given in
1353 error.
1354
1355 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1356 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1357 in a Windows environment.
1358
1359 3. The pcregrep --count option prints the count for each file even when it is
1360 zero, as does GNU grep. However, pcregrep was also printing all files when
1361 --files-with-matches was added. Now, when both options are given, it prints
1362 counts only for those files that have at least one match. (GNU grep just
1363 prints the file name in this circumstance, but including the count seems
1364 more useful - otherwise, why use --count?) Also ensured that the
1365 combination -clh just lists non-zero counts, with no names.
1366
1367 4. The long form of the pcregrep -F option was incorrectly implemented as
1368 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1369 but it seems right to fix it, and I didn't think it was worth preserving
1370 the old behaviour.
1371
1372 5. The command line items --regex=pattern and --regexp=pattern were not
1373 recognized by pcregrep, which required --regex pattern or --regexp pattern
1374 (with a space rather than an '='). The man page documented the '=' forms,
1375 which are compatible with GNU grep; these now work.
1376
1377 6. No libpcreposix.pc file was created for pkg-config; there was just
1378 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1379
1380 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1381 when UCP support is not needed, by modifying the Python script that
1382 generates it from Unicode data files. This should not matter if the module
1383 is correctly used as a library, but I received one complaint about 50K of
1384 unwanted data. My guess is that the person linked everything into his
1385 program rather than using a library. Anyway, it does no harm.
1386
1387 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1388 was a minimum greater than 1 for a wide character in a possessive
1389 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1390 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1391 character. Chaos in the form of incorrect output or a compiling loop could
1392 result.
1393
1394 9. The restrictions on what a pattern can contain when partial matching is
1395 requested for pcre_exec() have been removed. All patterns can now be
1396 partially matched by this function. In addition, if there are at least two
1397 slots in the offset vector, the offset of the earliest inspected character
1398 for the match and the offset of the end of the subject are set in them when
1399 PCRE_ERROR_PARTIAL is returned.
1400
1401 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1402 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1403 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1404 and may be more useful for multi-segment matching.
1405
1406 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1407 used to be given if ever the end of the subject was reached; now it is
1408 given only if matching could not proceed because another character was
1409 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1410 string "Z", which now yields "no match" instead of "partial match". In the
1411 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1412 final character ended with (*FAIL).
1413
1414 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1415 if the pattern had a "must contain" character that was already found in the
1416 earlier partial match, unless partial matching was again requested. For
1417 example, with the pattern /dog.(body)?/, the "must contain" character is
1418 "g". If the first part-match was for the string "dog", restarting with
1419 "sbody" failed. This bug has been fixed.
1420
1421 13. The string returned by pcre_dfa_exec() after a partial match has been
1422 changed so that it starts at the first inspected character rather than the
1423 first character of the match. This makes a difference only if the pattern
1424 starts with a lookbehind assertion or \b or \B (\K is not supported by
1425 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1426 matching functions compatible, and I think it's the right thing to do.
1427
1428 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1429 so that the demonstration program is easily available in environments where
1430 PCRE has not been installed from source.
1431
1432 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1433 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1434 library.
1435
1436 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1437 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1438 is not the first non-POSIX option to be added. Clearly some people find
1439 these options useful.
1440
1441 17. If a caller to the POSIX matching function regexec() passes a non-zero
1442 value for nmatch with a NULL value for pmatch, the value of
1443 nmatch is forced to zero.
1444
1445 18. RunGrepTest did not have a test for the availability of the -u option of
1446 the diff command, as RunTest does. It now checks in the same way as
1447 RunTest, and also checks for the -b option.
1448
1449 19. If an odd number of negated classes containing just a single character
1450 interposed, within parentheses, between a forward reference to a named
1451 subpattern and the definition of the subpattern, compilation crashed with
1452 an internal error, complaining that it could not find the referenced
1453 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1454 [The bug was that it was starting one character too far in when skipping
1455 over the character class, thus treating the ] as data rather than
1456 terminating the class. This meant it could skip too much.]
1457
1458 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1459 /g option in pcretest when the pattern contains \K, which makes it possible
1460 to have an empty string match not at the start, even when the pattern is
1461 anchored. Updated pcretest and pcredemo to use this option.
1462
1463 21. If the maximum number of capturing subpatterns in a recursion was greater
1464 than the maximum at the outer level, the higher number was returned, but
1465 with unset values at the outer level. The correct (outer level) value is
1466 now given.
1467
1468 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1469 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1470 make it do so. The string so far is captured, making this feature
1471 compatible with Perl.
1472
1473 23. The tests have been re-organized, adding tests 11 and 12, to make it
1474 possible to check the Perl 5.10 features against Perl 5.10.
1475
1476 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1477 pattern matches a fixed length string. PCRE did not allow this; now it
1478 does. Neither allows recursion.
1479
1480 25. I finally figured out how to implement a request to provide the minimum
1481 length of subject string that was needed in order to match a given pattern.
1482 (It was back references and recursion that I had previously got hung up
1483 on.) This code has now been added to pcre_study(); it finds a lower bound
1484 to the length of subject needed. It is not necessarily the greatest lower
1485 bound, but using it to avoid searching strings that are too short does give
1486 some useful speed-ups. The value is available to calling programs via
1487 pcre_fullinfo().
1488
1489 26. While implementing 25, I discovered to my embarrassment that pcretest had
1490 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1491 study optimizations had never been tested with that matching function.
1492 Oops. What is worse, even when it was passed study data, there was a bug in
1493 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1494 were also very few tests of studied patterns with pcre_dfa_exec().
1495
1496 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1497 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1498 on the other side of the coin, they are no longer allowed to have different
1499 names, because these cannot be distinguished in PCRE, and this has caused
1500 confusion. (This is a difference from Perl.)
1501
1502 28. When duplicate subpattern names are present (necessarily with different
1503 numbers, as required by 27 above), and a test is made by name in a
1504 conditional pattern, either for a subpattern having been matched, or for
1505 recursion in such a pattern, all the associated numbered subpatterns are
1506 tested, and the overall condition is true if the condition is true for any
1507 one of them. This is the way Perl works, and is also more like the way
1508 testing by number works.
1509
1510
1511 Version 7.9 11-Apr-09
1512 ---------------------
1513
1514 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1515 (pcretest), all targets were linked against these libraries. This included
1516 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1517 libraries. This caused unwanted dependencies to be created. This problem
1518 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1519 pcretest is linked with readline.
1520
1521 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1522 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1523 moved outside it again, because FALSE and TRUE are already defined in AIX,
1524 but BOOL is not.
1525
1526 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1527 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1528
1529 4. The pcregrep documentation said spaces were inserted as well as colons (or
1530 hyphens) following file names and line numbers when outputting matching
1531 lines. This is not true; no spaces are inserted. I have also clarified the
1532 wording for the --colour (or --color) option.
1533
1534 5. In pcregrep, when --colour was used with -o, the list of matching strings
1535 was not coloured; this is different to GNU grep, so I have changed it to be
1536 the same.
1537
1538 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1539 each matching line was coloured. Now it goes on to look for further matches
1540 of any of the test patterns, which is the same behaviour as GNU grep.
1541
1542 7. A pattern that could match an empty string could cause pcregrep to loop; it
1543 doesn't make sense to accept an empty string match in pcregrep, so I have
1544 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1545 seems to be how GNU grep behaves.
1546
1547 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1548 start or after a newline", because the conditional assertion was not being
1549 correctly handled. The rule now is that both the assertion and what follows
1550 in the first alternative must satisfy the test.
1551
1552 9. If auto-callout was enabled in a pattern with a conditional group whose
1553 condition was an assertion, PCRE could crash during matching, both with
1554 pcre_exec() and pcre_dfa_exec().
1555
1556 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1557 used for matching.
1558
1559 11. Unicode property support in character classes was not working for
1560 characters (bytes) greater than 127 when not in UTF-8 mode.
1561
1562 12. Added the -M command line option to pcretest.
1563
1564 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1565
1566 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1567
1568 16. Added comments and documentation about mis-use of no_arg in the C++
1569 wrapper.
1570
1571 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1572 from Martin Jerabek that uses macro names for all relevant character and
1573 string constants.
1574
1575 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1576 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1577 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1578 these, but not everybody uses configure.
1579
1580 19. A conditional group that had only one branch was not being correctly
1581 recognized as an item that could match an empty string. This meant that an
1582 enclosing group might also not be so recognized, causing infinite looping
1583 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1584 with the subject "ab", where knowledge that the repeated group can match
1585 nothing is needed in order to break the loop.
1586
1587 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1588 exec(), but without supplying a callout function, matching went wrong.
1589
1590 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1591 leak if the size of the offset vector was greater than 30. When the vector
1592 is smaller, the saved offsets during recursion go onto a local stack
1593 vector, but for larger vectors malloc() is used. It was failing to free
1594 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1595 error, in fact).
1596
1597 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1598 heapframe that is used only when UTF-8 support is enabled. This caused no
1599 problem, but was untidy.
1600
1601 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1602 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1603 included within another project.
1604
1605 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1606 slightly modified by me:
1607
1608 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1609 not building pcregrep.
1610
1611 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1612 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1613
1614 25. Forward references, both numeric and by name, in patterns that made use of
1615 duplicate group numbers, could behave incorrectly or give incorrect errors,
1616 because when scanning forward to find the reference group, PCRE was not
1617 taking into account the duplicate group numbers. A pattern such as
1618 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1619
1620 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1621 the feature of a custom pointer more persuasive (as requested by a user).
1622
1623 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1624 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1625 pre-defined.
1626
1627 28. Added support for (*UTF8) at the start of a pattern.
1628
1629 29. Arrange for flags added by the "release type" setting in CMake to be shown
1630 in the configuration summary.
1631
1632
1633 Version 7.8 05-Sep-08
1634 ---------------------
1635
1636 1. Replaced UCP searching code with optimized version as implemented for Ad
1637 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1638 stage table and inline lookup instead of a function, giving speed ups of 2
1639 to 5 times on some simple patterns that I tested. Permission was given to
1640 distribute the MultiStage2.py script that generates the tables (it's not in
1641 the tarball, but is in the Subversion repository).
1642
1643 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1644 scripts.
1645
1646 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1647 a group with a zero qualifier. The result of the study could be incorrect,
1648 or the function might crash, depending on the pattern.
1649
1650 4. Caseless matching was not working for non-ASCII characters in back
1651 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1652 It now works when Unicode Property Support is available.
1653
1654 5. In pcretest, an escape such as \x{de} in the data was always generating
1655 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1656 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1657 truncation.
1658
1659 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1660
1661 7. Added two (int) casts to pcregrep when printing the difference of two
1662 pointers, in case they are 64-bit values.
1663
1664 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1665 test 2 if it fails.
1666
1667 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1668 and a #define of that name to empty if it is not externally set. This is to
1669 allow users of MSVC to set it if necessary.
1670
1671 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1672 the convenience functions in the pcre_get.c source file.
1673
1674 11. An option change at the start of a pattern that had top-level alternatives
1675 could cause overwriting and/or a crash. This command provoked a crash in
1676 some environments:
1677
1678 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1679
1680 This potential security problem was recorded as CVE-2008-2371.
1681
1682 12. For a pattern where the match had to start at the beginning or immediately
1683 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1684 pcre_dfa_exec() could read past the end of the passed subject if there was
1685 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1686 pcretest so that it places the subject at the end of its malloc-ed buffer.
1687
1688 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1689 exec() might read past the end of the data buffer in UTF-8 mode.
1690
1691 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1692 the data contained the byte 0x85 as part of a UTF-8 character within its
1693 first line. This applied both to normal and DFA matching.
1694
1695 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1696 /^[^d]*?$/8 failed to match "abc".
1697
1698 16. Added a missing copyright notice to pcrecpp_internal.h.
1699
1700 17. Make it more clear in the documentation that values returned from
1701 pcre_exec() in ovector are byte offsets, not character counts.
1702
1703 18. Tidied a few places to stop certain compilers from issuing warnings.
1704
1705 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1706 supplied by Stefan Weber. I made a further small update for 7.8 because
1707 there is a change of source arrangements: the pcre_searchfuncs.c module is
1708 replaced by pcre_ucd.c.
1709
1710
1711 Version 7.7 07-May-08
1712 ---------------------
1713
1714 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1715 a string to a long long, pretend we don't even have a long long." This is
1716 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1717
1718 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1719 pre-7.6 versions, which defined a global no_arg variable instead of putting
1720 it in the RE class. (See also #8 below.)
1721
1722 3. Remove a line of dead code, identified by coverity and reported by Nuno
1723 Lopes.
1724
1725 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1726
1727 (1) The include/exclude patterns were being applied to the whole pathnames
1728 of files, instead of just to the final components.
1729
1730 (2) If there was more than one level of directory, the subdirectories were
1731 skipped unless they satisfied the include/exclude conditions. This is
1732 inconsistent with GNU grep (and could even be seen as contrary to the
1733 pcregrep specification - which I improved to make it absolutely clear).
1734 The action now is always to scan all levels of directory, and just
1735 apply the include/exclude patterns to regular files.
1736
1737 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1738 --exclude_dir in the tests to avoid scanning .svn directories.
1739
1740 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1741 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1742 doesn't support NULs in patterns.
1743
1744 7. Added some missing "const"s to declarations of static tables in
1745 pcre_compile.c and pcre_dfa_exec.c.
1746
1747 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1748 caused by fix #2 above. (Subsequently also a second patch to fix the
1749 first patch. And a third patch - this was a messy problem.)
1750
1751 9. Applied Craig's patch to remove the use of push_back().
1752
1753 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1754 matching function regexec().
1755
1756 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1757 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1758 references. PCRE supports relative numbers with this syntax (I don't think
1759 Oniguruma does).
1760
1761 12. Previously, a group with a zero repeat such as (...){0} was completely
1762 omitted from the compiled regex. However, this means that if the group
1763 was called as a subroutine from elsewhere in the pattern, things went wrong
1764 (an internal error was given). Such groups are now left in the compiled
1765 pattern, with a new opcode that causes them to be skipped at execution
1766 time.
1767
1768 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1769 to the way PCRE behaves:
1770
1771 (a) A lone ] character is dis-allowed (Perl treats it as data).
1772
1773 (b) A back reference to an unmatched subpattern matches an empty string
1774 (Perl fails the current match path).
1775
1776 (c) A data ] in a character class must be notated as \] because if the
1777 first data character in a class is ], it defines an empty class. (In
1778 Perl it is not possible to have an empty class.) The empty class []
1779 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1780 The negative empty class [^] matches any one character, independently
1781 of the DOTALL setting.
1782
1783 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1784 non-existent subpattern following a character class starting with ']' and
1785 containing () gave an internal compiling error instead of "reference to
1786 non-existent subpattern". Fortunately, when the pattern did exist, the
1787 compiled code was correct. (When scanning forwards to check for the
1788 existencd of the subpattern, it was treating the data ']' as terminating
1789 the class, so got the count wrong. When actually compiling, the reference
1790 was subsequently set up correctly.)
1791
1792 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1793 it was being rejected as not supported by pcre_dfa_exec(), even though
1794 other assertions are supported. I have made pcre_dfa_exec() support
1795 (*FAIL).
1796
1797 16. The implementation of 13c above involved the invention of a new opcode,
1798 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1799 cannot be changed at match time, I realized I could make a small
1800 improvement to matching performance by compiling OP_ALLANY instead of
1801 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1802 on the OP_ANY path.
1803
1804 17. Compiling pcretest on Windows with readline support failed without the
1805 following two fixes: (1) Make the unistd.h include conditional on
1806 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1807
1808 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1809 ncurses library to be included for pcretest when ReadLine support is
1810 requested, but also to allow for it to be overridden. This patch came from
1811 Daniel Bergström.
1812
1813 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1814 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1815 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1816 spotting this.
1817
1818
1819 Version 7.6 28-Jan-08
1820 ---------------------
1821
1822 1. A character class containing a very large number of characters with
1823 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1824 overflow.
1825
1826 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1827 HAVE_LONG_LONG is not defined.
1828
1829 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1830 bring them up to date and include new features. This patch includes:
1831
1832 - Fixed PH's badly added libz and libbz2 support.
1833 - Fixed a problem with static linking.
1834 - Added pcredemo. [But later removed - see 7 below.]
1835 - Fixed dftables problem and added an option.
1836 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1837 HAVE_LONG_LONG.
1838 - Added readline support for pcretest.
1839 - Added an listing of the option settings after cmake has run.
1840
1841 4. A user submitted a patch to Makefile that makes it easy to create
1842 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1843 Makefile.am that cause it to include this special target, without
1844 affecting anything else. Note that the same mingw target plus all
1845 the other distribution libraries and programs are now supported
1846 when configuring with CMake (see 6 below) instead of with
1847 Configure/Make.
1848
1849 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1850 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1851 exported in the Windows port". It has not yet been confirmed that the patch
1852 solves the problem, but it does no harm.
1853
1854 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1855 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1856 with CMake, and also correct the comment about stack recursion.
1857
1858 7. Remove the automatic building of pcredemo from the ./configure system and
1859 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1860 of a program that users should build themselves after PCRE is installed, so
1861 building it automatically is not really right. What is more, it gave
1862 trouble in some build environments.
1863
1864 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1865
1866
1867 Version 7.5 10-Jan-08
1868 ---------------------
1869
1870 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1871 values in parens when parsing an RE using the C++ wrapper."
1872
1873 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1874 Characters greater than 255 were excluded from the class instead of being
1875 included.
1876
1877 3. The same bug as (2) above applied to negated POSIX classes such as
1878 [:^space:].
1879
1880 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1881 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1882 I have changed it.
1883
1884 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1885 first named subpattern) and a construct such as (?&a) would reference the
1886 first named subpattern whose name started with "a" (in other words, the
1887 length check was missing). Both these problems are fixed. "Subpattern name
1888 expected" is now given for (?&) (a zero-length name), and this patch also
1889 makes it give the same error for \k'' (previously it complained that that
1890 was a reference to a non-existent subpattern).
1891
1892 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1893 this is right because (?- can be followed by option settings as well as by
1894 digits. I have, however, made the messages clearer.
1895
1896 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1897 than the number used in the conditional) now cause a compile-time error.
1898 This is actually not compatible with Perl, which accepts such patterns, but
1899 treats the conditional as always being FALSE (as PCRE used to), but it
1900 seems to me that giving a diagnostic is better.
1901
1902 8. Change "alphameric" to the more common word "alphanumeric" in comments
1903 and messages.
1904
1905 9. Fix two occurrences of "backslash" in comments that should have been
1906 "backspace".
1907
1908 10. Remove two redundant lines of code that can never be obeyed (their function
1909 was moved elsewhere).
1910
1911 11. The program that makes PCRE's Unicode character property table had a bug
1912 which caused it to generate incorrect table entries for sequences of
1913 characters that have the same character type, but are in different scripts.
1914 It amalgamated them into a single range, with the script of the first of
1915 them. In other words, some characters were in the wrong script. There were
1916 thirteen such cases, affecting characters in the following ranges:
1917
1918 U+002b0 - U+002c1
1919 U+0060c - U+0060d
1920 U+0061e - U+00612
1921 U+0064b - U+0065e
1922 U+0074d - U+0076d
1923 U+01800 - U+01805
1924 U+01d00 - U+01d77
1925 U+01d9b - U+01dbf
1926 U+0200b - U+0200f
1927 U+030fc - U+030fe
1928 U+03260 - U+0327f
1929 U+0fb46 - U+0fbb1
1930 U+10450 - U+1049d
1931
1932 12. The -o option (show only the matching part of a line) for pcregrep was not
1933 compatible with GNU grep in that, if there was more than one match in a
1934 line, it showed only the first of them. It now behaves in the same way as
1935 GNU grep.
1936
1937 13. If the -o and -v options were combined for pcregrep, it printed a blank
1938 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1939 does the same. The return code can be used to tell if there were any
1940 non-matching lines.
1941
1942 14. Added --file-offsets and --line-offsets to pcregrep.
1943
1944 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1945 infinitely looping recursion. The bug was that positive lookaheads were not
1946 being skipped when checking for a possible empty match (negative lookaheads
1947 and both kinds of lookbehind were skipped).
1948
1949 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1950 inclusion of <windows.h> to before rather than after the definition of
1951 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1952
1953 17. Specifying a possessive quantifier with a specific limit for a Unicode
1954 character property caused pcre_compile() to compile bad code, which led at
1955 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1956 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1957 caused the error; without that there was no problem.
1958
1959 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1960
1961 19. Added --enable-pcretest-libreadline.
1962
1963 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1964 RE::GlobalReplace(). As a result, the number of replacements returned was
1965 double what it should be. I removed one of the increments, but Craig sent a
1966 later patch that removed the other one (the right fix) and added unit tests
1967 that check the return values (which was not done before).
1968
1969 21. Several CMake things:
1970
1971 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1972 the names libpcre and libpcreposix, not just pcre and pcreposix.
1973
1974 (2) The above change means that pcretest and pcregrep are now correctly
1975 linked with the newly-built libraries, not previously installed ones.
1976
1977 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1978
1979 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1980 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1981 UTF-8 newline character). The key issue is that the pattern starts .*;
1982 this means that the match must be either at the beginning, or after a
1983 newline. The bug was in the code for advancing after a failed match and
1984 checking that the new position followed a newline. It was not taking
1985 account of UTF-8 characters correctly.
1986
1987 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1988 character classes. PCRE was not treating the sequence [:...:] as a
1989 character class unless the ... were all letters. Perl, however, seems to
1990 allow any characters between [: and :], though of course it rejects as
1991 unknown any "names" that contain non-letters, because all the known class
1992 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1993 for example, whereas PCRE did not - it did not recognize a POSIX character
1994 class. This seemed a bit dangerous, so the code has been changed to be
1995 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1996 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1997 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1998 Perl does, and where it didn't before.
1999
2000 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2001 Windows environments %n is disabled by default.
2002
2003
2004 Version 7.4 21-Sep-07
2005 ---------------------
2006
2007 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2008 means that a class such as [\s] counted as "explicit reference to CR or
2009 LF". That isn't really right - the whole point of the change was to try to
2010 help when there was an actual mention of one of the two characters. So now
2011 the change happens only if \r or \n (or a literal CR or LF) character is
2012 encountered.
2013
2014 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2015 of both had grown to the point where there were only 3 bits left.
2016 Fortunately, there was spare space in the data structure, and so I have
2017 moved the internal flags into a new 16-bit field to free up more option
2018 bits.
2019
2020 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2021 but did not set the internal JCHANGED flag - either of these is enough to
2022 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2023 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2024 start sets both bits.
2025
2026 4. Added options (at build time, compile time, exec time) to change \R from
2027 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2028
2029 5. doc/pcresyntax.html was missing from the distribution.
2030
2031 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2032 compatibility, even though it is no longer used.
2033
2034 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2035 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2036 windows.h file is present (where different names are used). [This was
2037 reversed later after testing - see 16 below.]
2038
2039 8. Changed all #include <config.h> to #include "config.h". There were also
2040 some further <pcre.h> cases that I changed to "pcre.h".
2041
2042 9. When pcregrep was used with the --colour option, it missed the line ending
2043 sequence off the lines that it output.
2044
2045 10. It was pointed out to me that arrays of string pointers cause lots of
2046 relocations when a shared library is dynamically loaded. A technique of
2047 using a single long string with a table of offsets can drastically reduce
2048 these. I have refactored PCRE in four places to do this. The result is
2049 dramatic:
2050
2051 Originally: 290
2052 After changing UCP table: 187
2053 After changing error message table: 43
2054 After changing table of "verbs" 36
2055 After changing table of Posix names 22
2056
2057 Thanks to the folks working on Gregex for glib for this insight.
2058
2059 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2060 unicode-properties was also set.
2061
2062 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2063
2064 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2065 checked only for CRLF.
2066
2067 14. Added casts to pcretest.c to avoid compiler warnings.
2068
2069 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2070
2071 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2072 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2073 entirely. This removes changes made in 7 above.
2074
2075 17. The CMake files have been updated, and there is now more information about
2076 building with CMake in the NON-UNIX-USE document.
2077
2078
2079 Version 7.3 28-Aug-07
2080 ---------------------
2081
2082 1. In the rejigging of the build system that eventually resulted in 7.1, the
2083 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2084 brackets there is not right, since it causes compilers to look for an
2085 installed pcre.h, not the version that is in the source that is being
2086 compiled (which of course may be different). I have changed it back to:
2087
2088 #include "pcre.h"
2089
2090 I have a vague recollection that the change was concerned with compiling in
2091 different directories, but in the new build system, that is taken care of
2092 by the VPATH setting the Makefile.
2093
2094 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2095 when the subject happened to end in the byte 0x85 (e.g. if the last
2096 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2097 characters but of course it shouldn't be taken as a newline when it is part
2098 of another character. The bug was that, for an unlimited repeat of . in
2099 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2100 characters when looking for a newline.
2101
2102 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2103
2104 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2105 in debug output.
2106
2107 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2108 long printing in the pcrecpp unittest when running under MinGW.
2109
2110 6. ESC_K was left out of the EBCDIC table.
2111
2112 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2113 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2114 limit also applies to "virtual nesting" when a pattern is recursive, and in
2115 this case 1000 isn't so big. I have been able to remove this limit at the
2116 expense of backing off one optimization in certain circumstances. Normally,
2117 when pcre_exec() would call its internal match() function recursively and
2118 immediately return the result unconditionally, it uses a "tail recursion"
2119 feature to save stack. However, when a subpattern that can match an empty
2120 string has an unlimited repetition quantifier, it no longer makes this
2121 optimization. That gives it a stack frame in which to save the data for
2122 checking that an empty string has been matched. Previously this was taken
2123 from the 1000-entry workspace that had been reserved. So now there is no
2124 explicit limit, but more stack is used.
2125
2126 8. Applied Daniel's patches to solve problems with the import/export magic
2127 syntax that is required for Windows, and which was going wrong for the
2128 pcreposix and pcrecpp parts of the library. These were overlooked when this
2129 problem was solved for the main library.
2130
2131 9. There were some crude static tests to avoid integer overflow when computing
2132 the size of patterns that contain repeated groups with explicit upper
2133 limits. As the maximum quantifier is 65535, the maximum group length was
2134 set at 30,000 so that the product of these two numbers did not overflow a
2135 32-bit integer. However, it turns out that people want to use groups that
2136 are longer than 30,000 bytes (though not repeat them that many times).
2137 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2138 made it possible to implement the integer overflow checks in a much more
2139 dynamic way, which I have now done. The artificial limitation on group
2140 length has been removed - we now have only the limit on the total length of
2141 the compiled pattern, which depends on the LINK_SIZE setting.
2142
2143 10. Fixed a bug in the documentation for get/copy named substring when
2144 duplicate names are permitted. If none of the named substrings are set, the
2145 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2146 empty string.
2147
2148 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2149 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2150 because the ] is interpreted as the first data character and the
2151 terminating ] is not found. PCRE has been made compatible with Perl in this
2152 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2153 cause memory overwriting.
2154
2155 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2156 string has been matched (to stop an infinite loop). It was not recognizing
2157 a conditional subpattern that could match an empty string if that
2158 subpattern was within another subpattern. For example, it looped when
2159 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2160 condition was not nested. This bug has been fixed.
2161
2162 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2163 past the start of the subject in the presence of bytes with the top bit
2164 set, for example "\x8aBCD".
2165
2166 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2167 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2168
2169 14. Optimized (?!) to (*FAIL).
2170
2171 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2172 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2173 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2174 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2175 does: it's just the validity check that is more restrictive.
2176
2177 16. Inserted checks for integer overflows during escape sequence (backslash)
2178 processing, and also fixed erroneous offset values for syntax errors during
2179 backslash processing.
2180
2181 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2182 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2183
2184 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2185 caused an overrun.
2186
2187 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2188 something other than just ASCII characters) inside a group that had an
2189 unlimited repeat caused a loop at compile time (while checking to see
2190 whether the group could match an empty string).
2191
2192 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2193 [\P{Any}] did so. (Error in the code for printing property names.)
2194
2195 21. An orphan \E inside a character class could cause a crash.
2196
2197 22. A repeated capturing bracket such as (A)? could cause a wild memory
2198 reference during compilation.
2199
2200 23. There are several functions in pcre_compile() that scan along a compiled
2201 expression for various reasons (e.g. to see if it's fixed length for look
2202 behind). There were bugs in these functions when a repeated \p or \P was
2203 present in the pattern. These operators have additional parameters compared
2204 with \d, etc, and these were not being taken into account when moving along
2205 the compiled data. Specifically:
2206
2207 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2208 length.
2209
2210 (b) An item such as \pL+ within a repeated group could cause crashes or
2211 loops.
2212
2213 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2214 "reference to non-existent subpattern" error.
2215
2216 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2217
2218 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2219 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2220
2221 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2222
2223 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2224 character were causing crashes (broken optimization).
2225
2226 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2227 \p or \P) caused a compile-time loop.
2228
2229 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2230 break. For example, the unstudied pattern [\r\n]A does not match the string
2231 "\r\nA" because change 7.0/46 below moves the current point on by two
2232 characters after failing to match at the start. However, the pattern \nA
2233 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2234 the same is true. There doesn't seem any very clean way out of this, but
2235 what I have chosen to do makes the common cases work: PCRE now takes note
2236 of whether there can be an explicit match for \r or \n anywhere in the
2237 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2238 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2239 pattern has explicit CR or LF references.
2240
2241 29. Added (*CR) etc for changing newline setting at start of pattern.
2242
2243
2244 Version 7.2 19-Jun-07
2245 ---------------------
2246
2247 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2248 which is apparently normally available under Windows.
2249
2250 2. Re-jig the pcregrep tests with different newline settings in an attempt
2251 to make them independent of the local environment's newline setting.
2252
2253 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2254
2255 4. Some of the "internals" tests were previously cut out when the link size
2256 was not 2, because the output contained actual offsets. The recent new
2257 "Z" feature of pcretest means that these can be cut out, making the tests
2258 usable with all link sizes.
2259
2260 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2261 stack recursion. This gives a massive performance boost under BSD, but just
2262 a small improvement under Linux. However, it saves one field in the frame
2263 in all cases.
2264
2265 6. Added more features from the forthcoming Perl 5.10:
2266
2267 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2268 recursion call. It refers to the nth most recently opened parentheses.
2269
2270 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2271 to be opened parentheses.
2272
2273 (c) Conditions that refer to capturing parentheses can be specified
2274 relatively, for example, (?(-2)... or (?(+3)...
2275
2276 (d) \K resets the start of the current match so that everything before
2277 is not part of it.
2278
2279 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2280
2281 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2282 reference syntax.
2283
2284 (g) (?| introduces a group in which the numbering of parentheses in each
2285 alternative starts with the same number.
2286
2287 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2288
2289 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2290 PCRE_INFO_JCHANGED.
2291
2292 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2293 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2294 for detecting groups that can match an empty string.
2295
2296 9. A pattern with a very large number of alternatives (more than several
2297 hundred) was running out of internal workspace during the pre-compile
2298 phase, where pcre_compile() figures out how much memory will be needed. A
2299 bit of new cunning has reduced the workspace needed for groups with
2300 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2301 workspace instead of running out of the 4096 that are available.
2302
2303 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2304
2305 11. Applied patch from Google to remove an optimization that didn't quite work.
2306 The report of the bug said:
2307
2308 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2309 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2310 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2311
2312 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2313 it matched the wrong number of bytes.
2314
2315
2316 Version 7.1 24-Apr-07
2317 ---------------------
2318
2319 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2320 that is more "standard", making use of automake and other Autotools. There
2321 is some re-arrangement of the files and adjustment of comments consequent
2322 on this.
2323
2324 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2325 for recursive directory scanning broke on some systems because the files
2326 are not scanned in any specific order and on different systems the order
2327 was different. A call to "sort" has been inserted into RunGrepTest for the
2328 approprate test as a short-term fix. In the longer term there may be an
2329 alternative.
2330
2331 3. I had an email from Eric Raymond about problems translating some of PCRE's
2332 man pages to HTML (despite the fact that I distribute HTML pages, some
2333 people do their own conversions for various reasons). The problems
2334 concerned the use of low-level troff macros .br and .in. I have therefore
2335 removed all such uses from the man pages (some were redundant, some could
2336 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2337 HTML has been updated to handle .nf/.fi and to complain if it encounters
2338 .br or .in.
2339
2340 4. Updated comments in configure.ac that get placed in config.h.in and also
2341 arranged for config.h to be included in the distribution, with the name
2342 config.h.generic, for the benefit of those who have to compile without
2343 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2344
2345 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2346 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2347 makevp.bat for the current PCRE, using the additional files
2348 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2349
2350 6. A Windows user reported a minor discrepancy with test 2, which turned out
2351 to be caused by a trailing space on an input line that had got lost in his
2352 copy. The trailing space was an accident, so I've just removed it.
2353
2354 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2355 that is needed.
2356
2357 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2358 as "const" (a) because they are and (b) because it helps the PHP
2359 maintainers who have recently made a script to detect big data structures
2360 in the php code that should be moved to the .rodata section. I remembered
2361 to update Builducptable as well, so it won't revert if ucptable.h is ever
2362 re-created.
2363
2364 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2365 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2366 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2367 support is not required. This saves 1.5-2K of code, which is important in
2368 some applications.
2369
2370 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2371 so as not to refer to the tables, even though these functions will never be
2372 called when UTF-8 support is disabled. Otherwise there are problems with a
2373 shared library.
2374
2375 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2376
2377 (a) It was defining its arguments as char * instead of void *.
2378
2379 (b) It was assuming that all moves were upwards in memory; this was true
2380 a long time ago when I wrote it, but is no longer the case.
2381
2382 The emulated memove() is provided for those environments that have neither
2383 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2384 is clearly not the case, as these two bugs were recently reported.
2385
2386 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2387 and Detrail to create the HTML documentation, the .txt form of the man
2388 pages, and it removes trailing spaces from listed files. It also creates
2389 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2390 case, it wraps all the #defines with #ifndefs. This script should be run
2391 before "make dist".
2392
2393 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2394 with Unicode property support.
2395
2396 (a) For a maximizing quantifier, if the two different cases of the
2397 character were of different lengths in their UTF-8 codings (there are
2398 some cases like this - I found 11), and the matching function had to
2399 back up over a mixture of the two cases, it incorrectly assumed they
2400 were both the same length.
2401
2402 (b) When PCRE was configured to use the heap rather than the stack for
2403 recursion during matching, it was not correctly preserving the data for
2404 the other case of a UTF-8 character when checking ahead for a match
2405 while processing a minimizing repeat. If the check also involved
2406 matching a wide character, but failed, corruption could cause an
2407 erroneous result when trying to check for a repeat of the original
2408 character.
2409
2410 13. Some tidying changes to the testing mechanism:
2411
2412 (a) The RunTest script now detects the internal link size and whether there
2413 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2414 values substituted by "configure". (The RunGrepTest script already did
2415 this for UTF-8.) The configure.ac script no longer substitutes the
2416 relevant variables.
2417
2418 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2419 with length and offset values. This means that the output is different
2420 for different internal link sizes. Test 2 is skipped for link sizes
2421 other than 2 because of this, bypassing the problem. Unfortunately,
2422 there was also a test in test 3 (the locale tests) that used /B and
2423 failed for link sizes other than 2. Rather than cut the whole test out,
2424 I have added a new /Z option to pcretest that replaces the length and
2425 offset values with spaces. This is now used to make test 3 independent
2426 of link size. (Test 2 will be tidied up later.)
2427
2428 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2429 segmentation fault instead of returning the appropriate error message.
2430
2431 15. In multiline mode when the newline sequence was set to "any", the pattern
2432 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2433 This doesn't seem right; it now treats the CRLF combination as the line
2434 ending, and so does not match in that case. It's only a pattern such as ^$
2435 that would hit this one: something like ^ABC$ would have failed after \r
2436 and then tried again after \r\n.
2437
2438 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2439 in an attempt to make files that differ only in their line terminators
2440 compare equal. This works on Linux.
2441
2442 17. Under certain error circumstances pcregrep might try to free random memory
2443 as it exited. This is now fixed, thanks to valgrind.
2444
2445 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2446 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2447 was because its rules for how to advance for /g after matching an empty
2448 string at the end of a line did not allow for this case. They now check for
2449 it specially.
2450
2451 20. pcretest is supposed to handle patterns and data of any length, by
2452 extending its buffers when necessary. It was getting this wrong when the
2453 buffer for a data line had to be extended.
2454
2455 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2456 CRLF as a newline sequence.
2457
2458 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2459 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2460 I have nevertheless tidied it up.
2461
2462 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2463
2464 24. Added a man page for pcre-config.
2465
2466
2467 Version 7.0 19-Dec-06
2468 ---------------------
2469
2470 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2471 moving to gcc 4.1.1.
2472
2473 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2474 sys/time.h, which is documented as needed for this function. It doesn't
2475 seem to matter on Linux, but it showed up on some releases of OS X.
2476
2477 3. It seems that there are systems where bytes whose values are greater than
2478 127 match isprint() in the "C" locale. The "C" locale should be the
2479 default when a C program starts up. In most systems, only ASCII printing
2480 characters match isprint(). This difference caused the output from pcretest
2481 to vary, making some of the tests fail. I have changed pcretest so that:
2482
2483 (a) When it is outputting text in the compiled version of a pattern, bytes
2484 other than 32-126 are always shown as hex escapes.
2485
2486 (b) When it is outputting text that is a matched part of a subject string,
2487 it does the same, unless a different locale has been set for the match
2488 (using the /L modifier). In this case, it uses isprint() to decide.
2489
2490 4. Fixed a major bug that caused incorrect computation of the amount of memory
2491 required for a compiled pattern when options that changed within the
2492 pattern affected the logic of the preliminary scan that determines the
2493 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2494 that the computed length was too small. The symptoms of this bug were
2495 either the PCRE error "internal error: code overflow" from pcre_compile(),
2496 or a glibc crash with a message such as "pcretest: free(): invalid next
2497 size (fast)". Examples of patterns that provoked this bug (shown in
2498 pcretest format) are:
2499
2500 /(?-x: )/x
2501 /(?x)(?-x: \s*#\s*)/
2502 /((?i)[\x{c0}])/8
2503 /(?i:[\x{c0}])/8
2504
2505 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2506 is now done differently.
2507
2508 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2509 wrapper classes; (b) implement a new function in the C++ scanner that is
2510 more efficient than the old way of doing things because it avoids levels of
2511 recursion in the regex matching; (c) add a paragraph to the documentation
2512 for the FullMatch() function.
2513
2514 6. The escape sequence \n was being treated as whatever was defined as
2515 "newline". Not only was this contrary to the documentation, which states
2516 that \n is character 10 (hex 0A), but it also went horribly wrong when
2517 "newline" was defined as CRLF. This has been fixed.
2518
2519 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2520 was being set to -1 for the "end of line" case (supposedly a value that no
2521 character can have). Though this value is never used (the check for end of
2522 line is "zero bytes in current character"), it caused compiler complaints.
2523 I've changed it to 0xffffffff.
2524
2525 8. In pcre_version.c, the version string was being built by a sequence of
2526 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2527 string (as it is for production releases) called a macro with an empty
2528 argument. The C standard says the result of this is undefined. The gcc
2529 compiler treats it as an empty string (which was what was wanted) but it is
2530 reported that Visual C gives an error. The source has been hacked around to
2531 avoid this problem.
2532
2533 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2534 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2535 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2536 of them did).
2537
2538 10. Originally, pcretest opened its input and output without "b"; then I was
2539 told that "b" was needed in some environments, so it was added for release
2540 5.0 to both the input and output. (It makes no difference on Unix-like
2541 systems.) Later I was told that it is wrong for the input on Windows. I've
2542 now abstracted the modes into two macros, to make it easier to fiddle with
2543 them, and removed "b" from the input mode under Windows.
2544
2545 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2546
2547 12. Added -help and --help to pcretest as an official way of being reminded
2548 of the options.
2549
2550 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2551 and pcrecpp.cc because they annoy compilers at high warning levels.
2552
2553 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2554
2555 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2556 scripts are not C programs :-) and which was not noticed because it works
2557 on Linux.
2558
2559 16. pcretest is supposed to handle any length of pattern and data line (as one
2560 line or as a continued sequence of lines) by extending its input buffer if
2561 necessary. This feature was broken for very long pattern lines, leading to
2562 a string of junk being passed to pcre_compile() if the pattern was longer
2563 than about 50K.
2564
2565 17. I have done a major re-factoring of the way pcre_compile() computes the
2566 amount of memory needed for a compiled pattern. Previously, there was code
2567 that made a preliminary scan of the pattern in order to do this. That was
2568 OK when PCRE was new, but as the facilities have expanded, it has become
2569 harder and harder to keep it in step with the real compile phase, and there
2570 have been a number of bugs (see for example, 4 above). I have now found a
2571 cunning way of running the real compile function in a "fake" mode that
2572 enables it to compute how much memory it would need, while actually only
2573 ever using a few hundred bytes of working memory and without too many
2574 tests of the mode. This should make future maintenance and development
2575 easier. A side effect of this work is that the limit of 200 on the nesting
2576 depth of parentheses has been removed (though this was never a serious
2577 limitation, I suspect). However, there is a downside: pcre_compile() now
2578 runs more slowly than before (30% or more, depending on the pattern). I
2579 hope this isn't a big issue. There is no effect on runtime performance.
2580
2581 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2582 newline (only possible for the last line of a file) and it was a
2583 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2584
2585 19. Added additional timing features to pcretest. (1) The -tm option now times
2586 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2587 separate command line item, by a number that specifies the number of
2588 repeats to use when timing. The default is 50000; this gives better
2589 precision, but takes uncomfortably long for very large patterns.
2590
2591 20. Extended pcre_study() to be more clever in cases where a branch of a
2592 subpattern has no definite first character. For example, (a*|b*)[cd] would
2593 previously give no result from pcre_study(). Now it recognizes that the
2594 first character must be a, b, c, or d.
2595
2596 21. There was an incorrect error "recursive call could loop indefinitely" if
2597 a subpattern (or the entire pattern) that was being tested for matching an
2598 empty string contained only one non-empty item after a nested subpattern.
2599 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2600 incorrectly, because the \d was being skipped in the check.
2601
2602 22. The pcretest program now has a new pattern option /B and a command line
2603 option -b, which is equivalent to adding /B to every pattern. This causes
2604 it to show the compiled bytecode, without the additional information that
2605 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2606 is the same as /B/I).
2607
2608 23. A new optimization is now able automatically to treat some sequences such
2609 as a*b as a*+b. More specifically, if something simple (such as a character
2610 or a simple class like \d) has an unlimited quantifier, and is followed by
2611 something that cannot possibly match the quantified thing, the quantifier
2612 is automatically "possessified".
2613
2614 24. A recursive reference to a subpattern whose number was greater than 39
2615 went wrong under certain circumstances in UTF-8 mode. This bug could also
2616 have affected the operation of pcre_study().
2617
2618 25. Realized that a little bit of performance could be had by replacing
2619 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2620
2621 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2622
2623 27. Possessive quantifiers such as a++ were previously implemented by turning
2624 them into atomic groups such as ($>a+). Now they have their own opcodes,
2625 which improves performance. This includes the automatically created ones
2626 from 23 above.
2627
2628 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2629 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2630 the first matched character to be a colon. This applied both to named and
2631 numbered groups.
2632
2633 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2634
2635 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2636 building PCRE on Windows easier, so I have included it in the distribution.
2637
2638 31. There is now a check in pcretest against a ridiculously large number being
2639 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2640 loop, the loop is abandoned.
2641
2642 32. Forward references to subpatterns in conditions such as (?(2)...) where
2643 subpattern 2 is defined later cause pcre_compile() to search forwards in
2644 the pattern for the relevant set of parentheses. This search went wrong
2645 when there were unescaped parentheses in a character class, parentheses
2646 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2647
2648 33. "Subroutine" calls and backreferences were previously restricted to
2649 referencing subpatterns earlier in the regex. This restriction has now
2650 been removed.
2651
2652 34. Added a number of extra features that are going to be in Perl 5.10. On the
2653 whole, these are just syntactic alternatives for features that PCRE had
2654 previously implemented using the Python syntax or my own invention. The
2655 other formats are all retained for compatibility.
2656
2657 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2658 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2659 also .NET compatible.
2660
2661 (b) A recursion or subroutine call to a named group can now be defined as
2662 (?&name) as well as (?P>name).
2663
2664 (c) A backreference to a named group can now be defined as \k<name> or
2665 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2666 5.10, are also .NET compatible.
2667
2668 (d) A conditional reference to a named group can now use the syntax
2669 (?(<name>) or (?('name') as well as (?(name).
2670
2671 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2672 groups (named and numbered) that are never evaluated inline, but can be
2673 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2674 is always false. There may be only one alternative in such a group.
2675
2676 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2677 as the simple (?(R). The condition is true only if the most recent
2678 recursion is that of the given number or name. It does not search out
2679 through the entire recursion stack.
2680
2681 (g) The escape \gN or \g{N} has been added, where N is a positive or
2682 negative number, specifying an absolute or relative reference.
2683
2684 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2685 some "unreachable code" warnings.
2686
2687 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2688 things, this adds five new scripts.
2689
2690 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2691 There were also incompatibilities regarding the handling of \Q..\E inside
2692 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2693 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2694
2695 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2696 matches an empty string, and forcibly breaks the loop. There were bugs in
2697 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2698 against aaaa the result was just "a" rather than "aaaa", for example. Two
2699 separate and independent bugs (that affected different cases) have been
2700 fixed.
2701
2702 39. Refactored the code to abolish the use of different opcodes for small
2703 capturing bracket numbers. This is a tidy that I avoided doing when I
2704 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2705 The new approach is not only tidier, it makes it possible to reduce the
2706 memory needed to fix the previous bug (38).
2707
2708 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2709 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2710 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2711 mode.
2712
2713 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2714 report.
2715
2716 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2717 copy construction and assignment in the C++ wrapper.
2718
2719 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2720 couple of bugs that could have given wrong results in the "--newline=crlf"
2721 case.
2722
2723 44. Added a number of casts and did some reorganization of signed/unsigned int
2724 variables following suggestions from Dair Grant. Also renamed the variable
2725 "this" as "item" because it is a C++ keyword.
2726
2727 45. Arranged for dftables to add
2728
2729 #include "pcre_internal.h"
2730
2731 to pcre_chartables.c because without it, gcc 4.x may remove the array
2732 definition from the final binary if PCRE is built into a static library and
2733 dead code stripping is activated.
2734
2735 46. For an unanchored pattern, if a match attempt fails at the start of a
2736 newline sequence, and the newline setting is CRLF or ANY, and the next two
2737 characters are CRLF, advance by two characters instead of one.
2738
2739
2740 Version 6.7 04-Jul-06
2741 ---------------------
2742
2743 1. In order to handle tests when input lines are enormously long, pcretest has
2744 been re-factored so that it automatically extends its buffers when
2745 necessary. The code is crude, but this _is_ just a test program. The
2746 default size has been increased from 32K to 50K.
2747
2748 2. The code in pcre_study() was using the value of the re argument before
2749 testing it for NULL. (Of course, in any sensible call of the function, it
2750 won't be NULL.)
2751
2752 3. The memmove() emulation function in pcre_internal.h, which is used on
2753 systems that lack both memmove() and bcopy() - that is, hardly ever -
2754 was missing a "static" storage class specifier.
2755
2756 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2757 containing an extended class (one that cannot be represented by a bitmap
2758 because it contains high-valued characters or Unicode property items, e.g.
2759 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2760 pattern, but PCRE should not loop if you do not (it no longer does).
2761 [Detail: two cases were found: (a) a repeated subpattern containing an
2762 extended class; (b) a recursive reference to a subpattern that followed a
2763 previous extended class. It wasn't skipping over the extended class
2764 correctly when UTF-8 mode was not set.]
2765
2766 5. A negated single-character class was not being recognized as fixed-length
2767 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2768 compile error "lookbehind assertion is not fixed length".
2769
2770 6. The RunPerlTest auxiliary script was showing an unexpected difference
2771 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2772 write a Perl script that can interpret lines of an input file either as
2773 byte characters or as UTF-8, which is what "perltest" was being required to
2774 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2775 can't do is switch easily at run time between having the "use utf8;" pragma
2776 or not. In the end, I fudged it by using the RunPerlTest script to insert
2777 "use utf8;" explicitly for the UTF-8 tests.
2778
2779 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2780 the end of the subject string, contrary to the documentation and to what
2781 Perl does. This was true of both matching functions. Now it matches only at
2782 the start of the subject and immediately after *internal* newlines.
2783
2784 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2785 a pointer to an int instead of a pointer to an unsigned long int. This
2786 caused problems on 64-bit systems.
2787
2788 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2789 instance of the 'standard' template library not being so standard".
2790
2791 10. There was no check on the number of named subpatterns nor the maximum
2792 length of a subpattern name. The product of these values is used to compute
2793 the size of the memory block for a compiled pattern. By supplying a very
2794 long subpattern name and a large number of named subpatterns, the size
2795 computation could be caused to overflow. This is now prevented by limiting
2796 the length of names to 32 characters, and the number of named subpatterns
2797 to 10,000.
2798
2799 11. Subpatterns that are repeated with specific counts have to be replicated in
2800 the compiled pattern. The size of memory for this was computed from the
2801 length of the subpattern and the repeat count. The latter is limited to
2802 65535, but there was no limit on the former, meaning that integer overflow
2803 could in principle occur. The compiled length of a repeated subpattern is
2804 now limited to 30,000 bytes in order to prevent this.
2805
2806 12. Added the optional facility to have named substrings with the same name.
2807
2808 13. Added the ability to use a named substring as a condition, using the
2809 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2810 are numbers (not recommended). Forward references are permitted.
2811
2812 14. Added forward references in named backreferences (if you see what I mean).
2813
2814 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2815 pattern could run off the end of the subject. For example, the pattern
2816 "(?s)(.{1,5})"8 did this with the subject "ab".
2817
2818 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2819 PCRE_CASELESS was set when matching characters that were quantified with ?
2820 or *.
2821
2822 17. A character class other than a single negated character that had a minimum
2823 but no maximum quantifier - for example [ab]{6,} - was not handled
2824 correctly by pce_dfa_exec(). It would match only one character.
2825
2826 18. A valid (though odd) pattern that looked like a POSIX character
2827 class but used an invalid character after [ (for example [[,abc,]]) caused
2828 pcre_compile() to give the error "Failed: internal error: code overflow" or
2829 in some cases to crash with a glibc free() error. This could even happen if
2830 the pattern terminated after [[ but there just happened to be a sequence of
2831 letters, a binary zero, and a closing ] in the memory that followed.
2832
2833 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2834 over the years. Originally (before any Unicode support), just the bottom 8
2835 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2836 output from "man perlunicode" includes this:
2837
2838 The regular expression compiler produces polymorphic opcodes. That
2839 is, the pattern adapts to the data and automatically switches to
2840 the Unicode character scheme when presented with Unicode data--or
2841 instead uses a traditional byte scheme when presented with byte
2842 data.
2843
2844 Sadly, a wide octal escape does not cause a switch, and in a string with
2845 no other multibyte characters, these octal escapes are treated as before.
2846 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2847 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2848 Unicode string.
2849
2850 I have not perpetrated such confusion in PCRE. Up till now, it took just
2851 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2852 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2853 translate to the appropriate multibyte character.
2854
2855 29. Applied some refactoring to reduce the number of warnings from Microsoft
2856 and Borland compilers. This has included removing the fudge introduced
2857 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2858 a warning about an unused variable.
2859
2860 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2861 characters since release 4.0, because Perl (from release 5.004) does not.
2862 [Or at least, is documented not to: some releases seem to be in conflict
2863 with the documentation.] However, when a pattern was studied with
2864 pcre_study() and all its branches started with \s, PCRE still included VT
2865 as a possible starting character. Of course, this did no harm; it just
2866 caused an unnecessary match attempt.
2867
2868 22. Removed a now-redundant internal flag bit that recorded the fact that case
2869 dependency changed within the pattern. This was once needed for "required
2870 byte" processing, but is no longer used. This recovers a now-scarce options
2871 bit. Also moved the least significant internal flag bit to the most-
2872 significant bit of the word, which was not previously used (hangover from
2873 the days when it was an int rather than a uint) to free up another bit for
2874 the future.
2875
2876 23. Added support for CRLF line endings as well as CR and LF. As well as the
2877 default being selectable at build time, it can now be changed at runtime
2878 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2879 specify that it is scanning data with non-default line endings.
2880
2881 24. Changed the definition of CXXLINK to make it agree with the definition of
2882 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2883
2884 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2885 recursions. This makes a big different to stack usage for some patterns.
2886
2887 26. If a subpattern containing a named recursion or subroutine reference such
2888 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2889 the space required for the compiled pattern went wrong and gave too small a
2890 value. Depending on the environment, this could lead to "Failed: internal
2891 error: code overflow at offset 49" or "glibc detected double free or
2892 corruption" errors.
2893
2894 27. Applied patches from Google (a) to support the new newline modes and (b) to
2895 advance over multibyte UTF-8 characters in GlobalReplace.
2896
2897 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2898 difference for some implementation of PCRE in some Windows version.
2899
2900 29. Added some extra testing facilities to pcretest:
2901
2902 \q<number> in a data line sets the "match limit" value
2903 \Q<number> in a data line sets the "match recursion limt" value
2904 -S <number> sets the stack size, where <number> is in megabytes
2905
2906 The -S option isn't available for Windows.
2907
2908
2909 Version 6.6 06-Feb-06
2910 ---------------------
2911
2912 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2913 in pcreposix.h. I have copied the definition from pcre.h.
2914
2915 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2916 because pcre.h is no longer a built file.
2917
2918 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2919 not normally included in the compiled code.
2920
2921
2922 Version 6.5 01-Feb-06
2923 ---------------------
2924
2925 1. When using the partial match feature with pcre_dfa_exec(), it was not
2926 anchoring the second and subsequent partial matches at the new starting
2927 point. This could lead to incorrect results. For example, with the pattern
2928 /1234/, partially matching against "123" and then "a4" gave a match.
2929
2930 2. Changes to pcregrep:
2931
2932 (a) All non-match returns from pcre_exec() were being treated as failures
2933 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2934 error message is output. Some extra information is given for the
2935 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2936 probably the only errors that are likely to be caused by users (by
2937 specifying a regex that has nested indefinite repeats, for instance).
2938 If there are more than 20 of these errors, pcregrep is abandoned.
2939
2940 (b) A binary zero was treated as data while matching, but terminated the
2941 output line if it was written out. This has been fixed: binary zeroes
2942 are now no different to any other data bytes.
2943
2944 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2945 used to set a locale for matching. The --locale=xxxx long option has
2946 been added (no short equivalent) to specify a locale explicitly on the
2947 pcregrep command, overriding the environment variables.
2948
2949 (d) When -B was used with -n, some line numbers in the output were one less
2950 than they should have been.
2951
2952 (e) Added the -o (--only-matching) option.
2953
2954 (f) If -A or -C was used with -c (count only), some lines of context were
2955 accidentally printed for the final match.
2956
2957 (g) Added the -H (--with-filename) option.
2958
2959 (h) The combination of options -rh failed to suppress file names for files
2960 that were found from directory arguments.
2961
2962 (i) Added the -D (--devices) and -d (--directories) options.
2963
2964 (j) Added the -F (--fixed-strings) option.
2965
2966 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2967
2968 (l) Added the --colo(u)r option.
2969
2970 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2971 is not present by default.
2972
2973 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2974 items such as (?R) or (?1), when the recursion could match a number of
2975 alternatives. If it matched one of the alternatives, but subsequently,
2976 outside the recursion, there was a failure, the code tried to back up into
2977 the recursion. However, because of the way PCRE is implemented, this is not
2978 possible, and the result was an incorrect result from the match.
2979
2980 In order to prevent this happening, the specification of recursion has
2981 been changed so that all such subpatterns are automatically treated as
2982 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2983
2984 4. I had overlooked the fact that, in some locales, there are characters for
2985 which isalpha() is true but neither isupper() nor islower() are true. In
2986 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2987 and ordfeminine) are like this. This affected the treatment of \w and \W
2988 when they appeared in character classes, but not when they appeared outside
2989 a character class. The bit map for "word" characters is now created
2990 separately from the results of isalnum() instead of just taking it from the
2991 upper, lower, and digit maps. (Plus the underscore character, of course.)
2992
2993 5. The above bug also affected the handling of POSIX character classes such as
2994 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2995 permanent tables. Instead, the bit maps for such a class were previously
2996 created as the appropriate unions of the upper, lower, and digit bitmaps.
2997 Now they are created by subtraction from the [[:word:]] class, which has
2998 its own bitmap.
2999
3000 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3001 It is created by subtracting the vertical space characters (\x09, \x0a,
3002 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3003 subtraction was done in the overall bitmap for a character class, meaning
3004 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3005 be recognized. This bug has been fixed.
3006
3007 7. Patches from the folks at Google:
3008
3009 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3010 real life, but is still worth protecting against".
3011
3012 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3013 regular expressions".
3014
3015 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3016 have it.
3017
3018 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3019 "configure" and the latter not, in order to fix a problem somebody had
3020 with compiling the Arg class on HP-UX.
3021
3022 (e) Improve the error-handling of the C++ wrapper a little bit.
3023
3024 (f) New tests for checking recursion limiting.
3025
3026 8. The pcre_memmove() function, which is used only if the environment does not
3027 have a standard memmove() function (and is therefore rarely compiled),
3028 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3029 returning a result (though PCRE never actually uses the result).
3030
3031 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3032 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3033 returned instead of calling malloc() with an overflowing number that would
3034 most likely cause subsequent chaos.
3035
3036 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3037
3038 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3039 with this option is matched, the nmatch and pmatch options of regexec() are
3040 ignored.
3041
3042 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3043 provided in case anyone wants to the the POSIX interface with UTF-8
3044 strings.
3045
3046 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3047 C++ linking (needed for some HP-UX environments).
3048
3049 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3050 (unused parameter) and in the pcre_printint() function (omitted "default"
3051 switch label when the default is to do nothing).
3052
3053 15. Added some code to make it possible, when PCRE is compiled as a C++
3054 library, to replace subject pointers for pcre_exec() with a smart pointer
3055 class, thus making it possible to process discontinuous strings.
3056
3057 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3058 much the same function. They were added by different people who were trying
3059 to make PCRE easy to compile on non-Unix systems. It has been suggested
3060 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3061 for compiling on Windows systems. I have therefore replaced it with
3062 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3063 defaults to "extern" for C or "extern C" for C++, which works fine on
3064 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3065 SCOPE with something explicit in config.h. In addition:
3066
3067 (a) pcreposix.h still had just "extern" instead of either of these macros;
3068 I have replaced it with PCRE_DATA_SCOPE.
3069
3070 (b) Functions such as _pcre_xclass(), which are internal to the library,
3071 but external in the C sense, all had PCRE_EXPORT in their definitions.
3072 This is apparently wrong for the Windows case, so I have removed it.
3073 (It makes no difference on Unix-like systems.)
3074
3075 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3076 of recursive calls to match(). This is different to MATCH_LIMIT because
3077 that limits the total number of calls to match(), not all of which increase
3078 the depth of recursion. Limiting the recursion depth limits the amount of
3079 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3080 when PCRE is compiled, and changed at run time. A patch from Google adds
3081 this functionality to the C++ interface.
3082
3083 18. Changes to the handling of Unicode character properties:
3084
3085 (a) Updated the table to Unicode 4.1.0.
3086
3087 (b) Recognize characters that are not in the table as "Cn" (undefined).
3088
3089 (c) I revised the way the table is implemented to a much improved format
3090 which includes recognition of ranges. It now supports the ranges that
3091 are defined in UnicodeData.txt, and it also amalgamates other
3092 characters into ranges. This has reduced the number of entries in the
3093 table from around 16,000 to around 3,000, thus reducing its size
3094 considerably. I realized I did not need to use a tree structure after
3095 all - a binary chop search is just as efficient. Having reduced the
3096 number of entries, I extended their size from 6 bytes to 8 bytes to
3097 allow for more data.
3098
3099 (d) Added support for Unicode script names via properties such as \p{Han}.
3100
3101 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3102 matching that character.
3103
3104 20. When matching a repeated Unicode property with a minimum greater than zero,
3105 (for example \pL{2,}), PCRE could look past the end of the subject if it
3106 reached it while seeking the minimum number of characters. This could
3107 happen only if some of the characters were more than one byte long, because
3108 there is a check for at least the minimum number of bytes.
3109
3110 21. Refactored the implementation of \p and \P so as to be more general, to
3111 allow for more different types of property in future. This has changed the
3112 compiled form incompatibly. Anybody with saved compiled patterns that use
3113 \p or \P will have to recompile them.
3114
3115 22. Added "Any" and "L&" to the supported property types.
3116
3117 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3118 but give a compile time error if the value is greater than 0xff.
3119
3120 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3121 accidentally not being installed or uninstalled.
3122
3123 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3124 made were to insert the current release number. This seemed silly, because
3125 it made things harder for people building PCRE on systems that don't run
3126 "configure". I have turned pcre.h into a distributed file, no longer built
3127 by "configure", with the version identification directly included. There is
3128 no longer a pcre.h.in file.
3129
3130 However, this change necessitated a change to the pcre-config script as
3131 well. It is built from pcre-config.in, and one of the substitutions was the
3132 release number. I have updated configure.ac so that ./configure now finds
3133 the release number by grepping pcre.h.
3134
3135 26. Added the ability to run the tests under valgrind.
3136
3137
3138 Version 6.4 05-Sep-05
3139 ---------------------
3140
3141 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3142 "--" to be printed when multiple files were scanned, even when none of the
3143 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3144 consider it to be a bug, and have restored the previous behaviour.
3145
3146 2. A couple of code tidies to get rid of compiler warnings.
3147
3148 3. The pcretest program used to cheat by referring to symbols in the library
3149 whose names begin with _pcre_. These are internal symbols that are not
3150 really supposed to be visible externally, and in some environments it is
3151 possible to suppress them. The cheating is now confined to including
3152 certain files from the library's source, which is a bit cleaner.
3153
3154 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3155 file's purpose clearer.
3156
3157 5. Reorganized pcre_ucp_findchar().
3158
3159
3160 Version 6.3 15-Aug-05
3161 ---------------------
3162
3163 1. The file libpcre.pc.in did not have general read permission in the tarball.
3164
3165 2. There were some problems when building without C++ support:
3166
3167 (a) If C++ support was not built, "make install" and "make test" still
3168 tried to test it.
3169
3170 (b) There were problems when the value of CXX was explicitly set. Some
3171 changes have been made to try to fix these, and ...
3172
3173 (c) --disable-cpp can now be used to explicitly disable C++ support.
3174
3175 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3176 backslash in a target when C++ was disabled. This confuses some
3177 versions of "make", apparently. Using an intermediate variable solves
3178 this. (Same for CPP_LOBJ.)
3179
3180 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3181 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3182 necessary on certain architectures.
3183
3184 4. Added a setting of -export-symbols-regex to the link command to remove
3185 those symbols that are exported in the C sense, but actually are local
3186 within the library, and not documented. Their names all begin with
3187 "_pcre_". This is not a perfect job, because (a) we have to except some
3188 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3189 available (and never for static libraries). I have made a note to try to
3190 find a way round (a) in the future.
3191
3192
3193 Version 6.2 01-Aug-05
3194 ---------------------
3195
3196 1. There was no test for integer overflow of quantifier values. A construction
3197 such as {1111111111111111} would give undefined results. What is worse, if
3198 a minimum quantifier for a parenthesized subpattern overflowed and became
3199 negative, the calculation of the memory size went wrong. This could have
3200 led to memory overwriting.
3201
3202 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3203
3204 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3205 operating environments where this matters.
3206
3207 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3208 PCRE options from within the C++ wrapper.
3209
3210 5. Named capturing subpatterns were not being correctly counted when a pattern
3211 was compiled. This caused two problems: (a) If there were more than 100
3212 such subpatterns, the calculation of the memory needed for the whole
3213 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3214 back references of the form \12, where the number was greater than 9, were
3215 not recognized as back references, even though there were sufficient
3216 previous subpatterns.
3217
3218 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3219 versions of gcc, e.g. 2.95.4.
3220
3221
3222 Version 6.1 21-Jun-05
3223 ---------------------
3224
3225 1. There was one reference to the variable "posix" in pcretest.c that was not
3226 surrounded by "#if !defined NOPOSIX".
3227
3228 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3229 the cross-check on the old pcre_info() function, for the benefit of the
3230 cut-down version of PCRE that is currently imported into Exim.
3231
3232 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3233 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3234 patterns that start (?i) but I don't think that matters. The use of (?i) is
3235 just an example; this all applies to the other options as well.
3236
3237 4. Since libtool seems to echo the compile commands it is issuing, the output
3238 from "make" can be reduced a bit by putting "@" in front of each libtool
3239 compile command.
3240
3241 5. Patch from the folks at Google for configure.in to be a bit more thorough
3242 in checking for a suitable C++ installation before trying to compile the
3243 C++ stuff. This should fix a reported problem when a compiler was present,
3244 but no suitable headers.
3245
3246 6. The man pages all had just "PCRE" as their title. I have changed them to
3247 be the relevant file name. I have also arranged that these names are
3248 retained in the file doc/pcre.txt, which is a concatenation in text format
3249 of all the man pages except the little individual ones for each function.
3250
3251 7. The NON-UNIX-USE file had not been updated for the different set of source
3252 files that come with release 6. I also added a few comments about the C++
3253 wrapper.
3254
3255
3256 Version 6.0 07-Jun-05
3257 ---------------------
3258
3259 1. Some minor internal re-organization to help with my DFA experiments.
3260
3261 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3262 didn't matter for the library itself when fully configured, but did matter
3263 when compiling without UCP support, or within Exim, where the ucp files are
3264 not imported.
3265
3266 3. Refactoring of the library code to split up the various functions into
3267 different source modules. The addition of the new DFA matching code (see
3268 below) to a single monolithic source would have made it really too
3269 unwieldy, quite apart from causing all the code to be include in a
3270 statically linked application, when only some functions are used. This is
3271 relevant even without the DFA addition now that patterns can be compiled in
3272 one application and matched in another.
3273
3274 The downside of splitting up is that there have to be some external
3275 functions and data tables that are used internally in different modules of
3276 the library but which are not part of the API. These have all had their
3277 names changed to start with "_pcre_" so that they are unlikely to clash
3278 with other external names.
3279
3280 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3281 a different (DFA) algorithm. Although it is slower than the original
3282 function, it does have some advantages for certain types of matching
3283 problem.
3284
3285 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3286 including restarting after a partial match.
3287
3288 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3289 defined when compiling for Windows was sent to me. I have put it into the
3290 code, though I have no means of testing or verifying it.
3291
3292 7. Added the pcre_refcount() auxiliary function.
3293
3294 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3295 match before or at the first newline in the subject string. In pcretest,
3296 the /f option on a pattern can be used to set this.
3297
3298 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3299 would behave wrongly. This has been present in PCRE since release 4.0.
3300
3301 10. A number of changes to the pcregrep command:
3302
3303 (a) Refactored how -x works; insert ^(...)$ instead of setting
3304 PCRE_ANCHORED and checking the length, in preparation for adding
3305 something similar for -w.
3306
3307 (b) Added the -w (match as a word) option.
3308
3309 (c) Refactored the way lines are read and buffered so as to have more
3310 than one at a time available.
3311
3312 (d) Implemented a pcregrep test script.
3313
3314 (e) Added the -M (multiline match) option. This allows patterns to match
3315 over several lines of the subject. The buffering ensures that at least
3316 8K, or the rest of the document (whichever is the shorter) is available
3317 for matching (and similarly the previous 8K for lookbehind assertions).
3318
3319 (f) Changed the --help output so that it now says
3320
3321 -w, --word-regex(p)
3322
3323 instead of two lines, one with "regex" and the other with "regexp"
3324 because that confused at least one person since the short forms are the
3325 same. (This required a bit of code, as the output is generated
3326 automatically from a table. It wasn't just a text change.)
3327
3328 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3329 option but starts with a hyphen. Could be a pattern or a path name
3330 starting with a hyphen, for instance.
3331
3332 (h) "-" can be given as a file name to represent stdin.
3333
3334 (i) When file names are being printed, "(standard input)" is used for
3335 the standard input, for compatibility with GNU grep. Previously
3336 "<stdin>" was used.
3337
3338 (j) The option --label=xxx can be used to supply a name to be used for
3339 stdin when file names are being printed. There is no short form.
3340
3341 (k) Re-factored the options decoding logic because we are going to add
3342 two more options that take data. Such options can now be given in four
3343 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3344
3345 (l) Added the -A, -B, and -C options for requesting that lines of context
3346 around matches be printed.
3347
3348 (m) Added the -L option to print the names of files that do not contain
3349 any matching lines, that is, the complement of -l.
3350
3351 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3352 continue to scan other files.
3353
3354 (o) The -s option was incorrectly implemented. For compatibility with other
3355 greps, it now suppresses the error message for a non-existent or non-
3356 accessible file (but not the return code). There is a new option called
3357 -q that suppresses the output of matching lines, which was what -s was
3358 previously doing.
3359
3360 (p) Added --include and --exclude options to specify files for inclusion
3361 and exclusion when recursing.
3362
3363 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3364 Hopefully, it now does.
3365
3366 12. Missing cast in pcre_study().
3367
3368 13. Added an "uninstall" target to the makefile.
3369
3370 14. Replaced "extern" in the function prototypes in Makefile.in with
3371 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3372 world, but is set differently for Windows.
3373
3374 15. Added a second compiling function called pcre_compile2(). The only
3375 difference is that it has an extra argument, which is a pointer to an
3376 integer error code. When there is a compile-time failure, this is set
3377 non-zero, in addition to the error test pointer being set to point to an
3378 error message. The new argument may be NULL if no error number is required
3379 (but then you may as well call pcre_compile(), which is now just a
3380 wrapper). This facility is provided because some applications need a
3381 numeric error indication, but it has also enabled me to tidy up the way
3382 compile-time errors are handled in the POSIX wrapper.
3383
3384 16. Added VPATH=.libs to the makefile; this should help when building with one
3385 prefix path and installing with another. (Or so I'm told by someone who
3386 knows more about this stuff than I do.)
3387
3388 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3389 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3390 match everything, including newlines. This is not POSIX-compatible, but
3391 somebody wanted the feature. From pcretest it can be activated by using
3392 both the P and the s flags.
3393
3394 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3395
3396 19. libpcre.pc was being incorrectly installed as executable.
3397
3398 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3399 it now also looks for '\r' so that it will work unmodified on Windows.
3400
3401 21. Added Google's contributed C++ wrapper to the distribution.
3402
3403 22. Added some untidy missing memory free() calls in pcretest, to keep
3404 Electric Fence happy when testing.
3405
3406
3407
3408 Version 5.0 13-Sep-04
3409 ---------------------
3410
3411 1. Internal change: literal characters are no longer packed up into items
3412 containing multiple characters in a single byte-string. Each character
3413 is now matched using a separate opcode. However, there may be more than one
3414 byte in the character in UTF-8 mode.
3415
3416 2. The pcre_callout_block structure has two new fields: pattern_position and
3417 next_item_length. These contain the offset in the pattern to the next match
3418 item, and its length, respectively.
3419
3420 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3421 insertion of callouts before each pattern item. Added the /C option to
3422 pcretest to make use of this.
3423
3424 4. On the advice of a Windows user, the lines
3425
3426 #if defined(_WIN32) || defined(WIN32)
3427 _setmode( _fileno( stdout ), 0x8000 );
3428 #endif /* defined(_WIN32) || defined(WIN32) */
3429
3430 have been added to the source of pcretest. This apparently does useful
3431 magic in relation to line terminators.
3432
3433 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3434 for the benefit of those environments where the "b" makes a difference.
3435
3436 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3437 to know about it. I have put a hack into configure.in that adds in code
3438 to set GCC=yes if CC=icc. This seems to end up at a point in the
3439 generated configure script that is early enough to affect the setting of
3440 compiler options, which is what is needed, but I have no means of testing
3441 whether it really works. (The user who reported this had patched the
3442 generated configure script, which of course I cannot do.)
3443
3444 LATER: After change 22 below (new libtool files), the configure script
3445 seems to know about icc (and also ecc). Therefore, I have commented out
3446 this hack in configure.in.
3447
3448 7. Added support for pkg-config (2 patches were sent in).
3449
3450 8. Negated POSIX character classes that used a combination of internal tables
3451 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3452 [[:^ascii]]. Typically, they would match almost any characters. The other
3453 POSIX classes were not broken in this way.
3454
3455 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3456 to find the match, as PCRE was deluded into thinking that the match had to
3457 start at the start point or following a newline. The same bug applied to
3458 patterns with negative forward assertions or any backward assertions
3459 preceding ".*" at the start, unless the pattern required a fixed first
3460 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3461
3462 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3463 starting at the last subject character, bytes beyond the end of the subject
3464 string were read.
3465
3466 11. Renamed the variable "class" as "classbits" to make life easier for C++
3467 users. (Previously there was a macro definition, but it apparently wasn't
3468 enough.)
3469
3470 12. Added the new field "tables" to the extra data so that tables can be passed
3471 in at exec time, or the internal tables can be re-selected. This allows
3472 a compiled regex to be saved and re-used at a later time by a different
3473 program that might have everything at different addresses.
3474
3475 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3476 -R library as well as a -L library.
3477
3478 14. The debugging options of pcretest (-d on the command line or D on a
3479 pattern) showed incorrect output for anything following an extended class
3480 that contained multibyte characters and which was followed by a quantifier.
3481
3482 15. Added optional support for general category Unicode character properties
3483 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3484 support. It adds about 90K to the size of the library. The meanings of the
3485 inbuilt class escapes such as \d and \s have NOT been changed.
3486
3487 16. Updated pcredemo.c to include calls to free() to release the memory for the
3488 compiled pattern.
3489
3490 17. The generated file chartables.c was being created in the source directory
3491 instead of in the building directory. This caused the build to fail if the
3492 source directory was different from the building directory, and was
3493 read-only.
3494
3495 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3496 file. No doubt somebody will tell me if they don't make sense... Also added
3497 Dan Mooney's comments about building on OpenVMS.
3498
3499 19. Added support for partial matching via the PCRE_PARTIAL option for
3500 pcre_exec() and the \P data escape in pcretest.
3501
3502 20. Extended pcretest with 3 new pattern features:
3503
3504 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3505 write the compiled pattern to the file whose name is "rest-of-line".
3506 This is a straight binary dump of the data, with the saved pointer to
3507 the character tables forced to be NULL. The study data, if any, is
3508 written too. After writing, pcretest reads a new pattern.
3509
3510 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3511 compiled pattern from the given file. There must not be any
3512 occurrences of "<" in the file name (pretty unlikely); if there are,
3513 pcretest will instead treat the initial "<" as a pattern delimiter.
3514 After reading in the pattern, pcretest goes on to read data lines as
3515 usual.
3516
3517 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3518 and 16-bit fields in a compiled pattern, to simulate a pattern that
3519 was compiled on a host of opposite endianness.
3520
3521 21. The pcre-exec() function can now cope with patterns that were compiled on
3522 hosts of opposite endianness, with this restriction:
3523
3524 As for any compiled expression that is saved and used later, the tables
3525 pointer field cannot be preserved; the extra_data field in the arguments
3526 to pcre_exec() should be used to pass in a tables address if a value
3527 other than the default internal tables were used at compile time.
3528
3529 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3530 now diagnosed as an error. Previously, most of the time, a negative number
3531 would have been treated as zero, but if in addition "ovector" was passed as
3532 NULL, a crash could occur.
3533
3534 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3535 new versions from the libtool 1.5 distribution (the last one is a copy of
3536 a file called libtool.m4). This seems to have fixed the need to patch
3537 "configure" to support Darwin 1.3 (which I used to do). However, I still
3538 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3539 workstation).
3540
3541 24. Changed the PCRE licence to be the more standard "BSD" licence.
3542
3543
3544 Version 4.5 01-Dec-03
3545 ---------------------
3546
3547 1. There has been some re-arrangement of the code for the match() function so
3548 that it can be compiled in a version that does not call itself recursively.
3549 Instead, it keeps those local variables that need separate instances for
3550 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3551 needs to "recurse". Keeping track of where control must go is done by means
3552 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3553 hide most of the details from the main code, and operates only if
3554 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3555 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3556 operating.
3557
3558 To make it easier for callers to provide specially tailored get/free
3559 functions for this usage, two new functions, pcre_stack_malloc, and
3560 pcre_stack_free, are used. They are always called in strict stacking order,
3561 and the size of block requested is always the same.
3562
3563 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3564 PCRE has been compiled to use the stack or the heap for recursion. The
3565 -C option of pcretest uses this to show which version is compiled.
3566
3567 A new data escape \S, is added to pcretest; it causes the amounts of store
3568 obtained and freed by both kinds of malloc/free at match time to be added
3569 to the output.
3570
3571 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3572 what's available on my current Linux desktop machine.
3573
3574 3. When matching a UTF-8 string, the test for a valid string at the start has
3575 been extended. If start_offset is not zero, PCRE now checks that it points
3576 to a byte that is the start of a UTF-8 character. If not, it returns
3577 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3578 this is necessary because there may be backward assertions in the pattern.
3579 When matching the same subject several times, it may save resources to use
3580 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3581
3582 4. The code for checking the validity of UTF-8 strings has been tightened so
3583 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3584 containing "overlong sequences".
3585
3586 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3587 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3588 should have been "&", but it just so happened that all the cases this let
3589 through by mistake were picked up later in the function.
3590
3591 6. I had used a variable called "isblank" - this is a C99 function, causing
3592 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3593
3594 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3595 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3596 so the version is shown; (c) stop "make test" echoing "./RunTest".
3597
3598 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3599
3600 9. The prototype for memmove() for systems that don't have it was using
3601 size_t, but the inclusion of the header that defines size_t was later. I've
3602 moved the #includes for the C headers earlier to avoid this.
3603
3604 10. Added some adjustments to the code to make it easier to compiler on certain
3605 special systems:
3606
3607 (a) Some "const" qualifiers were missing.
3608 (b) Added the macro EXPORT before all exported functions; by default this
3609 is defined to be empty.
3610 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3611 that it reads its output file name as an argument instead of writing
3612 to the standard output and assuming this can be redirected.
3613
3614 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3615 class containing characters with values greater than 255, PCRE compilation
3616 went into a loop.
3617
3618 12. A recursive reference to a subpattern that was within another subpattern
3619 that had a minimum quantifier of zero caused PCRE to crash. For example,
3620 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3621 recursion. If the recursively-called subpattern itself had a zero repeat,
3622 that was OK.
3623
3624 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3625 buffer into which it was copied (for escape processing) was still set at
3626 1024, so long lines caused crashes.
3627
3628 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3629 "internal error: code overflow...". This applied to any character class
3630 that was followed by a possessive quantifier.
3631
3632 15. Modified the Makefile to add libpcre.la as a prerequisite for
3633 libpcreposix.la because I was told this is needed for a parallel build to
3634 work.
3635
3636 16. If a pattern that contained .* following optional items at the start was
3637 studied, the wrong optimizing data was generated, leading to matching
3638 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3639 matching string must start with a or b or c. The correct conclusion for
3640 this pattern is that a match can start with any character.
3641
3642
3643 Version 4.4 13-Aug-03
3644 ---------------------
3645
3646 1. In UTF-8 mode, a character class containing characters with values between
3647 127 and 255 was not handled correctly if the compiled pattern was studied.
3648 In fixing this, I have also improved the studying algorithm for such
3649 classes (slightly).
3650
3651 2. Three internal functions had redundant arguments passed to them. Removal
3652 might give a very teeny performance improvement.
3653
3654 3. Documentation bug: the value of the capture_top field in a callout is *one
3655 more than* the number of the hightest numbered captured substring.
3656
3657 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3658 in incorrectly linking with a previously installed version. They now link
3659 explicitly with libpcre.la.
3660
3661 5. configure.in no longer needs to recognize Cygwin specially.
3662
3663 6. A problem in pcre.in for Windows platforms is fixed.
3664
3665 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3666 pcretest, it used to include the size of the study block as part of its
3667 output. Unfortunately, the structure contains a field that has a different
3668 size on different hardware architectures. This meant that the tests that
3669 showed this size failed. As the block is currently always of a fixed size,
3670 this information isn't actually particularly useful in pcretest output, so
3671 I have just removed it.
3672
3673 8. Three pre-processor statements accidentally did not start in column 1.
3674 Sadly, there are *still* compilers around that complain, even though
3675 standard C has not required this for well over a decade. Sigh.
3676
3677 9. In pcretest, the code for checking callouts passed small integers in the
3678 callout_data field, which is a void * field. However, some picky compilers
3679 complained about the casts involved for this on 64-bit systems. Now
3680 pcretest passes the address of the small integer instead, which should get
3681 rid of the warnings.
3682
3683 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3684 both compile and run time, and gives an error if an invalid UTF-8 sequence
3685 is found. There is a option for disabling this check in cases where the
3686 string is known to be correct and/or the maximum performance is wanted.
3687
3688 11. In response to a bug report, I changed one line in Makefile.in from
3689
3690 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3691 to
3692 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3693
3694 to look similar to other lines, but I have no way of telling whether this
3695 is the right thing to do, as I do not use Windows. No doubt I'll get told
3696 if it's wrong...
3697
3698
3699 Version 4.3 21-May-03
3700 ---------------------
3701
3702 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3703 Makefile.
3704
3705 2. Some refactoring to improve the quality of the code:
3706
3707 (i) The utf8_table... variables are now declared "const".
3708
3709 (ii) The code for \cx, which used the "case flipping" table to upper case
3710 lower case letters, now just substracts 32. This is ASCII-specific,
3711 but the whole concept of \cx is ASCII-specific, so it seems
3712 reasonable.
3713
3714 (iii) PCRE was using its character types table to recognize decimal and
3715 hexadecimal digits in the pattern. This is silly, because it handles
3716 only 0-9, a-f, and A-F, but the character types table is locale-
3717 specific, which means strange things might happen. A private
3718 table is now used for this - though it costs 256 bytes, a table is
3719 much faster than multiple explicit tests. Of course, the standard
3720 character types table is still used for matching digits in subject
3721 strings against \d.
3722
3723 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3724 ending in _t are). So I've renamed it as ESC_tee.
3725
3726 3. The first argument for regexec() in the POSIX wrapper should have been
3727 defined as "const".
3728
3729 4. Changed pcretest to use malloc() for its buffers so that they can be
3730 Electric Fenced for debugging.
3731
3732 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3733 to read one or more bytes before the start of the subject string. Often this
3734 had no effect on PCRE's behaviour, but in some circumstances it could
3735 provoke a segmentation fault.
3736
3737 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3738 to try to read one or more bytes before the start of the subject string.
3739
3740 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3741 UTF-8 support could misbehave in various ways if the subject string
3742 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3743 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3744 back over UTF-8 characters.)
3745
3746
3747 Version 4.2 14-Apr-03
3748 ---------------------
3749
3750 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3751
3752 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3753 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3754 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3755 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3756 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3757 and BUILD_EXEEXT
3758 Note: automatic setting of the BUILD variables is not yet working
3759 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3760 compile-time but not at link-time
3761 [LINK]: use for linking executables only
3762 make different versions for Windows and non-Windows
3763 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3764 libraries
3765 [LINK_FOR_BUILD]: new variable
3766 [OBJEXT]: use throughout
3767 [EXEEXT]: use throughout
3768 <winshared>: new target
3769 <wininstall>: new target
3770 <dftables.o>: use native compiler
3771 <dftables>: use native linker
3772 <install>: handle Windows platform correctly
3773 <clean>: ditto
3774 <check>: ditto
3775 copy DLL to top builddir before testing
3776
3777 As part of these changes, -no-undefined was removed again. This was reported
3778 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3779 in any case.
3780
3781 3. Some tidies to get rid of compiler warnings:
3782
3783 . In the match_data structure, match_limit was an unsigned long int, whereas
3784 match_call_count was an int. I've made them both unsigned long ints.
3785
3786 . In pcretest the fact that a const uschar * doesn't automatically cast to
3787 a void * provoked a warning.
3788
3789 . Turning on some more compiler warnings threw up some "shadow" variables
3790 and a few more missing casts.
3791
3792 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3793 option, a class that contained a single character with a value between 128
3794 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3795
3796 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3797 option, a class that contained several characters, but with at least one
3798 whose value was between 128 and 255 caused PCRE to crash.
3799
3800
3801 Version 4.1 12-Mar-03
3802 ---------------------
3803
3804 1. Compiling with gcc -pedantic found a couple of places where casts were
3805 needed, and a string in dftables.c that was longer than standard compilers are
3806 required to support.
3807
3808 2. Compiling with Sun's compiler found a few more places where the code could
3809 be tidied up in order to avoid warnings.
3810
3811 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3812 first of these names is deprecated in the latest Autoconf in favour of the name
3813 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3814 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3815 analogy I have changed it to CFLAGS_FOR_BUILD.
3816
3817 4. Added -no-undefined to the linking command in the Makefile, because this is
3818 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3819 linking step for the pcreposix library.
3820
3821 5. PCRE was failing to diagnose the case of two named groups with the same
3822 name.
3823
3824 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3825 literal character that is needed in the subject for a match, and scans along to
3826 ensure that it is present before embarking on the full matching process. This
3827 saves time in cases of nested unlimited repeats that are never going to match.
3828 Problem: the scan can take a lot of time if the subject is very long (e.g.
3829 megabytes), thus penalizing straightforward matches. It is now done only if the
3830 amount of subject to be scanned is less than 1000 bytes.
3831
3832 7. A lesser problem with the same optimization is that it was recording the
3833 first character of an anchored pattern as "needed", thus provoking a search
3834 right along the subject, even when the first match of the pattern was going to
3835 fail. The "needed" character is now not set for anchored patterns, unless it
3836 follows something in the pattern that is of non-fixed length. Thus, it still
3837 fulfils its original purpose of finding quick non-matches in cases of nested
3838 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3839
3840
3841 Version 4.0 17-Feb-03
3842 ---------------------
3843
3844 1. If a comment in an extended regex that started immediately after a meta-item
3845 extended to the end of string, PCRE compiled incorrect data. This could lead to
3846 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3847
3848 2. Moved to autoconf 2.53 and libtool 1.4.2.
3849
3850 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3851 the special perltest8 script is no longer needed - all the tests can be run
3852 from a single perltest script.
3853
3854 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3855 by \s. It has now been removed in PCRE. This means it isn't recognized as
3856 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3857 class [:space:] *does* include VT, thereby creating a mess.
3858
3859 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3860 space and tab.
3861
3862 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3863 its new features into the main test script, reducing the number of scripts.
3864
3865 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3866 were backward compatible, and made the (?i) apply to the whole pattern, as if
3867 /i were given. Now it behaves more logically, and applies the option setting
3868 only to what follows. PCRE has been changed to follow suit. However, if it
3869 finds options settings right at the start of the pattern, it extracts them into
3870 the global options, as before. Thus, they show up in the info data.
3871
3872 8. Added support for the \Q...\E escape sequence. Characters in between are
3873 treated as literals. This is slightly different from Perl in that $ and @ are
3874 also handled as literals inside the quotes. In Perl, they will cause variable
3875 interpolation. Note the following examples:
3876
3877 Pattern PCRE matches Perl matches
3878
3879 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3880 \Qabc\$xyz\E abc\$xyz abc\$xyz
3881 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3882
3883 For compatibility with Perl, \Q...\E sequences are recognized inside character
3884 classes as well as outside them.
3885
3886 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3887 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3888 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3889 signed/unsigned warnings.
3890
3891 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3892 option for pcretest, so I've replaced it by a simple function that does just
3893 that job.
3894
3895 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3896 "pcregrep -".
3897
3898 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3899 Java package. This provides some syntactic sugar for simple cases of what my
3900 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3901 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3902 item, you can use this simplified notation. Note that only makes sense with
3903 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3904 greediness, whatever the setting of the PCRE_UNGREEDY option.
3905
3906 13. A change of greediness default within a pattern was not taking effect at
3907 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3908 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3909 was abstracted outside.
3910
3911 14. PCRE now supports the \G assertion. It is true when the current matching
3912 position is at the start point of the match. This differs from \A when the
3913 starting offset is non-zero. Used with the /g option of pcretest (or similar
3914 code), it works in the same way as it does for Perl's /g option. If all
3915 alternatives of a regex begin with \G, the expression is anchored to the start
3916 match position, and the "anchored" flag is set in the compiled expression.
3917
3918 15. Some bugs concerning the handling of certain option changes within patterns
3919 have been fixed. These applied to options other than (?ims). For example,
3920 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3921 been the other way round. Some of this was related to change 7 above.
3922
3923 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3924 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3925 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3926 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3927
3928 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3929 mode. Unlike ".", it always matches newline, whatever the setting of
3930 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3931 assertions. Perl allows it, but it doesn't (in general) work because it can't
3932 calculate the length of the lookbehind. At least, that's the case for Perl
3933 5.8.0 - I've been told they are going to document that it doesn't work in
3934 future.
3935
3936 18. Added an error diagnosis for escapes that PCRE does not support: these are
3937 \L, \l, \N, \P, \p, \U, \u, and \X.
3938
3939 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3940 reading past the end of the pattern in cases such as /[abcd/.
3941
3942 20. PCRE was getting more memory than necessary for patterns with classes that
3943 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3944
3945 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3946 compiling PCRE for use with Virtual Pascal.
3947
3948 22. Small fix to the Makefile to make it work properly if the build is done
3949 outside the source tree.
3950
3951 23. Added a new extension: a condition to go with recursion. If a conditional
3952 subpattern starts with (?(R) the "true" branch is used if recursion has
3953 happened, whereas the "false" branch is used only at the top level.
3954
3955 24. When there was a very long string of literal characters (over 255 bytes
3956 without UTF support, over 250 bytes with UTF support), the computation of how
3957 much memory was required could be incorrect, leading to segfaults or other
3958 strange effects.
3959
3960 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3961 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3962 there was a subsequent back reference to those brackets. This meant that, for
3963 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3964 possible to check for precisely this case. All we can do is abandon the
3965 optimization if .* occurs inside capturing brackets when there are any back
3966 references whatsoever. (See below for a better fix that came later.)
3967
3968 26. The handling of the optimization for finding the first character of a
3969 non-anchored pattern, and for finding a character that is required later in the
3970 match were failing in some cases. This didn't break the matching; it just
3971 failed to optimize when it could. The way this is done has been re-implemented.
3972
3973 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3974
3975 28. Added a new feature that provides some of the functionality that Perl
3976 provides with (?{...}). The facility is termed a "callout". The way it is done
3977 in PCRE is for the caller to provide an optional function, by setting
3978 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3979 global variable. By default it is unset, which disables all calling out. To get
3980 the function called, the regex must include (?C) at appropriate points. This
3981 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3982 This provides a means of identifying different callout points. When PCRE
3983 reaches such a point in the regex, if pcre_callout has been set, the external
3984 function is called. It is provided with data in a structure called
3985 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3986 matching continues; if it returns a non-zero value, the match at the current
3987 point fails. However, backtracking will occur if possible. [This was changed
3988 later and other features added - see item 49 below.]
3989
3990 29. pcretest is upgraded to test the callout functionality. It provides a
3991 callout function that displays information. By default, it shows the start of
3992 the match and the current position in the text. There are some new data escapes
3993 to vary what happens:
3994
3995 \C+ in addition, show current contents of captured substrings
3996 \C- do not supply a callout function
3997 \C!n return 1 when callout number n is reached
3998 \C!n!m return 1 when callout number n is reached for the mth time
3999
4000 30. If pcregrep was called with the -l option and just a single file name, it
4001 output "<stdin>" if a match was found, instead of the file name.
4002
4003 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4004 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4005 pcre_exec(). This saves a malloc/free per call. The default value of
4006 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4007 when configuring.
4008
4009 32. The default maximum size of a compiled pattern is 64K. There have been a
4010 few cases of people hitting this limit. The code now uses macros to handle the
4011 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4012 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4013 configuring. Tests 2 and 5 work only with 2-byte links because they output
4014 debugging information about compiled patterns.
4015
4016 33. Internal code re-arrangements:
4017
4018 (a) Moved the debugging function for printing out a compiled regex into
4019 its own source file (printint.c) and used #include to pull it into
4020 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4021 separate copies.
4022
4023 (b) Defined the list of op-code names for debugging as a macro in
4024 internal.h so that it is next to the definition of the opcodes.
4025
4026 (c) Defined a table of op-code lengths for simpler skipping along compiled
4027 code. This is again a macro in internal.h so that it is next to the
4028 definition of the opcodes.
4029
4030 34. Added support for recursive calls to individual subpatterns, along the
4031 lines of Robin Houston's patch (but implemented somewhat differently).
4032
4033 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4034 allow it to read and process whole directories in Win32. This code was
4035 contributed by Lionel Fourquaux; it has not been tested by me.
4036
4037 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4038 used to name a group. Names consist of alphanumerics and underscores, and must
4039 be unique. Back references use the syntax (?P=name) and recursive calls use
4040 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4041 numbers. The function pcre_fullinfo() can be used after compilation to extract
4042 a name/number map. There are three relevant calls:
4043
4044 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4045 PCRE_INFO_NAMECOUNT yields the number of entries
4046 PCRE_INFO_NAMETABLE yields a pointer to the map.
4047
4048 The map is a vector of fixed-size entries. The size of each entry depends on
4049 the length of the longest name used. The first two bytes of each entry are the
4050 group number, most significant byte first. There follows the corresponding
4051 name, zero terminated. The names are in alphabetical order.
4052
4053 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4054 case instead of 255. Making it the same both with and without UTF-8 support
4055 means that the same test output works with both.
4056
4057 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4058 calling malloc() with a zero argument.
4059
4060 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4061 optimization. I've improved things by keeping a bitmap of backreferences with
4062 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4063 fact referenced, the optimization can be applied. It is unlikely that a
4064 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4065 the match to follow \n) will appear inside brackets with a number greater than
4066 31, but if it does, any back reference > 31 suppresses the optimization.
4067
4068 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4069 of disabling numbered capturing parentheses. Any opening parenthesis that is
4070 not followed by ? behaves as if it were followed by ?: but named parentheses
4071 can still be used for capturing (and they will acquire numbers in the usual
4072 way).
4073
4074 41. Redesigned the return codes from the match() function into yes/no/error so
4075 that errors can be passed back from deep inside the nested calls. A malloc
4076 failure while inside a recursive subpattern call now causes the
4077 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4078
4079 42. It is now possible to set a limit on the number of times the match()
4080 function is called in a call to pcre_exec(). This facility makes it possible to
4081 limit the amount of recursion and backtracking, though not in a directly
4082 obvious way, because the match() function is used in a number of different
4083 circumstances. The count starts from zero for each position in the subject
4084 string (for non-anchored patterns). The default limit is, for compatibility, a
4085 large number, namely 10 000 000. You can change this in two ways:
4086
4087 (a) When configuring PCRE before making, you can use --with-match-limit=n
4088 to set a default value for the compiled library.
4089
4090 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4091 a different value is set. See 45 below.
4092
4093 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4094
4095 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4096 of things that can be changed at compile time. The first argument specifies
4097 what is wanted and the second points to where the information is to be placed.
4098 The current list of available information is:
4099
4100 PCRE_CONFIG_UTF8
4101
4102 The output is an integer that is set to one if UTF-8 support is available;
4103 otherwise it is set to zero.
4104
4105 PCRE_CONFIG_NEWLINE
4106
4107 The output is an integer that it set to the value of the code that is used for
4108 newline. It is either LF (10) or CR (13).
4109
4110 PCRE_CONFIG_LINK_SIZE
4111
4112 The output is an integer that contains the number of bytes used for internal
4113 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4114
4115 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4116
4117 The output is an integer that contains the threshold above which the POSIX
4118 interface uses malloc() for output vectors. See item 31 above.
4119
4120 PCRE_CONFIG_MATCH_LIMIT
4121
4122 The output is an unsigned integer that contains the default limit of the number
4123 of match() calls in a pcre_exec() execution. See 42 above.
4124
4125 44. pcretest has been upgraded by the addition of the -C option. This causes it
4126 to extract all the available output from the new pcre_config() function, and to
4127 output it. The program then exits immediately.
4128
4129 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4130 order to support additional features. One way would have been to define
4131 pcre_exec2() (for example) with extra arguments, but this would not have been
4132 extensible, and would also have required all calls to the original function to
4133 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4134 is used for passing in "extra" data from pcre_study().
4135
4136 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4137 contains the following fields:
4138
4139 flags a bitmap indicating which of the following fields are set
4140 study_data opaque data from pcre_study()
4141 match_limit a way of specifying a limit on match() calls for a specific
4142 call to pcre_exec()
4143 callout_data data for callouts (see 49 below)
4144
4145 The flag bits are also defined in pcre.h, and are
4146
4147 PCRE_EXTRA_STUDY_DATA
4148 PCRE_EXTRA_MATCH_LIMIT
4149 PCRE_EXTRA_CALLOUT_DATA
4150
4151 The pcre_study() function now returns one of these new pcre_extra blocks, with
4152 the actual study data pointed to by the study_data field, and the
4153 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4154 before. That is, this change is entirely upwards-compatible and requires no
4155 change to existing code.
4156
4157 If you want to pass in additional data to pcre_exec(), you can either place it
4158 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4159 block.
4160
4161 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4162 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4163 times with different match limits, until it finds the minimum value needed for
4164 pcre_exec() to complete. The value is then output. This can be instructive; for
4165 most simple matches the number is quite small, but for pathological cases it
4166 gets very large very quickly.
4167
4168 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4169 returns the size of the data block pointed to by the study_data field in a
4170 pcre_extra block, that is, the value that was passed as the argument to
4171 pcre_malloc() when PCRE was getting memory in which to place the information
4172 created by pcre_study(). The fourth argument should point to a size_t variable.
4173 pcretest has been extended so that this information is shown after a successful
4174 pcre_study() call when information about the compiled regex is being displayed.
4175
4176 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4177 because what follows is always an absolute path. (Later: it turns out that this
4178 is more than cosmetic for MinGW, because it doesn't like empty path
4179 components.)
4180
4181 49. Some changes have been made to the callout feature (see 28 above):
4182
4183 (i) A callout function now has three choices for what it returns:
4184
4185 0 => success, carry on matching
4186 > 0 => failure at this point, but backtrack if possible
4187 < 0 => serious error, return this value from pcre_exec()
4188
4189 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4190 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4191 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4192 use by callout functions. It will never be used by PCRE itself.
4193
4194 (ii) The pcre_extra structure (see 45 above) has a void * field called
4195 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4196 pcre_callout_block structure has a field of the same name. The contents of
4197 the field passed in the pcre_extra structure are passed to the callout
4198 function in the corresponding field in the callout block. This makes it
4199 easier to use the same callout-containing regex from multiple threads. For
4200 testing, the pcretest program has a new data escape
4201
4202 \C*n pass the number n (may be negative) as callout_data
4203
4204 If the callout function in pcretest receives a non-zero value as
4205 callout_data, it returns that value.
4206
4207 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4208 there were some redundant $(CFLAGS) in commands that are now specified as
4209 $(LINK), which already includes $(CFLAGS).
4210
4211 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4212 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4213 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4214 one-byte characters throughout. Note that case-insensitive matching applies
4215 only to characters whose values are less than 256. PCRE doesn't support the
4216 notion of cases for higher-valued characters.
4217
4218 (i) A character class whose characters are all within 0-255 is handled as
4219 a bit map, and the map is inverted for negative classes. Previously, a
4220 character > 255 always failed to match such a class; however it should
4221 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4222
4223 (ii) A negated character class with a single character < 255 is coded as
4224 "not this character" (OP_NOT). This wasn't working properly when the test
4225 character was multibyte, either singly or repeated.
4226
4227 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4228 mode, for example: \x{100}{2,3}.
4229
4230 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4231 singly or repeated) now correctly test multibyte characters. However,
4232 PCRE doesn't recognize any characters with values greater than 255 as
4233 digits, spaces, or word characters. Such characters always match \D, \S,
4234 and \W, and never match \d, \s, or \w.
4235
4236 (v) Classes may now contain characters and character ranges with values
4237 greater than 255. For example: [ab\x{100}-\x{400}].
4238
4239 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4240 PCRE in UTF-8 mode.
4241
4242 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4243 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4244 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4245 value.)
4246
4247 53. The single man page has become too large. I have therefore split it up into
4248 a number of separate man pages. These also give rise to individual HTML pages;
4249 these are now put in a separate directory, and there is an index.html page that
4250 lists them all. Some hyperlinking between the pages has been installed.
4251
4252 54. Added convenience functions for handling named capturing parentheses.
4253
4254 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4255 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4256 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4257 are faulted.
4258
4259 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4260 calling configure. These values are used when compiling the dftables.c program
4261 which is run to generate the source of the default character tables. They
4262 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4263 you will need to set these values.
4264
4265 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4266
4267
4268 Version 3.9 02-Jan-02
4269 ---------------------
4270
4271 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4272
4273 2. If --disable-static was given, the building process failed when trying to
4274 build pcretest and pcregrep. (For some reason it was using libtool to compile
4275 them, which is not right, as they aren't part of the library.)
4276
4277
4278 Version 3.8 18-Dec-01
4279 ---------------------
4280
4281 1. The experimental UTF-8 code was completely screwed up. It was packing the
4282 bytes in the wrong order. How dumb can you get?
4283
4284
4285 Version 3.7 29-Oct-01
4286 ---------------------
4287
4288 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4289 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4290 this didn't happen under Solaris 8, where I normally test things.
4291
4292 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4293 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4294 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4295 this fix an hour or so after the initial 3.7 release.)
4296
4297
4298 Version 3.6 23-Oct-01
4299 ---------------------
4300
4301 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4302 offsets passed as NULL with zero offset count.
4303
4304 2. The config.guess and config.sub files had not been updated when I moved to
4305 the latest autoconf.
4306
4307
4308 Version 3.5 15-Aug-01
4309 ---------------------
4310
4311 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4312 had been forgotten.
4313
4314 2. By using declared but undefined structures, we can avoid using "void"
4315 definitions in pcre.h while keeping the internal definitions of the structures
4316 private.
4317
4318 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4319 user point of view, this means that both static and shared libraries are built
4320 by default, but this can be individually controlled. More of the work of
4321 handling this static/shared cases is now inside libtool instead of PCRE's make
4322 file.
4323
4324 4. The pcretest utility is now installed along with pcregrep because it is
4325 useful for users (to test regexs) and by doing this, it automatically gets
4326 relinked by libtool. The documentation has been turned into a man page, so
4327 there are now .1, .txt, and .html versions in /doc.
4328
4329 5. Upgrades to pcregrep:
4330 (i) Added long-form option names like gnu grep.
4331 (ii) Added --help to list all options with an explanatory phrase.
4332 (iii) Added -r, --recursive to recurse into sub-directories.
4333 (iv) Added -f, --file to read patterns from a file.
4334
4335 6. pcre_exec() was referring to its "code" argument before testing that
4336 argument for NULL (and giving an error if it was NULL).
4337
4338 7. Upgraded Makefile.in to allow for compiling in a different directory from
4339 the source directory.
4340
4341 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4342 options bits, the pointer it was passed was to an int instead of to an unsigned
4343 long int. This mattered only on 64-bit systems.
4344
4345 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4346 generated) instead of pcre.in, which it its source. Also made the same change
4347 in several of the .c files.
4348
4349 10. A new release of gcc defines printf() as a macro, which broke pcretest
4350 because it had an ifdef in the middle of a string argument for printf(). Fixed
4351 by using separate calls to printf().
4352
4353 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4354 script, to force use of CR or LF instead of \n in the source. On non-Unix
4355 systems, the value can be set in config.h.
4356
4357 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4358 absolute limit. Changed the text of the error message to make this clear, and
4359 likewise updated the man page.
4360
4361 13. The limit of 99 on the number of capturing subpatterns has been removed.
4362 The new limit is 65535, which I hope will not be a "real" limit.
4363
4364
4365 Version 3.4 22-Aug-00
4366 ---------------------
4367
4368 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4369
4370 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4371
4372
4373 Version 3.3 01-Aug-00
4374 ---------------------
4375
4376 1. If an octal character was given, but the value was greater than \377, it
4377 was not getting masked to the least significant bits, as documented. This could
4378 lead to crashes in some systems.
4379
4380 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4381 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4382
4383 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4384 These just pass their arguments on to (pcre_free)(), but they are provided
4385 because some uses of PCRE bind it to non-C systems that can call its functions,
4386 but cannot call free() or pcre_free() directly.
4387
4388 4. Add "make test" as a synonym for "make check". Corrected some comments in
4389 the Makefile.
4390
4391 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4392 Makefile.
4393
4394 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4395 command called pgrep for grepping around the active processes.
4396
4397 7. Added the beginnings of support for UTF-8 character strings.
4398
4399 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4400 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4401 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4402 out for the ar command.)
4403
4404
4405 Version 3.2 12-May-00
4406 ---------------------
4407
4408 This is purely a bug fixing release.
4409
4410 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4411 of ZA. This was just one example of several cases that could provoke this bug,
4412 which was introduced by change 9 of version 2.00. The code for breaking
4413 infinite loops after an iteration that matches an empty string was't working
4414 correctly.
4415
4416 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4417 when matched against abbab (for example). After matching an empty string, it
4418 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4419 caused it to match further down the string than it should.
4420
4421 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4422 was there because it doesn't seem to be needed, and it causes trouble on some
4423 systems, as it is not a Standard C header. It has been removed.
4424
4425 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4426 were reported on the Macintosh. The changes were from
4427
4428 while ((c = *(++ptr)) != 0 && c != '\n');
4429 to
4430 while ((c = *(++ptr)) != 0 && c != '\n') ;
4431
4432 Totally extraordinary, but if that's what it takes...
4433
4434 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4435 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4436 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4437 assumes the way PCRE uses memmove() (always moving upwards).
4438
4439 6. PCRE is being used in one environment where strchr() is not available. There
4440 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4441 faster code anyway.
4442
4443
4444 Version 3.1 09-Feb-00
4445 ---------------------
4446
4447 The only change in this release is the fixing of some bugs in Makefile.in for
4448 the "install" target:
4449
4450 (1) It was failing to install pcreposix.h.
4451
4452 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4453
4454
4455 Version 3.0 01-Feb-00
4456 ---------------------
4457
4458 1. Add support for the /+ modifier to perltest (to output $` like it does in
4459 pcretest).
4460
4461 2. Add support for the /g modifier to perltest.
4462
4463 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4464 matches null strings.
4465
4466 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4467 pattern. Perl treats empty patterns specially - it reuses the most recent
4468 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4469 effect.
4470
4471 5. The POSIX interface was broken in that it was just handing over the POSIX
4472 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4473 required a bigger vector, with some working space on the end. This means that
4474 the POSIX wrapper now has to get and free some memory, and copy the results.
4475
4476 6. Added some simple autoconf support, placing the test data and the
4477 documentation in separate directories, re-organizing some of the
4478 information files, and making it build pcre-config (a GNU standard). Also added
4479 libtool support for building PCRE as a shared library, which is now the
4480 default.
4481
4482 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4483 09 are not valid octal constants. Single digits will be used for minor values
4484 less than 10.
4485
4486 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4487 existing programs that set these in the POSIX interface can use PCRE without
4488 modification.
4489
4490 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4491 return all that pcre_info() returns, plus additional data. The pcre_info()
4492 function is retained for compatibility, but is considered to be obsolete.
4493
4494 10. Added experimental recursion feature (?R) to handle one common case that
4495 Perl 5.6 will be able to do with (?p{...}).
4496
4497 11. Added support for POSIX character classes like [:alpha:], which Perl is
4498 adopting.
4499
4500
4501 Version 2.08 31-Aug-99
4502 ----------------------
4503
4504 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4505 trying to match at the startoffset position, but instead was moving forward to
4506 the next newline as if a previous match had failed.
4507
4508 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4509 and could get into a loop if a null string was matched other than at the start
4510 of the subject.
4511
4512 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4513 be distinguished at compile time, and for completeness also added PCRE_DATE.
4514
4515 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4516 in GnuWin32 environments.
4517
4518
4519 Version 2.07 29-Jul-99
4520 ----------------------
4521
4522 1. The documentation is now supplied in plain text form and HTML as well as in
4523 the form of man page sources.
4524
4525 2. C++ compilers don't like assigning (void *) values to other pointer types.
4526 In particular this affects malloc(). Although there is no problem in Standard
4527 C, I've put in casts to keep C++ compilers happy.
4528
4529 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4530 should be (const char *).
4531
4532 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4533 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4534 However, I haven't made this a standard facility. The documentation doesn't
4535 mention it, and the Makefile doesn't support it.
4536
4537 5. The Makefile now contains an "install" target, with editable destinations at
4538 the top of the file. The pcretest program is not installed.
4539
4540 6. pgrep -V now gives the PCRE version number and date.
4541
4542 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4543 causing the entire string to be ignored, instead of just the last character.
4544
4545 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4546 non-matching string, it can take a very, very long time, even for strings of
4547 quite modest length, because of the nested recursion. PCRE now does better in
4548 some of these cases. It does this by remembering the last required literal
4549 character in the pattern, and pre-searching the subject to ensure it is present
4550 before running the real match. In other words, it applies a heuristic to detect
4551 some types of certain failure quickly, and in the above example, if presented
4552 with a string that has no trailing " it gives "no match" very quickly.
4553
4554 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4555 other alternatives are tried instead.
4556
4557
4558 Version 2.06 09-Jun-99
4559 ----------------------
4560
4561 1. Change pcretest's output for amount of store used to show just the code
4562 space, because the remainder (the data block) varies in size between 32-bit and
4563 64-bit systems.
4564
4565 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4566 start matching at. This allows lookbehinds to work when searching for multiple
4567 occurrences in a string.
4568
4569 3. Added additional options to pcretest for testing multiple occurrences:
4570
4571 /+ outputs the rest of the string that follows a match
4572 /g loops for multiple occurrences, using the new startoffset argument
4573 /G loops for multiple occurrences by passing an incremented pointer
4574
4575 4. PCRE wasn't doing the "first character" optimization for patterns starting
4576 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4577 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4578 the letter 'x'. On long subject strings, this gives a significant speed-up.
4579
4580
4581 Version 2.05 21-Apr-99
4582 ----------------------
4583
4584 1. Changed the type of magic_number from int to long int so that it works
4585 properly on 16-bit systems.
4586
4587 2. Fixed a bug which caused patterns starting with .* not to work correctly
4588 when the subject string contained newline characters. PCRE was assuming
4589 anchoring for such patterns in all cases, which is not correct because .* will
4590 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4591 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4592 must be retried after every newline in the subject.
4593
4594
4595 Version 2.04 18-Feb-99
4596 ----------------------
4597
4598 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4599 computation of the store needed to hold the pattern was incorrect (too large).
4600 If such patterns were nested a few deep, this could multiply and become a real
4601 problem.
4602
4603 2. Added /M option to pcretest to show the memory requirement of a specific
4604 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4605
4606 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4607 compiled in such a way that the backtracking after subsequent failure was
4608 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4609 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4610
4611
4612 Version 2.03 02-Feb-99
4613 ----------------------
4614
4615 1. Fixed typo and small mistake in man page.
4616
4617 2. Added 4th condition (GPL supersedes if conflict) and created separate
4618 LICENCE file containing the conditions.
4619
4620 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4621 Perl, that is the internal \ allows the delimiter to be included in the
4622 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4623 the final delimiter, add \ to the end of the pattern (to test the error).
4624
4625 4. Added the convenience functions for extracting substrings after a successful
4626 match. Updated pcretest to make it able to test these functions.
4627
4628
4629 Version 2.02 14-Jan-99
4630 ----------------------
4631
4632 1. Initialized the working variables associated with each extraction so that
4633 their saving and restoring doesn't refer to uninitialized store.
4634
4635 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4636 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4637 fix the problem.
4638
4639 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4640 calls, and wasn't printing the correct value for compiling calls. Increased the
4641 default value of LOOPREPEAT, and the number of significant figures in the
4642 times.
4643
4644 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4645
4646 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4647 a building problem on Windows NT with a FAT file system.
4648
4649
4650 Version 2.01 21-Oct-98
4651 ----------------------
4652
4653 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4654 to character tables built by pcre_maketables() in the current locale. If NULL
4655 is passed, the default tables are used.
4656
4657
4658 Version 2.00 24-Sep-98
4659 ----------------------
4660
4661 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4662 it any more.
4663
4664 2. Allow quantification of (?>) groups, and make it work correctly.
4665
4666 3. The first character computation wasn't working for (?>) groups.
4667
4668 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4669 end of the subject) and add 5.005's \z, which really does match only at the
4670 very end of the subject.
4671
4672 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4673
4674 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4675 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4676 localized options. All options to pcre_study() were also removed.
4677
4678 7. Add other new features from 5.005:
4679
4680 $(?<= positive lookbehind
4681 $(?<! negative lookbehind
4682 (?imsx-imsx) added the unsetting capability
4683 such a setting is global if at outer level; local otherwise
4684 (?imsx-imsx:) non-capturing groups with option setting
4685 (?(cond)re|re) conditional pattern matching
4686
4687 A backreference to itself in a repeated group matches the previous
4688 captured string.
4689
4690 8. General tidying up of studying (both automatic and via "study")
4691 consequential on the addition of new assertions.
4692
4693 9. As in 5.005, unlimited repeated groups that could match an empty substring
4694 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4695 runtime if any iteration does actually match an empty substring.
4696
4697 10. Include the RunTest script in the distribution.
4698
4699 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4700 discrepancies, some of which were old and were also with respect to 5.004. They
4701 have now been fixed.
4702
4703
4704 Version 1.09 28-Apr-98
4705 ----------------------
4706
4707 1. A negated single character class followed by a quantifier with a minimum
4708 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4709 program crashes, or just wrong answers. This did not apply to negated classes
4710 containing more than one character, or to minima other than one.
4711
4712
4713 Version 1.08 27-Mar-98
4714 ----------------------
4715
4716 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4717
4718 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4719 latter must appear before anything that relies on it in the pattern.
4720
4721
4722 Version 1.07 16-Feb-98
4723 ----------------------
4724
4725 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4726 repeat of a potentially empty string).
4727
4728
4729 Version 1.06 23-Jan-98
4730 ----------------------
4731
4732 1. Added Markus Oberhumer's little patches for C++.
4733
4734 2. Literal strings longer than 255 characters were broken.
4735
4736
4737 Version 1.05 23-Dec-97
4738 ----------------------
4739
4740 1. Negated character classes containing more than one character were failing if
4741 PCRE_CASELESS was set at run time.
4742
4743
4744 Version 1.04 19-Dec-97
4745 ----------------------
4746
4747 1. Corrected the man page, where some "const" qualifiers had been omitted.
4748
4749 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4750 input syntax.
4751
4752 3. Fixed memory leak which occurred when a regex with back references was
4753 matched with an offsets vector that wasn't big enough. The temporary memory
4754 that is used in this case wasn't being freed if the match failed.
4755
4756 4. Tidied pcretest to ensure it frees memory that it gets.
4757
4758 5. Temporary memory was being obtained in the case where the passed offsets
4759 vector was exactly big enough.
4760
4761 6. Corrected definition of offsetof() from change 5 below.
4762
4763 7. I had screwed up change 6 below and broken the rules for the use of
4764 setjmp(). Now fixed.
4765
4766
4767 Version 1.03 18-Dec-97
4768 ----------------------
4769
4770 1. A erroneous regex with a missing opening parenthesis was correctly
4771 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4772 on some systems.
4773
4774 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4775 it was reported that one broken compiler failed on the former because "code" is
4776 also an independent variable.
4777
4778 3. The erroneous regex a[]b caused an array overrun reference.
4779
4780 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4781 fail on data ending with that character. (It was going on too far, and checking
4782 the next character, typically a binary zero.) This was specific to the
4783 optimized code for single-character negative classes.
4784
4785 5. Added a contributed patch from the TIN world which does the following:
4786
4787 + Add an undef for memmove, in case the the system defines a macro for it.
4788
4789 + Add a definition of offsetof(), in case there isn't one. (I don't know
4790 the reason behind this - offsetof() is part of the ANSI standard - but
4791 it does no harm).
4792
4793 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4794 most of the places where whitespace preceded '#'. I have given up and
4795 allowed the remaining 2 cases to be at the margin.
4796
4797 + Rename some variables in pcre to eliminate shadowing. This seems very
4798 pedantic, but does no harm, of course.
4799
4800 6. Moved the call to setjmp() into its own function, to get rid of warnings
4801 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4802
4803 7. Constructs such as \d{8,} were compiling into the equivalent of
4804 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4805 outcome, but in this particular case used more store than had been allocated,
4806 which caused the bug to be discovered because it threw up an internal error.
4807
4808 8. The debugging code in both pcre and pcretest for outputting the compiled
4809 form of a regex was going wrong in the case of back references followed by
4810 curly-bracketed repeats.
4811
4812
4813 Version 1.02 12-Dec-97
4814 ----------------------
4815
4816 1. Typos in pcre.3 and comments in the source fixed.
4817
4818 2. Applied a contributed patch to get rid of places where it used to remove
4819 'const' from variables, and fixed some signed/unsigned and uninitialized
4820 variable warnings.
4821
4822 3. Added the "runtest" target to Makefile.
4823
4824 4. Set default compiler flag to -O2 rather than just -O.
4825
4826
4827 Version 1.01 19-Nov-97
4828 ----------------------
4829
4830 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4831 like /([ab]*)*/, that is, for classes with more than one character in them.
4832
4833 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4834 as /((?>a*))*/ (a PCRE_EXTRA facility).
4835
4836
4837 Version 1.00 18-Nov-97
4838 ----------------------
4839
4840 1. Added compile-time macros to support systems such as SunOS4 which don't have
4841 memmove() or strerror() but have other things that can be used instead.
4842
4843 2. Arranged that "make clean" removes the executables.
4844
4845
4846 Version 0.99 27-Oct-97
4847 ----------------------
4848
4849 1. Fixed bug in code for optimizing classes with only one character. It was
4850 initializing a 32-byte map regardless, which could cause it to run off the end
4851 of the memory it had got.
4852
4853 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4854
4855
4856 Version 0.98 22-Oct-97
4857 ----------------------
4858
4859 1. Fixed bug in code for handling temporary memory usage when there are more
4860 back references than supplied space in the ovector. This could cause segfaults.
4861
4862
4863 Version 0.97 21-Oct-97
4864 ----------------------
4865
4866 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4867
4868 2. Optimized negated single characters not to use a bit map.
4869
4870 3. Brought error texts together as macro definitions; clarified some of them;
4871 fixed one that was wrong - it said "range out of order" when it meant "invalid
4872 escape sequence".
4873
4874 4. Changed some char * arguments to const char *.
4875
4876 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4877
4878 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4879 pcretest.
4880
4881
4882 Version 0.96 16-Oct-97
4883 ----------------------
4884
4885 1. Added a simple "pgrep" utility to the distribution.
4886
4887 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4888 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4889 where "ddd" means "one or more decimal digits".
4890
4891 3. Fixed serious bug. If a pattern had a back reference, but the call to
4892 pcre_exec() didn't supply a large enough ovector to record the related
4893 identifying subpattern, the match always failed. PCRE now remembers the number
4894 of the largest back reference, and gets some temporary memory in which to save
4895 the offsets during matching if necessary, in order to ensure that
4896 backreferences always work.
4897
4898 4. Increased the compatibility with Perl in a number of ways:
4899
4900 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4901 to request this handling. The option can be set at compile or exec time.
4902
4903 (b) $ matches before a terminating newline by default; an option
4904 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4905 mode). The option can be set at compile or exec time.
4906
4907 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4908 the same as Perl's. If the decimal number it represents is less than 10
4909 or there aren't that many previous left capturing parentheses, an octal
4910 escape is read. Inside a character class, it's always an octal escape,
4911 even if it is a single digit.
4912
4913 (d) An escaped but undefined alphabetic character is taken as a literal,
4914 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4915 escapes.
4916
4917 (e) {0} is now permitted. (The previous item is removed from the compiled
4918 pattern).
4919
4920 5. Changed all the names of code files so that the basic parts are no longer
4921 than 10 characters, and abolished the teeny "globals.c" file.
4922
4923 6. Changed the handling of character classes; they are now done with a 32-byte
4924 bit map always.
4925
4926 7. Added the -d and /D options to pcretest to make it possible to look at the
4927 internals of compilation without having to recompile pcre.
4928
4929
4930 Version 0.95 23-Sep-97
4931 ----------------------
4932
4933 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4934 \x20 at the start of a run of normal characters. These were being treated as
4935 real characters, instead of the source characters being re-checked.
4936
4937
4938 Version 0.94 18-Sep-97
4939 ----------------------
4940
4941 1. The functions are now thread-safe, with the caveat that the global variables
4942 containing pointers to malloc() and free() or alternative functions are the
4943 same for all threads.
4944
4945 2. Get pcre_study() to generate a bitmap of initial characters for non-
4946 anchored patterns when this is possible, and use it if passed to pcre_exec().
4947
4948
4949 Version 0.93 15-Sep-97
4950 ----------------------
4951
4952 1. /(b)|(:+)/ was computing an incorrect first character.
4953
4954 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4955 but not actually doing anything yet.
4956
4957 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4958 as Perl does (e.g. [-az] or [az-]).
4959
4960 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4961 all possible positions.
4962
4963 5. Split up into different modules to avoid including unneeded functions in a
4964 compiled binary. However, compile and exec are still in one module. The "study"
4965 function is split off.
4966
4967 6. The character tables are now in a separate module whose source is generated
4968 by an auxiliary program - but can then be edited by hand if required. There are
4969 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4970 toupper() in the code.
4971
4972 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4973 make them global. Abolish the function for setting them, as the caller can now
4974 set them directly.
4975
4976
4977 Version 0.92 11-Sep-97
4978 ----------------------
4979
4980 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4981 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4982
4983 2. Caseless matching was not working in character classes if the characters in
4984 the pattern were in upper case.
4985
4986 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4987
4988 4. Make PCRE_ANCHORED public and accept as a compile option.
4989
4990 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4991 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4992 pass them.
4993
4994 6. Give an error if bad option bits passed at compile or run time.
4995
4996 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4997 pcretest to cause it to pass that flag.
4998
4999 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5000 options, and the first character, if set.
5001
5002 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5003
5004
5005 Version 0.91 10-Sep-97
5006 ----------------------
5007
5008 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5009 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5010
5011 2. PCRE was looping on encountering an indefinitely repeated back reference to
5012 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5013 Perl does - treats the match as successful.
5014
5015 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5