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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1240 - (show annotations)
Wed Jan 23 16:36:41 2013 UTC (6 years, 8 months ago) by ph10
File size: 241177 byte(s)
Error occurred while calculating annotation data.
Deal with warnings given by automake 1.12.
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
36 Version 8.32 30-November-2012
37 -----------------------------
38
39 1. Improved JIT compiler optimizations for first character search and single
40 character iterators.
41
42 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
43 Patch by Daniel Richard G.
44
45 3. Single character iterator optimizations in the JIT compiler.
46
47 4. Improved JIT compiler optimizations for character ranges.
48
49 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
50 Reported by Giuseppe D'Angelo.
51
52 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
53 of a line, was being set incorrectly in cases where .* appeared inside
54 atomic brackets at the start of a pattern, or where there was a subsequent
55 *PRUNE or *SKIP.
56
57 7. Improved instruction cache flush for POWER/PowerPC.
58 Patch by Daniel Richard G.
59
60 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
61 grep:
62
63 (a) There is now no limit to the number of patterns to be matched.
64
65 (b) An error is given if a pattern is too long.
66
67 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
68 are now supported.
69
70 (d) --exclude-from and --include-from (multiple use) have been added.
71
72 (e) Exclusions and inclusions now apply to all files and directories, not
73 just to those obtained from scanning a directory recursively.
74
75 (f) Multiple uses of -f and --file-list are now supported.
76
77 (g) In a Windows environment, the default for -d has been changed from
78 "read" (the GNU grep default) to "skip", because otherwise the presence
79 of a directory in the file list provokes an error.
80
81 (h) The documentation has been revised and clarified in places.
82
83 9. Improve the matching speed of capturing brackets.
84
85 10. Changed the meaning of \X so that it now matches a Unicode extended
86 grapheme cluster.
87
88 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
89 out POSIX threads when JIT support is configured.
90
91 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
92
93 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
94 structure could go wrong in environments where size_t is not the same size
95 as int.
96
97 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
98 set.
99
100 15. The EBCDIC support had decayed; later updates to the code had included
101 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
102 general tidy up of EBCDIC-related issues, and the documentation was also
103 not quite right. There is now a test that can be run on ASCII systems to
104 check some of the EBCDIC-related things (but is it not a full test).
105
106 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
107 in a small tidy to the code.
108
109 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
110
111 18. If the --only-matching (-o) option in pcregrep is specified multiple
112 times, each one causes appropriate output. For example, -o1 -o2 outputs the
113 substrings matched by the 1st and 2nd capturing parentheses. A separating
114 string can be specified by --om-separator (default empty).
115
116 19. Improving the first n character searches.
117
118 20. Turn case lists for horizontal and vertical white space into macros so that
119 they are defined only once.
120
121 21. This set of changes together give more compatible Unicode case-folding
122 behaviour for characters that have more than one other case when UCP
123 support is available.
124
125 (a) The Unicode property table now has offsets into a new table of sets of
126 three or more characters that are case-equivalent. The MultiStage2.py
127 script that generates these tables (the pcre_ucd.c file) now scans
128 CaseFolding.txt instead of UnicodeData.txt for character case
129 information.
130
131 (b) The code for adding characters or ranges of characters to a character
132 class has been abstracted into a generalized function that also handles
133 case-independence. In UTF-mode with UCP support, this uses the new data
134 to handle characters with more than one other case.
135
136 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
137 whose other case is greater than 256 are now correctly matched
138 caselessly. Previously, the high codepoint matched the low one, but not
139 vice versa.
140
141 (d) The processing of \h, \H, \v, and \ in character classes now makes use
142 of the new class addition function, using character lists defined as
143 macros alongside the case definitions of 20 above.
144
145 (e) Caseless back references now work with characters that have more than
146 one other case.
147
148 (f) General caseless matching of characters with more than one other case
149 is supported.
150
151 22. Unicode character properties were updated from Unicode 6.2.0
152
153 23. Improved CMake support under Windows. Patch by Daniel Richard G.
154
155 24. Add support for 32-bit character strings, and UTF-32
156
157 25. Major JIT compiler update (code refactoring and bugfixing).
158 Experimental Sparc 32 support is added.
159
160 26. Applied a modified version of Daniel Richard G's patch to create
161 pcre.h.generic and config.h.generic by "make" instead of in the
162 PrepareRelease script.
163
164 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
165 pcre_compile.c when checking for a zero character.
166
167 28. Introducing a native interface for JIT. Through this interface, the compiled
168 machine code can be directly executed. The purpose of this interface is to
169 provide fast pattern matching, so several sanity checks are not performed.
170 However, feature tests are still performed. The new interface provides
171 1.4x speedup compared to the old one.
172
173 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
174 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
175 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
176
177 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
178 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
179 the "old" RFC 2279). Instead, it ended up passing a negative length to
180 pcre_exec().
181
182 31. Add support for GCC's visibility feature to hide internal functions.
183
184 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
185 "unknown -C option" after outputting 0 or 1.
186
187 33. There is now support for generating a code coverage report for the test
188 suite in environments where gcc is the compiler and lcov is installed. This
189 is mainly for the benefit of the developers.
190
191 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
192 unaddressable using valgrind annotations, allowing valgrind to detect
193 invalid memory accesses. This is mainly for the benefit of the developers.
194
195 25. (*UTF) can now be used to start a pattern in any of the three libraries.
196
197 26. Give configure error if --enable-cpp but no C++ compiler found.
198
199
200 Version 8.31 06-July-2012
201 -------------------------
202
203 1. Fixing a wrong JIT test case and some compiler warnings.
204
205 2. Removed a bashism from the RunTest script.
206
207 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
208 to unsigned type, result still unsigned" that was given by an MS compiler
209 on encountering the code "-sizeof(xxx)".
210
211 4. Partial matching support is added to the JIT compiler.
212
213 5. Fixed several bugs concerned with partial matching of items that consist
214 of more than one character:
215
216 (a) /^(..)\1/ did not partially match "aba" because checking references was
217 done on an "all or nothing" basis. This also applied to repeated
218 references.
219
220 (b) \R did not give a hard partial match if \r was found at the end of the
221 subject.
222
223 (c) \X did not give a hard partial match after matching one or more
224 characters at the end of the subject.
225
226 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
227 a partial match for the string "\r".
228
229 (e) When newline was set to CRLF, the metacharacter "." did not recognize
230 a partial match for a CR character at the end of the subject string.
231
232 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
233 running pcretest, the text "(JIT)" added to the output whenever JIT is
234 actually used to run the match.
235
236 7. Individual JIT compile options can be set in pcretest by following -s+[+]
237 or /S+[+] with a digit between 1 and 7.
238
239 8. OP_NOT now supports any UTF character not just single-byte ones.
240
241 9. (*MARK) control verb is now supported by the JIT compiler.
242
243 10. The command "./RunTest list" lists the available tests without actually
244 running any of them. (Because I keep forgetting what they all are.)
245
246 11. Add PCRE_INFO_MAXLOOKBEHIND.
247
248 12. Applied a (slightly modified) user-supplied patch that improves performance
249 when the heap is used for recursion (compiled with --disable-stack-for-
250 recursion). Instead of malloc and free for each heap frame each time a
251 logical recursion happens, frames are retained on a chain and re-used where
252 possible. This sometimes gives as much as 30% improvement.
253
254 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
255 call.
256
257 14. As documented, (*COMMIT) is now confined to within a positive assertion.
258
259 15. It is now possible to link pcretest with libedit as an alternative to
260 libreadline.
261
262 16. (*COMMIT) control verb is now supported by the JIT compiler.
263
264 17. The Unicode data tables have been updated to Unicode 6.1.0.
265
266 18. Added --file-list option to pcregrep.
267
268 19. Added binary file support to pcregrep, including the -a, --binary-files,
269 -I, and --text options.
270
271 20. The madvise function is renamed for posix_madvise for QNX compatibility
272 reasons. Fixed by Giuseppe D'Angelo.
273
274 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
275
276 22. Changed the option for creating symbolic links for 16-bit man pages from
277 -s to -sf so that re-installing does not cause issues.
278
279 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
280
281 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
282 matches in certain environments (the workspace was not being correctly
283 retained). Also added to pcre_dfa_exec() a simple plausibility check on
284 some of the workspace data at the beginning of a restart.
285
286 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
287 was not doing so when it should - probably a typo introduced by SVN 528
288 (change 8.10/14).
289
290 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
291 \w+ when the character tables indicated that \x{c4} was a word character.
292 There were several related cases, all because the tests for doing a table
293 lookup were testing for characters less than 127 instead of 255.
294
295 27. If a pattern contains capturing parentheses that are not used in a match,
296 their slots in the ovector are set to -1. For those that are higher than
297 any matched groups, this happens at the end of processing. In the case when
298 there were back references that the ovector was too small to contain
299 (causing temporary malloc'd memory to be used during matching), and the
300 highest capturing number was not used, memory off the end of the ovector
301 was incorrectly being set to -1. (It was using the size of the temporary
302 memory instead of the true size.)
303
304 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
305 ovector size, it uses memory at the end of the block that it has got.
306
307 29. Check for an overlong MARK name and give an error at compile time. The
308 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
309
310 30. JIT compiler update.
311
312 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
313 Rill for the patch.
314
315 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
316
317 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
318
319 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
320 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
321 was enabled.
322
323 35. Improve JIT code generation for greedy plus quantifier.
324
325 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
326 1 to "aa" instead of to an empty string. The bug affected repeated groups
327 that could potentially match an empty string.
328
329 37. Optimizing single character iterators in JIT.
330
331 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
332 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
333 codepoints that are too big for the mode are faulted, and in a UTF mode,
334 disallowed codepoints are also faulted.
335
336 39. If PCRE was compiled with UTF support, in three places in the DFA
337 matcher there was code that should only have been obeyed in UTF mode, but
338 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
339 processing when bytes with values greater than 127 were present. In 16-bit
340 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
341 both cases the values are those that cannot be the first data item in a UTF
342 character. The three items that might have provoked this were recursions,
343 possessively repeated groups, and atomic groups.
344
345 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
346 and pcregrep, because some OS require shared objects to be explicitly
347 passed to ld, causing the link step to fail if they are not.
348
349 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
350 mode, patterns that started with \h* or \R* might be incorrectly matched.
351
352
353 Version 8.30 04-February-2012
354 -----------------------------
355
356 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
357 name is defined in ctype.h.
358
359 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
360 only in quite long subpatterns.
361
362 3. Removed the function pcre_info(), which has been obsolete and deprecated
363 since it was replaced by pcre_fullinfo() in February 2000.
364
365 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
366 match a (*MARK), and the match failed at the start of the subject, a
367 reference to memory before the start of the subject could occur. This bug
368 was introduced by fix 17 of release 8.21.
369
370 5. A reference to an unset group with zero minimum repetition was giving
371 totally wrong answers (in non-JavaScript-compatibility mode). For example,
372 /(another)?(\1?)test/ matched against "hello world test". This bug was
373 introduced in release 8.13.
374
375 6. Add support for 16-bit character strings (a large amount of work involving
376 many changes and refactorings).
377
378 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
379 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
380 from a file.
381
382 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
383 rounding is not applied in this particular case).
384
385 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
386 if they appear, or are escaped, in patterns.
387
388 10. Get rid of a number of -Wunused-but-set-variable warnings.
389
390 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
391 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
392 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
393 also returns the mark "x". This bug applied to capturing parentheses,
394 non-capturing parentheses, and atomic parentheses. It also applied to some
395 assertions.
396
397 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
398 information out of configure.ac instead of relying on pcre.h.generic, which
399 is not stored in the repository.
400
401 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
402 -lreadline.
403
404 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
405
406 15. Applied Graycode's patch to put the top-level frame on the stack rather
407 than the heap when not using the stack for recursion. This gives a
408 performance improvement in many cases when recursion is not deep.
409
410 16. Experimental code added to "pcretest -C" to output the stack frame size.
411
412
413 Version 8.21 12-Dec-2011
414 ------------------------
415
416 1. Updating the JIT compiler.
417
418 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
419 are added as well.
420
421 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
422 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
423 calling _pcre_jit_exec. Some extra comments are added.
424
425 4. (*MARK) settings inside atomic groups that do not contain any capturing
426 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
427 was introduced by change 18 for 8.20.
428
429 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
430 ECMA-262 standard.
431
432 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
433 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
434 This bug was probably introduced by change 9 of 8.13.
435
436 7. While fixing 6 above, I noticed that a number of other items were being
437 incorrectly rejected as "not fixed length". This arose partly because newer
438 opcodes had not been added to the fixed-length checking code. I have (a)
439 corrected the bug and added tests for these items, and (b) arranged for an
440 error to occur if an unknown opcode is encountered while checking for fixed
441 length instead of just assuming "not fixed length". The items that were
442 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
443 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
444 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
445
446 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
447 being incorrectly compiled and would have given unpredicatble results.
448
449 9. A possessively repeated subpattern with minimum repeat count greater than
450 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
451 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
452 the first (A) could occur when it should not.
453
454 10. Add a cast and remove a redundant test from the code.
455
456 11. JIT should use pcre_malloc/pcre_free for allocation.
457
458 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
459 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
460 is anything other than /usr, -L is still shown).
461
462 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
463
464 14. Perl does not support \N without a following name in a [] class; PCRE now
465 also gives an error.
466
467 15. If a forward reference was repeated with an upper limit of around 2000,
468 it caused the error "internal error: overran compiling workspace". The
469 maximum number of forward references (including repeats) was limited by the
470 internal workspace, and dependent on the LINK_SIZE. The code has been
471 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
472 the default depends on LINK_SIZE. There is a new upper limit (for safety)
473 of around 200,000 forward references. While doing this, I also speeded up
474 the filling in of repeated forward references.
475
476 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
477 incorrectly expecting the subject to contain another "a" after the start.
478
479 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
480 in the match, the SKIP should be ignored. This was not happening; instead
481 the SKIP was being treated as NOMATCH. For patterns such as
482 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
483 tested.
484
485 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
486 now much more compatible with Perl, in particular in cases where the result
487 is a non-match for a non-anchored pattern. For example, if
488 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
489 "m", where previously it did not return a name. A side effect of this
490 change is that for partial matches, the last encountered mark name is
491 returned, as for non matches. A number of tests that were previously not
492 Perl-compatible have been moved into the Perl-compatible test files. The
493 refactoring has had the pleasing side effect of removing one argument from
494 the match() function, thus reducing its stack requirements.
495
496 19. If the /S+ option was used in pcretest to study a pattern using JIT,
497 subsequent uses of /S (without +) incorrectly behaved like /S+.
498
499 21. Retrieve executable code size support for the JIT compiler and fixing
500 some warnings.
501
502 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
503 not work when the shorter character appeared right at the end of the
504 subject string.
505
506 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
507 systems.
508
509 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
510 output it when the /M option is used in pcretest.
511
512 25. The CheckMan script was not being included in the distribution. Also, added
513 an explicit "perl" to run Perl scripts from the PrepareRelease script
514 because this is reportedly needed in Windows.
515
516 26. If study data was being save in a file and studying had not found a set of
517 "starts with" bytes for the pattern, the data written to the file (though
518 never used) was taken from uninitialized memory and so caused valgrind to
519 complain.
520
521 27. Updated RunTest.bat as provided by Sheri Pierce.
522
523 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
524
525 29. Computation of memory usage for the table of capturing group names was
526 giving an unnecessarily large value.
527
528
529 Version 8.20 21-Oct-2011
530 ------------------------
531
532 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
533 a POSIX class. After further experiments with Perl, which convinced me that
534 Perl has bugs and confusions, a closing square bracket is no longer allowed
535 in a POSIX name. This bug also affected patterns with classes that started
536 with full stops.
537
538 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
539 captured substring, but while checking the failing first alternative,
540 substring 1 is temporarily captured. If the output vector supplied to
541 pcre_exec() was not big enough for this capture, the yield of the function
542 was still zero ("insufficient space for captured substrings"). This cannot
543 be totally fixed without adding another stack variable, which seems a lot
544 of expense for a edge case. However, I have improved the situation in cases
545 such as /(a)(b)x|abc/ matched against "abc", where the return code
546 indicates that fewer than the maximum number of slots in the ovector have
547 been set.
548
549 3. Related to (2) above: when there are more back references in a pattern than
550 slots in the output vector, pcre_exec() uses temporary memory during
551 matching, and copies in the captures as far as possible afterwards. It was
552 using the entire output vector, but this conflicts with the specification
553 that only 2/3 is used for passing back captured substrings. Now it uses
554 only the first 2/3, for compatibility. This is, of course, another edge
555 case.
556
557 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
558 main code base, and can be used by building with --enable-jit. When this is
559 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
560 runtime --no-jit option is given.
561
562 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
563 ovector, the return from the function was zero, implying that there were
564 other matches that did not fit. The correct "exactly full" value is now
565 returned.
566
567 6. If a subpattern that was called recursively or as a subroutine contained
568 (*PRUNE) or any other control that caused it to give a non-standard return,
569 invalid errors such as "Error -26 (nested recursion at the same subject
570 position)" or even infinite loops could occur.
571
572 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
573 computing the minimum length on reaching *ACCEPT, and so ended up with the
574 wrong value of 1 rather than 0. Further investigation indicates that
575 computing a minimum subject length in the presence of *ACCEPT is difficult
576 (think back references, subroutine calls), and so I have changed the code
577 so that no minimum is registered for a pattern that contains *ACCEPT.
578
579 8. If (*THEN) was present in the first (true) branch of a conditional group,
580 it was not handled as intended. [But see 16 below.]
581
582 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
583 Sheri Pierce.
584
585 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
586 the first byte in a match must be "a".
587
588 11. Change 17 for 8.13 increased the recursion depth for patterns like
589 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
590 pattern contains any instances of (*THEN). If it does not, the old
591 optimizations are restored. It would be nice to do this on a per-group
592 basis, but at the moment that is not feasible.
593
594 12. In some environments, the output of pcretest -C is CRLF terminated. This
595 broke RunTest's code that checks for the link size. A single white space
596 character after the value is now allowed for.
597
598 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
599 For "fr", it uses the Windows-specific input and output files.
600
601 14. If (*THEN) appeared in a group that was called recursively or as a
602 subroutine, it did not work as intended. [But see next item.]
603
604 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
605 pattern fragments (but not containing any | characters). If A and B are
606 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
607 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
608 D. In other words, Perl considers parentheses that do not contain any |
609 characters to be part of a surrounding alternative, whereas PCRE was
610 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
611 differently. PCRE now behaves in the same way as Perl, except in the case
612 of subroutine/recursion calls such as (?1) which have in any case always
613 been different (but PCRE had them first :-).
614
615 16. Related to 15 above: Perl does not treat the | in a conditional group as
616 creating alternatives. Such a group is treated in the same way as an
617 ordinary group without any | characters when processing (*THEN). PCRE has
618 been changed to match Perl's behaviour.
619
620 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
621 RunGrepTest script failed.
622
623 18. Change 22 for version 13 caused atomic groups to use more stack. This is
624 inevitable for groups that contain captures, but it can lead to a lot of
625 stack use in large patterns. The old behaviour has been restored for atomic
626 groups that do not contain any capturing parentheses.
627
628 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
629 suppress the check for a minimum subject length at run time. (If it was
630 given to pcre_exec() or pcre_dfa_exec() it did work.)
631
632 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
633 fail to work when decoding hex characters in data strings in EBCDIC
634 environments.
635
636 21. It appears that in at least one Mac OS environment, the isxdigit() function
637 is implemented as a macro that evaluates to its argument more than once,
638 contravening the C 90 Standard (I haven't checked a later standard). There
639 was an instance in pcretest which caused it to go wrong when processing
640 \x{...} escapes in subject strings. The has been rewritten to avoid using
641 things like p++ in the argument of isxdigit().
642
643
644 Version 8.13 16-Aug-2011
645 ------------------------
646
647 1. The Unicode data tables have been updated to Unicode 6.0.0.
648
649 2. Two minor typos in pcre_internal.h have been fixed.
650
651 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
652 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
653 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
654
655 4. There were a number of related bugs in the code for matching backrefences
656 caselessly in UTF-8 mode when codes for the characters concerned were
657 different numbers of bytes. For example, U+023A and U+2C65 are an upper
658 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
659 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
660 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
661 2-byte code at the end of the subject (it thought there wasn't enough data
662 left).
663
664 5. Comprehensive information about what went wrong is now returned by
665 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
666 as the output vector has at least 2 elements. The offset of the start of
667 the failing character and a reason code are placed in the vector.
668
669 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
670 now returned is for the first byte of the failing character, instead of the
671 last byte inspected. This is an incompatible change, but I hope it is small
672 enough not to be a problem. It makes the returned offset consistent with
673 pcre_exec() and pcre_dfa_exec().
674
675 7. pcretest now gives a text phrase as well as the error number when
676 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
677 failure, the offset and reason code are output.
678
679 8. When \R was used with a maximizing quantifier it failed to skip backwards
680 over a \r\n pair if the subsequent match failed. Instead, it just skipped
681 back over a single character (\n). This seems wrong (because it treated the
682 two characters as a single entity when going forwards), conflicts with the
683 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
684 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
685 has been changed.
686
687 9. Some internal refactoring has changed the processing so that the handling
688 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
689 time (the PCRE_DOTALL option was changed this way some time ago: version
690 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
691 which was always a bit of a fudge. It also means that there is one less
692 argument for the match() function, which reduces its stack requirements
693 slightly. This change also fixes an incompatibility with Perl: the pattern
694 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
695
696 10. More internal refactoring has drastically reduced the number of recursive
697 calls to match() for possessively repeated groups such as (abc)++ when
698 using pcre_exec().
699
700 11. While implementing 10, a number of bugs in the handling of groups were
701 discovered and fixed:
702
703 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
704 (a|)*(?1) gave a compile-time internal error.
705 ((a|)+)+ did not notice that the outer group could match an empty string.
706 (^a|^)+ was not marked as anchored.
707 (.*a|.*)+ was not marked as matching at start or after a newline.
708
709 12. Yet more internal refactoring has removed another argument from the match()
710 function. Special calls to this function are now indicated by setting a
711 value in a variable in the "match data" data block.
712
713 13. Be more explicit in pcre_study() instead of relying on "default" for
714 opcodes that mean there is no starting character; this means that when new
715 ones are added and accidentally left out of pcre_study(), testing should
716 pick them up.
717
718 14. The -s option of pcretest has been documented for ages as being an old
719 synonym of -m (show memory usage). I have changed it to mean "force study
720 for every regex", that is, assume /S for every regex. This is similar to -i
721 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
722 using it. It makes it easier to run collections of tests with and without
723 study enabled, and thereby test pcre_study() more easily. All the standard
724 tests are now run with and without -s (but some patterns can be marked as
725 "never study" - see 20 below).
726
727 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
728 restoration of the capturing data to the outer values was not happening
729 correctly.
730
731 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
732 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
733 pattern had matched an empty string, and so incorrectly returned a no
734 match.
735
736 17. There was optimizing code for the last branch of non-capturing parentheses,
737 and also for the obeyed branch of a conditional subexpression, which used
738 tail recursion to cut down on stack usage. Unfortunately, now that there is
739 the possibility of (*THEN) occurring in these branches, tail recursion is
740 no longer possible because the return has to be checked for (*THEN). These
741 two optimizations have therefore been removed. [But see 8.20/11 above.]
742
743 18. If a pattern containing \R was studied, it was assumed that \R always
744 matched two bytes, thus causing the minimum subject length to be
745 incorrectly computed because \R can also match just one byte.
746
747 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
748 was incorrectly computed.
749
750 20. If /S is present twice on a test pattern in pcretest input, it now
751 *disables* studying, thereby overriding the use of -s on the command line
752 (see 14 above). This is necessary for one or two tests to keep the output
753 identical in both cases.
754
755 21. When (*ACCEPT) was used in an assertion that matched an empty string and
756 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
757
758 22. When an atomic group that contained a capturing parenthesis was
759 successfully matched, but the branch in which it appeared failed, the
760 capturing was not being forgotten if a higher numbered group was later
761 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
762 group 1 to "a", when in fact it should be unset. This applied to multi-
763 branched capturing and non-capturing groups, repeated or not, and also to
764 positive assertions (capturing in negative assertions does not happen
765 in PCRE) and also to nested atomic groups.
766
767 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
768 subject after a captured substring, to make it easier to tell which of a
769 number of identical substrings has been captured.
770
771 24. The way atomic groups are processed by pcre_exec() has been changed so that
772 if they are repeated, backtracking one repetition now resets captured
773 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
774 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
775 "aaa". Previously, it would have been "a". As part of this code
776 refactoring, the way recursive calls are handled has also been changed.
777
778 25. If an assertion condition captured any substrings, they were not passed
779 back unless some other capturing happened later. For example, if
780 (?(?=(a))a) was matched against "a", no capturing was returned.
781
782 26. When studying a pattern that contained subroutine calls or assertions,
783 the code for finding the minimum length of a possible match was handling
784 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
785 group 1 called group 2 while simultaneously a separate group 2 called group
786 1). A stack overflow occurred in this case. I have fixed this by limiting
787 the recursion depth to 10.
788
789 27. Updated RunTest.bat in the distribution to the version supplied by Tom
790 Fortmann. This supports explicit test numbers on the command line, and has
791 argument validation and error reporting.
792
793 28. An instance of \X with an unlimited repeat could fail if at any point the
794 first character it looked at was a mark character.
795
796 29. Some minor code refactoring concerning Unicode properties and scripts
797 should reduce the stack requirement of match() slightly.
798
799 30. Added the '=' option to pcretest to check the setting of unused capturing
800 slots at the end of the pattern, which are documented as being -1, but are
801 not included in the return count.
802
803 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
804 compiled something random. Now it gives a compile-time error (as does
805 Perl).
806
807 32. A *MARK encountered during the processing of a positive assertion is now
808 recorded and passed back (compatible with Perl).
809
810 33. If --only-matching or --colour was set on a pcregrep call whose pattern
811 had alternative anchored branches, the search for a second match in a line
812 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
813 matched the line "0102" twice. The same bug affected patterns that started
814 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
815 twice.
816
817 34. Previously, PCRE did not allow quantification of assertions. However, Perl
818 does, and because of capturing effects, quantifying parenthesized
819 assertions may at times be useful. Quantifiers are now allowed for
820 parenthesized assertions.
821
822 35. A minor code tidy in pcre_compile() when checking options for \R usage.
823
824 36. \g was being checked for fancy things in a character class, when it should
825 just be a literal "g".
826
827 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
828 appearance of a nested POSIX class supersedes an apparent external class.
829 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
830 unescaped square brackets may also appear as part of class names. For
831 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
832 more like Perl. (But see 8.20/1 above.)
833
834 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
835 was because it thought it was \N{name}, which is not supported).
836
837 39. Add minix to OS list not supporting the -S option in pcretest.
838
839 40. PCRE tries to detect cases of infinite recursion at compile time, but it
840 cannot analyze patterns in sufficient detail to catch mutual recursions
841 such as ((?1))((?2)). There is now a runtime test that gives an error if a
842 subgroup is called recursively as a subpattern for a second time at the
843 same position in the subject string. In previous releases this might have
844 been caught by the recursion limit, or it might have run out of stack.
845
846 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
847 happen only once. PCRE was, however incorrectly giving a compile time error
848 "recursive call could loop indefinitely" because it cannot analyze the
849 pattern in sufficient detail. The compile time test no longer happens when
850 PCRE is compiling a conditional subpattern, but actual runaway loops are
851 now caught at runtime (see 40 above).
852
853 42. It seems that Perl allows any characters other than a closing parenthesis
854 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
855 has been changed to be the same.
856
857 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
858 as not to get warnings when autogen.sh is called. Also changed
859 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
860
861 44. To help people who use pcregrep to scan files containing exceedingly long
862 lines, the following changes have been made:
863
864 (a) The default value of the buffer size parameter has been increased from
865 8K to 20K. (The actual buffer used is three times this size.)
866
867 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
868 PCRE is built.
869
870 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
871 to be set at run time.
872
873 (d) Numerical values in pcregrep options can be followed by K or M, for
874 example --buffer-size=50K.
875
876 (e) If a line being scanned overflows pcregrep's buffer, an error is now
877 given and the return code is set to 2.
878
879 45. Add a pointer to the latest mark to the callout data block.
880
881 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
882 partial match of an empty string instead of no match. This was specific to
883 the use of ".".
884
885 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
886 complete match instead of a partial match. This bug was dependent on both
887 the PCRE_UTF8 and PCRE_DOTALL options being set.
888
889 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
890 starting byte set, because \b was not being ignored.
891
892
893 Version 8.12 15-Jan-2011
894 ------------------------
895
896 1. Fixed some typos in the markup of the man pages, and wrote a script that
897 checks for such things as part of the documentation building process.
898
899 2. On a big-endian 64-bit system, pcregrep did not correctly process the
900 --match-limit and --recursion-limit options (added for 8.11). In
901 particular, this made one of the standard tests fail. (The integer value
902 went into the wrong half of a long int.)
903
904 3. If the --colour option was given to pcregrep with -v (invert match), it
905 did strange things, either producing crazy output, or crashing. It should,
906 of course, ignore a request for colour when reporting lines that do not
907 match.
908
909 4. Another pcregrep bug caused similar problems if --colour was specified with
910 -M (multiline) and the pattern match finished with a line ending.
911
912 5. In pcregrep, when a pattern that ended with a literal newline sequence was
913 matched in multiline mode, the following line was shown as part of the
914 match. This seems wrong, so I have changed it.
915
916 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
917 the check for further matches in the same line (so they could be coloured)
918 to overrun the end of the current line. If another match was found, it was
919 incorrectly shown (and then shown again when found in the next line).
920
921 7. If pcregrep was compiled under Windows, there was a reference to the
922 function pcregrep_exit() before it was defined. I am assuming this was
923 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
924 reported by a user. I've moved the definition above the reference.
925
926
927 Version 8.11 10-Dec-2010
928 ------------------------
929
930 1. (*THEN) was not working properly if there were untried alternatives prior
931 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
932 backtracked to try for "b" instead of moving to the next alternative branch
933 at the same level (in this case, to look for "c"). The Perl documentation
934 is clear that when (*THEN) is backtracked onto, it goes to the "next
935 alternative in the innermost enclosing group".
936
937 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
938 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
939 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
940 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
941 (*THEN).
942
943 3. If \s appeared in a character class, it removed the VT character from
944 the class, even if it had been included by some previous item, for example
945 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
946 of \s, but is part of the POSIX "space" class.)
947
948 4. A partial match never returns an empty string (because you can always
949 match an empty string at the end of the subject); however the checking for
950 an empty string was starting at the "start of match" point. This has been
951 changed to the "earliest inspected character" point, because the returned
952 data for a partial match starts at this character. This means that, for
953 example, /(?<=abc)def/ gives a partial match for the subject "abc"
954 (previously it gave "no match").
955
956 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
957 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
958 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
959 has an implication that the given string is incomplete (because a partial
960 match is preferred over a full match). For this reason, these items now
961 give a partial match in this situation. [Aside: previously, the one case
962 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
963 match rather than a full match, which was wrong by the old rules, but is
964 now correct.]
965
966 6. There was a bug in the handling of #-introduced comments, recognized when
967 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
968 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
969 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
970 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
971 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
972 places in pcre_compile().
973
974 7. Related to (6) above, when pcre_compile() was skipping #-introduced
975 comments when looking ahead for named forward references to subpatterns,
976 the only newline sequence it recognized was NL. It now handles newlines
977 according to the set newline convention.
978
979 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
980 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
981 cater for a lack of strerror(). These oversights have been fixed.
982
983 9. Added --match-limit and --recursion-limit to pcregrep.
984
985 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
986
987 11. When the -o option was used, pcregrep was setting a return code of 1, even
988 when matches were found, and --line-buffered was not being honoured.
989
990 12. Added an optional parentheses number to the -o and --only-matching options
991 of pcregrep.
992
993 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
994 can match an empty string. The code to do it in pcretest and pcredemo
995 needed fixing:
996
997 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
998 only one byte after an empty string match just before CRLF (this case
999 just got forgotten; "any" and "anycrlf" were OK).
1000
1001 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1002 mode when an empty string match preceded an ASCII character followed by
1003 a non-ASCII character. (The code for advancing by one character rather
1004 than one byte was nonsense.)
1005
1006 (c) The pcredemo.c sample program did not have any code at all to handle
1007 the cases when CRLF is a valid newline sequence.
1008
1009 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1010 as a starting offset was within the subject string. There is now a new
1011 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1012 negative or greater than the length of the string. In order to test this,
1013 pcretest is extended to allow the setting of negative starting offsets.
1014
1015 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1016 starting offset points to the beginning of a UTF-8 character was
1017 unnecessarily clumsy. I tidied it up.
1018
1019 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1020 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1021
1022 17. Nobody had reported that the --include_dir option, which was added in
1023 release 7.7 should have been called --include-dir (hyphen, not underscore)
1024 for compatibility with GNU grep. I have changed it to --include-dir, but
1025 left --include_dir as an undocumented synonym, and the same for
1026 --exclude-dir, though that is not available in GNU grep, at least as of
1027 release 2.5.4.
1028
1029 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1030 characters from a string of bytes) have been redefined so as not to use
1031 loops, in order to improve performance in some environments. At the same
1032 time, I abstracted some of the common code into auxiliary macros to save
1033 repetition (this should not affect the compiled code).
1034
1035 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1036 compile-time error is now given if \c is not followed by an ASCII
1037 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1038 different, and any byte value is allowed.)
1039
1040 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1041 START_OPTIMIZE option, which is now allowed at compile time - but just
1042 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1043 to pcregrep and other applications that have no direct access to PCRE
1044 options. The new /Y option in pcretest sets this option when calling
1045 pcre_compile().
1046
1047 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1048 back references. Groups containing recursive back references were forced to
1049 be atomic by that change, but in the case of named groups, the amount of
1050 memory required was incorrectly computed, leading to "Failed: internal
1051 error: code overflow". This has been fixed.
1052
1053 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1054 pcretest.c, to avoid build problems in some Borland environments.
1055
1056
1057 Version 8.10 25-Jun-2010
1058 ------------------------
1059
1060 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1061 THEN.
1062
1063 2. (*ACCEPT) was not working when inside an atomic group.
1064
1065 3. Inside a character class, \B is treated as a literal by default, but
1066 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1067 causes the error). The code is unchanged, but I tidied the documentation.
1068
1069 4. Inside a character class, PCRE always treated \R and \X as literals,
1070 whereas Perl faults them if its -w option is set. I have changed PCRE so
1071 that it faults them when PCRE_EXTRA is set.
1072
1073 5. Added support for \N, which always matches any character other than
1074 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1075
1076 6. When compiling pcregrep with newer versions of gcc which may have
1077 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1078 declared with attribute warn_unused_result" were given. Just casting the
1079 result to (void) does not stop the warnings; a more elaborate fudge is
1080 needed. I've used a macro to implement this.
1081
1082 7. Minor change to pcretest.c to avoid a compiler warning.
1083
1084 8. Added four artifical Unicode properties to help with an option to make
1085 \s etc use properties (see next item). The new properties are: Xan
1086 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1087
1088 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1089 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1090 this option. Modified pcretest to add /W to test this facility. Added
1091 REG_UCP to make it available via the POSIX interface.
1092
1093 10. Added --line-buffered to pcregrep.
1094
1095 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1096 studied, and the match started with a letter with a code point greater than
1097 127 whose first byte was different to the first byte of the other case of
1098 the letter, the other case of this starting letter was not recognized
1099 (#976).
1100
1101 12. If a pattern that was studied started with a repeated Unicode property
1102 test, for example, \p{Nd}+, there was the theoretical possibility of
1103 setting up an incorrect bitmap of starting bytes, but fortunately it could
1104 not have actually happened in practice until change 8 above was made (it
1105 added property types that matched character-matching opcodes).
1106
1107 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1108 possible starting bytes for non-anchored patterns.
1109
1110 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1111 \R, and also a number of cases that involve Unicode properties, both
1112 explicit and implicit when PCRE_UCP is set.
1113
1114 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1115 input, it could crash or give wrong results if characters with values
1116 greater than 0xc0 were present in the subject string. (Detail: it assumed
1117 UTF-8 input when processing these items.)
1118
1119 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1120 size_t is 64-bit (#991).
1121
1122 17. Added a check for running out of memory when PCRE is compiled with
1123 --disable-stack-for-recursion (#990).
1124
1125 18. If the last data line in a file for pcretest does not have a newline on
1126 the end, a newline was missing in the output.
1127
1128 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1129 less than 128) in its various bitmaps. However, there is a facility for
1130 generating tables according to the current locale when PCRE is compiled. It
1131 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1132 characters, are recognized by isspace() and therefore were getting set in
1133 these tables, and indeed these tables seem to approximate to ISO 8859. This
1134 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1135 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1136 which of course cannot start UTF-8 characters. I have changed the code so
1137 that only real ASCII characters (less than 128) and the correct starting
1138 bytes for UTF-8 encodings are set for characters greater than 127 when in
1139 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1140 altogether.)
1141
1142 20. Added the /T option to pcretest so as to be able to run tests with non-
1143 standard character tables, thus making it possible to include the tests
1144 used for 19 above in the standard set of tests.
1145
1146 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1147 reference to a subpattern the other side of a comment that contains an
1148 opening parenthesis caused either an internal compiling error, or a
1149 reference to the wrong subpattern.
1150
1151
1152 Version 8.02 19-Mar-2010
1153 ------------------------
1154
1155 1. The Unicode data tables have been updated to Unicode 5.2.0.
1156
1157 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1158 configured.
1159
1160 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1161 original author of that file, following a query about its status.
1162
1163 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1164 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1165
1166 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1167 quantifier applied to a forward-referencing subroutine call, could compile
1168 incorrect code or give the error "internal error: previously-checked
1169 referenced subpattern not found".
1170
1171 6. Both MS Visual Studio and Symbian OS have problems with initializing
1172 variables to point to external functions. For these systems, therefore,
1173 pcre_malloc etc. are now initialized to local functions that call the
1174 relevant global functions.
1175
1176 7. There were two entries missing in the vectors called coptable and poptable
1177 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1178 I've fixed the data, and added a kludgy way of testing at compile time that
1179 the lengths are correct (equal to the number of opcodes).
1180
1181 8. Following on from 7, I added a similar kludge to check the length of the
1182 eint vector in pcreposix.c.
1183
1184 9. Error texts for pcre_compile() are held as one long string to avoid too
1185 much relocation at load time. To find a text, the string is searched,
1186 counting zeros. There was no check for running off the end of the string,
1187 which could happen if a new error number was added without updating the
1188 string.
1189
1190 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1191
1192 11. \K was not working if it appeared in an atomic group or in a group that
1193 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1194 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1195 the assertion is positive, but not if it is negative.
1196
1197 12. Change 11 fortuitously reduced the size of the stack frame used in the
1198 "match()" function of pcre_exec.c by one pointer. Forthcoming
1199 implementation of support for (*MARK) will need an extra pointer on the
1200 stack; I have reserved it now, so that the stack frame size does not
1201 decrease.
1202
1203 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1204 item in branch that calls a recursion is a subroutine call - as in the
1205 second branch in the above example - was incorrectly given the compile-
1206 time error "recursive call could loop indefinitely" because pcre_compile()
1207 was not correctly checking the subroutine for matching a non-empty string.
1208
1209 14. The checks for overrunning compiling workspace could trigger after an
1210 overrun had occurred. This is a "should never occur" error, but it can be
1211 triggered by pathological patterns such as hundreds of nested parentheses.
1212 The checks now trigger 100 bytes before the end of the workspace.
1213
1214 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1215
1216
1217 Version 8.01 19-Jan-2010
1218 ------------------------
1219
1220 1. If a pattern contained a conditional subpattern with only one branch (in
1221 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1222 computed the wrong minimum data length (which is of course zero for such
1223 subpatterns). This could cause incorrect "no match" results.
1224
1225 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1226 the pattern is reset in the first branch, pcre_compile() failed with
1227 "internal error: code overflow at offset...". This happened only when
1228 the reset was to the original external option setting. (An optimization
1229 abstracts leading options settings into an external setting, which was the
1230 cause of this.)
1231
1232 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1233 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1234 assertion pattern did not match (meaning that the assertion was true), it
1235 was incorrectly treated as false if the SKIP had been reached during the
1236 matching. This also applied to assertions used as conditions.
1237
1238 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1239 assertion subpattern, including such a pattern used as a condition,
1240 unpredictable results occurred, instead of the error return
1241 PCRE_ERROR_DFA_UITEM.
1242
1243 5. The C++ GlobalReplace function was not working like Perl for the special
1244 situation when an empty string is matched. It now does the fancy magic
1245 stuff that is necessary.
1246
1247 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1248 removed. (These were left over from very, very early versions of PCRE.)
1249
1250 7. Some cosmetic changes to the code to make life easier when compiling it
1251 as part of something else:
1252
1253 (a) Change DEBUG to PCRE_DEBUG.
1254
1255 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1256 called "current" as "current_branch", to prevent a collision with the
1257 Linux macro when compiled as a kernel module.
1258
1259 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1260 prevent a collision with the Linux macro when compiled as a kernel
1261 module.
1262
1263 8. In pcre_compile() there are some checks for integer overflows that used to
1264 cast potentially large values to (double). This has been changed to that
1265 when building, a check for int64_t is made, and if it is found, it is used
1266 instead, thus avoiding the use of floating point arithmetic. (There is no
1267 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1268 double.
1269
1270 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1271 2005 (difference between two addresses compared to an unsigned value).
1272
1273 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1274 custom one, because of the following reported problem in Windows:
1275
1276 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1277 under Win32.
1278 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1279 therefore missing the function definition.
1280 - The compiler thus generates a "C" signature for the test function.
1281 - The linker fails to find the "C" function.
1282 - PCRE fails to configure if asked to do so against libbz2.
1283
1284 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1285 messages were output:
1286
1287 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1288 rerunning libtoolize, to keep the correct libtool macros in-tree.
1289 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1290
1291 I have done both of these things.
1292
1293 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1294 most of the time, it *can* run out if it is given a pattern that contains a
1295 runaway infinite recursion. I updated the discussion in the pcrestack man
1296 page.
1297
1298 13. Now that we have gone to the x.xx style of version numbers, the minor
1299 version may start with zero. Using 08 or 09 is a bad idea because users
1300 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1301 interpreted as invalid octal numbers. I've updated the previous comment in
1302 configure.ac, and also added a check that gives an error if 08 or 09 are
1303 used.
1304
1305 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1306 causing partial matching to fail when the end of the subject matched \W
1307 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1308
1309 15. There were some discrepancies between the declarations in pcre_internal.h
1310 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1311 their definitions. The declarations used "const uschar *" and the
1312 definitions used USPTR. Even though USPTR is normally defined as "const
1313 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1314 reported that: "This difference in casting confuses some C++ compilers, for
1315 example, SunCC recognizes above declarations as different functions and
1316 generates broken code for hbpcre." I have changed the declarations to use
1317 USPTR.
1318
1319 16. GNU libtool is named differently on some systems. The autogen.sh script now
1320 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1321 (FreeBSD).
1322
1323 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1324 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1325 comment: "Figure out how to create a longlong from a string: strtoll and
1326 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1327 instance, but it only takes 2 args instead of 3!"
1328
1329 18. A subtle bug concerned with back references has been fixed by a change of
1330 specification, with a corresponding code fix. A pattern such as
1331 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1332 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1333 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1334 same bug. Such groups have to be quantified to be useful, or contained
1335 inside another quantified group. (If there's no repetition, the reference
1336 can never match.) The problem arises because, having left the group and
1337 moved on to the rest of the pattern, a later failure that backtracks into
1338 the group uses the captured value from the final iteration of the group
1339 rather than the correct earlier one. I have fixed this in PCRE by forcing
1340 any group that contains a reference to itself to be an atomic group; that
1341 is, there cannot be any backtracking into it once it has completed. This is
1342 similar to recursive and subroutine calls.
1343
1344
1345 Version 8.00 19-Oct-09
1346 ----------------------
1347
1348 1. The table for translating pcre_compile() error codes into POSIX error codes
1349 was out-of-date, and there was no check on the pcre_compile() error code
1350 being within the table. This could lead to an OK return being given in
1351 error.
1352
1353 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1354 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1355 in a Windows environment.
1356
1357 3. The pcregrep --count option prints the count for each file even when it is
1358 zero, as does GNU grep. However, pcregrep was also printing all files when
1359 --files-with-matches was added. Now, when both options are given, it prints
1360 counts only for those files that have at least one match. (GNU grep just
1361 prints the file name in this circumstance, but including the count seems
1362 more useful - otherwise, why use --count?) Also ensured that the
1363 combination -clh just lists non-zero counts, with no names.
1364
1365 4. The long form of the pcregrep -F option was incorrectly implemented as
1366 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1367 but it seems right to fix it, and I didn't think it was worth preserving
1368 the old behaviour.
1369
1370 5. The command line items --regex=pattern and --regexp=pattern were not
1371 recognized by pcregrep, which required --regex pattern or --regexp pattern
1372 (with a space rather than an '='). The man page documented the '=' forms,
1373 which are compatible with GNU grep; these now work.
1374
1375 6. No libpcreposix.pc file was created for pkg-config; there was just
1376 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1377
1378 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1379 when UCP support is not needed, by modifying the Python script that
1380 generates it from Unicode data files. This should not matter if the module
1381 is correctly used as a library, but I received one complaint about 50K of
1382 unwanted data. My guess is that the person linked everything into his
1383 program rather than using a library. Anyway, it does no harm.
1384
1385 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1386 was a minimum greater than 1 for a wide character in a possessive
1387 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1388 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1389 character. Chaos in the form of incorrect output or a compiling loop could
1390 result.
1391
1392 9. The restrictions on what a pattern can contain when partial matching is
1393 requested for pcre_exec() have been removed. All patterns can now be
1394 partially matched by this function. In addition, if there are at least two
1395 slots in the offset vector, the offset of the earliest inspected character
1396 for the match and the offset of the end of the subject are set in them when
1397 PCRE_ERROR_PARTIAL is returned.
1398
1399 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1400 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1401 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1402 and may be more useful for multi-segment matching.
1403
1404 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1405 used to be given if ever the end of the subject was reached; now it is
1406 given only if matching could not proceed because another character was
1407 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1408 string "Z", which now yields "no match" instead of "partial match". In the
1409 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1410 final character ended with (*FAIL).
1411
1412 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1413 if the pattern had a "must contain" character that was already found in the
1414 earlier partial match, unless partial matching was again requested. For
1415 example, with the pattern /dog.(body)?/, the "must contain" character is
1416 "g". If the first part-match was for the string "dog", restarting with
1417 "sbody" failed. This bug has been fixed.
1418
1419 13. The string returned by pcre_dfa_exec() after a partial match has been
1420 changed so that it starts at the first inspected character rather than the
1421 first character of the match. This makes a difference only if the pattern
1422 starts with a lookbehind assertion or \b or \B (\K is not supported by
1423 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1424 matching functions compatible, and I think it's the right thing to do.
1425
1426 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1427 so that the demonstration program is easily available in environments where
1428 PCRE has not been installed from source.
1429
1430 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1431 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1432 library.
1433
1434 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1435 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1436 is not the first non-POSIX option to be added. Clearly some people find
1437 these options useful.
1438
1439 17. If a caller to the POSIX matching function regexec() passes a non-zero
1440 value for nmatch with a NULL value for pmatch, the value of
1441 nmatch is forced to zero.
1442
1443 18. RunGrepTest did not have a test for the availability of the -u option of
1444 the diff command, as RunTest does. It now checks in the same way as
1445 RunTest, and also checks for the -b option.
1446
1447 19. If an odd number of negated classes containing just a single character
1448 interposed, within parentheses, between a forward reference to a named
1449 subpattern and the definition of the subpattern, compilation crashed with
1450 an internal error, complaining that it could not find the referenced
1451 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1452 [The bug was that it was starting one character too far in when skipping
1453 over the character class, thus treating the ] as data rather than
1454 terminating the class. This meant it could skip too much.]
1455
1456 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1457 /g option in pcretest when the pattern contains \K, which makes it possible
1458 to have an empty string match not at the start, even when the pattern is
1459 anchored. Updated pcretest and pcredemo to use this option.
1460
1461 21. If the maximum number of capturing subpatterns in a recursion was greater
1462 than the maximum at the outer level, the higher number was returned, but
1463 with unset values at the outer level. The correct (outer level) value is
1464 now given.
1465
1466 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1467 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1468 make it do so. The string so far is captured, making this feature
1469 compatible with Perl.
1470
1471 23. The tests have been re-organized, adding tests 11 and 12, to make it
1472 possible to check the Perl 5.10 features against Perl 5.10.
1473
1474 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1475 pattern matches a fixed length string. PCRE did not allow this; now it
1476 does. Neither allows recursion.
1477
1478 25. I finally figured out how to implement a request to provide the minimum
1479 length of subject string that was needed in order to match a given pattern.
1480 (It was back references and recursion that I had previously got hung up
1481 on.) This code has now been added to pcre_study(); it finds a lower bound
1482 to the length of subject needed. It is not necessarily the greatest lower
1483 bound, but using it to avoid searching strings that are too short does give
1484 some useful speed-ups. The value is available to calling programs via
1485 pcre_fullinfo().
1486
1487 26. While implementing 25, I discovered to my embarrassment that pcretest had
1488 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1489 study optimizations had never been tested with that matching function.
1490 Oops. What is worse, even when it was passed study data, there was a bug in
1491 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1492 were also very few tests of studied patterns with pcre_dfa_exec().
1493
1494 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1495 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1496 on the other side of the coin, they are no longer allowed to have different
1497 names, because these cannot be distinguished in PCRE, and this has caused
1498 confusion. (This is a difference from Perl.)
1499
1500 28. When duplicate subpattern names are present (necessarily with different
1501 numbers, as required by 27 above), and a test is made by name in a
1502 conditional pattern, either for a subpattern having been matched, or for
1503 recursion in such a pattern, all the associated numbered subpatterns are
1504 tested, and the overall condition is true if the condition is true for any
1505 one of them. This is the way Perl works, and is also more like the way
1506 testing by number works.
1507
1508
1509 Version 7.9 11-Apr-09
1510 ---------------------
1511
1512 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1513 (pcretest), all targets were linked against these libraries. This included
1514 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1515 libraries. This caused unwanted dependencies to be created. This problem
1516 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1517 pcretest is linked with readline.
1518
1519 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1520 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1521 moved outside it again, because FALSE and TRUE are already defined in AIX,
1522 but BOOL is not.
1523
1524 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1525 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1526
1527 4. The pcregrep documentation said spaces were inserted as well as colons (or
1528 hyphens) following file names and line numbers when outputting matching
1529 lines. This is not true; no spaces are inserted. I have also clarified the
1530 wording for the --colour (or --color) option.
1531
1532 5. In pcregrep, when --colour was used with -o, the list of matching strings
1533 was not coloured; this is different to GNU grep, so I have changed it to be
1534 the same.
1535
1536 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1537 each matching line was coloured. Now it goes on to look for further matches
1538 of any of the test patterns, which is the same behaviour as GNU grep.
1539
1540 7. A pattern that could match an empty string could cause pcregrep to loop; it
1541 doesn't make sense to accept an empty string match in pcregrep, so I have
1542 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1543 seems to be how GNU grep behaves.
1544
1545 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1546 start or after a newline", because the conditional assertion was not being
1547 correctly handled. The rule now is that both the assertion and what follows
1548 in the first alternative must satisfy the test.
1549
1550 9. If auto-callout was enabled in a pattern with a conditional group whose
1551 condition was an assertion, PCRE could crash during matching, both with
1552 pcre_exec() and pcre_dfa_exec().
1553
1554 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1555 used for matching.
1556
1557 11. Unicode property support in character classes was not working for
1558 characters (bytes) greater than 127 when not in UTF-8 mode.
1559
1560 12. Added the -M command line option to pcretest.
1561
1562 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1563
1564 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1565
1566 16. Added comments and documentation about mis-use of no_arg in the C++
1567 wrapper.
1568
1569 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1570 from Martin Jerabek that uses macro names for all relevant character and
1571 string constants.
1572
1573 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1574 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1575 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1576 these, but not everybody uses configure.
1577
1578 19. A conditional group that had only one branch was not being correctly
1579 recognized as an item that could match an empty string. This meant that an
1580 enclosing group might also not be so recognized, causing infinite looping
1581 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1582 with the subject "ab", where knowledge that the repeated group can match
1583 nothing is needed in order to break the loop.
1584
1585 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1586 exec(), but without supplying a callout function, matching went wrong.
1587
1588 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1589 leak if the size of the offset vector was greater than 30. When the vector
1590 is smaller, the saved offsets during recursion go onto a local stack
1591 vector, but for larger vectors malloc() is used. It was failing to free
1592 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1593 error, in fact).
1594
1595 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1596 heapframe that is used only when UTF-8 support is enabled. This caused no
1597 problem, but was untidy.
1598
1599 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1600 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1601 included within another project.
1602
1603 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1604 slightly modified by me:
1605
1606 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1607 not building pcregrep.
1608
1609 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1610 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1611
1612 25. Forward references, both numeric and by name, in patterns that made use of
1613 duplicate group numbers, could behave incorrectly or give incorrect errors,
1614 because when scanning forward to find the reference group, PCRE was not
1615 taking into account the duplicate group numbers. A pattern such as
1616 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1617
1618 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1619 the feature of a custom pointer more persuasive (as requested by a user).
1620
1621 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1622 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1623 pre-defined.
1624
1625 28. Added support for (*UTF8) at the start of a pattern.
1626
1627 29. Arrange for flags added by the "release type" setting in CMake to be shown
1628 in the configuration summary.
1629
1630
1631 Version 7.8 05-Sep-08
1632 ---------------------
1633
1634 1. Replaced UCP searching code with optimized version as implemented for Ad
1635 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1636 stage table and inline lookup instead of a function, giving speed ups of 2
1637 to 5 times on some simple patterns that I tested. Permission was given to
1638 distribute the MultiStage2.py script that generates the tables (it's not in
1639 the tarball, but is in the Subversion repository).
1640
1641 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1642 scripts.
1643
1644 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1645 a group with a zero qualifier. The result of the study could be incorrect,
1646 or the function might crash, depending on the pattern.
1647
1648 4. Caseless matching was not working for non-ASCII characters in back
1649 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1650 It now works when Unicode Property Support is available.
1651
1652 5. In pcretest, an escape such as \x{de} in the data was always generating
1653 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1654 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1655 truncation.
1656
1657 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1658
1659 7. Added two (int) casts to pcregrep when printing the difference of two
1660 pointers, in case they are 64-bit values.
1661
1662 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1663 test 2 if it fails.
1664
1665 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1666 and a #define of that name to empty if it is not externally set. This is to
1667 allow users of MSVC to set it if necessary.
1668
1669 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1670 the convenience functions in the pcre_get.c source file.
1671
1672 11. An option change at the start of a pattern that had top-level alternatives
1673 could cause overwriting and/or a crash. This command provoked a crash in
1674 some environments:
1675
1676 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1677
1678 This potential security problem was recorded as CVE-2008-2371.
1679
1680 12. For a pattern where the match had to start at the beginning or immediately
1681 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1682 pcre_dfa_exec() could read past the end of the passed subject if there was
1683 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1684 pcretest so that it places the subject at the end of its malloc-ed buffer.
1685
1686 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1687 exec() might read past the end of the data buffer in UTF-8 mode.
1688
1689 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1690 the data contained the byte 0x85 as part of a UTF-8 character within its
1691 first line. This applied both to normal and DFA matching.
1692
1693 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1694 /^[^d]*?$/8 failed to match "abc".
1695
1696 16. Added a missing copyright notice to pcrecpp_internal.h.
1697
1698 17. Make it more clear in the documentation that values returned from
1699 pcre_exec() in ovector are byte offsets, not character counts.
1700
1701 18. Tidied a few places to stop certain compilers from issuing warnings.
1702
1703 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1704 supplied by Stefan Weber. I made a further small update for 7.8 because
1705 there is a change of source arrangements: the pcre_searchfuncs.c module is
1706 replaced by pcre_ucd.c.
1707
1708
1709 Version 7.7 07-May-08
1710 ---------------------
1711
1712 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1713 a string to a long long, pretend we don't even have a long long." This is
1714 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1715
1716 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1717 pre-7.6 versions, which defined a global no_arg variable instead of putting
1718 it in the RE class. (See also #8 below.)
1719
1720 3. Remove a line of dead code, identified by coverity and reported by Nuno
1721 Lopes.
1722
1723 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1724
1725 (1) The include/exclude patterns were being applied to the whole pathnames
1726 of files, instead of just to the final components.
1727
1728 (2) If there was more than one level of directory, the subdirectories were
1729 skipped unless they satisfied the include/exclude conditions. This is
1730 inconsistent with GNU grep (and could even be seen as contrary to the
1731 pcregrep specification - which I improved to make it absolutely clear).
1732 The action now is always to scan all levels of directory, and just
1733 apply the include/exclude patterns to regular files.
1734
1735 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1736 --exclude_dir in the tests to avoid scanning .svn directories.
1737
1738 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1739 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1740 doesn't support NULs in patterns.
1741
1742 7. Added some missing "const"s to declarations of static tables in
1743 pcre_compile.c and pcre_dfa_exec.c.
1744
1745 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1746 caused by fix #2 above. (Subsequently also a second patch to fix the
1747 first patch. And a third patch - this was a messy problem.)
1748
1749 9. Applied Craig's patch to remove the use of push_back().
1750
1751 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1752 matching function regexec().
1753
1754 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1755 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1756 references. PCRE supports relative numbers with this syntax (I don't think
1757 Oniguruma does).
1758
1759 12. Previously, a group with a zero repeat such as (...){0} was completely
1760 omitted from the compiled regex. However, this means that if the group
1761 was called as a subroutine from elsewhere in the pattern, things went wrong
1762 (an internal error was given). Such groups are now left in the compiled
1763 pattern, with a new opcode that causes them to be skipped at execution
1764 time.
1765
1766 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1767 to the way PCRE behaves:
1768
1769 (a) A lone ] character is dis-allowed (Perl treats it as data).
1770
1771 (b) A back reference to an unmatched subpattern matches an empty string
1772 (Perl fails the current match path).
1773
1774 (c) A data ] in a character class must be notated as \] because if the
1775 first data character in a class is ], it defines an empty class. (In
1776 Perl it is not possible to have an empty class.) The empty class []
1777 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1778 The negative empty class [^] matches any one character, independently
1779 of the DOTALL setting.
1780
1781 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1782 non-existent subpattern following a character class starting with ']' and
1783 containing () gave an internal compiling error instead of "reference to
1784 non-existent subpattern". Fortunately, when the pattern did exist, the
1785 compiled code was correct. (When scanning forwards to check for the
1786 existencd of the subpattern, it was treating the data ']' as terminating
1787 the class, so got the count wrong. When actually compiling, the reference
1788 was subsequently set up correctly.)
1789
1790 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1791 it was being rejected as not supported by pcre_dfa_exec(), even though
1792 other assertions are supported. I have made pcre_dfa_exec() support
1793 (*FAIL).
1794
1795 16. The implementation of 13c above involved the invention of a new opcode,
1796 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1797 cannot be changed at match time, I realized I could make a small
1798 improvement to matching performance by compiling OP_ALLANY instead of
1799 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1800 on the OP_ANY path.
1801
1802 17. Compiling pcretest on Windows with readline support failed without the
1803 following two fixes: (1) Make the unistd.h include conditional on
1804 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1805
1806 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1807 ncurses library to be included for pcretest when ReadLine support is
1808 requested, but also to allow for it to be overridden. This patch came from
1809 Daniel Bergström.
1810
1811 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1812 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1813 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1814 spotting this.
1815
1816
1817 Version 7.6 28-Jan-08
1818 ---------------------
1819
1820 1. A character class containing a very large number of characters with
1821 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1822 overflow.
1823
1824 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1825 HAVE_LONG_LONG is not defined.
1826
1827 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1828 bring them up to date and include new features. This patch includes:
1829
1830 - Fixed PH's badly added libz and libbz2 support.
1831 - Fixed a problem with static linking.
1832 - Added pcredemo. [But later removed - see 7 below.]
1833 - Fixed dftables problem and added an option.
1834 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1835 HAVE_LONG_LONG.
1836 - Added readline support for pcretest.
1837 - Added an listing of the option settings after cmake has run.
1838
1839 4. A user submitted a patch to Makefile that makes it easy to create
1840 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1841 Makefile.am that cause it to include this special target, without
1842 affecting anything else. Note that the same mingw target plus all
1843 the other distribution libraries and programs are now supported
1844 when configuring with CMake (see 6 below) instead of with
1845 Configure/Make.
1846
1847 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1848 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1849 exported in the Windows port". It has not yet been confirmed that the patch
1850 solves the problem, but it does no harm.
1851
1852 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1853 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1854 with CMake, and also correct the comment about stack recursion.
1855
1856 7. Remove the automatic building of pcredemo from the ./configure system and
1857 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1858 of a program that users should build themselves after PCRE is installed, so
1859 building it automatically is not really right. What is more, it gave
1860 trouble in some build environments.
1861
1862 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1863
1864
1865 Version 7.5 10-Jan-08
1866 ---------------------
1867
1868 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1869 values in parens when parsing an RE using the C++ wrapper."
1870
1871 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1872 Characters greater than 255 were excluded from the class instead of being
1873 included.
1874
1875 3. The same bug as (2) above applied to negated POSIX classes such as
1876 [:^space:].
1877
1878 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1879 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1880 I have changed it.
1881
1882 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1883 first named subpattern) and a construct such as (?&a) would reference the
1884 first named subpattern whose name started with "a" (in other words, the
1885 length check was missing). Both these problems are fixed. "Subpattern name
1886 expected" is now given for (?&) (a zero-length name), and this patch also
1887 makes it give the same error for \k'' (previously it complained that that
1888 was a reference to a non-existent subpattern).
1889
1890 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1891 this is right because (?- can be followed by option settings as well as by
1892 digits. I have, however, made the messages clearer.
1893
1894 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1895 than the number used in the conditional) now cause a compile-time error.
1896 This is actually not compatible with Perl, which accepts such patterns, but
1897 treats the conditional as always being FALSE (as PCRE used to), but it
1898 seems to me that giving a diagnostic is better.
1899
1900 8. Change "alphameric" to the more common word "alphanumeric" in comments
1901 and messages.
1902
1903 9. Fix two occurrences of "backslash" in comments that should have been
1904 "backspace".
1905
1906 10. Remove two redundant lines of code that can never be obeyed (their function
1907 was moved elsewhere).
1908
1909 11. The program that makes PCRE's Unicode character property table had a bug
1910 which caused it to generate incorrect table entries for sequences of
1911 characters that have the same character type, but are in different scripts.
1912 It amalgamated them into a single range, with the script of the first of
1913 them. In other words, some characters were in the wrong script. There were
1914 thirteen such cases, affecting characters in the following ranges:
1915
1916 U+002b0 - U+002c1
1917 U+0060c - U+0060d
1918 U+0061e - U+00612
1919 U+0064b - U+0065e
1920 U+0074d - U+0076d
1921 U+01800 - U+01805
1922 U+01d00 - U+01d77
1923 U+01d9b - U+01dbf
1924 U+0200b - U+0200f
1925 U+030fc - U+030fe
1926 U+03260 - U+0327f
1927 U+0fb46 - U+0fbb1
1928 U+10450 - U+1049d
1929
1930 12. The -o option (show only the matching part of a line) for pcregrep was not
1931 compatible with GNU grep in that, if there was more than one match in a
1932 line, it showed only the first of them. It now behaves in the same way as
1933 GNU grep.
1934
1935 13. If the -o and -v options were combined for pcregrep, it printed a blank
1936 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1937 does the same. The return code can be used to tell if there were any
1938 non-matching lines.
1939
1940 14. Added --file-offsets and --line-offsets to pcregrep.
1941
1942 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1943 infinitely looping recursion. The bug was that positive lookaheads were not
1944 being skipped when checking for a possible empty match (negative lookaheads
1945 and both kinds of lookbehind were skipped).
1946
1947 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1948 inclusion of <windows.h> to before rather than after the definition of
1949 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1950
1951 17. Specifying a possessive quantifier with a specific limit for a Unicode
1952 character property caused pcre_compile() to compile bad code, which led at
1953 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1954 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1955 caused the error; without that there was no problem.
1956
1957 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1958
1959 19. Added --enable-pcretest-libreadline.
1960
1961 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1962 RE::GlobalReplace(). As a result, the number of replacements returned was
1963 double what it should be. I removed one of the increments, but Craig sent a
1964 later patch that removed the other one (the right fix) and added unit tests
1965 that check the return values (which was not done before).
1966
1967 21. Several CMake things:
1968
1969 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1970 the names libpcre and libpcreposix, not just pcre and pcreposix.
1971
1972 (2) The above change means that pcretest and pcregrep are now correctly
1973 linked with the newly-built libraries, not previously installed ones.
1974
1975 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1976
1977 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1978 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1979 UTF-8 newline character). The key issue is that the pattern starts .*;
1980 this means that the match must be either at the beginning, or after a
1981 newline. The bug was in the code for advancing after a failed match and
1982 checking that the new position followed a newline. It was not taking
1983 account of UTF-8 characters correctly.
1984
1985 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1986 character classes. PCRE was not treating the sequence [:...:] as a
1987 character class unless the ... were all letters. Perl, however, seems to
1988 allow any characters between [: and :], though of course it rejects as
1989 unknown any "names" that contain non-letters, because all the known class
1990 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1991 for example, whereas PCRE did not - it did not recognize a POSIX character
1992 class. This seemed a bit dangerous, so the code has been changed to be
1993 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1994 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1995 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1996 Perl does, and where it didn't before.
1997
1998 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1999 Windows environments %n is disabled by default.
2000
2001
2002 Version 7.4 21-Sep-07
2003 ---------------------
2004
2005 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2006 means that a class such as [\s] counted as "explicit reference to CR or
2007 LF". That isn't really right - the whole point of the change was to try to
2008 help when there was an actual mention of one of the two characters. So now
2009 the change happens only if \r or \n (or a literal CR or LF) character is
2010 encountered.
2011
2012 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2013 of both had grown to the point where there were only 3 bits left.
2014 Fortunately, there was spare space in the data structure, and so I have
2015 moved the internal flags into a new 16-bit field to free up more option
2016 bits.
2017
2018 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2019 but did not set the internal JCHANGED flag - either of these is enough to
2020 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2021 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2022 start sets both bits.
2023
2024 4. Added options (at build time, compile time, exec time) to change \R from
2025 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2026
2027 5. doc/pcresyntax.html was missing from the distribution.
2028
2029 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2030 compatibility, even though it is no longer used.
2031
2032 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2033 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2034 windows.h file is present (where different names are used). [This was
2035 reversed later after testing - see 16 below.]
2036
2037 8. Changed all #include <config.h> to #include "config.h". There were also
2038 some further <pcre.h> cases that I changed to "pcre.h".
2039
2040 9. When pcregrep was used with the --colour option, it missed the line ending
2041 sequence off the lines that it output.
2042
2043 10. It was pointed out to me that arrays of string pointers cause lots of
2044 relocations when a shared library is dynamically loaded. A technique of
2045 using a single long string with a table of offsets can drastically reduce
2046 these. I have refactored PCRE in four places to do this. The result is
2047 dramatic:
2048
2049 Originally: 290
2050 After changing UCP table: 187
2051 After changing error message table: 43
2052 After changing table of "verbs" 36
2053 After changing table of Posix names 22
2054
2055 Thanks to the folks working on Gregex for glib for this insight.
2056
2057 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2058 unicode-properties was also set.
2059
2060 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2061
2062 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2063 checked only for CRLF.
2064
2065 14. Added casts to pcretest.c to avoid compiler warnings.
2066
2067 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2068
2069 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2070 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2071 entirely. This removes changes made in 7 above.
2072
2073 17. The CMake files have been updated, and there is now more information about
2074 building with CMake in the NON-UNIX-USE document.
2075
2076
2077 Version 7.3 28-Aug-07
2078 ---------------------
2079
2080 1. In the rejigging of the build system that eventually resulted in 7.1, the
2081 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2082 brackets there is not right, since it causes compilers to look for an
2083 installed pcre.h, not the version that is in the source that is being
2084 compiled (which of course may be different). I have changed it back to:
2085
2086 #include "pcre.h"
2087
2088 I have a vague recollection that the change was concerned with compiling in
2089 different directories, but in the new build system, that is taken care of
2090 by the VPATH setting the Makefile.
2091
2092 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2093 when the subject happened to end in the byte 0x85 (e.g. if the last
2094 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2095 characters but of course it shouldn't be taken as a newline when it is part
2096 of another character. The bug was that, for an unlimited repeat of . in
2097 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2098 characters when looking for a newline.
2099
2100 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2101
2102 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2103 in debug output.
2104
2105 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2106 long printing in the pcrecpp unittest when running under MinGW.
2107
2108 6. ESC_K was left out of the EBCDIC table.
2109
2110 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2111 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2112 limit also applies to "virtual nesting" when a pattern is recursive, and in
2113 this case 1000 isn't so big. I have been able to remove this limit at the
2114 expense of backing off one optimization in certain circumstances. Normally,
2115 when pcre_exec() would call its internal match() function recursively and
2116 immediately return the result unconditionally, it uses a "tail recursion"
2117 feature to save stack. However, when a subpattern that can match an empty
2118 string has an unlimited repetition quantifier, it no longer makes this
2119 optimization. That gives it a stack frame in which to save the data for
2120 checking that an empty string has been matched. Previously this was taken
2121 from the 1000-entry workspace that had been reserved. So now there is no
2122 explicit limit, but more stack is used.
2123
2124 8. Applied Daniel's patches to solve problems with the import/export magic
2125 syntax that is required for Windows, and which was going wrong for the
2126 pcreposix and pcrecpp parts of the library. These were overlooked when this
2127 problem was solved for the main library.
2128
2129 9. There were some crude static tests to avoid integer overflow when computing
2130 the size of patterns that contain repeated groups with explicit upper
2131 limits. As the maximum quantifier is 65535, the maximum group length was
2132 set at 30,000 so that the product of these two numbers did not overflow a
2133 32-bit integer. However, it turns out that people want to use groups that
2134 are longer than 30,000 bytes (though not repeat them that many times).
2135 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2136 made it possible to implement the integer overflow checks in a much more
2137 dynamic way, which I have now done. The artificial limitation on group
2138 length has been removed - we now have only the limit on the total length of
2139 the compiled pattern, which depends on the LINK_SIZE setting.
2140
2141 10. Fixed a bug in the documentation for get/copy named substring when
2142 duplicate names are permitted. If none of the named substrings are set, the
2143 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2144 empty string.
2145
2146 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2147 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2148 because the ] is interpreted as the first data character and the
2149 terminating ] is not found. PCRE has been made compatible with Perl in this
2150 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2151 cause memory overwriting.
2152
2153 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2154 string has been matched (to stop an infinite loop). It was not recognizing
2155 a conditional subpattern that could match an empty string if that
2156 subpattern was within another subpattern. For example, it looped when
2157 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2158 condition was not nested. This bug has been fixed.
2159
2160 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2161 past the start of the subject in the presence of bytes with the top bit
2162 set, for example "\x8aBCD".
2163
2164 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2165 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2166
2167 14. Optimized (?!) to (*FAIL).
2168
2169 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2170 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2171 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2172 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2173 does: it's just the validity check that is more restrictive.
2174
2175 16. Inserted checks for integer overflows during escape sequence (backslash)
2176 processing, and also fixed erroneous offset values for syntax errors during
2177 backslash processing.
2178
2179 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2180 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2181
2182 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2183 caused an overrun.
2184
2185 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2186 something other than just ASCII characters) inside a group that had an
2187 unlimited repeat caused a loop at compile time (while checking to see
2188 whether the group could match an empty string).
2189
2190 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2191 [\P{Any}] did so. (Error in the code for printing property names.)
2192
2193 21. An orphan \E inside a character class could cause a crash.
2194
2195 22. A repeated capturing bracket such as (A)? could cause a wild memory
2196 reference during compilation.
2197
2198 23. There are several functions in pcre_compile() that scan along a compiled
2199 expression for various reasons (e.g. to see if it's fixed length for look
2200 behind). There were bugs in these functions when a repeated \p or \P was
2201 present in the pattern. These operators have additional parameters compared
2202 with \d, etc, and these were not being taken into account when moving along
2203 the compiled data. Specifically:
2204
2205 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2206 length.
2207
2208 (b) An item such as \pL+ within a repeated group could cause crashes or
2209 loops.
2210
2211 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2212 "reference to non-existent subpattern" error.
2213
2214 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2215
2216 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2217 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2218
2219 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2220
2221 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2222 character were causing crashes (broken optimization).
2223
2224 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2225 \p or \P) caused a compile-time loop.
2226
2227 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2228 break. For example, the unstudied pattern [\r\n]A does not match the string
2229 "\r\nA" because change 7.0/46 below moves the current point on by two
2230 characters after failing to match at the start. However, the pattern \nA
2231 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2232 the same is true. There doesn't seem any very clean way out of this, but
2233 what I have chosen to do makes the common cases work: PCRE now takes note
2234 of whether there can be an explicit match for \r or \n anywhere in the
2235 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2236 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2237 pattern has explicit CR or LF references.
2238
2239 29. Added (*CR) etc for changing newline setting at start of pattern.
2240
2241
2242 Version 7.2 19-Jun-07
2243 ---------------------
2244
2245 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2246 which is apparently normally available under Windows.
2247
2248 2. Re-jig the pcregrep tests with different newline settings in an attempt
2249 to make them independent of the local environment's newline setting.
2250
2251 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2252
2253 4. Some of the "internals" tests were previously cut out when the link size
2254 was not 2, because the output contained actual offsets. The recent new
2255 "Z" feature of pcretest means that these can be cut out, making the tests
2256 usable with all link sizes.
2257
2258 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2259 stack recursion. This gives a massive performance boost under BSD, but just
2260 a small improvement under Linux. However, it saves one field in the frame
2261 in all cases.
2262
2263 6. Added more features from the forthcoming Perl 5.10:
2264
2265 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2266 recursion call. It refers to the nth most recently opened parentheses.
2267
2268 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2269 to be opened parentheses.
2270
2271 (c) Conditions that refer to capturing parentheses can be specified
2272 relatively, for example, (?(-2)... or (?(+3)...
2273
2274 (d) \K resets the start of the current match so that everything before
2275 is not part of it.
2276
2277 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2278
2279 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2280 reference syntax.
2281
2282 (g) (?| introduces a group in which the numbering of parentheses in each
2283 alternative starts with the same number.
2284
2285 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2286
2287 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2288 PCRE_INFO_JCHANGED.
2289
2290 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2291 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2292 for detecting groups that can match an empty string.
2293
2294 9. A pattern with a very large number of alternatives (more than several
2295 hundred) was running out of internal workspace during the pre-compile
2296 phase, where pcre_compile() figures out how much memory will be needed. A
2297 bit of new cunning has reduced the workspace needed for groups with
2298 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2299 workspace instead of running out of the 4096 that are available.
2300
2301 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2302
2303 11. Applied patch from Google to remove an optimization that didn't quite work.
2304 The report of the bug said:
2305
2306 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2307 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2308 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2309
2310 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2311 it matched the wrong number of bytes.
2312
2313
2314 Version 7.1 24-Apr-07
2315 ---------------------
2316
2317 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2318 that is more "standard", making use of automake and other Autotools. There
2319 is some re-arrangement of the files and adjustment of comments consequent
2320 on this.
2321
2322 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2323 for recursive directory scanning broke on some systems because the files
2324 are not scanned in any specific order and on different systems the order
2325 was different. A call to "sort" has been inserted into RunGrepTest for the
2326 approprate test as a short-term fix. In the longer term there may be an
2327 alternative.
2328
2329 3. I had an email from Eric Raymond about problems translating some of PCRE's
2330 man pages to HTML (despite the fact that I distribute HTML pages, some
2331 people do their own conversions for various reasons). The problems
2332 concerned the use of low-level troff macros .br and .in. I have therefore
2333 removed all such uses from the man pages (some were redundant, some could
2334 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2335 HTML has been updated to handle .nf/.fi and to complain if it encounters
2336 .br or .in.
2337
2338 4. Updated comments in configure.ac that get placed in config.h.in and also
2339 arranged for config.h to be included in the distribution, with the name
2340 config.h.generic, for the benefit of those who have to compile without
2341 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2342
2343 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2344 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2345 makevp.bat for the current PCRE, using the additional files
2346 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2347
2348 6. A Windows user reported a minor discrepancy with test 2, which turned out
2349 to be caused by a trailing space on an input line that had got lost in his
2350 copy. The trailing space was an accident, so I've just removed it.
2351
2352 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2353 that is needed.
2354
2355 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2356 as "const" (a) because they are and (b) because it helps the PHP
2357 maintainers who have recently made a script to detect big data structures
2358 in the php code that should be moved to the .rodata section. I remembered
2359 to update Builducptable as well, so it won't revert if ucptable.h is ever
2360 re-created.
2361
2362 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2363 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2364 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2365 support is not required. This saves 1.5-2K of code, which is important in
2366 some applications.
2367
2368 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2369 so as not to refer to the tables, even though these functions will never be
2370 called when UTF-8 support is disabled. Otherwise there are problems with a
2371 shared library.
2372
2373 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2374
2375 (a) It was defining its arguments as char * instead of void *.
2376
2377 (b) It was assuming that all moves were upwards in memory; this was true
2378 a long time ago when I wrote it, but is no longer the case.
2379
2380 The emulated memove() is provided for those environments that have neither
2381 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2382 is clearly not the case, as these two bugs were recently reported.
2383
2384 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2385 and Detrail to create the HTML documentation, the .txt form of the man
2386 pages, and it removes trailing spaces from listed files. It also creates
2387 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2388 case, it wraps all the #defines with #ifndefs. This script should be run
2389 before "make dist".
2390
2391 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2392 with Unicode property support.
2393
2394 (a) For a maximizing quantifier, if the two different cases of the
2395 character were of different lengths in their UTF-8 codings (there are
2396 some cases like this - I found 11), and the matching function had to
2397 back up over a mixture of the two cases, it incorrectly assumed they
2398 were both the same length.
2399
2400 (b) When PCRE was configured to use the heap rather than the stack for
2401 recursion during matching, it was not correctly preserving the data for
2402 the other case of a UTF-8 character when checking ahead for a match
2403 while processing a minimizing repeat. If the check also involved
2404 matching a wide character, but failed, corruption could cause an
2405 erroneous result when trying to check for a repeat of the original
2406 character.
2407
2408 13. Some tidying changes to the testing mechanism:
2409
2410 (a) The RunTest script now detects the internal link size and whether there
2411 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2412 values substituted by "configure". (The RunGrepTest script already did
2413 this for UTF-8.) The configure.ac script no longer substitutes the
2414 relevant variables.
2415
2416 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2417 with length and offset values. This means that the output is different
2418 for different internal link sizes. Test 2 is skipped for link sizes
2419 other than 2 because of this, bypassing the problem. Unfortunately,
2420 there was also a test in test 3 (the locale tests) that used /B and
2421 failed for link sizes other than 2. Rather than cut the whole test out,
2422 I have added a new /Z option to pcretest that replaces the length and
2423 offset values with spaces. This is now used to make test 3 independent
2424 of link size. (Test 2 will be tidied up later.)
2425
2426 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2427 segmentation fault instead of returning the appropriate error message.
2428
2429 15. In multiline mode when the newline sequence was set to "any", the pattern
2430 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2431 This doesn't seem right; it now treats the CRLF combination as the line
2432 ending, and so does not match in that case. It's only a pattern such as ^$
2433 that would hit this one: something like ^ABC$ would have failed after \r
2434 and then tried again after \r\n.
2435
2436 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2437 in an attempt to make files that differ only in their line terminators
2438 compare equal. This works on Linux.
2439
2440 17. Under certain error circumstances pcregrep might try to free random memory
2441 as it exited. This is now fixed, thanks to valgrind.
2442
2443 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2444 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2445 was because its rules for how to advance for /g after matching an empty
2446 string at the end of a line did not allow for this case. They now check for
2447 it specially.
2448
2449 20. pcretest is supposed to handle patterns and data of any length, by
2450 extending its buffers when necessary. It was getting this wrong when the
2451 buffer for a data line had to be extended.
2452
2453 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2454 CRLF as a newline sequence.
2455
2456 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2457 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2458 I have nevertheless tidied it up.
2459
2460 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2461
2462 24. Added a man page for pcre-config.
2463
2464
2465 Version 7.0 19-Dec-06
2466 ---------------------
2467
2468 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2469 moving to gcc 4.1.1.
2470
2471 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2472 sys/time.h, which is documented as needed for this function. It doesn't
2473 seem to matter on Linux, but it showed up on some releases of OS X.
2474
2475 3. It seems that there are systems where bytes whose values are greater than
2476 127 match isprint() in the "C" locale. The "C" locale should be the
2477 default when a C program starts up. In most systems, only ASCII printing
2478 characters match isprint(). This difference caused the output from pcretest
2479 to vary, making some of the tests fail. I have changed pcretest so that:
2480
2481 (a) When it is outputting text in the compiled version of a pattern, bytes
2482 other than 32-126 are always shown as hex escapes.
2483
2484 (b) When it is outputting text that is a matched part of a subject string,
2485 it does the same, unless a different locale has been set for the match
2486 (using the /L modifier). In this case, it uses isprint() to decide.
2487
2488 4. Fixed a major bug that caused incorrect computation of the amount of memory
2489 required for a compiled pattern when options that changed within the
2490 pattern affected the logic of the preliminary scan that determines the
2491 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2492 that the computed length was too small. The symptoms of this bug were
2493 either the PCRE error "internal error: code overflow" from pcre_compile(),
2494 or a glibc crash with a message such as "pcretest: free(): invalid next
2495 size (fast)". Examples of patterns that provoked this bug (shown in
2496 pcretest format) are:
2497
2498 /(?-x: )/x
2499 /(?x)(?-x: \s*#\s*)/
2500 /((?i)[\x{c0}])/8
2501 /(?i:[\x{c0}])/8
2502
2503 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2504 is now done differently.
2505
2506 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2507 wrapper classes; (b) implement a new function in the C++ scanner that is
2508 more efficient than the old way of doing things because it avoids levels of
2509 recursion in the regex matching; (c) add a paragraph to the documentation
2510 for the FullMatch() function.
2511
2512 6. The escape sequence \n was being treated as whatever was defined as
2513 "newline". Not only was this contrary to the documentation, which states
2514 that \n is character 10 (hex 0A), but it also went horribly wrong when
2515 "newline" was defined as CRLF. This has been fixed.
2516
2517 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2518 was being set to -1 for the "end of line" case (supposedly a value that no
2519 character can have). Though this value is never used (the check for end of
2520 line is "zero bytes in current character"), it caused compiler complaints.
2521 I've changed it to 0xffffffff.
2522
2523 8. In pcre_version.c, the version string was being built by a sequence of
2524 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2525 string (as it is for production releases) called a macro with an empty
2526 argument. The C standard says the result of this is undefined. The gcc
2527 compiler treats it as an empty string (which was what was wanted) but it is
2528 reported that Visual C gives an error. The source has been hacked around to
2529 avoid this problem.
2530
2531 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2532 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2533 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2534 of them did).
2535
2536 10. Originally, pcretest opened its input and output without "b"; then I was
2537 told that "b" was needed in some environments, so it was added for release
2538 5.0 to both the input and output. (It makes no difference on Unix-like
2539 systems.) Later I was told that it is wrong for the input on Windows. I've
2540 now abstracted the modes into two macros, to make it easier to fiddle with
2541 them, and removed "b" from the input mode under Windows.
2542
2543 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2544
2545 12. Added -help and --help to pcretest as an official way of being reminded
2546 of the options.
2547
2548 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2549 and pcrecpp.cc because they annoy compilers at high warning levels.
2550
2551 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2552
2553 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2554 scripts are not C programs :-) and which was not noticed because it works
2555 on Linux.
2556
2557 16. pcretest is supposed to handle any length of pattern and data line (as one
2558 line or as a continued sequence of lines) by extending its input buffer if
2559 necessary. This feature was broken for very long pattern lines, leading to
2560 a string of junk being passed to pcre_compile() if the pattern was longer
2561 than about 50K.
2562
2563 17. I have done a major re-factoring of the way pcre_compile() computes the
2564 amount of memory needed for a compiled pattern. Previously, there was code
2565 that made a preliminary scan of the pattern in order to do this. That was
2566 OK when PCRE was new, but as the facilities have expanded, it has become
2567 harder and harder to keep it in step with the real compile phase, and there
2568 have been a number of bugs (see for example, 4 above). I have now found a
2569 cunning way of running the real compile function in a "fake" mode that
2570 enables it to compute how much memory it would need, while actually only
2571 ever using a few hundred bytes of working memory and without too many
2572 tests of the mode. This should make future maintenance and development
2573 easier. A side effect of this work is that the limit of 200 on the nesting
2574 depth of parentheses has been removed (though this was never a serious
2575 limitation, I suspect). However, there is a downside: pcre_compile() now
2576 runs more slowly than before (30% or more, depending on the pattern). I
2577 hope this isn't a big issue. There is no effect on runtime performance.
2578
2579 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2580 newline (only possible for the last line of a file) and it was a
2581 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2582
2583 19. Added additional timing features to pcretest. (1) The -tm option now times
2584 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2585 separate command line item, by a number that specifies the number of
2586 repeats to use when timing. The default is 50000; this gives better
2587 precision, but takes uncomfortably long for very large patterns.
2588
2589 20. Extended pcre_study() to be more clever in cases where a branch of a
2590 subpattern has no definite first character. For example, (a*|b*)[cd] would
2591 previously give no result from pcre_study(). Now it recognizes that the
2592 first character must be a, b, c, or d.
2593
2594 21. There was an incorrect error "recursive call could loop indefinitely" if
2595 a subpattern (or the entire pattern) that was being tested for matching an
2596 empty string contained only one non-empty item after a nested subpattern.
2597 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2598 incorrectly, because the \d was being skipped in the check.
2599
2600 22. The pcretest program now has a new pattern option /B and a command line
2601 option -b, which is equivalent to adding /B to every pattern. This causes
2602 it to show the compiled bytecode, without the additional information that
2603 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2604 is the same as /B/I).
2605
2606 23. A new optimization is now able automatically to treat some sequences such
2607 as a*b as a*+b. More specifically, if something simple (such as a character
2608 or a simple class like \d) has an unlimited quantifier, and is followed by
2609 something that cannot possibly match the quantified thing, the quantifier
2610 is automatically "possessified".
2611
2612 24. A recursive reference to a subpattern whose number was greater than 39
2613 went wrong under certain circumstances in UTF-8 mode. This bug could also
2614 have affected the operation of pcre_study().
2615
2616 25. Realized that a little bit of performance could be had by replacing
2617 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2618
2619 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2620
2621 27. Possessive quantifiers such as a++ were previously implemented by turning
2622 them into atomic groups such as ($>a+). Now they have their own opcodes,
2623 which improves performance. This includes the automatically created ones
2624 from 23 above.
2625
2626 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2627 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2628 the first matched character to be a colon. This applied both to named and
2629 numbered groups.
2630
2631 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2632
2633 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2634 building PCRE on Windows easier, so I have included it in the distribution.
2635
2636 31. There is now a check in pcretest against a ridiculously large number being
2637 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2638 loop, the loop is abandoned.
2639
2640 32. Forward references to subpatterns in conditions such as (?(2)...) where
2641 subpattern 2 is defined later cause pcre_compile() to search forwards in
2642 the pattern for the relevant set of parentheses. This search went wrong
2643 when there were unescaped parentheses in a character class, parentheses
2644 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2645
2646 33. "Subroutine" calls and backreferences were previously restricted to
2647 referencing subpatterns earlier in the regex. This restriction has now
2648 been removed.
2649
2650 34. Added a number of extra features that are going to be in Perl 5.10. On the
2651 whole, these are just syntactic alternatives for features that PCRE had
2652 previously implemented using the Python syntax or my own invention. The
2653 other formats are all retained for compatibility.
2654
2655 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2656 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2657 also .NET compatible.
2658
2659 (b) A recursion or subroutine call to a named group can now be defined as
2660 (?&name) as well as (?P>name).
2661
2662 (c) A backreference to a named group can now be defined as \k<name> or
2663 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2664 5.10, are also .NET compatible.
2665
2666 (d) A conditional reference to a named group can now use the syntax
2667 (?(<name>) or (?('name') as well as (?(name).
2668
2669 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2670 groups (named and numbered) that are never evaluated inline, but can be
2671 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2672 is always false. There may be only one alternative in such a group.
2673
2674 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2675 as the simple (?(R). The condition is true only if the most recent
2676 recursion is that of the given number or name. It does not search out
2677 through the entire recursion stack.
2678
2679 (g) The escape \gN or \g{N} has been added, where N is a positive or
2680 negative number, specifying an absolute or relative reference.
2681
2682 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2683 some "unreachable code" warnings.
2684
2685 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2686 things, this adds five new scripts.
2687
2688 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2689 There were also incompatibilities regarding the handling of \Q..\E inside
2690 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2691 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2692
2693 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2694 matches an empty string, and forcibly breaks the loop. There were bugs in
2695 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2696 against aaaa the result was just "a" rather than "aaaa", for example. Two
2697 separate and independent bugs (that affected different cases) have been
2698 fixed.
2699
2700 39. Refactored the code to abolish the use of different opcodes for small
2701 capturing bracket numbers. This is a tidy that I avoided doing when I
2702 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2703 The new approach is not only tidier, it makes it possible to reduce the
2704 memory needed to fix the previous bug (38).
2705
2706 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2707 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2708 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2709 mode.
2710
2711 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2712 report.
2713
2714 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2715 copy construction and assignment in the C++ wrapper.
2716
2717 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2718 couple of bugs that could have given wrong results in the "--newline=crlf"
2719 case.
2720
2721 44. Added a number of casts and did some reorganization of signed/unsigned int
2722 variables following suggestions from Dair Grant. Also renamed the variable
2723 "this" as "item" because it is a C++ keyword.
2724
2725 45. Arranged for dftables to add
2726
2727 #include "pcre_internal.h"
2728
2729 to pcre_chartables.c because without it, gcc 4.x may remove the array
2730 definition from the final binary if PCRE is built into a static library and
2731 dead code stripping is activated.
2732
2733 46. For an unanchored pattern, if a match attempt fails at the start of a
2734 newline sequence, and the newline setting is CRLF or ANY, and the next two
2735 characters are CRLF, advance by two characters instead of one.
2736
2737
2738 Version 6.7 04-Jul-06
2739 ---------------------
2740
2741 1. In order to handle tests when input lines are enormously long, pcretest has
2742 been re-factored so that it automatically extends its buffers when
2743 necessary. The code is crude, but this _is_ just a test program. The
2744 default size has been increased from 32K to 50K.
2745
2746 2. The code in pcre_study() was using the value of the re argument before
2747 testing it for NULL. (Of course, in any sensible call of the function, it
2748 won't be NULL.)
2749
2750 3. The memmove() emulation function in pcre_internal.h, which is used on
2751 systems that lack both memmove() and bcopy() - that is, hardly ever -
2752 was missing a "static" storage class specifier.
2753
2754 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2755 containing an extended class (one that cannot be represented by a bitmap
2756 because it contains high-valued characters or Unicode property items, e.g.
2757 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2758 pattern, but PCRE should not loop if you do not (it no longer does).
2759 [Detail: two cases were found: (a) a repeated subpattern containing an
2760 extended class; (b) a recursive reference to a subpattern that followed a
2761 previous extended class. It wasn't skipping over the extended class
2762 correctly when UTF-8 mode was not set.]
2763
2764 5. A negated single-character class was not being recognized as fixed-length
2765 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2766 compile error "lookbehind assertion is not fixed length".
2767
2768 6. The RunPerlTest auxiliary script was showing an unexpected difference
2769 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2770 write a Perl script that can interpret lines of an input file either as
2771 byte characters or as UTF-8, which is what "perltest" was being required to
2772 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2773 can't do is switch easily at run time between having the "use utf8;" pragma
2774 or not. In the end, I fudged it by using the RunPerlTest script to insert
2775 "use utf8;" explicitly for the UTF-8 tests.
2776
2777 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2778 the end of the subject string, contrary to the documentation and to what
2779 Perl does. This was true of both matching functions. Now it matches only at
2780 the start of the subject and immediately after *internal* newlines.
2781
2782 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2783 a pointer to an int instead of a pointer to an unsigned long int. This
2784 caused problems on 64-bit systems.
2785
2786 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2787 instance of the 'standard' template library not being so standard".
2788
2789 10. There was no check on the number of named subpatterns nor the maximum
2790 length of a subpattern name. The product of these values is used to compute
2791 the size of the memory block for a compiled pattern. By supplying a very
2792 long subpattern name and a large number of named subpatterns, the size
2793 computation could be caused to overflow. This is now prevented by limiting
2794 the length of names to 32 characters, and the number of named subpatterns
2795 to 10,000.
2796
2797 11. Subpatterns that are repeated with specific counts have to be replicated in
2798 the compiled pattern. The size of memory for this was computed from the
2799 length of the subpattern and the repeat count. The latter is limited to
2800 65535, but there was no limit on the former, meaning that integer overflow
2801 could in principle occur. The compiled length of a repeated subpattern is
2802 now limited to 30,000 bytes in order to prevent this.
2803
2804 12. Added the optional facility to have named substrings with the same name.
2805
2806 13. Added the ability to use a named substring as a condition, using the
2807 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2808 are numbers (not recommended). Forward references are permitted.
2809
2810 14. Added forward references in named backreferences (if you see what I mean).
2811
2812 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2813 pattern could run off the end of the subject. For example, the pattern
2814 "(?s)(.{1,5})"8 did this with the subject "ab".
2815
2816 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2817 PCRE_CASELESS was set when matching characters that were quantified with ?
2818 or *.
2819
2820 17. A character class other than a single negated character that had a minimum
2821 but no maximum quantifier - for example [ab]{6,} - was not handled
2822 correctly by pce_dfa_exec(). It would match only one character.
2823
2824 18. A valid (though odd) pattern that looked like a POSIX character
2825 class but used an invalid character after [ (for example [[,abc,]]) caused
2826 pcre_compile() to give the error "Failed: internal error: code overflow" or
2827 in some cases to crash with a glibc free() error. This could even happen if
2828 the pattern terminated after [[ but there just happened to be a sequence of
2829 letters, a binary zero, and a closing ] in the memory that followed.
2830
2831 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2832 over the years. Originally (before any Unicode support), just the bottom 8
2833 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2834 output from "man perlunicode" includes this:
2835
2836 The regular expression compiler produces polymorphic opcodes. That
2837 is, the pattern adapts to the data and automatically switches to
2838 the Unicode character scheme when presented with Unicode data--or
2839 instead uses a traditional byte scheme when presented with byte
2840 data.
2841
2842 Sadly, a wide octal escape does not cause a switch, and in a string with
2843 no other multibyte characters, these octal escapes are treated as before.
2844 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2845 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2846 Unicode string.
2847
2848 I have not perpetrated such confusion in PCRE. Up till now, it took just
2849 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2850 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2851 translate to the appropriate multibyte character.
2852
2853 29. Applied some refactoring to reduce the number of warnings from Microsoft
2854 and Borland compilers. This has included removing the fudge introduced
2855 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2856 a warning about an unused variable.
2857
2858 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2859 characters since release 4.0, because Perl (from release 5.004) does not.
2860 [Or at least, is documented not to: some releases seem to be in conflict
2861 with the documentation.] However, when a pattern was studied with
2862 pcre_study() and all its branches started with \s, PCRE still included VT
2863 as a possible starting character. Of course, this did no harm; it just
2864 caused an unnecessary match attempt.
2865
2866 22. Removed a now-redundant internal flag bit that recorded the fact that case
2867 dependency changed within the pattern. This was once needed for "required
2868 byte" processing, but is no longer used. This recovers a now-scarce options
2869 bit. Also moved the least significant internal flag bit to the most-
2870 significant bit of the word, which was not previously used (hangover from
2871 the days when it was an int rather than a uint) to free up another bit for
2872 the future.
2873
2874 23. Added support for CRLF line endings as well as CR and LF. As well as the
2875 default being selectable at build time, it can now be changed at runtime
2876 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2877 specify that it is scanning data with non-default line endings.
2878
2879 24. Changed the definition of CXXLINK to make it agree with the definition of
2880 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2881
2882 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2883 recursions. This makes a big different to stack usage for some patterns.
2884
2885 26. If a subpattern containing a named recursion or subroutine reference such
2886 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2887 the space required for the compiled pattern went wrong and gave too small a
2888 value. Depending on the environment, this could lead to "Failed: internal
2889 error: code overflow at offset 49" or "glibc detected double free or
2890 corruption" errors.
2891
2892 27. Applied patches from Google (a) to support the new newline modes and (b) to
2893 advance over multibyte UTF-8 characters in GlobalReplace.
2894
2895 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2896 difference for some implementation of PCRE in some Windows version.
2897
2898 29. Added some extra testing facilities to pcretest:
2899
2900 \q<number> in a data line sets the "match limit" value
2901 \Q<number> in a data line sets the "match recursion limt" value
2902 -S <number> sets the stack size, where <number> is in megabytes
2903
2904 The -S option isn't available for Windows.
2905
2906
2907 Version 6.6 06-Feb-06
2908 ---------------------
2909
2910 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2911 in pcreposix.h. I have copied the definition from pcre.h.
2912
2913 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2914 because pcre.h is no longer a built file.
2915
2916 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2917 not normally included in the compiled code.
2918
2919
2920 Version 6.5 01-Feb-06
2921 ---------------------
2922
2923 1. When using the partial match feature with pcre_dfa_exec(), it was not
2924 anchoring the second and subsequent partial matches at the new starting
2925 point. This could lead to incorrect results. For example, with the pattern
2926 /1234/, partially matching against "123" and then "a4" gave a match.
2927
2928 2. Changes to pcregrep:
2929
2930 (a) All non-match returns from pcre_exec() were being treated as failures
2931 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2932 error message is output. Some extra information is given for the
2933 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2934 probably the only errors that are likely to be caused by users (by
2935 specifying a regex that has nested indefinite repeats, for instance).
2936 If there are more than 20 of these errors, pcregrep is abandoned.
2937
2938 (b) A binary zero was treated as data while matching, but terminated the
2939 output line if it was written out. This has been fixed: binary zeroes
2940 are now no different to any other data bytes.
2941
2942 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2943 used to set a locale for matching. The --locale=xxxx long option has
2944 been added (no short equivalent) to specify a locale explicitly on the
2945 pcregrep command, overriding the environment variables.
2946
2947 (d) When -B was used with -n, some line numbers in the output were one less
2948 than they should have been.
2949
2950 (e) Added the -o (--only-matching) option.
2951
2952 (f) If -A or -C was used with -c (count only), some lines of context were
2953 accidentally printed for the final match.
2954
2955 (g) Added the -H (--with-filename) option.
2956
2957 (h) The combination of options -rh failed to suppress file names for files
2958 that were found from directory arguments.
2959
2960 (i) Added the -D (--devices) and -d (--directories) options.
2961
2962 (j) Added the -F (--fixed-strings) option.
2963
2964 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2965
2966 (l) Added the --colo(u)r option.
2967
2968 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2969 is not present by default.
2970
2971 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2972 items such as (?R) or (?1), when the recursion could match a number of
2973 alternatives. If it matched one of the alternatives, but subsequently,
2974 outside the recursion, there was a failure, the code tried to back up into
2975 the recursion. However, because of the way PCRE is implemented, this is not
2976 possible, and the result was an incorrect result from the match.
2977
2978 In order to prevent this happening, the specification of recursion has
2979 been changed so that all such subpatterns are automatically treated as
2980 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2981
2982 4. I had overlooked the fact that, in some locales, there are characters for
2983 which isalpha() is true but neither isupper() nor islower() are true. In
2984 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2985 and ordfeminine) are like this. This affected the treatment of \w and \W
2986 when they appeared in character classes, but not when they appeared outside
2987 a character class. The bit map for "word" characters is now created
2988 separately from the results of isalnum() instead of just taking it from the
2989 upper, lower, and digit maps. (Plus the underscore character, of course.)
2990
2991 5. The above bug also affected the handling of POSIX character classes such as
2992 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2993 permanent tables. Instead, the bit maps for such a class were previously
2994 created as the appropriate unions of the upper, lower, and digit bitmaps.
2995 Now they are created by subtraction from the [[:word:]] class, which has
2996 its own bitmap.
2997
2998 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2999 It is created by subtracting the vertical space characters (\x09, \x0a,
3000 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3001 subtraction was done in the overall bitmap for a character class, meaning
3002 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3003 be recognized. This bug has been fixed.
3004
3005 7. Patches from the folks at Google:
3006
3007 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3008 real life, but is still worth protecting against".
3009
3010 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3011 regular expressions".
3012
3013 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3014 have it.
3015
3016 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3017 "configure" and the latter not, in order to fix a problem somebody had
3018 with compiling the Arg class on HP-UX.
3019
3020 (e) Improve the error-handling of the C++ wrapper a little bit.
3021
3022 (f) New tests for checking recursion limiting.
3023
3024 8. The pcre_memmove() function, which is used only if the environment does not
3025 have a standard memmove() function (and is therefore rarely compiled),
3026 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3027 returning a result (though PCRE never actually uses the result).
3028
3029 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3030 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3031 returned instead of calling malloc() with an overflowing number that would
3032 most likely cause subsequent chaos.
3033
3034 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3035
3036 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3037 with this option is matched, the nmatch and pmatch options of regexec() are
3038 ignored.
3039
3040 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3041 provided in case anyone wants to the the POSIX interface with UTF-8
3042 strings.
3043
3044 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3045 C++ linking (needed for some HP-UX environments).
3046
3047 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3048 (unused parameter) and in the pcre_printint() function (omitted "default"
3049 switch label when the default is to do nothing).
3050
3051 15. Added some code to make it possible, when PCRE is compiled as a C++
3052 library, to replace subject pointers for pcre_exec() with a smart pointer
3053 class, thus making it possible to process discontinuous strings.
3054
3055 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3056 much the same function. They were added by different people who were trying
3057 to make PCRE easy to compile on non-Unix systems. It has been suggested
3058 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3059 for compiling on Windows systems. I have therefore replaced it with
3060 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3061 defaults to "extern" for C or "extern C" for C++, which works fine on
3062 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3063 SCOPE with something explicit in config.h. In addition:
3064
3065 (a) pcreposix.h still had just "extern" instead of either of these macros;
3066 I have replaced it with PCRE_DATA_SCOPE.
3067
3068 (b) Functions such as _pcre_xclass(), which are internal to the library,
3069 but external in the C sense, all had PCRE_EXPORT in their definitions.
3070 This is apparently wrong for the Windows case, so I have removed it.
3071 (It makes no difference on Unix-like systems.)
3072
3073 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3074 of recursive calls to match(). This is different to MATCH_LIMIT because
3075 that limits the total number of calls to match(), not all of which increase
3076 the depth of recursion. Limiting the recursion depth limits the amount of
3077 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3078 when PCRE is compiled, and changed at run time. A patch from Google adds
3079 this functionality to the C++ interface.
3080
3081 18. Changes to the handling of Unicode character properties:
3082
3083 (a) Updated the table to Unicode 4.1.0.
3084
3085 (b) Recognize characters that are not in the table as "Cn" (undefined).
3086
3087 (c) I revised the way the table is implemented to a much improved format
3088 which includes recognition of ranges. It now supports the ranges that
3089 are defined in UnicodeData.txt, and it also amalgamates other
3090 characters into ranges. This has reduced the number of entries in the
3091 table from around 16,000 to around 3,000, thus reducing its size
3092 considerably. I realized I did not need to use a tree structure after
3093 all - a binary chop search is just as efficient. Having reduced the
3094 number of entries, I extended their size from 6 bytes to 8 bytes to
3095 allow for more data.
3096
3097 (d) Added support for Unicode script names via properties such as \p{Han}.
3098
3099 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3100 matching that character.
3101
3102 20. When matching a repeated Unicode property with a minimum greater than zero,
3103 (for example \pL{2,}), PCRE could look past the end of the subject if it
3104 reached it while seeking the minimum number of characters. This could
3105 happen only if some of the characters were more than one byte long, because
3106 there is a check for at least the minimum number of bytes.
3107
3108 21. Refactored the implementation of \p and \P so as to be more general, to
3109 allow for more different types of property in future. This has changed the
3110 compiled form incompatibly. Anybody with saved compiled patterns that use
3111 \p or \P will have to recompile them.
3112
3113 22. Added "Any" and "L&" to the supported property types.
3114
3115 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3116 but give a compile time error if the value is greater than 0xff.
3117
3118 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3119 accidentally not being installed or uninstalled.
3120
3121 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3122 made were to insert the current release number. This seemed silly, because
3123 it made things harder for people building PCRE on systems that don't run
3124 "configure". I have turned pcre.h into a distributed file, no longer built
3125 by "configure", with the version identification directly included. There is
3126 no longer a pcre.h.in file.
3127
3128 However, this change necessitated a change to the pcre-config script as
3129 well. It is built from pcre-config.in, and one of the substitutions was the
3130 release number. I have updated configure.ac so that ./configure now finds
3131 the release number by grepping pcre.h.
3132
3133 26. Added the ability to run the tests under valgrind.
3134
3135
3136 Version 6.4 05-Sep-05
3137 ---------------------
3138
3139 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3140 "--" to be printed when multiple files were scanned, even when none of the
3141 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3142 consider it to be a bug, and have restored the previous behaviour.
3143
3144 2. A couple of code tidies to get rid of compiler warnings.
3145
3146 3. The pcretest program used to cheat by referring to symbols in the library
3147 whose names begin with _pcre_. These are internal symbols that are not
3148 really supposed to be visible externally, and in some environments it is
3149 possible to suppress them. The cheating is now confined to including
3150 certain files from the library's source, which is a bit cleaner.
3151
3152 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3153 file's purpose clearer.
3154
3155 5. Reorganized pcre_ucp_findchar().
3156
3157
3158 Version 6.3 15-Aug-05
3159 ---------------------
3160
3161 1. The file libpcre.pc.in did not have general read permission in the tarball.
3162
3163 2. There were some problems when building without C++ support:
3164
3165 (a) If C++ support was not built, "make install" and "make test" still
3166 tried to test it.
3167
3168 (b) There were problems when the value of CXX was explicitly set. Some
3169 changes have been made to try to fix these, and ...
3170
3171 (c) --disable-cpp can now be used to explicitly disable C++ support.
3172
3173 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3174 backslash in a target when C++ was disabled. This confuses some
3175 versions of "make", apparently. Using an intermediate variable solves
3176 this. (Same for CPP_LOBJ.)
3177
3178 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3179 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3180 necessary on certain architectures.
3181
3182 4. Added a setting of -export-symbols-regex to the link command to remove
3183 those symbols that are exported in the C sense, but actually are local
3184 within the library, and not documented. Their names all begin with
3185 "_pcre_". This is not a perfect job, because (a) we have to except some
3186 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3187 available (and never for static libraries). I have made a note to try to
3188 find a way round (a) in the future.
3189
3190
3191 Version 6.2 01-Aug-05
3192 ---------------------
3193
3194 1. There was no test for integer overflow of quantifier values. A construction
3195 such as {1111111111111111} would give undefined results. What is worse, if
3196 a minimum quantifier for a parenthesized subpattern overflowed and became
3197 negative, the calculation of the memory size went wrong. This could have
3198 led to memory overwriting.
3199
3200 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3201
3202 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3203 operating environments where this matters.
3204
3205 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3206 PCRE options from within the C++ wrapper.
3207
3208 5. Named capturing subpatterns were not being correctly counted when a pattern
3209 was compiled. This caused two problems: (a) If there were more than 100
3210 such subpatterns, the calculation of the memory needed for the whole
3211 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3212 back references of the form \12, where the number was greater than 9, were
3213 not recognized as back references, even though there were sufficient
3214 previous subpatterns.
3215
3216 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3217 versions of gcc, e.g. 2.95.4.
3218
3219
3220 Version 6.1 21-Jun-05
3221 ---------------------
3222
3223 1. There was one reference to the variable "posix" in pcretest.c that was not
3224 surrounded by "#if !defined NOPOSIX".
3225
3226 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3227 the cross-check on the old pcre_info() function, for the benefit of the
3228 cut-down version of PCRE that is currently imported into Exim.
3229
3230 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3231 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3232 patterns that start (?i) but I don't think that matters. The use of (?i) is
3233 just an example; this all applies to the other options as well.
3234
3235 4. Since libtool seems to echo the compile commands it is issuing, the output
3236 from "make" can be reduced a bit by putting "@" in front of each libtool
3237 compile command.
3238
3239 5. Patch from the folks at Google for configure.in to be a bit more thorough
3240 in checking for a suitable C++ installation before trying to compile the
3241 C++ stuff. This should fix a reported problem when a compiler was present,
3242 but no suitable headers.
3243
3244 6. The man pages all had just "PCRE" as their title. I have changed them to
3245 be the relevant file name. I have also arranged that these names are
3246 retained in the file doc/pcre.txt, which is a concatenation in text format
3247 of all the man pages except the little individual ones for each function.
3248
3249 7. The NON-UNIX-USE file had not been updated for the different set of source
3250 files that come with release 6. I also added a few comments about the C++
3251 wrapper.
3252
3253
3254 Version 6.0 07-Jun-05
3255 ---------------------
3256
3257 1. Some minor internal re-organization to help with my DFA experiments.
3258
3259 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3260 didn't matter for the library itself when fully configured, but did matter
3261 when compiling without UCP support, or within Exim, where the ucp files are
3262 not imported.
3263
3264 3. Refactoring of the library code to split up the various functions into
3265 different source modules. The addition of the new DFA matching code (see
3266 below) to a single monolithic source would have made it really too
3267 unwieldy, quite apart from causing all the code to be include in a
3268 statically linked application, when only some functions are used. This is
3269 relevant even without the DFA addition now that patterns can be compiled in
3270 one application and matched in another.
3271
3272 The downside of splitting up is that there have to be some external
3273 functions and data tables that are used internally in different modules of
3274 the library but which are not part of the API. These have all had their
3275 names changed to start with "_pcre_" so that they are unlikely to clash
3276 with other external names.
3277
3278 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3279 a different (DFA) algorithm. Although it is slower than the original
3280 function, it does have some advantages for certain types of matching
3281 problem.
3282
3283 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3284 including restarting after a partial match.
3285
3286 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3287 defined when compiling for Windows was sent to me. I have put it into the
3288 code, though I have no means of testing or verifying it.
3289
3290 7. Added the pcre_refcount() auxiliary function.
3291
3292 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3293 match before or at the first newline in the subject string. In pcretest,
3294 the /f option on a pattern can be used to set this.
3295
3296 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3297 would behave wrongly. This has been present in PCRE since release 4.0.
3298
3299 10. A number of changes to the pcregrep command:
3300
3301 (a) Refactored how -x works; insert ^(...)$ instead of setting
3302 PCRE_ANCHORED and checking the length, in preparation for adding
3303 something similar for -w.
3304
3305 (b) Added the -w (match as a word) option.
3306
3307 (c) Refactored the way lines are read and buffered so as to have more
3308 than one at a time available.
3309
3310 (d) Implemented a pcregrep test script.
3311
3312 (e) Added the -M (multiline match) option. This allows patterns to match
3313 over several lines of the subject. The buffering ensures that at least
3314 8K, or the rest of the document (whichever is the shorter) is available
3315 for matching (and similarly the previous 8K for lookbehind assertions).
3316
3317 (f) Changed the --help output so that it now says
3318
3319 -w, --word-regex(p)
3320
3321 instead of two lines, one with "regex" and the other with "regexp"
3322 because that confused at least one person since the short forms are the
3323 same. (This required a bit of code, as the output is generated
3324 automatically from a table. It wasn't just a text change.)
3325
3326 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3327 option but starts with a hyphen. Could be a pattern or a path name
3328 starting with a hyphen, for instance.
3329
3330 (h) "-" can be given as a file name to represent stdin.
3331
3332 (i) When file names are being printed, "(standard input)" is used for
3333 the standard input, for compatibility with GNU grep. Previously
3334 "<stdin>" was used.
3335
3336 (j) The option --label=xxx can be used to supply a name to be used for
3337 stdin when file names are being printed. There is no short form.
3338
3339 (k) Re-factored the options decoding logic because we are going to add
3340 two more options that take data. Such options can now be given in four
3341 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3342
3343 (l) Added the -A, -B, and -C options for requesting that lines of context
3344 around matches be printed.
3345
3346 (m) Added the -L option to print the names of files that do not contain
3347 any matching lines, that is, the complement of -l.
3348
3349 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3350 continue to scan other files.
3351
3352 (o) The -s option was incorrectly implemented. For compatibility with other
3353 greps, it now suppresses the error message for a non-existent or non-
3354 accessible file (but not the return code). There is a new option called
3355 -q that suppresses the output of matching lines, which was what -s was
3356 previously doing.
3357
3358 (p) Added --include and --exclude options to specify files for inclusion
3359 and exclusion when recursing.
3360
3361 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3362 Hopefully, it now does.
3363
3364 12. Missing cast in pcre_study().
3365
3366 13. Added an "uninstall" target to the makefile.
3367
3368 14. Replaced "extern" in the function prototypes in Makefile.in with
3369 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3370 world, but is set differently for Windows.
3371
3372 15. Added a second compiling function called pcre_compile2(). The only
3373 difference is that it has an extra argument, which is a pointer to an
3374 integer error code. When there is a compile-time failure, this is set
3375 non-zero, in addition to the error test pointer being set to point to an
3376 error message. The new argument may be NULL if no error number is required
3377 (but then you may as well call pcre_compile(), which is now just a
3378 wrapper). This facility is provided because some applications need a
3379 numeric error indication, but it has also enabled me to tidy up the way
3380 compile-time errors are handled in the POSIX wrapper.
3381
3382 16. Added VPATH=.libs to the makefile; this should help when building with one
3383 prefix path and installing with another. (Or so I'm told by someone who
3384 knows more about this stuff than I do.)
3385
3386 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3387 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3388 match everything, including newlines. This is not POSIX-compatible, but
3389 somebody wanted the feature. From pcretest it can be activated by using
3390 both the P and the s flags.
3391
3392 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3393
3394 19. libpcre.pc was being incorrectly installed as executable.
3395
3396 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3397 it now also looks for '\r' so that it will work unmodified on Windows.
3398
3399 21. Added Google's contributed C++ wrapper to the distribution.
3400
3401 22. Added some untidy missing memory free() calls in pcretest, to keep
3402 Electric Fence happy when testing.
3403
3404
3405
3406 Version 5.0 13-Sep-04
3407 ---------------------
3408
3409 1. Internal change: literal characters are no longer packed up into items
3410 containing multiple characters in a single byte-string. Each character
3411 is now matched using a separate opcode. However, there may be more than one
3412 byte in the character in UTF-8 mode.
3413
3414 2. The pcre_callout_block structure has two new fields: pattern_position and
3415 next_item_length. These contain the offset in the pattern to the next match
3416 item, and its length, respectively.
3417
3418 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3419 insertion of callouts before each pattern item. Added the /C option to
3420 pcretest to make use of this.
3421
3422 4. On the advice of a Windows user, the lines
3423
3424 #if defined(_WIN32) || defined(WIN32)
3425 _setmode( _fileno( stdout ), 0x8000 );
3426 #endif /* defined(_WIN32) || defined(WIN32) */
3427
3428 have been added to the source of pcretest. This apparently does useful
3429 magic in relation to line terminators.
3430
3431 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3432 for the benefit of those environments where the "b" makes a difference.
3433
3434 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3435 to know about it. I have put a hack into configure.in that adds in code
3436 to set GCC=yes if CC=icc. This seems to end up at a point in the
3437 generated configure script that is early enough to affect the setting of
3438 compiler options, which is what is needed, but I have no means of testing
3439 whether it really works. (The user who reported this had patched the
3440 generated configure script, which of course I cannot do.)
3441
3442 LATER: After change 22 below (new libtool files), the configure script
3443 seems to know about icc (and also ecc). Therefore, I have commented out
3444 this hack in configure.in.
3445
3446 7. Added support for pkg-config (2 patches were sent in).
3447
3448 8. Negated POSIX character classes that used a combination of internal tables
3449 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3450 [[:^ascii]]. Typically, they would match almost any characters. The other
3451 POSIX classes were not broken in this way.
3452
3453 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3454 to find the match, as PCRE was deluded into thinking that the match had to
3455 start at the start point or following a newline. The same bug applied to
3456 patterns with negative forward assertions or any backward assertions
3457 preceding ".*" at the start, unless the pattern required a fixed first
3458 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3459
3460 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3461 starting at the last subject character, bytes beyond the end of the subject
3462 string were read.
3463
3464 11. Renamed the variable "class" as "classbits" to make life easier for C++
3465 users. (Previously there was a macro definition, but it apparently wasn't
3466 enough.)
3467
3468 12. Added the new field "tables" to the extra data so that tables can be passed
3469 in at exec time, or the internal tables can be re-selected. This allows
3470 a compiled regex to be saved and re-used at a later time by a different
3471 program that might have everything at different addresses.
3472
3473 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3474 -R library as well as a -L library.
3475
3476 14. The debugging options of pcretest (-d on the command line or D on a
3477 pattern) showed incorrect output for anything following an extended class
3478 that contained multibyte characters and which was followed by a quantifier.
3479
3480 15. Added optional support for general category Unicode character properties
3481 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3482 support. It adds about 90K to the size of the library. The meanings of the
3483 inbuilt class escapes such as \d and \s have NOT been changed.
3484
3485 16. Updated pcredemo.c to include calls to free() to release the memory for the
3486 compiled pattern.
3487
3488 17. The generated file chartables.c was being created in the source directory
3489 instead of in the building directory. This caused the build to fail if the
3490 source directory was different from the building directory, and was
3491 read-only.
3492
3493 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3494 file. No doubt somebody will tell me if they don't make sense... Also added
3495 Dan Mooney's comments about building on OpenVMS.
3496
3497 19. Added support for partial matching via the PCRE_PARTIAL option for
3498 pcre_exec() and the \P data escape in pcretest.
3499
3500 20. Extended pcretest with 3 new pattern features:
3501
3502 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3503 write the compiled pattern to the file whose name is "rest-of-line".
3504 This is a straight binary dump of the data, with the saved pointer to
3505 the character tables forced to be NULL. The study data, if any, is
3506 written too. After writing, pcretest reads a new pattern.
3507
3508 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3509 compiled pattern from the given file. There must not be any
3510 occurrences of "<" in the file name (pretty unlikely); if there are,
3511 pcretest will instead treat the initial "<" as a pattern delimiter.
3512 After reading in the pattern, pcretest goes on to read data lines as
3513 usual.
3514
3515 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3516 and 16-bit fields in a compiled pattern, to simulate a pattern that
3517 was compiled on a host of opposite endianness.
3518
3519 21. The pcre-exec() function can now cope with patterns that were compiled on
3520 hosts of opposite endianness, with this restriction:
3521
3522 As for any compiled expression that is saved and used later, the tables
3523 pointer field cannot be preserved; the extra_data field in the arguments
3524 to pcre_exec() should be used to pass in a tables address if a value
3525 other than the default internal tables were used at compile time.
3526
3527 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3528 now diagnosed as an error. Previously, most of the time, a negative number
3529 would have been treated as zero, but if in addition "ovector" was passed as
3530 NULL, a crash could occur.
3531
3532 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3533 new versions from the libtool 1.5 distribution (the last one is a copy of
3534 a file called libtool.m4). This seems to have fixed the need to patch
3535 "configure" to support Darwin 1.3 (which I used to do). However, I still
3536 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3537 workstation).
3538
3539 24. Changed the PCRE licence to be the more standard "BSD" licence.
3540
3541
3542 Version 4.5 01-Dec-03
3543 ---------------------
3544
3545 1. There has been some re-arrangement of the code for the match() function so
3546 that it can be compiled in a version that does not call itself recursively.
3547 Instead, it keeps those local variables that need separate instances for
3548 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3549 needs to "recurse". Keeping track of where control must go is done by means
3550 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3551 hide most of the details from the main code, and operates only if
3552 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3553 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3554 operating.
3555
3556 To make it easier for callers to provide specially tailored get/free
3557 functions for this usage, two new functions, pcre_stack_malloc, and
3558 pcre_stack_free, are used. They are always called in strict stacking order,
3559 and the size of block requested is always the same.
3560
3561 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3562 PCRE has been compiled to use the stack or the heap for recursion. The
3563 -C option of pcretest uses this to show which version is compiled.
3564
3565 A new data escape \S, is added to pcretest; it causes the amounts of store
3566 obtained and freed by both kinds of malloc/free at match time to be added
3567 to the output.
3568
3569 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3570 what's available on my current Linux desktop machine.
3571
3572 3. When matching a UTF-8 string, the test for a valid string at the start has
3573 been extended. If start_offset is not zero, PCRE now checks that it points
3574 to a byte that is the start of a UTF-8 character. If not, it returns
3575 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3576 this is necessary because there may be backward assertions in the pattern.
3577 When matching the same subject several times, it may save resources to use
3578 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3579
3580 4. The code for checking the validity of UTF-8 strings has been tightened so
3581 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3582 containing "overlong sequences".
3583
3584 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3585 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3586 should have been "&", but it just so happened that all the cases this let
3587 through by mistake were picked up later in the function.
3588
3589 6. I had used a variable called "isblank" - this is a C99 function, causing
3590 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3591
3592 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3593 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3594 so the version is shown; (c) stop "make test" echoing "./RunTest".
3595
3596 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3597
3598 9. The prototype for memmove() for systems that don't have it was using
3599 size_t, but the inclusion of the header that defines size_t was later. I've
3600 moved the #includes for the C headers earlier to avoid this.
3601
3602 10. Added some adjustments to the code to make it easier to compiler on certain
3603 special systems:
3604
3605 (a) Some "const" qualifiers were missing.
3606 (b) Added the macro EXPORT before all exported functions; by default this
3607 is defined to be empty.
3608 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3609 that it reads its output file name as an argument instead of writing
3610 to the standard output and assuming this can be redirected.
3611
3612 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3613 class containing characters with values greater than 255, PCRE compilation
3614 went into a loop.
3615
3616 12. A recursive reference to a subpattern that was within another subpattern
3617 that had a minimum quantifier of zero caused PCRE to crash. For example,
3618 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3619 recursion. If the recursively-called subpattern itself had a zero repeat,
3620 that was OK.
3621
3622 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3623 buffer into which it was copied (for escape processing) was still set at
3624 1024, so long lines caused crashes.
3625
3626 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3627 "internal error: code overflow...". This applied to any character class
3628 that was followed by a possessive quantifier.
3629
3630 15. Modified the Makefile to add libpcre.la as a prerequisite for
3631 libpcreposix.la because I was told this is needed for a parallel build to
3632 work.
3633
3634 16. If a pattern that contained .* following optional items at the start was
3635 studied, the wrong optimizing data was generated, leading to matching
3636 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3637 matching string must start with a or b or c. The correct conclusion for
3638 this pattern is that a match can start with any character.
3639
3640
3641 Version 4.4 13-Aug-03
3642 ---------------------
3643
3644 1. In UTF-8 mode, a character class containing characters with values between
3645 127 and 255 was not handled correctly if the compiled pattern was studied.
3646 In fixing this, I have also improved the studying algorithm for such
3647 classes (slightly).
3648
3649 2. Three internal functions had redundant arguments passed to them. Removal
3650 might give a very teeny performance improvement.
3651
3652 3. Documentation bug: the value of the capture_top field in a callout is *one
3653 more than* the number of the hightest numbered captured substring.
3654
3655 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3656 in incorrectly linking with a previously installed version. They now link
3657 explicitly with libpcre.la.
3658
3659 5. configure.in no longer needs to recognize Cygwin specially.
3660
3661 6. A problem in pcre.in for Windows platforms is fixed.
3662
3663 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3664 pcretest, it used to include the size of the study block as part of its
3665 output. Unfortunately, the structure contains a field that has a different
3666 size on different hardware architectures. This meant that the tests that
3667 showed this size failed. As the block is currently always of a fixed size,
3668 this information isn't actually particularly useful in pcretest output, so
3669 I have just removed it.
3670
3671 8. Three pre-processor statements accidentally did not start in column 1.
3672 Sadly, there are *still* compilers around that complain, even though
3673 standard C has not required this for well over a decade. Sigh.
3674
3675 9. In pcretest, the code for checking callouts passed small integers in the
3676 callout_data field, which is a void * field. However, some picky compilers
3677 complained about the casts involved for this on 64-bit systems. Now
3678 pcretest passes the address of the small integer instead, which should get
3679 rid of the warnings.
3680
3681 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3682 both compile and run time, and gives an error if an invalid UTF-8 sequence
3683 is found. There is a option for disabling this check in cases where the
3684 string is known to be correct and/or the maximum performance is wanted.
3685
3686 11. In response to a bug report, I changed one line in Makefile.in from
3687
3688 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3689 to
3690 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3691
3692 to look similar to other lines, but I have no way of telling whether this
3693 is the right thing to do, as I do not use Windows. No doubt I'll get told
3694 if it's wrong...
3695
3696
3697 Version 4.3 21-May-03
3698 ---------------------
3699
3700 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3701 Makefile.
3702
3703 2. Some refactoring to improve the quality of the code:
3704
3705 (i) The utf8_table... variables are now declared "const".
3706
3707 (ii) The code for \cx, which used the "case flipping" table to upper case
3708 lower case letters, now just substracts 32. This is ASCII-specific,
3709 but the whole concept of \cx is ASCII-specific, so it seems
3710 reasonable.
3711
3712 (iii) PCRE was using its character types table to recognize decimal and
3713 hexadecimal digits in the pattern. This is silly, because it handles
3714 only 0-9, a-f, and A-F, but the character types table is locale-
3715 specific, which means strange things might happen. A private
3716 table is now used for this - though it costs 256 bytes, a table is
3717 much faster than multiple explicit tests. Of course, the standard
3718 character types table is still used for matching digits in subject
3719 strings against \d.
3720
3721 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3722 ending in _t are). So I've renamed it as ESC_tee.
3723
3724 3. The first argument for regexec() in the POSIX wrapper should have been
3725 defined as "const".
3726
3727 4. Changed pcretest to use malloc() for its buffers so that they can be
3728 Electric Fenced for debugging.
3729
3730 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3731 to read one or more bytes before the start of the subject string. Often this
3732 had no effect on PCRE's behaviour, but in some circumstances it could
3733 provoke a segmentation fault.
3734
3735 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3736 to try to read one or more bytes before the start of the subject string.
3737
3738 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3739 UTF-8 support could misbehave in various ways if the subject string
3740 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3741 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3742 back over UTF-8 characters.)
3743
3744
3745 Version 4.2 14-Apr-03
3746 ---------------------
3747
3748 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3749
3750 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3751 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3752 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3753 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3754 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3755 and BUILD_EXEEXT
3756 Note: automatic setting of the BUILD variables is not yet working
3757 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3758 compile-time but not at link-time
3759 [LINK]: use for linking executables only
3760 make different versions for Windows and non-Windows
3761 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3762 libraries
3763 [LINK_FOR_BUILD]: new variable
3764 [OBJEXT]: use throughout
3765 [EXEEXT]: use throughout
3766 <winshared>: new target
3767 <wininstall>: new target
3768 <dftables.o>: use native compiler
3769 <dftables>: use native linker
3770 <install>: handle Windows platform correctly
3771 <clean>: ditto
3772 <check>: ditto
3773 copy DLL to top builddir before testing
3774
3775 As part of these changes, -no-undefined was removed again. This was reported
3776 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3777 in any case.
3778
3779 3. Some tidies to get rid of compiler warnings:
3780
3781 . In the match_data structure, match_limit was an unsigned long int, whereas
3782 match_call_count was an int. I've made them both unsigned long ints.
3783
3784 . In pcretest the fact that a const uschar * doesn't automatically cast to
3785 a void * provoked a warning.
3786
3787 . Turning on some more compiler warnings threw up some "shadow" variables
3788 and a few more missing casts.
3789
3790 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3791 option, a class that contained a single character with a value between 128
3792 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3793
3794 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3795 option, a class that contained several characters, but with at least one
3796 whose value was between 128 and 255 caused PCRE to crash.
3797
3798
3799 Version 4.1 12-Mar-03
3800 ---------------------
3801
3802 1. Compiling with gcc -pedantic found a couple of places where casts were
3803 needed, and a string in dftables.c that was longer than standard compilers are
3804 required to support.
3805
3806 2. Compiling with Sun's compiler found a few more places where the code could
3807 be tidied up in order to avoid warnings.
3808
3809 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3810 first of these names is deprecated in the latest Autoconf in favour of the name
3811 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3812 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3813 analogy I have changed it to CFLAGS_FOR_BUILD.
3814
3815 4. Added -no-undefined to the linking command in the Makefile, because this is
3816 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3817 linking step for the pcreposix library.
3818
3819 5. PCRE was failing to diagnose the case of two named groups with the same
3820 name.
3821
3822 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3823 literal character that is needed in the subject for a match, and scans along to
3824 ensure that it is present before embarking on the full matching process. This
3825 saves time in cases of nested unlimited repeats that are never going to match.
3826 Problem: the scan can take a lot of time if the subject is very long (e.g.
3827 megabytes), thus penalizing straightforward matches. It is now done only if the
3828 amount of subject to be scanned is less than 1000 bytes.
3829
3830 7. A lesser problem with the same optimization is that it was recording the
3831 first character of an anchored pattern as "needed", thus provoking a search
3832 right along the subject, even when the first match of the pattern was going to
3833 fail. The "needed" character is now not set for anchored patterns, unless it
3834 follows something in the pattern that is of non-fixed length. Thus, it still
3835 fulfils its original purpose of finding quick non-matches in cases of nested
3836 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3837
3838
3839 Version 4.0 17-Feb-03
3840 ---------------------
3841
3842 1. If a comment in an extended regex that started immediately after a meta-item
3843 extended to the end of string, PCRE compiled incorrect data. This could lead to
3844 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3845
3846 2. Moved to autoconf 2.53 and libtool 1.4.2.
3847
3848 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3849 the special perltest8 script is no longer needed - all the tests can be run
3850 from a single perltest script.
3851
3852 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3853 by \s. It has now been removed in PCRE. This means it isn't recognized as
3854 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3855 class [:space:] *does* include VT, thereby creating a mess.
3856
3857 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3858 space and tab.
3859
3860 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3861 its new features into the main test script, reducing the number of scripts.
3862
3863 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3864 were backward compatible, and made the (?i) apply to the whole pattern, as if
3865 /i were given. Now it behaves more logically, and applies the option setting
3866 only to what follows. PCRE has been changed to follow suit. However, if it
3867 finds options settings right at the start of the pattern, it extracts them into
3868 the global options, as before. Thus, they show up in the info data.
3869
3870 8. Added support for the \Q...\E escape sequence. Characters in between are
3871 treated as literals. This is slightly different from Perl in that $ and @ are
3872 also handled as literals inside the quotes. In Perl, they will cause variable
3873 interpolation. Note the following examples:
3874
3875 Pattern PCRE matches Perl matches
3876
3877 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3878 \Qabc\$xyz\E abc\$xyz abc\$xyz
3879 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3880
3881 For compatibility with Perl, \Q...\E sequences are recognized inside character
3882 classes as well as outside them.
3883
3884 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3885 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3886 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3887 signed/unsigned warnings.
3888
3889 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3890 option for pcretest, so I've replaced it by a simple function that does just
3891 that job.
3892
3893 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3894 "pcregrep -".
3895
3896 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3897 Java package. This provides some syntactic sugar for simple cases of what my
3898 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3899 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3900 item, you can use this simplified notation. Note that only makes sense with
3901 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3902 greediness, whatever the setting of the PCRE_UNGREEDY option.
3903
3904 13. A change of greediness default within a pattern was not taking effect at
3905 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3906 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3907 was abstracted outside.
3908
3909 14. PCRE now supports the \G assertion. It is true when the current matching
3910 position is at the start point of the match. This differs from \A when the
3911 starting offset is non-zero. Used with the /g option of pcretest (or similar
3912 code), it works in the same way as it does for Perl's /g option. If all
3913 alternatives of a regex begin with \G, the expression is anchored to the start
3914 match position, and the "anchored" flag is set in the compiled expression.
3915
3916 15. Some bugs concerning the handling of certain option changes within patterns
3917 have been fixed. These applied to options other than (?ims). For example,
3918 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3919 been the other way round. Some of this was related to change 7 above.
3920
3921 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3922 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3923 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3924 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3925
3926 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3927 mode. Unlike ".", it always matches newline, whatever the setting of
3928 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3929 assertions. Perl allows it, but it doesn't (in general) work because it can't
3930 calculate the length of the lookbehind. At least, that's the case for Perl
3931 5.8.0 - I've been told they are going to document that it doesn't work in
3932 future.
3933
3934 18. Added an error diagnosis for escapes that PCRE does not support: these are
3935 \L, \l, \N, \P, \p, \U, \u, and \X.
3936
3937 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3938 reading past the end of the pattern in cases such as /[abcd/.
3939
3940 20. PCRE was getting more memory than necessary for patterns with classes that
3941 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3942
3943 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3944 compiling PCRE for use with Virtual Pascal.
3945
3946 22. Small fix to the Makefile to make it work properly if the build is done
3947 outside the source tree.
3948
3949 23. Added a new extension: a condition to go with recursion. If a conditional
3950 subpattern starts with (?(R) the "true" branch is used if recursion has
3951 happened, whereas the "false" branch is used only at the top level.
3952
3953 24. When there was a very long string of literal characters (over 255 bytes
3954 without UTF support, over 250 bytes with UTF support), the computation of how
3955 much memory was required could be incorrect, leading to segfaults or other
3956 strange effects.
3957
3958 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3959 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3960 there was a subsequent back reference to those brackets. This meant that, for
3961 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3962 possible to check for precisely this case. All we can do is abandon the
3963 optimization if .* occurs inside capturing brackets when there are any back
3964 references whatsoever. (See below for a better fix that came later.)
3965
3966 26. The handling of the optimization for finding the first character of a
3967 non-anchored pattern, and for finding a character that is required later in the
3968 match were failing in some cases. This didn't break the matching; it just
3969 failed to optimize when it could. The way this is done has been re-implemented.
3970
3971 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3972
3973 28. Added a new feature that provides some of the functionality that Perl
3974 provides with (?{...}). The facility is termed a "callout". The way it is done
3975 in PCRE is for the caller to provide an optional function, by setting
3976 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3977 global variable. By default it is unset, which disables all calling out. To get
3978 the function called, the regex must include (?C) at appropriate points. This
3979 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3980 This provides a means of identifying different callout points. When PCRE
3981 reaches such a point in the regex, if pcre_callout has been set, the external
3982 function is called. It is provided with data in a structure called
3983 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3984 matching continues; if it returns a non-zero value, the match at the current
3985 point fails. However, backtracking will occur if possible. [This was changed
3986 later and other features added - see item 49 below.]
3987
3988 29. pcretest is upgraded to test the callout functionality. It provides a
3989 callout function that displays information. By default, it shows the start of
3990 the match and the current position in the text. There are some new data escapes
3991 to vary what happens:
3992
3993 \C+ in addition, show current contents of captured substrings
3994 \C- do not supply a callout function
3995 \C!n return 1 when callout number n is reached
3996 \C!n!m return 1 when callout number n is reached for the mth time
3997
3998 30. If pcregrep was called with the -l option and just a single file name, it
3999 output "<stdin>" if a match was found, instead of the file name.
4000
4001 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4002 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4003 pcre_exec(). This saves a malloc/free per call. The default value of
4004 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4005 when configuring.
4006
4007 32. The default maximum size of a compiled pattern is 64K. There have been a
4008 few cases of people hitting this limit. The code now uses macros to handle the
4009 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4010 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4011 configuring. Tests 2 and 5 work only with 2-byte links because they output
4012 debugging information about compiled patterns.
4013
4014 33. Internal code re-arrangements:
4015
4016 (a) Moved the debugging function for printing out a compiled regex into
4017 its own source file (printint.c) and used #include to pull it into
4018 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4019 separate copies.
4020
4021 (b) Defined the list of op-code names for debugging as a macro in
4022 internal.h so that it is next to the definition of the opcodes.
4023
4024 (c) Defined a table of op-code lengths for simpler skipping along compiled
4025 code. This is again a macro in internal.h so that it is next to the
4026 definition of the opcodes.
4027
4028 34. Added support for recursive calls to individual subpatterns, along the
4029 lines of Robin Houston's patch (but implemented somewhat differently).
4030
4031 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4032 allow it to read and process whole directories in Win32. This code was
4033 contributed by Lionel Fourquaux; it has not been tested by me.
4034
4035 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4036 used to name a group. Names consist of alphanumerics and underscores, and must
4037 be unique. Back references use the syntax (?P=name) and recursive calls use
4038 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4039 numbers. The function pcre_fullinfo() can be used after compilation to extract
4040 a name/number map. There are three relevant calls:
4041
4042 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4043 PCRE_INFO_NAMECOUNT yields the number of entries
4044 PCRE_INFO_NAMETABLE yields a pointer to the map.
4045
4046 The map is a vector of fixed-size entries. The size of each entry depends on
4047 the length of the longest name used. The first two bytes of each entry are the
4048 group number, most significant byte first. There follows the corresponding
4049 name, zero terminated. The names are in alphabetical order.
4050
4051 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4052 case instead of 255. Making it the same both with and without UTF-8 support
4053 means that the same test output works with both.
4054
4055 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4056 calling malloc() with a zero argument.
4057
4058 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4059 optimization. I've improved things by keeping a bitmap of backreferences with
4060 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4061 fact referenced, the optimization can be applied. It is unlikely that a
4062 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4063 the match to follow \n) will appear inside brackets with a number greater than
4064 31, but if it does, any back reference > 31 suppresses the optimization.
4065
4066 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4067 of disabling numbered capturing parentheses. Any opening parenthesis that is
4068 not followed by ? behaves as if it were followed by ?: but named parentheses
4069 can still be used for capturing (and they will acquire numbers in the usual
4070 way).
4071
4072 41. Redesigned the return codes from the match() function into yes/no/error so
4073 that errors can be passed back from deep inside the nested calls. A malloc
4074 failure while inside a recursive subpattern call now causes the
4075 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4076
4077 42. It is now possible to set a limit on the number of times the match()
4078 function is called in a call to pcre_exec(). This facility makes it possible to
4079 limit the amount of recursion and backtracking, though not in a directly
4080 obvious way, because the match() function is used in a number of different
4081 circumstances. The count starts from zero for each position in the subject
4082 string (for non-anchored patterns). The default limit is, for compatibility, a
4083 large number, namely 10 000 000. You can change this in two ways:
4084
4085 (a) When configuring PCRE before making, you can use --with-match-limit=n
4086 to set a default value for the compiled library.
4087
4088 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4089 a different value is set. See 45 below.
4090
4091 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4092
4093 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4094 of things that can be changed at compile time. The first argument specifies
4095 what is wanted and the second points to where the information is to be placed.
4096 The current list of available information is:
4097
4098 PCRE_CONFIG_UTF8
4099
4100 The output is an integer that is set to one if UTF-8 support is available;
4101 otherwise it is set to zero.
4102
4103 PCRE_CONFIG_NEWLINE
4104
4105 The output is an integer that it set to the value of the code that is used for
4106 newline. It is either LF (10) or CR (13).
4107
4108 PCRE_CONFIG_LINK_SIZE
4109
4110 The output is an integer that contains the number of bytes used for internal
4111 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4112
4113 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4114
4115 The output is an integer that contains the threshold above which the POSIX
4116 interface uses malloc() for output vectors. See item 31 above.
4117
4118 PCRE_CONFIG_MATCH_LIMIT
4119
4120 The output is an unsigned integer that contains the default limit of the number
4121 of match() calls in a pcre_exec() execution. See 42 above.
4122
4123 44. pcretest has been upgraded by the addition of the -C option. This causes it
4124 to extract all the available output from the new pcre_config() function, and to
4125 output it. The program then exits immediately.
4126
4127 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4128 order to support additional features. One way would have been to define
4129 pcre_exec2() (for example) with extra arguments, but this would not have been
4130 extensible, and would also have required all calls to the original function to
4131 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4132 is used for passing in "extra" data from pcre_study().
4133
4134 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4135 contains the following fields:
4136
4137 flags a bitmap indicating which of the following fields are set
4138 study_data opaque data from pcre_study()
4139 match_limit a way of specifying a limit on match() calls for a specific
4140 call to pcre_exec()
4141 callout_data data for callouts (see 49 below)
4142
4143 The flag bits are also defined in pcre.h, and are
4144
4145 PCRE_EXTRA_STUDY_DATA
4146 PCRE_EXTRA_MATCH_LIMIT
4147 PCRE_EXTRA_CALLOUT_DATA
4148
4149 The pcre_study() function now returns one of these new pcre_extra blocks, with
4150 the actual study data pointed to by the study_data field, and the
4151 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4152 before. That is, this change is entirely upwards-compatible and requires no
4153 change to existing code.
4154
4155 If you want to pass in additional data to pcre_exec(), you can either place it
4156 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4157 block.
4158
4159 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4160 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4161 times with different match limits, until it finds the minimum value needed for
4162 pcre_exec() to complete. The value is then output. This can be instructive; for
4163 most simple matches the number is quite small, but for pathological cases it
4164 gets very large very quickly.
4165
4166 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4167 returns the size of the data block pointed to by the study_data field in a
4168 pcre_extra block, that is, the value that was passed as the argument to
4169 pcre_malloc() when PCRE was getting memory in which to place the information
4170 created by pcre_study(). The fourth argument should point to a size_t variable.
4171 pcretest has been extended so that this information is shown after a successful
4172 pcre_study() call when information about the compiled regex is being displayed.
4173
4174 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4175 because what follows is always an absolute path. (Later: it turns out that this
4176 is more than cosmetic for MinGW, because it doesn't like empty path
4177 components.)
4178
4179 49. Some changes have been made to the callout feature (see 28 above):
4180
4181 (i) A callout function now has three choices for what it returns:
4182
4183 0 => success, carry on matching
4184 > 0 => failure at this point, but backtrack if possible
4185 < 0 => serious error, return this value from pcre_exec()
4186
4187 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4188 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4189 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4190 use by callout functions. It will never be used by PCRE itself.
4191
4192 (ii) The pcre_extra structure (see 45 above) has a void * field called
4193 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4194 pcre_callout_block structure has a field of the same name. The contents of
4195 the field passed in the pcre_extra structure are passed to the callout
4196 function in the corresponding field in the callout block. This makes it
4197 easier to use the same callout-containing regex from multiple threads. For
4198 testing, the pcretest program has a new data escape
4199
4200 \C*n pass the number n (may be negative) as callout_data
4201
4202 If the callout function in pcretest receives a non-zero value as
4203 callout_data, it returns that value.
4204
4205 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4206 there were some redundant $(CFLAGS) in commands that are now specified as
4207 $(LINK), which already includes $(CFLAGS).
4208
4209 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4210 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4211 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4212 one-byte characters throughout. Note that case-insensitive matching applies
4213 only to characters whose values are less than 256. PCRE doesn't support the
4214 notion of cases for higher-valued characters.
4215
4216 (i) A character class whose characters are all within 0-255 is handled as
4217 a bit map, and the map is inverted for negative classes. Previously, a
4218 character > 255 always failed to match such a class; however it should
4219 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4220
4221 (ii) A negated character class with a single character < 255 is coded as
4222 "not this character" (OP_NOT). This wasn't working properly when the test
4223 character was multibyte, either singly or repeated.
4224
4225 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4226 mode, for example: \x{100}{2,3}.
4227
4228 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4229 singly or repeated) now correctly test multibyte characters. However,
4230 PCRE doesn't recognize any characters with values greater than 255 as
4231 digits, spaces, or word characters. Such characters always match \D, \S,
4232 and \W, and never match \d, \s, or \w.
4233
4234 (v) Classes may now contain characters and character ranges with values
4235 greater than 255. For example: [ab\x{100}-\x{400}].
4236
4237 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4238 PCRE in UTF-8 mode.
4239
4240 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4241 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4242 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4243 value.)
4244
4245 53. The single man page has become too large. I have therefore split it up into
4246 a number of separate man pages. These also give rise to individual HTML pages;
4247 these are now put in a separate directory, and there is an index.html page that
4248 lists them all. Some hyperlinking between the pages has been installed.
4249
4250 54. Added convenience functions for handling named capturing parentheses.
4251
4252 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4253 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4254 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4255 are faulted.
4256
4257 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4258 calling configure. These values are used when compiling the dftables.c program
4259 which is run to generate the source of the default character tables. They
4260 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4261 you will need to set these values.
4262
4263 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4264
4265
4266 Version 3.9 02-Jan-02
4267 ---------------------
4268
4269 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4270
4271 2. If --disable-static was given, the building process failed when trying to
4272 build pcretest and pcregrep. (For some reason it was using libtool to compile
4273 them, which is not right, as they aren't part of the library.)
4274
4275
4276 Version 3.8 18-Dec-01
4277 ---------------------
4278
4279 1. The experimental UTF-8 code was completely screwed up. It was packing the
4280 bytes in the wrong order. How dumb can you get?
4281
4282
4283 Version 3.7 29-Oct-01
4284 ---------------------
4285
4286 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4287 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4288 this didn't happen under Solaris 8, where I normally test things.
4289
4290 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4291 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4292 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4293 this fix an hour or so after the initial 3.7 release.)
4294
4295
4296 Version 3.6 23-Oct-01
4297 ---------------------
4298
4299 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4300 offsets passed as NULL with zero offset count.
4301
4302 2. The config.guess and config.sub files had not been updated when I moved to
4303 the latest autoconf.
4304
4305
4306 Version 3.5 15-Aug-01
4307 ---------------------
4308
4309 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4310 had been forgotten.
4311
4312 2. By using declared but undefined structures, we can avoid using "void"
4313 definitions in pcre.h while keeping the internal definitions of the structures
4314 private.
4315
4316 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4317 user point of view, this means that both static and shared libraries are built
4318 by default, but this can be individually controlled. More of the work of
4319 handling this static/shared cases is now inside libtool instead of PCRE's make
4320 file.
4321
4322 4. The pcretest utility is now installed along with pcregrep because it is
4323 useful for users (to test regexs) and by doing this, it automatically gets
4324 relinked by libtool. The documentation has been turned into a man page, so
4325 there are now .1, .txt, and .html versions in /doc.
4326
4327 5. Upgrades to pcregrep:
4328 (i) Added long-form option names like gnu grep.
4329 (ii) Added --help to list all options with an explanatory phrase.
4330 (iii) Added -r, --recursive to recurse into sub-directories.
4331 (iv) Added -f, --file to read patterns from a file.
4332
4333 6. pcre_exec() was referring to its "code" argument before testing that
4334 argument for NULL (and giving an error if it was NULL).
4335
4336 7. Upgraded Makefile.in to allow for compiling in a different directory from
4337 the source directory.
4338
4339 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4340 options bits, the pointer it was passed was to an int instead of to an unsigned
4341 long int. This mattered only on 64-bit systems.
4342
4343 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4344 generated) instead of pcre.in, which it its source. Also made the same change
4345 in several of the .c files.
4346
4347 10. A new release of gcc defines printf() as a macro, which broke pcretest
4348 because it had an ifdef in the middle of a string argument for printf(). Fixed
4349 by using separate calls to printf().
4350
4351 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4352 script, to force use of CR or LF instead of \n in the source. On non-Unix
4353 systems, the value can be set in config.h.
4354
4355 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4356 absolute limit. Changed the text of the error message to make this clear, and
4357 likewise updated the man page.
4358
4359 13. The limit of 99 on the number of capturing subpatterns has been removed.
4360 The new limit is 65535, which I hope will not be a "real" limit.
4361
4362
4363 Version 3.4 22-Aug-00
4364 ---------------------
4365
4366 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4367
4368 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4369
4370
4371 Version 3.3 01-Aug-00
4372 ---------------------
4373
4374 1. If an octal character was given, but the value was greater than \377, it
4375 was not getting masked to the least significant bits, as documented. This could
4376 lead to crashes in some systems.
4377
4378 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4379 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4380
4381 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4382 These just pass their arguments on to (pcre_free)(), but they are provided
4383 because some uses of PCRE bind it to non-C systems that can call its functions,
4384 but cannot call free() or pcre_free() directly.
4385
4386 4. Add "make test" as a synonym for "make check". Corrected some comments in
4387 the Makefile.
4388
4389 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4390 Makefile.
4391
4392 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4393 command called pgrep for grepping around the active processes.
4394
4395 7. Added the beginnings of support for UTF-8 character strings.
4396
4397 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4398 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4399 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4400 out for the ar command.)
4401
4402
4403 Version 3.2 12-May-00
4404 ---------------------
4405
4406 This is purely a bug fixing release.
4407
4408 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4409 of ZA. This was just one example of several cases that could provoke this bug,
4410 which was introduced by change 9 of version 2.00. The code for breaking
4411 infinite loops after an iteration that matches an empty string was't working
4412 correctly.
4413
4414 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4415 when matched against abbab (for example). After matching an empty string, it
4416 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4417 caused it to match further down the string than it should.
4418
4419 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4420 was there because it doesn't seem to be needed, and it causes trouble on some
4421 systems, as it is not a Standard C header. It has been removed.
4422
4423 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4424 were reported on the Macintosh. The changes were from
4425
4426 while ((c = *(++ptr)) != 0 && c != '\n');
4427 to
4428 while ((c = *(++ptr)) != 0 && c != '\n') ;
4429
4430 Totally extraordinary, but if that's what it takes...
4431
4432 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4433 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4434 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4435 assumes the way PCRE uses memmove() (always moving upwards).
4436
4437 6. PCRE is being used in one environment where strchr() is not available. There
4438 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4439 faster code anyway.
4440
4441
4442 Version 3.1 09-Feb-00
4443 ---------------------
4444
4445 The only change in this release is the fixing of some bugs in Makefile.in for
4446 the "install" target:
4447
4448 (1) It was failing to install pcreposix.h.
4449
4450 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4451
4452
4453 Version 3.0 01-Feb-00
4454 ---------------------
4455
4456 1. Add support for the /+ modifier to perltest (to output $` like it does in
4457 pcretest).
4458
4459 2. Add support for the /g modifier to perltest.
4460
4461 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4462 matches null strings.
4463
4464 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4465 pattern. Perl treats empty patterns specially - it reuses the most recent
4466 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4467 effect.
4468
4469 5. The POSIX interface was broken in that it was just handing over the POSIX
4470 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4471 required a bigger vector, with some working space on the end. This means that
4472 the POSIX wrapper now has to get and free some memory, and copy the results.
4473
4474 6. Added some simple autoconf support, placing the test data and the
4475 documentation in separate directories, re-organizing some of the
4476 information files, and making it build pcre-config (a GNU standard). Also added
4477 libtool support for building PCRE as a shared library, which is now the
4478 default.
4479
4480 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4481 09 are not valid octal constants. Single digits will be used for minor values
4482 less than 10.
4483
4484 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4485 existing programs that set these in the POSIX interface can use PCRE without
4486 modification.
4487
4488 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4489 return all that pcre_info() returns, plus additional data. The pcre_info()
4490 function is retained for compatibility, but is considered to be obsolete.
4491
4492 10. Added experimental recursion feature (?R) to handle one common case that
4493 Perl 5.6 will be able to do with (?p{...}).
4494
4495 11. Added support for POSIX character classes like [:alpha:], which Perl is
4496 adopting.
4497
4498
4499 Version 2.08 31-Aug-99
4500 ----------------------
4501
4502 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4503 trying to match at the startoffset position, but instead was moving forward to
4504 the next newline as if a previous match had failed.
4505
4506 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4507 and could get into a loop if a null string was matched other than at the start
4508 of the subject.
4509
4510 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4511 be distinguished at compile time, and for completeness also added PCRE_DATE.
4512
4513 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4514 in GnuWin32 environments.
4515
4516
4517 Version 2.07 29-Jul-99
4518 ----------------------
4519
4520 1. The documentation is now supplied in plain text form and HTML as well as in
4521 the form of man page sources.
4522
4523 2. C++ compilers don't like assigning (void *) values to other pointer types.
4524 In particular this affects malloc(). Although there is no problem in Standard
4525 C, I've put in casts to keep C++ compilers happy.
4526
4527 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4528 should be (const char *).
4529
4530 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4531 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4532 However, I haven't made this a standard facility. The documentation doesn't
4533 mention it, and the Makefile doesn't support it.
4534
4535 5. The Makefile now contains an "install" target, with editable destinations at
4536 the top of the file. The pcretest program is not installed.
4537
4538 6. pgrep -V now gives the PCRE version number and date.
4539
4540 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4541 causing the entire string to be ignored, instead of just the last character.
4542
4543 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4544 non-matching string, it can take a very, very long time, even for strings of
4545 quite modest length, because of the nested recursion. PCRE now does better in
4546 some of these cases. It does this by remembering the last required literal
4547 character in the pattern, and pre-searching the subject to ensure it is present
4548 before running the real match. In other words, it applies a heuristic to detect
4549 some types of certain failure quickly, and in the above example, if presented
4550 with a string that has no trailing " it gives "no match" very quickly.
4551
4552 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4553 other alternatives are tried instead.
4554
4555
4556 Version 2.06 09-Jun-99
4557 ----------------------
4558
4559 1. Change pcretest's output for amount of store used to show just the code
4560 space, because the remainder (the data block) varies in size between 32-bit and
4561 64-bit systems.
4562
4563 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4564 start matching at. This allows lookbehinds to work when searching for multiple
4565 occurrences in a string.
4566
4567 3. Added additional options to pcretest for testing multiple occurrences:
4568
4569 /+ outputs the rest of the string that follows a match
4570 /g loops for multiple occurrences, using the new startoffset argument
4571 /G loops for multiple occurrences by passing an incremented pointer
4572
4573 4. PCRE wasn't doing the "first character" optimization for patterns starting
4574 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4575 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4576 the letter 'x'. On long subject strings, this gives a significant speed-up.
4577
4578
4579 Version 2.05 21-Apr-99
4580 ----------------------
4581
4582 1. Changed the type of magic_number from int to long int so that it works
4583 properly on 16-bit systems.
4584
4585 2. Fixed a bug which caused patterns starting with .* not to work correctly
4586 when the subject string contained newline characters. PCRE was assuming
4587 anchoring for such patterns in all cases, which is not correct because .* will
4588 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4589 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4590 must be retried after every newline in the subject.
4591
4592
4593 Version 2.04 18-Feb-99
4594 ----------------------
4595
4596 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4597 computation of the store needed to hold the pattern was incorrect (too large).
4598 If such patterns were nested a few deep, this could multiply and become a real
4599 problem.
4600
4601 2. Added /M option to pcretest to show the memory requirement of a specific
4602 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4603
4604 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4605 compiled in such a way that the backtracking after subsequent failure was
4606 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4607 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4608
4609
4610 Version 2.03 02-Feb-99
4611 ----------------------
4612
4613 1. Fixed typo and small mistake in man page.
4614
4615 2. Added 4th condition (GPL supersedes if conflict) and created separate
4616 LICENCE file containing the conditions.
4617
4618 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4619 Perl, that is the internal \ allows the delimiter to be included in the
4620 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4621 the final delimiter, add \ to the end of the pattern (to test the error).
4622
4623 4. Added the convenience functions for extracting substrings after a successful
4624 match. Updated pcretest to make it able to test these functions.
4625
4626
4627 Version 2.02 14-Jan-99
4628 ----------------------
4629
4630 1. Initialized the working variables associated with each extraction so that
4631 their saving and restoring doesn't refer to uninitialized store.
4632
4633 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4634 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4635 fix the problem.
4636
4637 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4638 calls, and wasn't printing the correct value for compiling calls. Increased the
4639 default value of LOOPREPEAT, and the number of significant figures in the
4640 times.
4641
4642 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4643
4644 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4645 a building problem on Windows NT with a FAT file system.
4646
4647
4648 Version 2.01 21-Oct-98
4649 ----------------------
4650
4651 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4652 to character tables built by pcre_maketables() in the current locale. If NULL
4653 is passed, the default tables are used.
4654
4655
4656 Version 2.00 24-Sep-98
4657 ----------------------
4658
4659 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4660 it any more.
4661
4662 2. Allow quantification of (?>) groups, and make it work correctly.
4663
4664 3. The first character computation wasn't working for (?>) groups.
4665
4666 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4667 end of the subject) and add 5.005's \z, which really does match only at the
4668 very end of the subject.
4669
4670 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4671
4672 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4673 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4674 localized options. All options to pcre_study() were also removed.
4675
4676 7. Add other new features from 5.005:
4677
4678 $(?<= positive lookbehind
4679 $(?<! negative lookbehind
4680 (?imsx-imsx) added the unsetting capability
4681 such a setting is global if at outer level; local otherwise
4682 (?imsx-imsx:) non-capturing groups with option setting
4683 (?(cond)re|re) conditional pattern matching
4684
4685 A backreference to itself in a repeated group matches the previous
4686 captured string.
4687
4688 8. General tidying up of studying (both automatic and via "study")
4689 consequential on the addition of new assertions.
4690
4691 9. As in 5.005, unlimited repeated groups that could match an empty substring
4692 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4693 runtime if any iteration does actually match an empty substring.
4694
4695 10. Include the RunTest script in the distribution.
4696
4697 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4698 discrepancies, some of which were old and were also with respect to 5.004. They
4699 have now been fixed.
4700
4701
4702 Version 1.09 28-Apr-98
4703 ----------------------
4704
4705 1. A negated single character class followed by a quantifier with a minimum
4706 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4707 program crashes, or just wrong answers. This did not apply to negated classes
4708 containing more than one character, or to minima other than one.
4709
4710
4711 Version 1.08 27-Mar-98
4712 ----------------------
4713
4714 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4715
4716 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4717 latter must appear before anything that relies on it in the pattern.
4718
4719
4720 Version 1.07 16-Feb-98
4721 ----------------------
4722
4723 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4724 repeat of a potentially empty string).
4725
4726
4727 Version 1.06 23-Jan-98
4728 ----------------------
4729
4730 1. Added Markus Oberhumer's little patches for C++.
4731
4732 2. Literal strings longer than 255 characters were broken.
4733
4734
4735 Version 1.05 23-Dec-97
4736 ----------------------
4737
4738 1. Negated character classes containing more than one character were failing if
4739 PCRE_CASELESS was set at run time.
4740
4741
4742 Version 1.04 19-Dec-97
4743 ----------------------
4744
4745 1. Corrected the man page, where some "const" qualifiers had been omitted.
4746
4747 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4748 input syntax.
4749
4750 3. Fixed memory leak which occurred when a regex with back references was
4751 matched with an offsets vector that wasn't big enough. The temporary memory
4752 that is used in this case wasn't being freed if the match failed.
4753
4754 4. Tidied pcretest to ensure it frees memory that it gets.
4755
4756 5. Temporary memory was being obtained in the case where the passed offsets
4757 vector was exactly big enough.
4758
4759 6. Corrected definition of offsetof() from change 5 below.
4760
4761 7. I had screwed up change 6 below and broken the rules for the use of
4762 setjmp(). Now fixed.
4763
4764
4765 Version 1.03 18-Dec-97
4766 ----------------------
4767
4768 1. A erroneous regex with a missing opening parenthesis was correctly
4769 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4770 on some systems.
4771
4772 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4773 it was reported that one broken compiler failed on the former because "code" is
4774 also an independent variable.
4775
4776 3. The erroneous regex a[]b caused an array overrun reference.
4777
4778 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4779 fail on data ending with that character. (It was going on too far, and checking
4780 the next character, typically a binary zero.) This was specific to the
4781 optimized code for single-character negative classes.
4782
4783 5. Added a contributed patch from the TIN world which does the following:
4784
4785 + Add an undef for memmove, in case the the system defines a macro for it.
4786
4787 + Add a definition of offsetof(), in case there isn't one. (I don't know
4788 the reason behind this - offsetof() is part of the ANSI standard - but
4789 it does no harm).
4790
4791 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4792 most of the places where whitespace preceded '#'. I have given up and
4793 allowed the remaining 2 cases to be at the margin.
4794
4795 + Rename some variables in pcre to eliminate shadowing. This seems very
4796 pedantic, but does no harm, of course.
4797
4798 6. Moved the call to setjmp() into its own function, to get rid of warnings
4799 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4800
4801 7. Constructs such as \d{8,} were compiling into the equivalent of
4802 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4803 outcome, but in this particular case used more store than had been allocated,
4804 which caused the bug to be discovered because it threw up an internal error.
4805
4806 8. The debugging code in both pcre and pcretest for outputting the compiled
4807 form of a regex was going wrong in the case of back references followed by
4808 curly-bracketed repeats.
4809
4810
4811 Version 1.02 12-Dec-97
4812 ----------------------
4813
4814 1. Typos in pcre.3 and comments in the source fixed.
4815
4816 2. Applied a contributed patch to get rid of places where it used to remove
4817 'const' from variables, and fixed some signed/unsigned and uninitialized
4818 variable warnings.
4819
4820 3. Added the "runtest" target to Makefile.
4821
4822 4. Set default compiler flag to -O2 rather than just -O.
4823
4824
4825 Version 1.01 19-Nov-97
4826 ----------------------
4827
4828 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4829 like /([ab]*)*/, that is, for classes with more than one character in them.
4830
4831 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4832 as /((?>a*))*/ (a PCRE_EXTRA facility).
4833
4834
4835 Version 1.00 18-Nov-97
4836 ----------------------
4837
4838 1. Added compile-time macros to support systems such as SunOS4 which don't have
4839 memmove() or strerror() but have other things that can be used instead.
4840
4841 2. Arranged that "make clean" removes the executables.
4842
4843
4844 Version 0.99 27-Oct-97
4845 ----------------------
4846
4847 1. Fixed bug in code for optimizing classes with only one character. It was
4848 initializing a 32-byte map regardless, which could cause it to run off the end
4849 of the memory it had got.
4850
4851 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4852
4853
4854 Version 0.98 22-Oct-97
4855 ----------------------
4856
4857 1. Fixed bug in code for handling temporary memory usage when there are more
4858 back references than supplied space in the ovector. This could cause segfaults.
4859
4860
4861 Version 0.97 21-Oct-97
4862 ----------------------
4863
4864 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4865
4866 2. Optimized negated single characters not to use a bit map.
4867
4868 3. Brought error texts together as macro definitions; clarified some of them;
4869 fixed one that was wrong - it said "range out of order" when it meant "invalid
4870 escape sequence".
4871
4872 4. Changed some char * arguments to const char *.
4873
4874 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4875
4876 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4877 pcretest.
4878
4879
4880 Version 0.96 16-Oct-97
4881 ----------------------
4882
4883 1. Added a simple "pgrep" utility to the distribution.
4884
4885 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4886 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4887 where "ddd" means "one or more decimal digits".
4888
4889 3. Fixed serious bug. If a pattern had a back reference, but the call to
4890 pcre_exec() didn't supply a large enough ovector to record the related
4891 identifying subpattern, the match always failed. PCRE now remembers the number
4892 of the largest back reference, and gets some temporary memory in which to save
4893 the offsets during matching if necessary, in order to ensure that
4894 backreferences always work.
4895
4896 4. Increased the compatibility with Perl in a number of ways:
4897
4898 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4899 to request this handling. The option can be set at compile or exec time.
4900
4901 (b) $ matches before a terminating newline by default; an option
4902 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4903 mode). The option can be set at compile or exec time.
4904
4905 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4906 the same as Perl's. If the decimal number it represents is less than 10
4907 or there aren't that many previous left capturing parentheses, an octal
4908 escape is read. Inside a character class, it's always an octal escape,
4909 even if it is a single digit.
4910
4911 (d) An escaped but undefined alphabetic character is taken as a literal,
4912 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4913 escapes.
4914
4915 (e) {0} is now permitted. (The previous item is removed from the compiled
4916 pattern).
4917
4918 5. Changed all the names of code files so that the basic parts are no longer
4919 than 10 characters, and abolished the teeny "globals.c" file.
4920
4921 6. Changed the handling of character classes; they are now done with a 32-byte
4922 bit map always.
4923
4924 7. Added the -d and /D options to pcretest to make it possible to look at the
4925 internals of compilation without having to recompile pcre.
4926
4927
4928 Version 0.95 23-Sep-97
4929 ----------------------
4930
4931 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4932 \x20 at the start of a run of normal characters. These were being treated as
4933 real characters, instead of the source characters being re-checked.
4934
4935
4936 Version 0.94 18-Sep-97
4937 ----------------------
4938
4939 1. The functions are now thread-safe, with the caveat that the global variables
4940 containing pointers to malloc() and free() or alternative functions are the
4941 same for all threads.
4942
4943 2. Get pcre_study() to generate a bitmap of initial characters for non-
4944 anchored patterns when this is possible, and use it if passed to pcre_exec().
4945
4946
4947 Version 0.93 15-Sep-97
4948 ----------------------
4949
4950 1. /(b)|(:+)/ was computing an incorrect first character.
4951
4952 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4953 but not actually doing anything yet.
4954
4955 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4956 as Perl does (e.g. [-az] or [az-]).
4957
4958 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4959 all possible positions.
4960
4961 5. Split up into different modules to avoid including unneeded functions in a
4962 compiled binary. However, compile and exec are still in one module. The "study"
4963 function is split off.
4964
4965 6. The character tables are now in a separate module whose source is generated
4966 by an auxiliary program - but can then be edited by hand if required. There are
4967 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4968 toupper() in the code.
4969
4970 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4971 make them global. Abolish the function for setting them, as the caller can now
4972 set them directly.
4973
4974
4975 Version 0.92 11-Sep-97
4976 ----------------------
4977
4978 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4979 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4980
4981 2. Caseless matching was not working in character classes if the characters in
4982 the pattern were in upper case.
4983
4984 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4985
4986 4. Make PCRE_ANCHORED public and accept as a compile option.
4987
4988 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4989 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4990 pass them.
4991
4992 6. Give an error if bad option bits passed at compile or run time.
4993
4994 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4995 pcretest to cause it to pass that flag.
4996
4997 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4998 options, and the first character, if set.
4999
5000 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5001
5002
5003 Version 0.91 10-Sep-97
5004 ----------------------
5005
5006 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5007 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5008
5009 2. PCRE was looping on encountering an indefinitely repeated back reference to
5010 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5011 Perl does - treats the match as successful.
5012
5013 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5