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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1238 - (show annotations)
Sat Jan 5 16:27:59 2013 UTC (6 years, 9 months ago) by ph10
File size: 240993 byte(s)
Small tweaks give performance improvements.
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
30 Version 8.32 30-November-2012
31 -----------------------------
32
33 1. Improved JIT compiler optimizations for first character search and single
34 character iterators.
35
36 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
37 Patch by Daniel Richard G.
38
39 3. Single character iterator optimizations in the JIT compiler.
40
41 4. Improved JIT compiler optimizations for character ranges.
42
43 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
44 Reported by Giuseppe D'Angelo.
45
46 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
47 of a line, was being set incorrectly in cases where .* appeared inside
48 atomic brackets at the start of a pattern, or where there was a subsequent
49 *PRUNE or *SKIP.
50
51 7. Improved instruction cache flush for POWER/PowerPC.
52 Patch by Daniel Richard G.
53
54 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
55 grep:
56
57 (a) There is now no limit to the number of patterns to be matched.
58
59 (b) An error is given if a pattern is too long.
60
61 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
62 are now supported.
63
64 (d) --exclude-from and --include-from (multiple use) have been added.
65
66 (e) Exclusions and inclusions now apply to all files and directories, not
67 just to those obtained from scanning a directory recursively.
68
69 (f) Multiple uses of -f and --file-list are now supported.
70
71 (g) In a Windows environment, the default for -d has been changed from
72 "read" (the GNU grep default) to "skip", because otherwise the presence
73 of a directory in the file list provokes an error.
74
75 (h) The documentation has been revised and clarified in places.
76
77 9. Improve the matching speed of capturing brackets.
78
79 10. Changed the meaning of \X so that it now matches a Unicode extended
80 grapheme cluster.
81
82 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
83 out POSIX threads when JIT support is configured.
84
85 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
86
87 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
88 structure could go wrong in environments where size_t is not the same size
89 as int.
90
91 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
92 set.
93
94 15. The EBCDIC support had decayed; later updates to the code had included
95 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
96 general tidy up of EBCDIC-related issues, and the documentation was also
97 not quite right. There is now a test that can be run on ASCII systems to
98 check some of the EBCDIC-related things (but is it not a full test).
99
100 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
101 in a small tidy to the code.
102
103 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
104
105 18. If the --only-matching (-o) option in pcregrep is specified multiple
106 times, each one causes appropriate output. For example, -o1 -o2 outputs the
107 substrings matched by the 1st and 2nd capturing parentheses. A separating
108 string can be specified by --om-separator (default empty).
109
110 19. Improving the first n character searches.
111
112 20. Turn case lists for horizontal and vertical white space into macros so that
113 they are defined only once.
114
115 21. This set of changes together give more compatible Unicode case-folding
116 behaviour for characters that have more than one other case when UCP
117 support is available.
118
119 (a) The Unicode property table now has offsets into a new table of sets of
120 three or more characters that are case-equivalent. The MultiStage2.py
121 script that generates these tables (the pcre_ucd.c file) now scans
122 CaseFolding.txt instead of UnicodeData.txt for character case
123 information.
124
125 (b) The code for adding characters or ranges of characters to a character
126 class has been abstracted into a generalized function that also handles
127 case-independence. In UTF-mode with UCP support, this uses the new data
128 to handle characters with more than one other case.
129
130 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
131 whose other case is greater than 256 are now correctly matched
132 caselessly. Previously, the high codepoint matched the low one, but not
133 vice versa.
134
135 (d) The processing of \h, \H, \v, and \ in character classes now makes use
136 of the new class addition function, using character lists defined as
137 macros alongside the case definitions of 20 above.
138
139 (e) Caseless back references now work with characters that have more than
140 one other case.
141
142 (f) General caseless matching of characters with more than one other case
143 is supported.
144
145 22. Unicode character properties were updated from Unicode 6.2.0
146
147 23. Improved CMake support under Windows. Patch by Daniel Richard G.
148
149 24. Add support for 32-bit character strings, and UTF-32
150
151 25. Major JIT compiler update (code refactoring and bugfixing).
152 Experimental Sparc 32 support is added.
153
154 26. Applied a modified version of Daniel Richard G's patch to create
155 pcre.h.generic and config.h.generic by "make" instead of in the
156 PrepareRelease script.
157
158 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
159 pcre_compile.c when checking for a zero character.
160
161 28. Introducing a native interface for JIT. Through this interface, the compiled
162 machine code can be directly executed. The purpose of this interface is to
163 provide fast pattern matching, so several sanity checks are not performed.
164 However, feature tests are still performed. The new interface provides
165 1.4x speedup compared to the old one.
166
167 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
168 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
169 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
170
171 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
172 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
173 the "old" RFC 2279). Instead, it ended up passing a negative length to
174 pcre_exec().
175
176 31. Add support for GCC's visibility feature to hide internal functions.
177
178 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
179 "unknown -C option" after outputting 0 or 1.
180
181 33. There is now support for generating a code coverage report for the test
182 suite in environments where gcc is the compiler and lcov is installed. This
183 is mainly for the benefit of the developers.
184
185 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
186 unaddressable using valgrind annotations, allowing valgrind to detect
187 invalid memory accesses. This is mainly for the benefit of the developers.
188
189 25. (*UTF) can now be used to start a pattern in any of the three libraries.
190
191 26. Give configure error if --enable-cpp but no C++ compiler found.
192
193
194 Version 8.31 06-July-2012
195 -------------------------
196
197 1. Fixing a wrong JIT test case and some compiler warnings.
198
199 2. Removed a bashism from the RunTest script.
200
201 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
202 to unsigned type, result still unsigned" that was given by an MS compiler
203 on encountering the code "-sizeof(xxx)".
204
205 4. Partial matching support is added to the JIT compiler.
206
207 5. Fixed several bugs concerned with partial matching of items that consist
208 of more than one character:
209
210 (a) /^(..)\1/ did not partially match "aba" because checking references was
211 done on an "all or nothing" basis. This also applied to repeated
212 references.
213
214 (b) \R did not give a hard partial match if \r was found at the end of the
215 subject.
216
217 (c) \X did not give a hard partial match after matching one or more
218 characters at the end of the subject.
219
220 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
221 a partial match for the string "\r".
222
223 (e) When newline was set to CRLF, the metacharacter "." did not recognize
224 a partial match for a CR character at the end of the subject string.
225
226 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
227 running pcretest, the text "(JIT)" added to the output whenever JIT is
228 actually used to run the match.
229
230 7. Individual JIT compile options can be set in pcretest by following -s+[+]
231 or /S+[+] with a digit between 1 and 7.
232
233 8. OP_NOT now supports any UTF character not just single-byte ones.
234
235 9. (*MARK) control verb is now supported by the JIT compiler.
236
237 10. The command "./RunTest list" lists the available tests without actually
238 running any of them. (Because I keep forgetting what they all are.)
239
240 11. Add PCRE_INFO_MAXLOOKBEHIND.
241
242 12. Applied a (slightly modified) user-supplied patch that improves performance
243 when the heap is used for recursion (compiled with --disable-stack-for-
244 recursion). Instead of malloc and free for each heap frame each time a
245 logical recursion happens, frames are retained on a chain and re-used where
246 possible. This sometimes gives as much as 30% improvement.
247
248 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
249 call.
250
251 14. As documented, (*COMMIT) is now confined to within a positive assertion.
252
253 15. It is now possible to link pcretest with libedit as an alternative to
254 libreadline.
255
256 16. (*COMMIT) control verb is now supported by the JIT compiler.
257
258 17. The Unicode data tables have been updated to Unicode 6.1.0.
259
260 18. Added --file-list option to pcregrep.
261
262 19. Added binary file support to pcregrep, including the -a, --binary-files,
263 -I, and --text options.
264
265 20. The madvise function is renamed for posix_madvise for QNX compatibility
266 reasons. Fixed by Giuseppe D'Angelo.
267
268 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
269
270 22. Changed the option for creating symbolic links for 16-bit man pages from
271 -s to -sf so that re-installing does not cause issues.
272
273 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
274
275 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
276 matches in certain environments (the workspace was not being correctly
277 retained). Also added to pcre_dfa_exec() a simple plausibility check on
278 some of the workspace data at the beginning of a restart.
279
280 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
281 was not doing so when it should - probably a typo introduced by SVN 528
282 (change 8.10/14).
283
284 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
285 \w+ when the character tables indicated that \x{c4} was a word character.
286 There were several related cases, all because the tests for doing a table
287 lookup were testing for characters less than 127 instead of 255.
288
289 27. If a pattern contains capturing parentheses that are not used in a match,
290 their slots in the ovector are set to -1. For those that are higher than
291 any matched groups, this happens at the end of processing. In the case when
292 there were back references that the ovector was too small to contain
293 (causing temporary malloc'd memory to be used during matching), and the
294 highest capturing number was not used, memory off the end of the ovector
295 was incorrectly being set to -1. (It was using the size of the temporary
296 memory instead of the true size.)
297
298 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
299 ovector size, it uses memory at the end of the block that it has got.
300
301 29. Check for an overlong MARK name and give an error at compile time. The
302 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
303
304 30. JIT compiler update.
305
306 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
307 Rill for the patch.
308
309 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
310
311 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
312
313 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
314 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
315 was enabled.
316
317 35. Improve JIT code generation for greedy plus quantifier.
318
319 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
320 1 to "aa" instead of to an empty string. The bug affected repeated groups
321 that could potentially match an empty string.
322
323 37. Optimizing single character iterators in JIT.
324
325 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
326 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
327 codepoints that are too big for the mode are faulted, and in a UTF mode,
328 disallowed codepoints are also faulted.
329
330 39. If PCRE was compiled with UTF support, in three places in the DFA
331 matcher there was code that should only have been obeyed in UTF mode, but
332 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
333 processing when bytes with values greater than 127 were present. In 16-bit
334 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
335 both cases the values are those that cannot be the first data item in a UTF
336 character. The three items that might have provoked this were recursions,
337 possessively repeated groups, and atomic groups.
338
339 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
340 and pcregrep, because some OS require shared objects to be explicitly
341 passed to ld, causing the link step to fail if they are not.
342
343 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
344 mode, patterns that started with \h* or \R* might be incorrectly matched.
345
346
347 Version 8.30 04-February-2012
348 -----------------------------
349
350 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
351 name is defined in ctype.h.
352
353 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
354 only in quite long subpatterns.
355
356 3. Removed the function pcre_info(), which has been obsolete and deprecated
357 since it was replaced by pcre_fullinfo() in February 2000.
358
359 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
360 match a (*MARK), and the match failed at the start of the subject, a
361 reference to memory before the start of the subject could occur. This bug
362 was introduced by fix 17 of release 8.21.
363
364 5. A reference to an unset group with zero minimum repetition was giving
365 totally wrong answers (in non-JavaScript-compatibility mode). For example,
366 /(another)?(\1?)test/ matched against "hello world test". This bug was
367 introduced in release 8.13.
368
369 6. Add support for 16-bit character strings (a large amount of work involving
370 many changes and refactorings).
371
372 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
373 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
374 from a file.
375
376 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
377 rounding is not applied in this particular case).
378
379 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
380 if they appear, or are escaped, in patterns.
381
382 10. Get rid of a number of -Wunused-but-set-variable warnings.
383
384 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
385 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
386 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
387 also returns the mark "x". This bug applied to capturing parentheses,
388 non-capturing parentheses, and atomic parentheses. It also applied to some
389 assertions.
390
391 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
392 information out of configure.ac instead of relying on pcre.h.generic, which
393 is not stored in the repository.
394
395 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
396 -lreadline.
397
398 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
399
400 15. Applied Graycode's patch to put the top-level frame on the stack rather
401 than the heap when not using the stack for recursion. This gives a
402 performance improvement in many cases when recursion is not deep.
403
404 16. Experimental code added to "pcretest -C" to output the stack frame size.
405
406
407 Version 8.21 12-Dec-2011
408 ------------------------
409
410 1. Updating the JIT compiler.
411
412 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
413 are added as well.
414
415 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
416 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
417 calling _pcre_jit_exec. Some extra comments are added.
418
419 4. (*MARK) settings inside atomic groups that do not contain any capturing
420 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
421 was introduced by change 18 for 8.20.
422
423 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
424 ECMA-262 standard.
425
426 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
427 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
428 This bug was probably introduced by change 9 of 8.13.
429
430 7. While fixing 6 above, I noticed that a number of other items were being
431 incorrectly rejected as "not fixed length". This arose partly because newer
432 opcodes had not been added to the fixed-length checking code. I have (a)
433 corrected the bug and added tests for these items, and (b) arranged for an
434 error to occur if an unknown opcode is encountered while checking for fixed
435 length instead of just assuming "not fixed length". The items that were
436 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
437 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
438 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
439
440 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
441 being incorrectly compiled and would have given unpredicatble results.
442
443 9. A possessively repeated subpattern with minimum repeat count greater than
444 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
445 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
446 the first (A) could occur when it should not.
447
448 10. Add a cast and remove a redundant test from the code.
449
450 11. JIT should use pcre_malloc/pcre_free for allocation.
451
452 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
453 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
454 is anything other than /usr, -L is still shown).
455
456 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
457
458 14. Perl does not support \N without a following name in a [] class; PCRE now
459 also gives an error.
460
461 15. If a forward reference was repeated with an upper limit of around 2000,
462 it caused the error "internal error: overran compiling workspace". The
463 maximum number of forward references (including repeats) was limited by the
464 internal workspace, and dependent on the LINK_SIZE. The code has been
465 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
466 the default depends on LINK_SIZE. There is a new upper limit (for safety)
467 of around 200,000 forward references. While doing this, I also speeded up
468 the filling in of repeated forward references.
469
470 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
471 incorrectly expecting the subject to contain another "a" after the start.
472
473 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
474 in the match, the SKIP should be ignored. This was not happening; instead
475 the SKIP was being treated as NOMATCH. For patterns such as
476 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
477 tested.
478
479 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
480 now much more compatible with Perl, in particular in cases where the result
481 is a non-match for a non-anchored pattern. For example, if
482 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
483 "m", where previously it did not return a name. A side effect of this
484 change is that for partial matches, the last encountered mark name is
485 returned, as for non matches. A number of tests that were previously not
486 Perl-compatible have been moved into the Perl-compatible test files. The
487 refactoring has had the pleasing side effect of removing one argument from
488 the match() function, thus reducing its stack requirements.
489
490 19. If the /S+ option was used in pcretest to study a pattern using JIT,
491 subsequent uses of /S (without +) incorrectly behaved like /S+.
492
493 21. Retrieve executable code size support for the JIT compiler and fixing
494 some warnings.
495
496 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
497 not work when the shorter character appeared right at the end of the
498 subject string.
499
500 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
501 systems.
502
503 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
504 output it when the /M option is used in pcretest.
505
506 25. The CheckMan script was not being included in the distribution. Also, added
507 an explicit "perl" to run Perl scripts from the PrepareRelease script
508 because this is reportedly needed in Windows.
509
510 26. If study data was being save in a file and studying had not found a set of
511 "starts with" bytes for the pattern, the data written to the file (though
512 never used) was taken from uninitialized memory and so caused valgrind to
513 complain.
514
515 27. Updated RunTest.bat as provided by Sheri Pierce.
516
517 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
518
519 29. Computation of memory usage for the table of capturing group names was
520 giving an unnecessarily large value.
521
522
523 Version 8.20 21-Oct-2011
524 ------------------------
525
526 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
527 a POSIX class. After further experiments with Perl, which convinced me that
528 Perl has bugs and confusions, a closing square bracket is no longer allowed
529 in a POSIX name. This bug also affected patterns with classes that started
530 with full stops.
531
532 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
533 captured substring, but while checking the failing first alternative,
534 substring 1 is temporarily captured. If the output vector supplied to
535 pcre_exec() was not big enough for this capture, the yield of the function
536 was still zero ("insufficient space for captured substrings"). This cannot
537 be totally fixed without adding another stack variable, which seems a lot
538 of expense for a edge case. However, I have improved the situation in cases
539 such as /(a)(b)x|abc/ matched against "abc", where the return code
540 indicates that fewer than the maximum number of slots in the ovector have
541 been set.
542
543 3. Related to (2) above: when there are more back references in a pattern than
544 slots in the output vector, pcre_exec() uses temporary memory during
545 matching, and copies in the captures as far as possible afterwards. It was
546 using the entire output vector, but this conflicts with the specification
547 that only 2/3 is used for passing back captured substrings. Now it uses
548 only the first 2/3, for compatibility. This is, of course, another edge
549 case.
550
551 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
552 main code base, and can be used by building with --enable-jit. When this is
553 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
554 runtime --no-jit option is given.
555
556 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
557 ovector, the return from the function was zero, implying that there were
558 other matches that did not fit. The correct "exactly full" value is now
559 returned.
560
561 6. If a subpattern that was called recursively or as a subroutine contained
562 (*PRUNE) or any other control that caused it to give a non-standard return,
563 invalid errors such as "Error -26 (nested recursion at the same subject
564 position)" or even infinite loops could occur.
565
566 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
567 computing the minimum length on reaching *ACCEPT, and so ended up with the
568 wrong value of 1 rather than 0. Further investigation indicates that
569 computing a minimum subject length in the presence of *ACCEPT is difficult
570 (think back references, subroutine calls), and so I have changed the code
571 so that no minimum is registered for a pattern that contains *ACCEPT.
572
573 8. If (*THEN) was present in the first (true) branch of a conditional group,
574 it was not handled as intended. [But see 16 below.]
575
576 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
577 Sheri Pierce.
578
579 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
580 the first byte in a match must be "a".
581
582 11. Change 17 for 8.13 increased the recursion depth for patterns like
583 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
584 pattern contains any instances of (*THEN). If it does not, the old
585 optimizations are restored. It would be nice to do this on a per-group
586 basis, but at the moment that is not feasible.
587
588 12. In some environments, the output of pcretest -C is CRLF terminated. This
589 broke RunTest's code that checks for the link size. A single white space
590 character after the value is now allowed for.
591
592 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
593 For "fr", it uses the Windows-specific input and output files.
594
595 14. If (*THEN) appeared in a group that was called recursively or as a
596 subroutine, it did not work as intended. [But see next item.]
597
598 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
599 pattern fragments (but not containing any | characters). If A and B are
600 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
601 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
602 D. In other words, Perl considers parentheses that do not contain any |
603 characters to be part of a surrounding alternative, whereas PCRE was
604 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
605 differently. PCRE now behaves in the same way as Perl, except in the case
606 of subroutine/recursion calls such as (?1) which have in any case always
607 been different (but PCRE had them first :-).
608
609 16. Related to 15 above: Perl does not treat the | in a conditional group as
610 creating alternatives. Such a group is treated in the same way as an
611 ordinary group without any | characters when processing (*THEN). PCRE has
612 been changed to match Perl's behaviour.
613
614 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
615 RunGrepTest script failed.
616
617 18. Change 22 for version 13 caused atomic groups to use more stack. This is
618 inevitable for groups that contain captures, but it can lead to a lot of
619 stack use in large patterns. The old behaviour has been restored for atomic
620 groups that do not contain any capturing parentheses.
621
622 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
623 suppress the check for a minimum subject length at run time. (If it was
624 given to pcre_exec() or pcre_dfa_exec() it did work.)
625
626 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
627 fail to work when decoding hex characters in data strings in EBCDIC
628 environments.
629
630 21. It appears that in at least one Mac OS environment, the isxdigit() function
631 is implemented as a macro that evaluates to its argument more than once,
632 contravening the C 90 Standard (I haven't checked a later standard). There
633 was an instance in pcretest which caused it to go wrong when processing
634 \x{...} escapes in subject strings. The has been rewritten to avoid using
635 things like p++ in the argument of isxdigit().
636
637
638 Version 8.13 16-Aug-2011
639 ------------------------
640
641 1. The Unicode data tables have been updated to Unicode 6.0.0.
642
643 2. Two minor typos in pcre_internal.h have been fixed.
644
645 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
646 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
647 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
648
649 4. There were a number of related bugs in the code for matching backrefences
650 caselessly in UTF-8 mode when codes for the characters concerned were
651 different numbers of bytes. For example, U+023A and U+2C65 are an upper
652 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
653 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
654 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
655 2-byte code at the end of the subject (it thought there wasn't enough data
656 left).
657
658 5. Comprehensive information about what went wrong is now returned by
659 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
660 as the output vector has at least 2 elements. The offset of the start of
661 the failing character and a reason code are placed in the vector.
662
663 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
664 now returned is for the first byte of the failing character, instead of the
665 last byte inspected. This is an incompatible change, but I hope it is small
666 enough not to be a problem. It makes the returned offset consistent with
667 pcre_exec() and pcre_dfa_exec().
668
669 7. pcretest now gives a text phrase as well as the error number when
670 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
671 failure, the offset and reason code are output.
672
673 8. When \R was used with a maximizing quantifier it failed to skip backwards
674 over a \r\n pair if the subsequent match failed. Instead, it just skipped
675 back over a single character (\n). This seems wrong (because it treated the
676 two characters as a single entity when going forwards), conflicts with the
677 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
678 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
679 has been changed.
680
681 9. Some internal refactoring has changed the processing so that the handling
682 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
683 time (the PCRE_DOTALL option was changed this way some time ago: version
684 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
685 which was always a bit of a fudge. It also means that there is one less
686 argument for the match() function, which reduces its stack requirements
687 slightly. This change also fixes an incompatibility with Perl: the pattern
688 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
689
690 10. More internal refactoring has drastically reduced the number of recursive
691 calls to match() for possessively repeated groups such as (abc)++ when
692 using pcre_exec().
693
694 11. While implementing 10, a number of bugs in the handling of groups were
695 discovered and fixed:
696
697 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
698 (a|)*(?1) gave a compile-time internal error.
699 ((a|)+)+ did not notice that the outer group could match an empty string.
700 (^a|^)+ was not marked as anchored.
701 (.*a|.*)+ was not marked as matching at start or after a newline.
702
703 12. Yet more internal refactoring has removed another argument from the match()
704 function. Special calls to this function are now indicated by setting a
705 value in a variable in the "match data" data block.
706
707 13. Be more explicit in pcre_study() instead of relying on "default" for
708 opcodes that mean there is no starting character; this means that when new
709 ones are added and accidentally left out of pcre_study(), testing should
710 pick them up.
711
712 14. The -s option of pcretest has been documented for ages as being an old
713 synonym of -m (show memory usage). I have changed it to mean "force study
714 for every regex", that is, assume /S for every regex. This is similar to -i
715 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
716 using it. It makes it easier to run collections of tests with and without
717 study enabled, and thereby test pcre_study() more easily. All the standard
718 tests are now run with and without -s (but some patterns can be marked as
719 "never study" - see 20 below).
720
721 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
722 restoration of the capturing data to the outer values was not happening
723 correctly.
724
725 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
726 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
727 pattern had matched an empty string, and so incorrectly returned a no
728 match.
729
730 17. There was optimizing code for the last branch of non-capturing parentheses,
731 and also for the obeyed branch of a conditional subexpression, which used
732 tail recursion to cut down on stack usage. Unfortunately, now that there is
733 the possibility of (*THEN) occurring in these branches, tail recursion is
734 no longer possible because the return has to be checked for (*THEN). These
735 two optimizations have therefore been removed. [But see 8.20/11 above.]
736
737 18. If a pattern containing \R was studied, it was assumed that \R always
738 matched two bytes, thus causing the minimum subject length to be
739 incorrectly computed because \R can also match just one byte.
740
741 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
742 was incorrectly computed.
743
744 20. If /S is present twice on a test pattern in pcretest input, it now
745 *disables* studying, thereby overriding the use of -s on the command line
746 (see 14 above). This is necessary for one or two tests to keep the output
747 identical in both cases.
748
749 21. When (*ACCEPT) was used in an assertion that matched an empty string and
750 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
751
752 22. When an atomic group that contained a capturing parenthesis was
753 successfully matched, but the branch in which it appeared failed, the
754 capturing was not being forgotten if a higher numbered group was later
755 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
756 group 1 to "a", when in fact it should be unset. This applied to multi-
757 branched capturing and non-capturing groups, repeated or not, and also to
758 positive assertions (capturing in negative assertions does not happen
759 in PCRE) and also to nested atomic groups.
760
761 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
762 subject after a captured substring, to make it easier to tell which of a
763 number of identical substrings has been captured.
764
765 24. The way atomic groups are processed by pcre_exec() has been changed so that
766 if they are repeated, backtracking one repetition now resets captured
767 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
768 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
769 "aaa". Previously, it would have been "a". As part of this code
770 refactoring, the way recursive calls are handled has also been changed.
771
772 25. If an assertion condition captured any substrings, they were not passed
773 back unless some other capturing happened later. For example, if
774 (?(?=(a))a) was matched against "a", no capturing was returned.
775
776 26. When studying a pattern that contained subroutine calls or assertions,
777 the code for finding the minimum length of a possible match was handling
778 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
779 group 1 called group 2 while simultaneously a separate group 2 called group
780 1). A stack overflow occurred in this case. I have fixed this by limiting
781 the recursion depth to 10.
782
783 27. Updated RunTest.bat in the distribution to the version supplied by Tom
784 Fortmann. This supports explicit test numbers on the command line, and has
785 argument validation and error reporting.
786
787 28. An instance of \X with an unlimited repeat could fail if at any point the
788 first character it looked at was a mark character.
789
790 29. Some minor code refactoring concerning Unicode properties and scripts
791 should reduce the stack requirement of match() slightly.
792
793 30. Added the '=' option to pcretest to check the setting of unused capturing
794 slots at the end of the pattern, which are documented as being -1, but are
795 not included in the return count.
796
797 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
798 compiled something random. Now it gives a compile-time error (as does
799 Perl).
800
801 32. A *MARK encountered during the processing of a positive assertion is now
802 recorded and passed back (compatible with Perl).
803
804 33. If --only-matching or --colour was set on a pcregrep call whose pattern
805 had alternative anchored branches, the search for a second match in a line
806 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
807 matched the line "0102" twice. The same bug affected patterns that started
808 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
809 twice.
810
811 34. Previously, PCRE did not allow quantification of assertions. However, Perl
812 does, and because of capturing effects, quantifying parenthesized
813 assertions may at times be useful. Quantifiers are now allowed for
814 parenthesized assertions.
815
816 35. A minor code tidy in pcre_compile() when checking options for \R usage.
817
818 36. \g was being checked for fancy things in a character class, when it should
819 just be a literal "g".
820
821 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
822 appearance of a nested POSIX class supersedes an apparent external class.
823 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
824 unescaped square brackets may also appear as part of class names. For
825 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
826 more like Perl. (But see 8.20/1 above.)
827
828 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
829 was because it thought it was \N{name}, which is not supported).
830
831 39. Add minix to OS list not supporting the -S option in pcretest.
832
833 40. PCRE tries to detect cases of infinite recursion at compile time, but it
834 cannot analyze patterns in sufficient detail to catch mutual recursions
835 such as ((?1))((?2)). There is now a runtime test that gives an error if a
836 subgroup is called recursively as a subpattern for a second time at the
837 same position in the subject string. In previous releases this might have
838 been caught by the recursion limit, or it might have run out of stack.
839
840 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
841 happen only once. PCRE was, however incorrectly giving a compile time error
842 "recursive call could loop indefinitely" because it cannot analyze the
843 pattern in sufficient detail. The compile time test no longer happens when
844 PCRE is compiling a conditional subpattern, but actual runaway loops are
845 now caught at runtime (see 40 above).
846
847 42. It seems that Perl allows any characters other than a closing parenthesis
848 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
849 has been changed to be the same.
850
851 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
852 as not to get warnings when autogen.sh is called. Also changed
853 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
854
855 44. To help people who use pcregrep to scan files containing exceedingly long
856 lines, the following changes have been made:
857
858 (a) The default value of the buffer size parameter has been increased from
859 8K to 20K. (The actual buffer used is three times this size.)
860
861 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
862 PCRE is built.
863
864 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
865 to be set at run time.
866
867 (d) Numerical values in pcregrep options can be followed by K or M, for
868 example --buffer-size=50K.
869
870 (e) If a line being scanned overflows pcregrep's buffer, an error is now
871 given and the return code is set to 2.
872
873 45. Add a pointer to the latest mark to the callout data block.
874
875 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
876 partial match of an empty string instead of no match. This was specific to
877 the use of ".".
878
879 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
880 complete match instead of a partial match. This bug was dependent on both
881 the PCRE_UTF8 and PCRE_DOTALL options being set.
882
883 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
884 starting byte set, because \b was not being ignored.
885
886
887 Version 8.12 15-Jan-2011
888 ------------------------
889
890 1. Fixed some typos in the markup of the man pages, and wrote a script that
891 checks for such things as part of the documentation building process.
892
893 2. On a big-endian 64-bit system, pcregrep did not correctly process the
894 --match-limit and --recursion-limit options (added for 8.11). In
895 particular, this made one of the standard tests fail. (The integer value
896 went into the wrong half of a long int.)
897
898 3. If the --colour option was given to pcregrep with -v (invert match), it
899 did strange things, either producing crazy output, or crashing. It should,
900 of course, ignore a request for colour when reporting lines that do not
901 match.
902
903 4. Another pcregrep bug caused similar problems if --colour was specified with
904 -M (multiline) and the pattern match finished with a line ending.
905
906 5. In pcregrep, when a pattern that ended with a literal newline sequence was
907 matched in multiline mode, the following line was shown as part of the
908 match. This seems wrong, so I have changed it.
909
910 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
911 the check for further matches in the same line (so they could be coloured)
912 to overrun the end of the current line. If another match was found, it was
913 incorrectly shown (and then shown again when found in the next line).
914
915 7. If pcregrep was compiled under Windows, there was a reference to the
916 function pcregrep_exit() before it was defined. I am assuming this was
917 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
918 reported by a user. I've moved the definition above the reference.
919
920
921 Version 8.11 10-Dec-2010
922 ------------------------
923
924 1. (*THEN) was not working properly if there were untried alternatives prior
925 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
926 backtracked to try for "b" instead of moving to the next alternative branch
927 at the same level (in this case, to look for "c"). The Perl documentation
928 is clear that when (*THEN) is backtracked onto, it goes to the "next
929 alternative in the innermost enclosing group".
930
931 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
932 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
933 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
934 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
935 (*THEN).
936
937 3. If \s appeared in a character class, it removed the VT character from
938 the class, even if it had been included by some previous item, for example
939 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
940 of \s, but is part of the POSIX "space" class.)
941
942 4. A partial match never returns an empty string (because you can always
943 match an empty string at the end of the subject); however the checking for
944 an empty string was starting at the "start of match" point. This has been
945 changed to the "earliest inspected character" point, because the returned
946 data for a partial match starts at this character. This means that, for
947 example, /(?<=abc)def/ gives a partial match for the subject "abc"
948 (previously it gave "no match").
949
950 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
951 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
952 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
953 has an implication that the given string is incomplete (because a partial
954 match is preferred over a full match). For this reason, these items now
955 give a partial match in this situation. [Aside: previously, the one case
956 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
957 match rather than a full match, which was wrong by the old rules, but is
958 now correct.]
959
960 6. There was a bug in the handling of #-introduced comments, recognized when
961 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
962 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
963 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
964 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
965 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
966 places in pcre_compile().
967
968 7. Related to (6) above, when pcre_compile() was skipping #-introduced
969 comments when looking ahead for named forward references to subpatterns,
970 the only newline sequence it recognized was NL. It now handles newlines
971 according to the set newline convention.
972
973 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
974 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
975 cater for a lack of strerror(). These oversights have been fixed.
976
977 9. Added --match-limit and --recursion-limit to pcregrep.
978
979 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
980
981 11. When the -o option was used, pcregrep was setting a return code of 1, even
982 when matches were found, and --line-buffered was not being honoured.
983
984 12. Added an optional parentheses number to the -o and --only-matching options
985 of pcregrep.
986
987 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
988 can match an empty string. The code to do it in pcretest and pcredemo
989 needed fixing:
990
991 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
992 only one byte after an empty string match just before CRLF (this case
993 just got forgotten; "any" and "anycrlf" were OK).
994
995 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
996 mode when an empty string match preceded an ASCII character followed by
997 a non-ASCII character. (The code for advancing by one character rather
998 than one byte was nonsense.)
999
1000 (c) The pcredemo.c sample program did not have any code at all to handle
1001 the cases when CRLF is a valid newline sequence.
1002
1003 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1004 as a starting offset was within the subject string. There is now a new
1005 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1006 negative or greater than the length of the string. In order to test this,
1007 pcretest is extended to allow the setting of negative starting offsets.
1008
1009 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1010 starting offset points to the beginning of a UTF-8 character was
1011 unnecessarily clumsy. I tidied it up.
1012
1013 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1014 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1015
1016 17. Nobody had reported that the --include_dir option, which was added in
1017 release 7.7 should have been called --include-dir (hyphen, not underscore)
1018 for compatibility with GNU grep. I have changed it to --include-dir, but
1019 left --include_dir as an undocumented synonym, and the same for
1020 --exclude-dir, though that is not available in GNU grep, at least as of
1021 release 2.5.4.
1022
1023 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1024 characters from a string of bytes) have been redefined so as not to use
1025 loops, in order to improve performance in some environments. At the same
1026 time, I abstracted some of the common code into auxiliary macros to save
1027 repetition (this should not affect the compiled code).
1028
1029 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1030 compile-time error is now given if \c is not followed by an ASCII
1031 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1032 different, and any byte value is allowed.)
1033
1034 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1035 START_OPTIMIZE option, which is now allowed at compile time - but just
1036 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1037 to pcregrep and other applications that have no direct access to PCRE
1038 options. The new /Y option in pcretest sets this option when calling
1039 pcre_compile().
1040
1041 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1042 back references. Groups containing recursive back references were forced to
1043 be atomic by that change, but in the case of named groups, the amount of
1044 memory required was incorrectly computed, leading to "Failed: internal
1045 error: code overflow". This has been fixed.
1046
1047 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1048 pcretest.c, to avoid build problems in some Borland environments.
1049
1050
1051 Version 8.10 25-Jun-2010
1052 ------------------------
1053
1054 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1055 THEN.
1056
1057 2. (*ACCEPT) was not working when inside an atomic group.
1058
1059 3. Inside a character class, \B is treated as a literal by default, but
1060 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1061 causes the error). The code is unchanged, but I tidied the documentation.
1062
1063 4. Inside a character class, PCRE always treated \R and \X as literals,
1064 whereas Perl faults them if its -w option is set. I have changed PCRE so
1065 that it faults them when PCRE_EXTRA is set.
1066
1067 5. Added support for \N, which always matches any character other than
1068 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1069
1070 6. When compiling pcregrep with newer versions of gcc which may have
1071 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1072 declared with attribute warn_unused_result" were given. Just casting the
1073 result to (void) does not stop the warnings; a more elaborate fudge is
1074 needed. I've used a macro to implement this.
1075
1076 7. Minor change to pcretest.c to avoid a compiler warning.
1077
1078 8. Added four artifical Unicode properties to help with an option to make
1079 \s etc use properties (see next item). The new properties are: Xan
1080 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1081
1082 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1083 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1084 this option. Modified pcretest to add /W to test this facility. Added
1085 REG_UCP to make it available via the POSIX interface.
1086
1087 10. Added --line-buffered to pcregrep.
1088
1089 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1090 studied, and the match started with a letter with a code point greater than
1091 127 whose first byte was different to the first byte of the other case of
1092 the letter, the other case of this starting letter was not recognized
1093 (#976).
1094
1095 12. If a pattern that was studied started with a repeated Unicode property
1096 test, for example, \p{Nd}+, there was the theoretical possibility of
1097 setting up an incorrect bitmap of starting bytes, but fortunately it could
1098 not have actually happened in practice until change 8 above was made (it
1099 added property types that matched character-matching opcodes).
1100
1101 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1102 possible starting bytes for non-anchored patterns.
1103
1104 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1105 \R, and also a number of cases that involve Unicode properties, both
1106 explicit and implicit when PCRE_UCP is set.
1107
1108 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1109 input, it could crash or give wrong results if characters with values
1110 greater than 0xc0 were present in the subject string. (Detail: it assumed
1111 UTF-8 input when processing these items.)
1112
1113 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1114 size_t is 64-bit (#991).
1115
1116 17. Added a check for running out of memory when PCRE is compiled with
1117 --disable-stack-for-recursion (#990).
1118
1119 18. If the last data line in a file for pcretest does not have a newline on
1120 the end, a newline was missing in the output.
1121
1122 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1123 less than 128) in its various bitmaps. However, there is a facility for
1124 generating tables according to the current locale when PCRE is compiled. It
1125 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1126 characters, are recognized by isspace() and therefore were getting set in
1127 these tables, and indeed these tables seem to approximate to ISO 8859. This
1128 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1129 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1130 which of course cannot start UTF-8 characters. I have changed the code so
1131 that only real ASCII characters (less than 128) and the correct starting
1132 bytes for UTF-8 encodings are set for characters greater than 127 when in
1133 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1134 altogether.)
1135
1136 20. Added the /T option to pcretest so as to be able to run tests with non-
1137 standard character tables, thus making it possible to include the tests
1138 used for 19 above in the standard set of tests.
1139
1140 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1141 reference to a subpattern the other side of a comment that contains an
1142 opening parenthesis caused either an internal compiling error, or a
1143 reference to the wrong subpattern.
1144
1145
1146 Version 8.02 19-Mar-2010
1147 ------------------------
1148
1149 1. The Unicode data tables have been updated to Unicode 5.2.0.
1150
1151 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1152 configured.
1153
1154 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1155 original author of that file, following a query about its status.
1156
1157 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1158 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1159
1160 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1161 quantifier applied to a forward-referencing subroutine call, could compile
1162 incorrect code or give the error "internal error: previously-checked
1163 referenced subpattern not found".
1164
1165 6. Both MS Visual Studio and Symbian OS have problems with initializing
1166 variables to point to external functions. For these systems, therefore,
1167 pcre_malloc etc. are now initialized to local functions that call the
1168 relevant global functions.
1169
1170 7. There were two entries missing in the vectors called coptable and poptable
1171 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1172 I've fixed the data, and added a kludgy way of testing at compile time that
1173 the lengths are correct (equal to the number of opcodes).
1174
1175 8. Following on from 7, I added a similar kludge to check the length of the
1176 eint vector in pcreposix.c.
1177
1178 9. Error texts for pcre_compile() are held as one long string to avoid too
1179 much relocation at load time. To find a text, the string is searched,
1180 counting zeros. There was no check for running off the end of the string,
1181 which could happen if a new error number was added without updating the
1182 string.
1183
1184 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1185
1186 11. \K was not working if it appeared in an atomic group or in a group that
1187 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1188 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1189 the assertion is positive, but not if it is negative.
1190
1191 12. Change 11 fortuitously reduced the size of the stack frame used in the
1192 "match()" function of pcre_exec.c by one pointer. Forthcoming
1193 implementation of support for (*MARK) will need an extra pointer on the
1194 stack; I have reserved it now, so that the stack frame size does not
1195 decrease.
1196
1197 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1198 item in branch that calls a recursion is a subroutine call - as in the
1199 second branch in the above example - was incorrectly given the compile-
1200 time error "recursive call could loop indefinitely" because pcre_compile()
1201 was not correctly checking the subroutine for matching a non-empty string.
1202
1203 14. The checks for overrunning compiling workspace could trigger after an
1204 overrun had occurred. This is a "should never occur" error, but it can be
1205 triggered by pathological patterns such as hundreds of nested parentheses.
1206 The checks now trigger 100 bytes before the end of the workspace.
1207
1208 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1209
1210
1211 Version 8.01 19-Jan-2010
1212 ------------------------
1213
1214 1. If a pattern contained a conditional subpattern with only one branch (in
1215 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1216 computed the wrong minimum data length (which is of course zero for such
1217 subpatterns). This could cause incorrect "no match" results.
1218
1219 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1220 the pattern is reset in the first branch, pcre_compile() failed with
1221 "internal error: code overflow at offset...". This happened only when
1222 the reset was to the original external option setting. (An optimization
1223 abstracts leading options settings into an external setting, which was the
1224 cause of this.)
1225
1226 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1227 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1228 assertion pattern did not match (meaning that the assertion was true), it
1229 was incorrectly treated as false if the SKIP had been reached during the
1230 matching. This also applied to assertions used as conditions.
1231
1232 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1233 assertion subpattern, including such a pattern used as a condition,
1234 unpredictable results occurred, instead of the error return
1235 PCRE_ERROR_DFA_UITEM.
1236
1237 5. The C++ GlobalReplace function was not working like Perl for the special
1238 situation when an empty string is matched. It now does the fancy magic
1239 stuff that is necessary.
1240
1241 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1242 removed. (These were left over from very, very early versions of PCRE.)
1243
1244 7. Some cosmetic changes to the code to make life easier when compiling it
1245 as part of something else:
1246
1247 (a) Change DEBUG to PCRE_DEBUG.
1248
1249 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1250 called "current" as "current_branch", to prevent a collision with the
1251 Linux macro when compiled as a kernel module.
1252
1253 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1254 prevent a collision with the Linux macro when compiled as a kernel
1255 module.
1256
1257 8. In pcre_compile() there are some checks for integer overflows that used to
1258 cast potentially large values to (double). This has been changed to that
1259 when building, a check for int64_t is made, and if it is found, it is used
1260 instead, thus avoiding the use of floating point arithmetic. (There is no
1261 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1262 double.
1263
1264 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1265 2005 (difference between two addresses compared to an unsigned value).
1266
1267 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1268 custom one, because of the following reported problem in Windows:
1269
1270 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1271 under Win32.
1272 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1273 therefore missing the function definition.
1274 - The compiler thus generates a "C" signature for the test function.
1275 - The linker fails to find the "C" function.
1276 - PCRE fails to configure if asked to do so against libbz2.
1277
1278 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1279 messages were output:
1280
1281 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1282 rerunning libtoolize, to keep the correct libtool macros in-tree.
1283 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1284
1285 I have done both of these things.
1286
1287 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1288 most of the time, it *can* run out if it is given a pattern that contains a
1289 runaway infinite recursion. I updated the discussion in the pcrestack man
1290 page.
1291
1292 13. Now that we have gone to the x.xx style of version numbers, the minor
1293 version may start with zero. Using 08 or 09 is a bad idea because users
1294 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1295 interpreted as invalid octal numbers. I've updated the previous comment in
1296 configure.ac, and also added a check that gives an error if 08 or 09 are
1297 used.
1298
1299 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1300 causing partial matching to fail when the end of the subject matched \W
1301 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1302
1303 15. There were some discrepancies between the declarations in pcre_internal.h
1304 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1305 their definitions. The declarations used "const uschar *" and the
1306 definitions used USPTR. Even though USPTR is normally defined as "const
1307 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1308 reported that: "This difference in casting confuses some C++ compilers, for
1309 example, SunCC recognizes above declarations as different functions and
1310 generates broken code for hbpcre." I have changed the declarations to use
1311 USPTR.
1312
1313 16. GNU libtool is named differently on some systems. The autogen.sh script now
1314 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1315 (FreeBSD).
1316
1317 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1318 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1319 comment: "Figure out how to create a longlong from a string: strtoll and
1320 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1321 instance, but it only takes 2 args instead of 3!"
1322
1323 18. A subtle bug concerned with back references has been fixed by a change of
1324 specification, with a corresponding code fix. A pattern such as
1325 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1326 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1327 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1328 same bug. Such groups have to be quantified to be useful, or contained
1329 inside another quantified group. (If there's no repetition, the reference
1330 can never match.) The problem arises because, having left the group and
1331 moved on to the rest of the pattern, a later failure that backtracks into
1332 the group uses the captured value from the final iteration of the group
1333 rather than the correct earlier one. I have fixed this in PCRE by forcing
1334 any group that contains a reference to itself to be an atomic group; that
1335 is, there cannot be any backtracking into it once it has completed. This is
1336 similar to recursive and subroutine calls.
1337
1338
1339 Version 8.00 19-Oct-09
1340 ----------------------
1341
1342 1. The table for translating pcre_compile() error codes into POSIX error codes
1343 was out-of-date, and there was no check on the pcre_compile() error code
1344 being within the table. This could lead to an OK return being given in
1345 error.
1346
1347 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1348 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1349 in a Windows environment.
1350
1351 3. The pcregrep --count option prints the count for each file even when it is
1352 zero, as does GNU grep. However, pcregrep was also printing all files when
1353 --files-with-matches was added. Now, when both options are given, it prints
1354 counts only for those files that have at least one match. (GNU grep just
1355 prints the file name in this circumstance, but including the count seems
1356 more useful - otherwise, why use --count?) Also ensured that the
1357 combination -clh just lists non-zero counts, with no names.
1358
1359 4. The long form of the pcregrep -F option was incorrectly implemented as
1360 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1361 but it seems right to fix it, and I didn't think it was worth preserving
1362 the old behaviour.
1363
1364 5. The command line items --regex=pattern and --regexp=pattern were not
1365 recognized by pcregrep, which required --regex pattern or --regexp pattern
1366 (with a space rather than an '='). The man page documented the '=' forms,
1367 which are compatible with GNU grep; these now work.
1368
1369 6. No libpcreposix.pc file was created for pkg-config; there was just
1370 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1371
1372 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1373 when UCP support is not needed, by modifying the Python script that
1374 generates it from Unicode data files. This should not matter if the module
1375 is correctly used as a library, but I received one complaint about 50K of
1376 unwanted data. My guess is that the person linked everything into his
1377 program rather than using a library. Anyway, it does no harm.
1378
1379 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1380 was a minimum greater than 1 for a wide character in a possessive
1381 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1382 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1383 character. Chaos in the form of incorrect output or a compiling loop could
1384 result.
1385
1386 9. The restrictions on what a pattern can contain when partial matching is
1387 requested for pcre_exec() have been removed. All patterns can now be
1388 partially matched by this function. In addition, if there are at least two
1389 slots in the offset vector, the offset of the earliest inspected character
1390 for the match and the offset of the end of the subject are set in them when
1391 PCRE_ERROR_PARTIAL is returned.
1392
1393 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1394 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1395 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1396 and may be more useful for multi-segment matching.
1397
1398 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1399 used to be given if ever the end of the subject was reached; now it is
1400 given only if matching could not proceed because another character was
1401 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1402 string "Z", which now yields "no match" instead of "partial match". In the
1403 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1404 final character ended with (*FAIL).
1405
1406 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1407 if the pattern had a "must contain" character that was already found in the
1408 earlier partial match, unless partial matching was again requested. For
1409 example, with the pattern /dog.(body)?/, the "must contain" character is
1410 "g". If the first part-match was for the string "dog", restarting with
1411 "sbody" failed. This bug has been fixed.
1412
1413 13. The string returned by pcre_dfa_exec() after a partial match has been
1414 changed so that it starts at the first inspected character rather than the
1415 first character of the match. This makes a difference only if the pattern
1416 starts with a lookbehind assertion or \b or \B (\K is not supported by
1417 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1418 matching functions compatible, and I think it's the right thing to do.
1419
1420 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1421 so that the demonstration program is easily available in environments where
1422 PCRE has not been installed from source.
1423
1424 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1425 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1426 library.
1427
1428 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1429 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1430 is not the first non-POSIX option to be added. Clearly some people find
1431 these options useful.
1432
1433 17. If a caller to the POSIX matching function regexec() passes a non-zero
1434 value for nmatch with a NULL value for pmatch, the value of
1435 nmatch is forced to zero.
1436
1437 18. RunGrepTest did not have a test for the availability of the -u option of
1438 the diff command, as RunTest does. It now checks in the same way as
1439 RunTest, and also checks for the -b option.
1440
1441 19. If an odd number of negated classes containing just a single character
1442 interposed, within parentheses, between a forward reference to a named
1443 subpattern and the definition of the subpattern, compilation crashed with
1444 an internal error, complaining that it could not find the referenced
1445 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1446 [The bug was that it was starting one character too far in when skipping
1447 over the character class, thus treating the ] as data rather than
1448 terminating the class. This meant it could skip too much.]
1449
1450 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1451 /g option in pcretest when the pattern contains \K, which makes it possible
1452 to have an empty string match not at the start, even when the pattern is
1453 anchored. Updated pcretest and pcredemo to use this option.
1454
1455 21. If the maximum number of capturing subpatterns in a recursion was greater
1456 than the maximum at the outer level, the higher number was returned, but
1457 with unset values at the outer level. The correct (outer level) value is
1458 now given.
1459
1460 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1461 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1462 make it do so. The string so far is captured, making this feature
1463 compatible with Perl.
1464
1465 23. The tests have been re-organized, adding tests 11 and 12, to make it
1466 possible to check the Perl 5.10 features against Perl 5.10.
1467
1468 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1469 pattern matches a fixed length string. PCRE did not allow this; now it
1470 does. Neither allows recursion.
1471
1472 25. I finally figured out how to implement a request to provide the minimum
1473 length of subject string that was needed in order to match a given pattern.
1474 (It was back references and recursion that I had previously got hung up
1475 on.) This code has now been added to pcre_study(); it finds a lower bound
1476 to the length of subject needed. It is not necessarily the greatest lower
1477 bound, but using it to avoid searching strings that are too short does give
1478 some useful speed-ups. The value is available to calling programs via
1479 pcre_fullinfo().
1480
1481 26. While implementing 25, I discovered to my embarrassment that pcretest had
1482 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1483 study optimizations had never been tested with that matching function.
1484 Oops. What is worse, even when it was passed study data, there was a bug in
1485 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1486 were also very few tests of studied patterns with pcre_dfa_exec().
1487
1488 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1489 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1490 on the other side of the coin, they are no longer allowed to have different
1491 names, because these cannot be distinguished in PCRE, and this has caused
1492 confusion. (This is a difference from Perl.)
1493
1494 28. When duplicate subpattern names are present (necessarily with different
1495 numbers, as required by 27 above), and a test is made by name in a
1496 conditional pattern, either for a subpattern having been matched, or for
1497 recursion in such a pattern, all the associated numbered subpatterns are
1498 tested, and the overall condition is true if the condition is true for any
1499 one of them. This is the way Perl works, and is also more like the way
1500 testing by number works.
1501
1502
1503 Version 7.9 11-Apr-09
1504 ---------------------
1505
1506 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1507 (pcretest), all targets were linked against these libraries. This included
1508 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1509 libraries. This caused unwanted dependencies to be created. This problem
1510 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1511 pcretest is linked with readline.
1512
1513 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1514 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1515 moved outside it again, because FALSE and TRUE are already defined in AIX,
1516 but BOOL is not.
1517
1518 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1519 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1520
1521 4. The pcregrep documentation said spaces were inserted as well as colons (or
1522 hyphens) following file names and line numbers when outputting matching
1523 lines. This is not true; no spaces are inserted. I have also clarified the
1524 wording for the --colour (or --color) option.
1525
1526 5. In pcregrep, when --colour was used with -o, the list of matching strings
1527 was not coloured; this is different to GNU grep, so I have changed it to be
1528 the same.
1529
1530 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1531 each matching line was coloured. Now it goes on to look for further matches
1532 of any of the test patterns, which is the same behaviour as GNU grep.
1533
1534 7. A pattern that could match an empty string could cause pcregrep to loop; it
1535 doesn't make sense to accept an empty string match in pcregrep, so I have
1536 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1537 seems to be how GNU grep behaves.
1538
1539 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1540 start or after a newline", because the conditional assertion was not being
1541 correctly handled. The rule now is that both the assertion and what follows
1542 in the first alternative must satisfy the test.
1543
1544 9. If auto-callout was enabled in a pattern with a conditional group whose
1545 condition was an assertion, PCRE could crash during matching, both with
1546 pcre_exec() and pcre_dfa_exec().
1547
1548 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1549 used for matching.
1550
1551 11. Unicode property support in character classes was not working for
1552 characters (bytes) greater than 127 when not in UTF-8 mode.
1553
1554 12. Added the -M command line option to pcretest.
1555
1556 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1557
1558 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1559
1560 16. Added comments and documentation about mis-use of no_arg in the C++
1561 wrapper.
1562
1563 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1564 from Martin Jerabek that uses macro names for all relevant character and
1565 string constants.
1566
1567 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1568 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1569 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1570 these, but not everybody uses configure.
1571
1572 19. A conditional group that had only one branch was not being correctly
1573 recognized as an item that could match an empty string. This meant that an
1574 enclosing group might also not be so recognized, causing infinite looping
1575 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1576 with the subject "ab", where knowledge that the repeated group can match
1577 nothing is needed in order to break the loop.
1578
1579 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1580 exec(), but without supplying a callout function, matching went wrong.
1581
1582 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1583 leak if the size of the offset vector was greater than 30. When the vector
1584 is smaller, the saved offsets during recursion go onto a local stack
1585 vector, but for larger vectors malloc() is used. It was failing to free
1586 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1587 error, in fact).
1588
1589 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1590 heapframe that is used only when UTF-8 support is enabled. This caused no
1591 problem, but was untidy.
1592
1593 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1594 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1595 included within another project.
1596
1597 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1598 slightly modified by me:
1599
1600 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1601 not building pcregrep.
1602
1603 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1604 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1605
1606 25. Forward references, both numeric and by name, in patterns that made use of
1607 duplicate group numbers, could behave incorrectly or give incorrect errors,
1608 because when scanning forward to find the reference group, PCRE was not
1609 taking into account the duplicate group numbers. A pattern such as
1610 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1611
1612 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1613 the feature of a custom pointer more persuasive (as requested by a user).
1614
1615 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1616 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1617 pre-defined.
1618
1619 28. Added support for (*UTF8) at the start of a pattern.
1620
1621 29. Arrange for flags added by the "release type" setting in CMake to be shown
1622 in the configuration summary.
1623
1624
1625 Version 7.8 05-Sep-08
1626 ---------------------
1627
1628 1. Replaced UCP searching code with optimized version as implemented for Ad
1629 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1630 stage table and inline lookup instead of a function, giving speed ups of 2
1631 to 5 times on some simple patterns that I tested. Permission was given to
1632 distribute the MultiStage2.py script that generates the tables (it's not in
1633 the tarball, but is in the Subversion repository).
1634
1635 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1636 scripts.
1637
1638 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1639 a group with a zero qualifier. The result of the study could be incorrect,
1640 or the function might crash, depending on the pattern.
1641
1642 4. Caseless matching was not working for non-ASCII characters in back
1643 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1644 It now works when Unicode Property Support is available.
1645
1646 5. In pcretest, an escape such as \x{de} in the data was always generating
1647 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1648 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1649 truncation.
1650
1651 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1652
1653 7. Added two (int) casts to pcregrep when printing the difference of two
1654 pointers, in case they are 64-bit values.
1655
1656 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1657 test 2 if it fails.
1658
1659 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1660 and a #define of that name to empty if it is not externally set. This is to
1661 allow users of MSVC to set it if necessary.
1662
1663 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1664 the convenience functions in the pcre_get.c source file.
1665
1666 11. An option change at the start of a pattern that had top-level alternatives
1667 could cause overwriting and/or a crash. This command provoked a crash in
1668 some environments:
1669
1670 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1671
1672 This potential security problem was recorded as CVE-2008-2371.
1673
1674 12. For a pattern where the match had to start at the beginning or immediately
1675 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1676 pcre_dfa_exec() could read past the end of the passed subject if there was
1677 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1678 pcretest so that it places the subject at the end of its malloc-ed buffer.
1679
1680 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1681 exec() might read past the end of the data buffer in UTF-8 mode.
1682
1683 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1684 the data contained the byte 0x85 as part of a UTF-8 character within its
1685 first line. This applied both to normal and DFA matching.
1686
1687 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1688 /^[^d]*?$/8 failed to match "abc".
1689
1690 16. Added a missing copyright notice to pcrecpp_internal.h.
1691
1692 17. Make it more clear in the documentation that values returned from
1693 pcre_exec() in ovector are byte offsets, not character counts.
1694
1695 18. Tidied a few places to stop certain compilers from issuing warnings.
1696
1697 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1698 supplied by Stefan Weber. I made a further small update for 7.8 because
1699 there is a change of source arrangements: the pcre_searchfuncs.c module is
1700 replaced by pcre_ucd.c.
1701
1702
1703 Version 7.7 07-May-08
1704 ---------------------
1705
1706 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1707 a string to a long long, pretend we don't even have a long long." This is
1708 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1709
1710 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1711 pre-7.6 versions, which defined a global no_arg variable instead of putting
1712 it in the RE class. (See also #8 below.)
1713
1714 3. Remove a line of dead code, identified by coverity and reported by Nuno
1715 Lopes.
1716
1717 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1718
1719 (1) The include/exclude patterns were being applied to the whole pathnames
1720 of files, instead of just to the final components.
1721
1722 (2) If there was more than one level of directory, the subdirectories were
1723 skipped unless they satisfied the include/exclude conditions. This is
1724 inconsistent with GNU grep (and could even be seen as contrary to the
1725 pcregrep specification - which I improved to make it absolutely clear).
1726 The action now is always to scan all levels of directory, and just
1727 apply the include/exclude patterns to regular files.
1728
1729 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1730 --exclude_dir in the tests to avoid scanning .svn directories.
1731
1732 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1733 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1734 doesn't support NULs in patterns.
1735
1736 7. Added some missing "const"s to declarations of static tables in
1737 pcre_compile.c and pcre_dfa_exec.c.
1738
1739 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1740 caused by fix #2 above. (Subsequently also a second patch to fix the
1741 first patch. And a third patch - this was a messy problem.)
1742
1743 9. Applied Craig's patch to remove the use of push_back().
1744
1745 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1746 matching function regexec().
1747
1748 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1749 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1750 references. PCRE supports relative numbers with this syntax (I don't think
1751 Oniguruma does).
1752
1753 12. Previously, a group with a zero repeat such as (...){0} was completely
1754 omitted from the compiled regex. However, this means that if the group
1755 was called as a subroutine from elsewhere in the pattern, things went wrong
1756 (an internal error was given). Such groups are now left in the compiled
1757 pattern, with a new opcode that causes them to be skipped at execution
1758 time.
1759
1760 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1761 to the way PCRE behaves:
1762
1763 (a) A lone ] character is dis-allowed (Perl treats it as data).
1764
1765 (b) A back reference to an unmatched subpattern matches an empty string
1766 (Perl fails the current match path).
1767
1768 (c) A data ] in a character class must be notated as \] because if the
1769 first data character in a class is ], it defines an empty class. (In
1770 Perl it is not possible to have an empty class.) The empty class []
1771 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1772 The negative empty class [^] matches any one character, independently
1773 of the DOTALL setting.
1774
1775 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1776 non-existent subpattern following a character class starting with ']' and
1777 containing () gave an internal compiling error instead of "reference to
1778 non-existent subpattern". Fortunately, when the pattern did exist, the
1779 compiled code was correct. (When scanning forwards to check for the
1780 existencd of the subpattern, it was treating the data ']' as terminating
1781 the class, so got the count wrong. When actually compiling, the reference
1782 was subsequently set up correctly.)
1783
1784 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1785 it was being rejected as not supported by pcre_dfa_exec(), even though
1786 other assertions are supported. I have made pcre_dfa_exec() support
1787 (*FAIL).
1788
1789 16. The implementation of 13c above involved the invention of a new opcode,
1790 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1791 cannot be changed at match time, I realized I could make a small
1792 improvement to matching performance by compiling OP_ALLANY instead of
1793 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1794 on the OP_ANY path.
1795
1796 17. Compiling pcretest on Windows with readline support failed without the
1797 following two fixes: (1) Make the unistd.h include conditional on
1798 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1799
1800 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1801 ncurses library to be included for pcretest when ReadLine support is
1802 requested, but also to allow for it to be overridden. This patch came from
1803 Daniel Bergström.
1804
1805 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1806 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1807 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1808 spotting this.
1809
1810
1811 Version 7.6 28-Jan-08
1812 ---------------------
1813
1814 1. A character class containing a very large number of characters with
1815 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1816 overflow.
1817
1818 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1819 HAVE_LONG_LONG is not defined.
1820
1821 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1822 bring them up to date and include new features. This patch includes:
1823
1824 - Fixed PH's badly added libz and libbz2 support.
1825 - Fixed a problem with static linking.
1826 - Added pcredemo. [But later removed - see 7 below.]
1827 - Fixed dftables problem and added an option.
1828 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1829 HAVE_LONG_LONG.
1830 - Added readline support for pcretest.
1831 - Added an listing of the option settings after cmake has run.
1832
1833 4. A user submitted a patch to Makefile that makes it easy to create
1834 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1835 Makefile.am that cause it to include this special target, without
1836 affecting anything else. Note that the same mingw target plus all
1837 the other distribution libraries and programs are now supported
1838 when configuring with CMake (see 6 below) instead of with
1839 Configure/Make.
1840
1841 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1842 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1843 exported in the Windows port". It has not yet been confirmed that the patch
1844 solves the problem, but it does no harm.
1845
1846 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1847 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1848 with CMake, and also correct the comment about stack recursion.
1849
1850 7. Remove the automatic building of pcredemo from the ./configure system and
1851 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1852 of a program that users should build themselves after PCRE is installed, so
1853 building it automatically is not really right. What is more, it gave
1854 trouble in some build environments.
1855
1856 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1857
1858
1859 Version 7.5 10-Jan-08
1860 ---------------------
1861
1862 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1863 values in parens when parsing an RE using the C++ wrapper."
1864
1865 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1866 Characters greater than 255 were excluded from the class instead of being
1867 included.
1868
1869 3. The same bug as (2) above applied to negated POSIX classes such as
1870 [:^space:].
1871
1872 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1873 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1874 I have changed it.
1875
1876 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1877 first named subpattern) and a construct such as (?&a) would reference the
1878 first named subpattern whose name started with "a" (in other words, the
1879 length check was missing). Both these problems are fixed. "Subpattern name
1880 expected" is now given for (?&) (a zero-length name), and this patch also
1881 makes it give the same error for \k'' (previously it complained that that
1882 was a reference to a non-existent subpattern).
1883
1884 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1885 this is right because (?- can be followed by option settings as well as by
1886 digits. I have, however, made the messages clearer.
1887
1888 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1889 than the number used in the conditional) now cause a compile-time error.
1890 This is actually not compatible with Perl, which accepts such patterns, but
1891 treats the conditional as always being FALSE (as PCRE used to), but it
1892 seems to me that giving a diagnostic is better.
1893
1894 8. Change "alphameric" to the more common word "alphanumeric" in comments
1895 and messages.
1896
1897 9. Fix two occurrences of "backslash" in comments that should have been
1898 "backspace".
1899
1900 10. Remove two redundant lines of code that can never be obeyed (their function
1901 was moved elsewhere).
1902
1903 11. The program that makes PCRE's Unicode character property table had a bug
1904 which caused it to generate incorrect table entries for sequences of
1905 characters that have the same character type, but are in different scripts.
1906 It amalgamated them into a single range, with the script of the first of
1907 them. In other words, some characters were in the wrong script. There were
1908 thirteen such cases, affecting characters in the following ranges:
1909
1910 U+002b0 - U+002c1
1911 U+0060c - U+0060d
1912 U+0061e - U+00612
1913 U+0064b - U+0065e
1914 U+0074d - U+0076d
1915 U+01800 - U+01805
1916 U+01d00 - U+01d77
1917 U+01d9b - U+01dbf
1918 U+0200b - U+0200f
1919 U+030fc - U+030fe
1920 U+03260 - U+0327f
1921 U+0fb46 - U+0fbb1
1922 U+10450 - U+1049d
1923
1924 12. The -o option (show only the matching part of a line) for pcregrep was not
1925 compatible with GNU grep in that, if there was more than one match in a
1926 line, it showed only the first of them. It now behaves in the same way as
1927 GNU grep.
1928
1929 13. If the -o and -v options were combined for pcregrep, it printed a blank
1930 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1931 does the same. The return code can be used to tell if there were any
1932 non-matching lines.
1933
1934 14. Added --file-offsets and --line-offsets to pcregrep.
1935
1936 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1937 infinitely looping recursion. The bug was that positive lookaheads were not
1938 being skipped when checking for a possible empty match (negative lookaheads
1939 and both kinds of lookbehind were skipped).
1940
1941 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1942 inclusion of <windows.h> to before rather than after the definition of
1943 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1944
1945 17. Specifying a possessive quantifier with a specific limit for a Unicode
1946 character property caused pcre_compile() to compile bad code, which led at
1947 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1948 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1949 caused the error; without that there was no problem.
1950
1951 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1952
1953 19. Added --enable-pcretest-libreadline.
1954
1955 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1956 RE::GlobalReplace(). As a result, the number of replacements returned was
1957 double what it should be. I removed one of the increments, but Craig sent a
1958 later patch that removed the other one (the right fix) and added unit tests
1959 that check the return values (which was not done before).
1960
1961 21. Several CMake things:
1962
1963 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1964 the names libpcre and libpcreposix, not just pcre and pcreposix.
1965
1966 (2) The above change means that pcretest and pcregrep are now correctly
1967 linked with the newly-built libraries, not previously installed ones.
1968
1969 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1970
1971 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1972 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1973 UTF-8 newline character). The key issue is that the pattern starts .*;
1974 this means that the match must be either at the beginning, or after a
1975 newline. The bug was in the code for advancing after a failed match and
1976 checking that the new position followed a newline. It was not taking
1977 account of UTF-8 characters correctly.
1978
1979 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1980 character classes. PCRE was not treating the sequence [:...:] as a
1981 character class unless the ... were all letters. Perl, however, seems to
1982 allow any characters between [: and :], though of course it rejects as
1983 unknown any "names" that contain non-letters, because all the known class
1984 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1985 for example, whereas PCRE did not - it did not recognize a POSIX character
1986 class. This seemed a bit dangerous, so the code has been changed to be
1987 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1988 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1989 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1990 Perl does, and where it didn't before.
1991
1992 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1993 Windows environments %n is disabled by default.
1994
1995
1996 Version 7.4 21-Sep-07
1997 ---------------------
1998
1999 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2000 means that a class such as [\s] counted as "explicit reference to CR or
2001 LF". That isn't really right - the whole point of the change was to try to
2002 help when there was an actual mention of one of the two characters. So now
2003 the change happens only if \r or \n (or a literal CR or LF) character is
2004 encountered.
2005
2006 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2007 of both had grown to the point where there were only 3 bits left.
2008 Fortunately, there was spare space in the data structure, and so I have
2009 moved the internal flags into a new 16-bit field to free up more option
2010 bits.
2011
2012 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2013 but did not set the internal JCHANGED flag - either of these is enough to
2014 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2015 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2016 start sets both bits.
2017
2018 4. Added options (at build time, compile time, exec time) to change \R from
2019 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2020
2021 5. doc/pcresyntax.html was missing from the distribution.
2022
2023 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2024 compatibility, even though it is no longer used.
2025
2026 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2027 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2028 windows.h file is present (where different names are used). [This was
2029 reversed later after testing - see 16 below.]
2030
2031 8. Changed all #include <config.h> to #include "config.h". There were also
2032 some further <pcre.h> cases that I changed to "pcre.h".
2033
2034 9. When pcregrep was used with the --colour option, it missed the line ending
2035 sequence off the lines that it output.
2036
2037 10. It was pointed out to me that arrays of string pointers cause lots of
2038 relocations when a shared library is dynamically loaded. A technique of
2039 using a single long string with a table of offsets can drastically reduce
2040 these. I have refactored PCRE in four places to do this. The result is
2041 dramatic:
2042
2043 Originally: 290
2044 After changing UCP table: 187
2045 After changing error message table: 43
2046 After changing table of "verbs" 36
2047 After changing table of Posix names 22
2048
2049 Thanks to the folks working on Gregex for glib for this insight.
2050
2051 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2052 unicode-properties was also set.
2053
2054 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2055
2056 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2057 checked only for CRLF.
2058
2059 14. Added casts to pcretest.c to avoid compiler warnings.
2060
2061 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2062
2063 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2064 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2065 entirely. This removes changes made in 7 above.
2066
2067 17. The CMake files have been updated, and there is now more information about
2068 building with CMake in the NON-UNIX-USE document.
2069
2070
2071 Version 7.3 28-Aug-07
2072 ---------------------
2073
2074 1. In the rejigging of the build system that eventually resulted in 7.1, the
2075 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2076 brackets there is not right, since it causes compilers to look for an
2077 installed pcre.h, not the version that is in the source that is being
2078 compiled (which of course may be different). I have changed it back to:
2079
2080 #include "pcre.h"
2081
2082 I have a vague recollection that the change was concerned with compiling in
2083 different directories, but in the new build system, that is taken care of
2084 by the VPATH setting the Makefile.
2085
2086 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2087 when the subject happened to end in the byte 0x85 (e.g. if the last
2088 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2089 characters but of course it shouldn't be taken as a newline when it is part
2090 of another character. The bug was that, for an unlimited repeat of . in
2091 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2092 characters when looking for a newline.
2093
2094 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2095
2096 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2097 in debug output.
2098
2099 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2100 long printing in the pcrecpp unittest when running under MinGW.
2101
2102 6. ESC_K was left out of the EBCDIC table.
2103
2104 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2105 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2106 limit also applies to "virtual nesting" when a pattern is recursive, and in
2107 this case 1000 isn't so big. I have been able to remove this limit at the
2108 expense of backing off one optimization in certain circumstances. Normally,
2109 when pcre_exec() would call its internal match() function recursively and
2110 immediately return the result unconditionally, it uses a "tail recursion"
2111 feature to save stack. However, when a subpattern that can match an empty
2112 string has an unlimited repetition quantifier, it no longer makes this
2113 optimization. That gives it a stack frame in which to save the data for
2114 checking that an empty string has been matched. Previously this was taken
2115 from the 1000-entry workspace that had been reserved. So now there is no
2116 explicit limit, but more stack is used.
2117
2118 8. Applied Daniel's patches to solve problems with the import/export magic
2119 syntax that is required for Windows, and which was going wrong for the
2120 pcreposix and pcrecpp parts of the library. These were overlooked when this
2121 problem was solved for the main library.
2122
2123 9. There were some crude static tests to avoid integer overflow when computing
2124 the size of patterns that contain repeated groups with explicit upper
2125 limits. As the maximum quantifier is 65535, the maximum group length was
2126 set at 30,000 so that the product of these two numbers did not overflow a
2127 32-bit integer. However, it turns out that people want to use groups that
2128 are longer than 30,000 bytes (though not repeat them that many times).
2129 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2130 made it possible to implement the integer overflow checks in a much more
2131 dynamic way, which I have now done. The artificial limitation on group
2132 length has been removed - we now have only the limit on the total length of
2133 the compiled pattern, which depends on the LINK_SIZE setting.
2134
2135 10. Fixed a bug in the documentation for get/copy named substring when
2136 duplicate names are permitted. If none of the named substrings are set, the
2137 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2138 empty string.
2139
2140 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2141 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2142 because the ] is interpreted as the first data character and the
2143 terminating ] is not found. PCRE has been made compatible with Perl in this
2144 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2145 cause memory overwriting.
2146
2147 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2148 string has been matched (to stop an infinite loop). It was not recognizing
2149 a conditional subpattern that could match an empty string if that
2150 subpattern was within another subpattern. For example, it looped when
2151 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2152 condition was not nested. This bug has been fixed.
2153
2154 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2155 past the start of the subject in the presence of bytes with the top bit
2156 set, for example "\x8aBCD".
2157
2158 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2159 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2160
2161 14. Optimized (?!) to (*FAIL).
2162
2163 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2164 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2165 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2166 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2167 does: it's just the validity check that is more restrictive.
2168
2169 16. Inserted checks for integer overflows during escape sequence (backslash)
2170 processing, and also fixed erroneous offset values for syntax errors during
2171 backslash processing.
2172
2173 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2174 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2175
2176 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2177 caused an overrun.
2178
2179 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2180 something other than just ASCII characters) inside a group that had an
2181 unlimited repeat caused a loop at compile time (while checking to see
2182 whether the group could match an empty string).
2183
2184 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2185 [\P{Any}] did so. (Error in the code for printing property names.)
2186
2187 21. An orphan \E inside a character class could cause a crash.
2188
2189 22. A repeated capturing bracket such as (A)? could cause a wild memory
2190 reference during compilation.
2191
2192 23. There are several functions in pcre_compile() that scan along a compiled
2193 expression for various reasons (e.g. to see if it's fixed length for look
2194 behind). There were bugs in these functions when a repeated \p or \P was
2195 present in the pattern. These operators have additional parameters compared
2196 with \d, etc, and these were not being taken into account when moving along
2197 the compiled data. Specifically:
2198
2199 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2200 length.
2201
2202 (b) An item such as \pL+ within a repeated group could cause crashes or
2203 loops.
2204
2205 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2206 "reference to non-existent subpattern" error.
2207
2208 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2209
2210 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2211 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2212
2213 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2214
2215 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2216 character were causing crashes (broken optimization).
2217
2218 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2219 \p or \P) caused a compile-time loop.
2220
2221 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2222 break. For example, the unstudied pattern [\r\n]A does not match the string
2223 "\r\nA" because change 7.0/46 below moves the current point on by two
2224 characters after failing to match at the start. However, the pattern \nA
2225 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2226 the same is true. There doesn't seem any very clean way out of this, but
2227 what I have chosen to do makes the common cases work: PCRE now takes note
2228 of whether there can be an explicit match for \r or \n anywhere in the
2229 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2230 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2231 pattern has explicit CR or LF references.
2232
2233 29. Added (*CR) etc for changing newline setting at start of pattern.
2234
2235
2236 Version 7.2 19-Jun-07
2237 ---------------------
2238
2239 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2240 which is apparently normally available under Windows.
2241
2242 2. Re-jig the pcregrep tests with different newline settings in an attempt
2243 to make them independent of the local environment's newline setting.
2244
2245 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2246
2247 4. Some of the "internals" tests were previously cut out when the link size
2248 was not 2, because the output contained actual offsets. The recent new
2249 "Z" feature of pcretest means that these can be cut out, making the tests
2250 usable with all link sizes.
2251
2252 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2253 stack recursion. This gives a massive performance boost under BSD, but just
2254 a small improvement under Linux. However, it saves one field in the frame
2255 in all cases.
2256
2257 6. Added more features from the forthcoming Perl 5.10:
2258
2259 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2260 recursion call. It refers to the nth most recently opened parentheses.
2261
2262 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2263 to be opened parentheses.
2264
2265 (c) Conditions that refer to capturing parentheses can be specified
2266 relatively, for example, (?(-2)... or (?(+3)...
2267
2268 (d) \K resets the start of the current match so that everything before
2269 is not part of it.
2270
2271 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2272
2273 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2274 reference syntax.
2275
2276 (g) (?| introduces a group in which the numbering of parentheses in each
2277 alternative starts with the same number.
2278
2279 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2280
2281 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2282 PCRE_INFO_JCHANGED.
2283
2284 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2285 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2286 for detecting groups that can match an empty string.
2287
2288 9. A pattern with a very large number of alternatives (more than several
2289 hundred) was running out of internal workspace during the pre-compile
2290 phase, where pcre_compile() figures out how much memory will be needed. A
2291 bit of new cunning has reduced the workspace needed for groups with
2292 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2293 workspace instead of running out of the 4096 that are available.
2294
2295 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2296
2297 11. Applied patch from Google to remove an optimization that didn't quite work.
2298 The report of the bug said:
2299
2300 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2301 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2302 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2303
2304 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2305 it matched the wrong number of bytes.
2306
2307
2308 Version 7.1 24-Apr-07
2309 ---------------------
2310
2311 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2312 that is more "standard", making use of automake and other Autotools. There
2313 is some re-arrangement of the files and adjustment of comments consequent
2314 on this.
2315
2316 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2317 for recursive directory scanning broke on some systems because the files
2318 are not scanned in any specific order and on different systems the order
2319 was different. A call to "sort" has been inserted into RunGrepTest for the
2320 approprate test as a short-term fix. In the longer term there may be an
2321 alternative.
2322
2323 3. I had an email from Eric Raymond about problems translating some of PCRE's
2324 man pages to HTML (despite the fact that I distribute HTML pages, some
2325 people do their own conversions for various reasons). The problems
2326 concerned the use of low-level troff macros .br and .in. I have therefore
2327 removed all such uses from the man pages (some were redundant, some could
2328 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2329 HTML has been updated to handle .nf/.fi and to complain if it encounters
2330 .br or .in.
2331
2332 4. Updated comments in configure.ac that get placed in config.h.in and also
2333 arranged for config.h to be included in the distribution, with the name
2334 config.h.generic, for the benefit of those who have to compile without
2335 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2336
2337 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2338 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2339 makevp.bat for the current PCRE, using the additional files
2340 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2341
2342 6. A Windows user reported a minor discrepancy with test 2, which turned out
2343 to be caused by a trailing space on an input line that had got lost in his
2344 copy. The trailing space was an accident, so I've just removed it.
2345
2346 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2347 that is needed.
2348
2349 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2350 as "const" (a) because they are and (b) because it helps the PHP
2351 maintainers who have recently made a script to detect big data structures
2352 in the php code that should be moved to the .rodata section. I remembered
2353 to update Builducptable as well, so it won't revert if ucptable.h is ever
2354 re-created.
2355
2356 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2357 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2358 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2359 support is not required. This saves 1.5-2K of code, which is important in
2360 some applications.
2361
2362 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2363 so as not to refer to the tables, even though these functions will never be
2364 called when UTF-8 support is disabled. Otherwise there are problems with a
2365 shared library.
2366
2367 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2368
2369 (a) It was defining its arguments as char * instead of void *.
2370
2371 (b) It was assuming that all moves were upwards in memory; this was true
2372 a long time ago when I wrote it, but is no longer the case.
2373
2374 The emulated memove() is provided for those environments that have neither
2375 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2376 is clearly not the case, as these two bugs were recently reported.
2377
2378 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2379 and Detrail to create the HTML documentation, the .txt form of the man
2380 pages, and it removes trailing spaces from listed files. It also creates
2381 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2382 case, it wraps all the #defines with #ifndefs. This script should be run
2383 before "make dist".
2384
2385 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2386 with Unicode property support.
2387
2388 (a) For a maximizing quantifier, if the two different cases of the
2389 character were of different lengths in their UTF-8 codings (there are
2390 some cases like this - I found 11), and the matching function had to
2391 back up over a mixture of the two cases, it incorrectly assumed they
2392 were both the same length.
2393
2394 (b) When PCRE was configured to use the heap rather than the stack for
2395 recursion during matching, it was not correctly preserving the data for
2396 the other case of a UTF-8 character when checking ahead for a match
2397 while processing a minimizing repeat. If the check also involved
2398 matching a wide character, but failed, corruption could cause an
2399 erroneous result when trying to check for a repeat of the original
2400 character.
2401
2402 13. Some tidying changes to the testing mechanism:
2403
2404 (a) The RunTest script now detects the internal link size and whether there
2405 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2406 values substituted by "configure". (The RunGrepTest script already did
2407 this for UTF-8.) The configure.ac script no longer substitutes the
2408 relevant variables.
2409
2410 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2411 with length and offset values. This means that the output is different
2412 for different internal link sizes. Test 2 is skipped for link sizes
2413 other than 2 because of this, bypassing the problem. Unfortunately,
2414 there was also a test in test 3 (the locale tests) that used /B and
2415 failed for link sizes other than 2. Rather than cut the whole test out,
2416 I have added a new /Z option to pcretest that replaces the length and
2417 offset values with spaces. This is now used to make test 3 independent
2418 of link size. (Test 2 will be tidied up later.)
2419
2420 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2421 segmentation fault instead of returning the appropriate error message.
2422
2423 15. In multiline mode when the newline sequence was set to "any", the pattern
2424 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2425 This doesn't seem right; it now treats the CRLF combination as the line
2426 ending, and so does not match in that case. It's only a pattern such as ^$
2427 that would hit this one: something like ^ABC$ would have failed after \r
2428 and then tried again after \r\n.
2429
2430 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2431 in an attempt to make files that differ only in their line terminators
2432 compare equal. This works on Linux.
2433
2434 17. Under certain error circumstances pcregrep might try to free random memory
2435 as it exited. This is now fixed, thanks to valgrind.
2436
2437 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2438 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2439 was because its rules for how to advance for /g after matching an empty
2440 string at the end of a line did not allow for this case. They now check for
2441 it specially.
2442
2443 20. pcretest is supposed to handle patterns and data of any length, by
2444 extending its buffers when necessary. It was getting this wrong when the
2445 buffer for a data line had to be extended.
2446
2447 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2448 CRLF as a newline sequence.
2449
2450 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2451 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2452 I have nevertheless tidied it up.
2453
2454 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2455
2456 24. Added a man page for pcre-config.
2457
2458
2459 Version 7.0 19-Dec-06
2460 ---------------------
2461
2462 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2463 moving to gcc 4.1.1.
2464
2465 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2466 sys/time.h, which is documented as needed for this function. It doesn't
2467 seem to matter on Linux, but it showed up on some releases of OS X.
2468
2469 3. It seems that there are systems where bytes whose values are greater than
2470 127 match isprint() in the "C" locale. The "C" locale should be the
2471 default when a C program starts up. In most systems, only ASCII printing
2472 characters match isprint(). This difference caused the output from pcretest
2473 to vary, making some of the tests fail. I have changed pcretest so that:
2474
2475 (a) When it is outputting text in the compiled version of a pattern, bytes
2476 other than 32-126 are always shown as hex escapes.
2477
2478 (b) When it is outputting text that is a matched part of a subject string,
2479 it does the same, unless a different locale has been set for the match
2480 (using the /L modifier). In this case, it uses isprint() to decide.
2481
2482 4. Fixed a major bug that caused incorrect computation of the amount of memory
2483 required for a compiled pattern when options that changed within the
2484 pattern affected the logic of the preliminary scan that determines the
2485 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2486 that the computed length was too small. The symptoms of this bug were
2487 either the PCRE error "internal error: code overflow" from pcre_compile(),
2488 or a glibc crash with a message such as "pcretest: free(): invalid next
2489 size (fast)". Examples of patterns that provoked this bug (shown in
2490 pcretest format) are:
2491
2492 /(?-x: )/x
2493 /(?x)(?-x: \s*#\s*)/
2494 /((?i)[\x{c0}])/8
2495 /(?i:[\x{c0}])/8
2496
2497 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2498 is now done differently.
2499
2500 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2501 wrapper classes; (b) implement a new function in the C++ scanner that is
2502 more efficient than the old way of doing things because it avoids levels of
2503 recursion in the regex matching; (c) add a paragraph to the documentation
2504 for the FullMatch() function.
2505
2506 6. The escape sequence \n was being treated as whatever was defined as
2507 "newline". Not only was this contrary to the documentation, which states
2508 that \n is character 10 (hex 0A), but it also went horribly wrong when
2509 "newline" was defined as CRLF. This has been fixed.
2510
2511 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2512 was being set to -1 for the "end of line" case (supposedly a value that no
2513 character can have). Though this value is never used (the check for end of
2514 line is "zero bytes in current character"), it caused compiler complaints.
2515 I've changed it to 0xffffffff.
2516
2517 8. In pcre_version.c, the version string was being built by a sequence of
2518 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2519 string (as it is for production releases) called a macro with an empty
2520 argument. The C standard says the result of this is undefined. The gcc
2521 compiler treats it as an empty string (which was what was wanted) but it is
2522 reported that Visual C gives an error. The source has been hacked around to
2523 avoid this problem.
2524
2525 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2526 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2527 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2528 of them did).
2529
2530 10. Originally, pcretest opened its input and output without "b"; then I was
2531 told that "b" was needed in some environments, so it was added for release
2532 5.0 to both the input and output. (It makes no difference on Unix-like
2533 systems.) Later I was told that it is wrong for the input on Windows. I've
2534 now abstracted the modes into two macros, to make it easier to fiddle with
2535 them, and removed "b" from the input mode under Windows.
2536
2537 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2538
2539 12. Added -help and --help to pcretest as an official way of being reminded
2540 of the options.
2541
2542 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2543 and pcrecpp.cc because they annoy compilers at high warning levels.
2544
2545 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2546
2547 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2548 scripts are not C programs :-) and which was not noticed because it works
2549 on Linux.
2550
2551 16. pcretest is supposed to handle any length of pattern and data line (as one
2552 line or as a continued sequence of lines) by extending its input buffer if
2553 necessary. This feature was broken for very long pattern lines, leading to
2554 a string of junk being passed to pcre_compile() if the pattern was longer
2555 than about 50K.
2556
2557 17. I have done a major re-factoring of the way pcre_compile() computes the
2558 amount of memory needed for a compiled pattern. Previously, there was code
2559 that made a preliminary scan of the pattern in order to do this. That was
2560 OK when PCRE was new, but as the facilities have expanded, it has become
2561 harder and harder to keep it in step with the real compile phase, and there
2562 have been a number of bugs (see for example, 4 above). I have now found a
2563 cunning way of running the real compile function in a "fake" mode that
2564 enables it to compute how much memory it would need, while actually only
2565 ever using a few hundred bytes of working memory and without too many
2566 tests of the mode. This should make future maintenance and development
2567 easier. A side effect of this work is that the limit of 200 on the nesting
2568 depth of parentheses has been removed (though this was never a serious
2569 limitation, I suspect). However, there is a downside: pcre_compile() now
2570 runs more slowly than before (30% or more, depending on the pattern). I
2571 hope this isn't a big issue. There is no effect on runtime performance.
2572
2573 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2574 newline (only possible for the last line of a file) and it was a
2575 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2576
2577 19. Added additional timing features to pcretest. (1) The -tm option now times
2578 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2579 separate command line item, by a number that specifies the number of
2580 repeats to use when timing. The default is 50000; this gives better
2581 precision, but takes uncomfortably long for very large patterns.
2582
2583 20. Extended pcre_study() to be more clever in cases where a branch of a
2584 subpattern has no definite first character. For example, (a*|b*)[cd] would
2585 previously give no result from pcre_study(). Now it recognizes that the
2586 first character must be a, b, c, or d.
2587
2588 21. There was an incorrect error "recursive call could loop indefinitely" if
2589 a subpattern (or the entire pattern) that was being tested for matching an
2590 empty string contained only one non-empty item after a nested subpattern.
2591 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2592 incorrectly, because the \d was being skipped in the check.
2593
2594 22. The pcretest program now has a new pattern option /B and a command line
2595 option -b, which is equivalent to adding /B to every pattern. This causes
2596 it to show the compiled bytecode, without the additional information that
2597 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2598 is the same as /B/I).
2599
2600 23. A new optimization is now able automatically to treat some sequences such
2601 as a*b as a*+b. More specifically, if something simple (such as a character
2602 or a simple class like \d) has an unlimited quantifier, and is followed by
2603 something that cannot possibly match the quantified thing, the quantifier
2604 is automatically "possessified".
2605
2606 24. A recursive reference to a subpattern whose number was greater than 39
2607 went wrong under certain circumstances in UTF-8 mode. This bug could also
2608 have affected the operation of pcre_study().
2609
2610 25. Realized that a little bit of performance could be had by replacing
2611 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2612
2613 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2614
2615 27. Possessive quantifiers such as a++ were previously implemented by turning
2616 them into atomic groups such as ($>a+). Now they have their own opcodes,
2617 which improves performance. This includes the automatically created ones
2618 from 23 above.
2619
2620 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2621 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2622 the first matched character to be a colon. This applied both to named and
2623 numbered groups.
2624
2625 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2626
2627 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2628 building PCRE on Windows easier, so I have included it in the distribution.
2629
2630 31. There is now a check in pcretest against a ridiculously large number being
2631 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2632 loop, the loop is abandoned.
2633
2634 32. Forward references to subpatterns in conditions such as (?(2)...) where
2635 subpattern 2 is defined later cause pcre_compile() to search forwards in
2636 the pattern for the relevant set of parentheses. This search went wrong
2637 when there were unescaped parentheses in a character class, parentheses
2638 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2639
2640 33. "Subroutine" calls and backreferences were previously restricted to
2641 referencing subpatterns earlier in the regex. This restriction has now
2642 been removed.
2643
2644 34. Added a number of extra features that are going to be in Perl 5.10. On the
2645 whole, these are just syntactic alternatives for features that PCRE had
2646 previously implemented using the Python syntax or my own invention. The
2647 other formats are all retained for compatibility.
2648
2649 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2650 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2651 also .NET compatible.
2652
2653 (b) A recursion or subroutine call to a named group can now be defined as
2654 (?&name) as well as (?P>name).
2655
2656 (c) A backreference to a named group can now be defined as \k<name> or
2657 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2658 5.10, are also .NET compatible.
2659
2660 (d) A conditional reference to a named group can now use the syntax
2661 (?(<name>) or (?('name') as well as (?(name).
2662
2663 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2664 groups (named and numbered) that are never evaluated inline, but can be
2665 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2666 is always false. There may be only one alternative in such a group.
2667
2668 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2669 as the simple (?(R). The condition is true only if the most recent
2670 recursion is that of the given number or name. It does not search out
2671 through the entire recursion stack.
2672
2673 (g) The escape \gN or \g{N} has been added, where N is a positive or
2674 negative number, specifying an absolute or relative reference.
2675
2676 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2677 some "unreachable code" warnings.
2678
2679 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2680 things, this adds five new scripts.
2681
2682 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2683 There were also incompatibilities regarding the handling of \Q..\E inside
2684 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2685 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2686
2687 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2688 matches an empty string, and forcibly breaks the loop. There were bugs in
2689 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2690 against aaaa the result was just "a" rather than "aaaa", for example. Two
2691 separate and independent bugs (that affected different cases) have been
2692 fixed.
2693
2694 39. Refactored the code to abolish the use of different opcodes for small
2695 capturing bracket numbers. This is a tidy that I avoided doing when I
2696 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2697 The new approach is not only tidier, it makes it possible to reduce the
2698 memory needed to fix the previous bug (38).
2699
2700 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2701 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2702 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2703 mode.
2704
2705 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2706 report.
2707
2708 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2709 copy construction and assignment in the C++ wrapper.
2710
2711 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2712 couple of bugs that could have given wrong results in the "--newline=crlf"
2713 case.
2714
2715 44. Added a number of casts and did some reorganization of signed/unsigned int
2716 variables following suggestions from Dair Grant. Also renamed the variable
2717 "this" as "item" because it is a C++ keyword.
2718
2719 45. Arranged for dftables to add
2720
2721 #include "pcre_internal.h"
2722
2723 to pcre_chartables.c because without it, gcc 4.x may remove the array
2724 definition from the final binary if PCRE is built into a static library and
2725 dead code stripping is activated.
2726
2727 46. For an unanchored pattern, if a match attempt fails at the start of a
2728 newline sequence, and the newline setting is CRLF or ANY, and the next two
2729 characters are CRLF, advance by two characters instead of one.
2730
2731
2732 Version 6.7 04-Jul-06
2733 ---------------------
2734
2735 1. In order to handle tests when input lines are enormously long, pcretest has
2736 been re-factored so that it automatically extends its buffers when
2737 necessary. The code is crude, but this _is_ just a test program. The
2738 default size has been increased from 32K to 50K.
2739
2740 2. The code in pcre_study() was using the value of the re argument before
2741 testing it for NULL. (Of course, in any sensible call of the function, it
2742 won't be NULL.)
2743
2744 3. The memmove() emulation function in pcre_internal.h, which is used on
2745 systems that lack both memmove() and bcopy() - that is, hardly ever -
2746 was missing a "static" storage class specifier.
2747
2748 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2749 containing an extended class (one that cannot be represented by a bitmap
2750 because it contains high-valued characters or Unicode property items, e.g.
2751 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2752 pattern, but PCRE should not loop if you do not (it no longer does).
2753 [Detail: two cases were found: (a) a repeated subpattern containing an
2754 extended class; (b) a recursive reference to a subpattern that followed a
2755 previous extended class. It wasn't skipping over the extended class
2756 correctly when UTF-8 mode was not set.]
2757
2758 5. A negated single-character class was not being recognized as fixed-length
2759 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2760 compile error "lookbehind assertion is not fixed length".
2761
2762 6. The RunPerlTest auxiliary script was showing an unexpected difference
2763 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2764 write a Perl script that can interpret lines of an input file either as
2765 byte characters or as UTF-8, which is what "perltest" was being required to
2766 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2767 can't do is switch easily at run time between having the "use utf8;" pragma
2768 or not. In the end, I fudged it by using the RunPerlTest script to insert
2769 "use utf8;" explicitly for the UTF-8 tests.
2770
2771 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2772 the end of the subject string, contrary to the documentation and to what
2773 Perl does. This was true of both matching functions. Now it matches only at
2774 the start of the subject and immediately after *internal* newlines.
2775
2776 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2777 a pointer to an int instead of a pointer to an unsigned long int. This
2778 caused problems on 64-bit systems.
2779
2780 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2781 instance of the 'standard' template library not being so standard".
2782
2783 10. There was no check on the number of named subpatterns nor the maximum
2784 length of a subpattern name. The product of these values is used to compute
2785 the size of the memory block for a compiled pattern. By supplying a very
2786 long subpattern name and a large number of named subpatterns, the size
2787 computation could be caused to overflow. This is now prevented by limiting
2788 the length of names to 32 characters, and the number of named subpatterns
2789 to 10,000.
2790
2791 11. Subpatterns that are repeated with specific counts have to be replicated in
2792 the compiled pattern. The size of memory for this was computed from the
2793 length of the subpattern and the repeat count. The latter is limited to
2794 65535, but there was no limit on the former, meaning that integer overflow
2795 could in principle occur. The compiled length of a repeated subpattern is
2796 now limited to 30,000 bytes in order to prevent this.
2797
2798 12. Added the optional facility to have named substrings with the same name.
2799
2800 13. Added the ability to use a named substring as a condition, using the
2801 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2802 are numbers (not recommended). Forward references are permitted.
2803
2804 14. Added forward references in named backreferences (if you see what I mean).
2805
2806 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2807 pattern could run off the end of the subject. For example, the pattern
2808 "(?s)(.{1,5})"8 did this with the subject "ab".
2809
2810 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2811 PCRE_CASELESS was set when matching characters that were quantified with ?
2812 or *.
2813
2814 17. A character class other than a single negated character that had a minimum
2815 but no maximum quantifier - for example [ab]{6,} - was not handled
2816 correctly by pce_dfa_exec(). It would match only one character.
2817
2818 18. A valid (though odd) pattern that looked like a POSIX character
2819 class but used an invalid character after [ (for example [[,abc,]]) caused
2820 pcre_compile() to give the error "Failed: internal error: code overflow" or
2821 in some cases to crash with a glibc free() error. This could even happen if
2822 the pattern terminated after [[ but there just happened to be a sequence of
2823 letters, a binary zero, and a closing ] in the memory that followed.
2824
2825 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2826 over the years. Originally (before any Unicode support), just the bottom 8
2827 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2828 output from "man perlunicode" includes this:
2829
2830 The regular expression compiler produces polymorphic opcodes. That
2831 is, the pattern adapts to the data and automatically switches to
2832 the Unicode character scheme when presented with Unicode data--or
2833 instead uses a traditional byte scheme when presented with byte
2834 data.
2835
2836 Sadly, a wide octal escape does not cause a switch, and in a string with
2837 no other multibyte characters, these octal escapes are treated as before.
2838 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2839 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2840 Unicode string.
2841
2842 I have not perpetrated such confusion in PCRE. Up till now, it took just
2843 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2844 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2845 translate to the appropriate multibyte character.
2846
2847 29. Applied some refactoring to reduce the number of warnings from Microsoft
2848 and Borland compilers. This has included removing the fudge introduced
2849 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2850 a warning about an unused variable.
2851
2852 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2853 characters since release 4.0, because Perl (from release 5.004) does not.
2854 [Or at least, is documented not to: some releases seem to be in conflict
2855 with the documentation.] However, when a pattern was studied with
2856 pcre_study() and all its branches started with \s, PCRE still included VT
2857 as a possible starting character. Of course, this did no harm; it just
2858 caused an unnecessary match attempt.
2859
2860 22. Removed a now-redundant internal flag bit that recorded the fact that case
2861 dependency changed within the pattern. This was once needed for "required
2862 byte" processing, but is no longer used. This recovers a now-scarce options
2863 bit. Also moved the least significant internal flag bit to the most-
2864 significant bit of the word, which was not previously used (hangover from
2865 the days when it was an int rather than a uint) to free up another bit for
2866 the future.
2867
2868 23. Added support for CRLF line endings as well as CR and LF. As well as the
2869 default being selectable at build time, it can now be changed at runtime
2870 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2871 specify that it is scanning data with non-default line endings.
2872
2873 24. Changed the definition of CXXLINK to make it agree with the definition of
2874 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2875
2876 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2877 recursions. This makes a big different to stack usage for some patterns.
2878
2879 26. If a subpattern containing a named recursion or subroutine reference such
2880 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2881 the space required for the compiled pattern went wrong and gave too small a
2882 value. Depending on the environment, this could lead to "Failed: internal
2883 error: code overflow at offset 49" or "glibc detected double free or
2884 corruption" errors.
2885
2886 27. Applied patches from Google (a) to support the new newline modes and (b) to
2887 advance over multibyte UTF-8 characters in GlobalReplace.
2888
2889 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2890 difference for some implementation of PCRE in some Windows version.
2891
2892 29. Added some extra testing facilities to pcretest:
2893
2894 \q<number> in a data line sets the "match limit" value
2895 \Q<number> in a data line sets the "match recursion limt" value
2896 -S <number> sets the stack size, where <number> is in megabytes
2897
2898 The -S option isn't available for Windows.
2899
2900
2901 Version 6.6 06-Feb-06
2902 ---------------------
2903
2904 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2905 in pcreposix.h. I have copied the definition from pcre.h.
2906
2907 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2908 because pcre.h is no longer a built file.
2909
2910 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2911 not normally included in the compiled code.
2912
2913
2914 Version 6.5 01-Feb-06
2915 ---------------------
2916
2917 1. When using the partial match feature with pcre_dfa_exec(), it was not
2918 anchoring the second and subsequent partial matches at the new starting
2919 point. This could lead to incorrect results. For example, with the pattern
2920 /1234/, partially matching against "123" and then "a4" gave a match.
2921
2922 2. Changes to pcregrep:
2923
2924 (a) All non-match returns from pcre_exec() were being treated as failures
2925 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2926 error message is output. Some extra information is given for the
2927 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2928 probably the only errors that are likely to be caused by users (by
2929 specifying a regex that has nested indefinite repeats, for instance).
2930 If there are more than 20 of these errors, pcregrep is abandoned.
2931
2932 (b) A binary zero was treated as data while matching, but terminated the
2933 output line if it was written out. This has been fixed: binary zeroes
2934 are now no different to any other data bytes.
2935
2936 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2937 used to set a locale for matching. The --locale=xxxx long option has
2938 been added (no short equivalent) to specify a locale explicitly on the
2939 pcregrep command, overriding the environment variables.
2940
2941 (d) When -B was used with -n, some line numbers in the output were one less
2942 than they should have been.
2943
2944 (e) Added the -o (--only-matching) option.
2945
2946 (f) If -A or -C was used with -c (count only), some lines of context were
2947 accidentally printed for the final match.
2948
2949 (g) Added the -H (--with-filename) option.
2950
2951 (h) The combination of options -rh failed to suppress file names for files
2952 that were found from directory arguments.
2953
2954 (i) Added the -D (--devices) and -d (--directories) options.
2955
2956 (j) Added the -F (--fixed-strings) option.
2957
2958 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2959
2960 (l) Added the --colo(u)r option.
2961
2962 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2963 is not present by default.
2964
2965 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2966 items such as (?R) or (?1), when the recursion could match a number of
2967 alternatives. If it matched one of the alternatives, but subsequently,
2968 outside the recursion, there was a failure, the code tried to back up into
2969 the recursion. However, because of the way PCRE is implemented, this is not
2970 possible, and the result was an incorrect result from the match.
2971
2972 In order to prevent this happening, the specification of recursion has
2973 been changed so that all such subpatterns are automatically treated as
2974 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2975
2976 4. I had overlooked the fact that, in some locales, there are characters for
2977 which isalpha() is true but neither isupper() nor islower() are true. In
2978 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2979 and ordfeminine) are like this. This affected the treatment of \w and \W
2980 when they appeared in character classes, but not when they appeared outside
2981 a character class. The bit map for "word" characters is now created
2982 separately from the results of isalnum() instead of just taking it from the
2983 upper, lower, and digit maps. (Plus the underscore character, of course.)
2984
2985 5. The above bug also affected the handling of POSIX character classes such as
2986 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2987 permanent tables. Instead, the bit maps for such a class were previously
2988 created as the appropriate unions of the upper, lower, and digit bitmaps.
2989 Now they are created by subtraction from the [[:word:]] class, which has
2990 its own bitmap.
2991
2992 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2993 It is created by subtracting the vertical space characters (\x09, \x0a,
2994 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2995 subtraction was done in the overall bitmap for a character class, meaning
2996 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2997 be recognized. This bug has been fixed.
2998
2999 7. Patches from the folks at Google:
3000
3001 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3002 real life, but is still worth protecting against".
3003
3004 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3005 regular expressions".
3006
3007 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3008 have it.
3009
3010 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3011 "configure" and the latter not, in order to fix a problem somebody had
3012 with compiling the Arg class on HP-UX.
3013
3014 (e) Improve the error-handling of the C++ wrapper a little bit.
3015
3016 (f) New tests for checking recursion limiting.
3017
3018 8. The pcre_memmove() function, which is used only if the environment does not
3019 have a standard memmove() function (and is therefore rarely compiled),
3020 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3021 returning a result (though PCRE never actually uses the result).
3022
3023 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3024 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3025 returned instead of calling malloc() with an overflowing number that would
3026 most likely cause subsequent chaos.
3027
3028 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3029
3030 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3031 with this option is matched, the nmatch and pmatch options of regexec() are
3032 ignored.
3033
3034 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3035 provided in case anyone wants to the the POSIX interface with UTF-8
3036 strings.
3037
3038 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3039 C++ linking (needed for some HP-UX environments).
3040
3041 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3042 (unused parameter) and in the pcre_printint() function (omitted "default"
3043 switch label when the default is to do nothing).
3044
3045 15. Added some code to make it possible, when PCRE is compiled as a C++
3046 library, to replace subject pointers for pcre_exec() with a smart pointer
3047 class, thus making it possible to process discontinuous strings.
3048
3049 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3050 much the same function. They were added by different people who were trying
3051 to make PCRE easy to compile on non-Unix systems. It has been suggested
3052 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3053 for compiling on Windows systems. I have therefore replaced it with
3054 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3055 defaults to "extern" for C or "extern C" for C++, which works fine on
3056 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3057 SCOPE with something explicit in config.h. In addition:
3058
3059 (a) pcreposix.h still had just "extern" instead of either of these macros;
3060 I have replaced it with PCRE_DATA_SCOPE.
3061
3062 (b) Functions such as _pcre_xclass(), which are internal to the library,
3063 but external in the C sense, all had PCRE_EXPORT in their definitions.
3064 This is apparently wrong for the Windows case, so I have removed it.
3065 (It makes no difference on Unix-like systems.)
3066
3067 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3068 of recursive calls to match(). This is different to MATCH_LIMIT because
3069 that limits the total number of calls to match(), not all of which increase
3070 the depth of recursion. Limiting the recursion depth limits the amount of
3071 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3072 when PCRE is compiled, and changed at run time. A patch from Google adds
3073 this functionality to the C++ interface.
3074
3075 18. Changes to the handling of Unicode character properties:
3076
3077 (a) Updated the table to Unicode 4.1.0.
3078
3079 (b) Recognize characters that are not in the table as "Cn" (undefined).
3080
3081 (c) I revised the way the table is implemented to a much improved format
3082 which includes recognition of ranges. It now supports the ranges that
3083 are defined in UnicodeData.txt, and it also amalgamates other
3084 characters into ranges. This has reduced the number of entries in the
3085 table from around 16,000 to around 3,000, thus reducing its size
3086 considerably. I realized I did not need to use a tree structure after
3087 all - a binary chop search is just as efficient. Having reduced the
3088 number of entries, I extended their size from 6 bytes to 8 bytes to
3089 allow for more data.
3090
3091 (d) Added support for Unicode script names via properties such as \p{Han}.
3092
3093 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3094 matching that character.
3095
3096 20. When matching a repeated Unicode property with a minimum greater than zero,
3097 (for example \pL{2,}), PCRE could look past the end of the subject if it
3098 reached it while seeking the minimum number of characters. This could
3099 happen only if some of the characters were more than one byte long, because
3100 there is a check for at least the minimum number of bytes.
3101
3102 21. Refactored the implementation of \p and \P so as to be more general, to
3103 allow for more different types of property in future. This has changed the
3104 compiled form incompatibly. Anybody with saved compiled patterns that use
3105 \p or \P will have to recompile them.
3106
3107 22. Added "Any" and "L&" to the supported property types.
3108
3109 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3110 but give a compile time error if the value is greater than 0xff.
3111
3112 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3113 accidentally not being installed or uninstalled.
3114
3115 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3116 made were to insert the current release number. This seemed silly, because
3117 it made things harder for people building PCRE on systems that don't run
3118 "configure". I have turned pcre.h into a distributed file, no longer built
3119 by "configure", with the version identification directly included. There is
3120 no longer a pcre.h.in file.
3121
3122 However, this change necessitated a change to the pcre-config script as
3123 well. It is built from pcre-config.in, and one of the substitutions was the
3124 release number. I have updated configure.ac so that ./configure now finds
3125 the release number by grepping pcre.h.
3126
3127 26. Added the ability to run the tests under valgrind.
3128
3129
3130 Version 6.4 05-Sep-05
3131 ---------------------
3132
3133 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3134 "--" to be printed when multiple files were scanned, even when none of the
3135 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3136 consider it to be a bug, and have restored the previous behaviour.
3137
3138 2. A couple of code tidies to get rid of compiler warnings.
3139
3140 3. The pcretest program used to cheat by referring to symbols in the library
3141 whose names begin with _pcre_. These are internal symbols that are not
3142 really supposed to be visible externally, and in some environments it is
3143 possible to suppress them. The cheating is now confined to including
3144 certain files from the library's source, which is a bit cleaner.
3145
3146 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3147 file's purpose clearer.
3148
3149 5. Reorganized pcre_ucp_findchar().
3150
3151
3152 Version 6.3 15-Aug-05
3153 ---------------------
3154
3155 1. The file libpcre.pc.in did not have general read permission in the tarball.
3156
3157 2. There were some problems when building without C++ support:
3158
3159 (a) If C++ support was not built, "make install" and "make test" still
3160 tried to test it.
3161
3162 (b) There were problems when the value of CXX was explicitly set. Some
3163 changes have been made to try to fix these, and ...
3164
3165 (c) --disable-cpp can now be used to explicitly disable C++ support.
3166
3167 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3168 backslash in a target when C++ was disabled. This confuses some
3169 versions of "make", apparently. Using an intermediate variable solves
3170 this. (Same for CPP_LOBJ.)
3171
3172 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3173 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3174 necessary on certain architectures.
3175
3176 4. Added a setting of -export-symbols-regex to the link command to remove
3177 those symbols that are exported in the C sense, but actually are local
3178 within the library, and not documented. Their names all begin with
3179 "_pcre_". This is not a perfect job, because (a) we have to except some
3180 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3181 available (and never for static libraries). I have made a note to try to
3182 find a way round (a) in the future.
3183
3184
3185 Version 6.2 01-Aug-05
3186 ---------------------
3187
3188 1. There was no test for integer overflow of quantifier values. A construction
3189 such as {1111111111111111} would give undefined results. What is worse, if
3190 a minimum quantifier for a parenthesized subpattern overflowed and became
3191 negative, the calculation of the memory size went wrong. This could have
3192 led to memory overwriting.
3193
3194 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3195
3196 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3197 operating environments where this matters.
3198
3199 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3200 PCRE options from within the C++ wrapper.
3201
3202 5. Named capturing subpatterns were not being correctly counted when a pattern
3203 was compiled. This caused two problems: (a) If there were more than 100
3204 such subpatterns, the calculation of the memory needed for the whole
3205 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3206 back references of the form \12, where the number was greater than 9, were
3207 not recognized as back references, even though there were sufficient
3208 previous subpatterns.
3209
3210 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3211 versions of gcc, e.g. 2.95.4.
3212
3213
3214 Version 6.1 21-Jun-05
3215 ---------------------
3216
3217 1. There was one reference to the variable "posix" in pcretest.c that was not
3218 surrounded by "#if !defined NOPOSIX".
3219
3220 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3221 the cross-check on the old pcre_info() function, for the benefit of the
3222 cut-down version of PCRE that is currently imported into Exim.
3223
3224 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3225 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3226 patterns that start (?i) but I don't think that matters. The use of (?i) is
3227 just an example; this all applies to the other options as well.
3228
3229 4. Since libtool seems to echo the compile commands it is issuing, the output
3230 from "make" can be reduced a bit by putting "@" in front of each libtool
3231 compile command.
3232
3233 5. Patch from the folks at Google for configure.in to be a bit more thorough
3234 in checking for a suitable C++ installation before trying to compile the
3235 C++ stuff. This should fix a reported problem when a compiler was present,
3236 but no suitable headers.
3237
3238 6. The man pages all had just "PCRE" as their title. I have changed them to
3239 be the relevant file name. I have also arranged that these names are
3240 retained in the file doc/pcre.txt, which is a concatenation in text format
3241 of all the man pages except the little individual ones for each function.
3242
3243 7. The NON-UNIX-USE file had not been updated for the different set of source
3244 files that come with release 6. I also added a few comments about the C++
3245 wrapper.
3246
3247
3248 Version 6.0 07-Jun-05
3249 ---------------------
3250
3251 1. Some minor internal re-organization to help with my DFA experiments.
3252
3253 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3254 didn't matter for the library itself when fully configured, but did matter
3255 when compiling without UCP support, or within Exim, where the ucp files are
3256 not imported.
3257
3258 3. Refactoring of the library code to split up the various functions into
3259 different source modules. The addition of the new DFA matching code (see
3260 below) to a single monolithic source would have made it really too
3261 unwieldy, quite apart from causing all the code to be include in a
3262 statically linked application, when only some functions are used. This is
3263 relevant even without the DFA addition now that patterns can be compiled in
3264 one application and matched in another.
3265
3266 The downside of splitting up is that there have to be some external
3267 functions and data tables that are used internally in different modules of
3268 the library but which are not part of the API. These have all had their
3269 names changed to start with "_pcre_" so that they are unlikely to clash
3270 with other external names.
3271
3272 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3273 a different (DFA) algorithm. Although it is slower than the original
3274 function, it does have some advantages for certain types of matching
3275 problem.
3276
3277 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3278 including restarting after a partial match.
3279
3280 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3281 defined when compiling for Windows was sent to me. I have put it into the
3282 code, though I have no means of testing or verifying it.
3283
3284 7. Added the pcre_refcount() auxiliary function.
3285
3286 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3287 match before or at the first newline in the subject string. In pcretest,
3288 the /f option on a pattern can be used to set this.
3289
3290 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3291 would behave wrongly. This has been present in PCRE since release 4.0.
3292
3293 10. A number of changes to the pcregrep command:
3294
3295 (a) Refactored how -x works; insert ^(...)$ instead of setting
3296 PCRE_ANCHORED and checking the length, in preparation for adding
3297 something similar for -w.
3298
3299 (b) Added the -w (match as a word) option.
3300
3301 (c) Refactored the way lines are read and buffered so as to have more
3302 than one at a time available.
3303
3304 (d) Implemented a pcregrep test script.
3305
3306 (e) Added the -M (multiline match) option. This allows patterns to match
3307 over several lines of the subject. The buffering ensures that at least
3308 8K, or the rest of the document (whichever is the shorter) is available
3309 for matching (and similarly the previous 8K for lookbehind assertions).
3310
3311 (f) Changed the --help output so that it now says
3312
3313 -w, --word-regex(p)
3314
3315 instead of two lines, one with "regex" and the other with "regexp"
3316 because that confused at least one person since the short forms are the
3317 same. (This required a bit of code, as the output is generated
3318 automatically from a table. It wasn't just a text change.)
3319
3320 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3321 option but starts with a hyphen. Could be a pattern or a path name
3322 starting with a hyphen, for instance.
3323
3324 (h) "-" can be given as a file name to represent stdin.
3325
3326 (i) When file names are being printed, "(standard input)" is used for
3327 the standard input, for compatibility with GNU grep. Previously
3328 "<stdin>" was used.
3329
3330 (j) The option --label=xxx can be used to supply a name to be used for
3331 stdin when file names are being printed. There is no short form.
3332
3333 (k) Re-factored the options decoding logic because we are going to add
3334 two more options that take data. Such options can now be given in four
3335 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3336
3337 (l) Added the -A, -B, and -C options for requesting that lines of context
3338 around matches be printed.
3339
3340 (m) Added the -L option to print the names of files that do not contain
3341 any matching lines, that is, the complement of -l.
3342
3343 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3344 continue to scan other files.
3345
3346 (o) The -s option was incorrectly implemented. For compatibility with other
3347 greps, it now suppresses the error message for a non-existent or non-
3348 accessible file (but not the return code). There is a new option called
3349 -q that suppresses the output of matching lines, which was what -s was
3350 previously doing.
3351
3352 (p) Added --include and --exclude options to specify files for inclusion
3353 and exclusion when recursing.
3354
3355 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3356 Hopefully, it now does.
3357
3358 12. Missing cast in pcre_study().
3359
3360 13. Added an "uninstall" target to the makefile.
3361
3362 14. Replaced "extern" in the function prototypes in Makefile.in with
3363 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3364 world, but is set differently for Windows.
3365
3366 15. Added a second compiling function called pcre_compile2(). The only
3367 difference is that it has an extra argument, which is a pointer to an
3368 integer error code. When there is a compile-time failure, this is set
3369 non-zero, in addition to the error test pointer being set to point to an
3370 error message. The new argument may be NULL if no error number is required
3371 (but then you may as well call pcre_compile(), which is now just a
3372 wrapper). This facility is provided because some applications need a
3373 numeric error indication, but it has also enabled me to tidy up the way
3374 compile-time errors are handled in the POSIX wrapper.
3375
3376 16. Added VPATH=.libs to the makefile; this should help when building with one
3377 prefix path and installing with another. (Or so I'm told by someone who
3378 knows more about this stuff than I do.)
3379
3380 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3381 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3382 match everything, including newlines. This is not POSIX-compatible, but
3383 somebody wanted the feature. From pcretest it can be activated by using
3384 both the P and the s flags.
3385
3386 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3387
3388 19. libpcre.pc was being incorrectly installed as executable.
3389
3390 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3391 it now also looks for '\r' so that it will work unmodified on Windows.
3392
3393 21. Added Google's contributed C++ wrapper to the distribution.
3394
3395 22. Added some untidy missing memory free() calls in pcretest, to keep
3396 Electric Fence happy when testing.
3397
3398
3399
3400 Version 5.0 13-Sep-04
3401 ---------------------
3402
3403 1. Internal change: literal characters are no longer packed up into items
3404 containing multiple characters in a single byte-string. Each character
3405 is now matched using a separate opcode. However, there may be more than one
3406 byte in the character in UTF-8 mode.
3407
3408 2. The pcre_callout_block structure has two new fields: pattern_position and
3409 next_item_length. These contain the offset in the pattern to the next match
3410 item, and its length, respectively.
3411
3412 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3413 insertion of callouts before each pattern item. Added the /C option to
3414 pcretest to make use of this.
3415
3416 4. On the advice of a Windows user, the lines
3417
3418 #if defined(_WIN32) || defined(WIN32)
3419 _setmode( _fileno( stdout ), 0x8000 );
3420 #endif /* defined(_WIN32) || defined(WIN32) */
3421
3422 have been added to the source of pcretest. This apparently does useful
3423 magic in relation to line terminators.
3424
3425 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3426 for the benefit of those environments where the "b" makes a difference.
3427
3428 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3429 to know about it. I have put a hack into configure.in that adds in code
3430 to set GCC=yes if CC=icc. This seems to end up at a point in the
3431 generated configure script that is early enough to affect the setting of
3432 compiler options, which is what is needed, but I have no means of testing
3433 whether it really works. (The user who reported this had patched the
3434 generated configure script, which of course I cannot do.)
3435
3436 LATER: After change 22 below (new libtool files), the configure script
3437 seems to know about icc (and also ecc). Therefore, I have commented out
3438 this hack in configure.in.
3439
3440 7. Added support for pkg-config (2 patches were sent in).
3441
3442 8. Negated POSIX character classes that used a combination of internal tables
3443 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3444 [[:^ascii]]. Typically, they would match almost any characters. The other
3445 POSIX classes were not broken in this way.
3446
3447 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3448 to find the match, as PCRE was deluded into thinking that the match had to
3449 start at the start point or following a newline. The same bug applied to
3450 patterns with negative forward assertions or any backward assertions
3451 preceding ".*" at the start, unless the pattern required a fixed first
3452 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3453
3454 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3455 starting at the last subject character, bytes beyond the end of the subject
3456 string were read.
3457
3458 11. Renamed the variable "class" as "classbits" to make life easier for C++
3459 users. (Previously there was a macro definition, but it apparently wasn't
3460 enough.)
3461
3462 12. Added the new field "tables" to the extra data so that tables can be passed
3463 in at exec time, or the internal tables can be re-selected. This allows
3464 a compiled regex to be saved and re-used at a later time by a different
3465 program that might have everything at different addresses.
3466
3467 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3468 -R library as well as a -L library.
3469
3470 14. The debugging options of pcretest (-d on the command line or D on a
3471 pattern) showed incorrect output for anything following an extended class
3472 that contained multibyte characters and which was followed by a quantifier.
3473
3474 15. Added optional support for general category Unicode character properties
3475 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3476 support. It adds about 90K to the size of the library. The meanings of the
3477 inbuilt class escapes such as \d and \s have NOT been changed.
3478
3479 16. Updated pcredemo.c to include calls to free() to release the memory for the
3480 compiled pattern.
3481
3482 17. The generated file chartables.c was being created in the source directory
3483 instead of in the building directory. This caused the build to fail if the
3484 source directory was different from the building directory, and was
3485 read-only.
3486
3487 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3488 file. No doubt somebody will tell me if they don't make sense... Also added
3489 Dan Mooney's comments about building on OpenVMS.
3490
3491 19. Added support for partial matching via the PCRE_PARTIAL option for
3492 pcre_exec() and the \P data escape in pcretest.
3493
3494 20. Extended pcretest with 3 new pattern features:
3495
3496 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3497 write the compiled pattern to the file whose name is "rest-of-line".
3498 This is a straight binary dump of the data, with the saved pointer to
3499 the character tables forced to be NULL. The study data, if any, is
3500 written too. After writing, pcretest reads a new pattern.
3501
3502 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3503 compiled pattern from the given file. There must not be any
3504 occurrences of "<" in the file name (pretty unlikely); if there are,
3505 pcretest will instead treat the initial "<" as a pattern delimiter.
3506 After reading in the pattern, pcretest goes on to read data lines as
3507 usual.
3508
3509 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3510 and 16-bit fields in a compiled pattern, to simulate a pattern that
3511 was compiled on a host of opposite endianness.
3512
3513 21. The pcre-exec() function can now cope with patterns that were compiled on
3514 hosts of opposite endianness, with this restriction:
3515
3516 As for any compiled expression that is saved and used later, the tables
3517 pointer field cannot be preserved; the extra_data field in the arguments
3518 to pcre_exec() should be used to pass in a tables address if a value
3519 other than the default internal tables were used at compile time.
3520
3521 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3522 now diagnosed as an error. Previously, most of the time, a negative number
3523 would have been treated as zero, but if in addition "ovector" was passed as
3524 NULL, a crash could occur.
3525
3526 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3527 new versions from the libtool 1.5 distribution (the last one is a copy of
3528 a file called libtool.m4). This seems to have fixed the need to patch
3529 "configure" to support Darwin 1.3 (which I used to do). However, I still
3530 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3531 workstation).
3532
3533 24. Changed the PCRE licence to be the more standard "BSD" licence.
3534
3535
3536 Version 4.5 01-Dec-03
3537 ---------------------
3538
3539 1. There has been some re-arrangement of the code for the match() function so
3540 that it can be compiled in a version that does not call itself recursively.
3541 Instead, it keeps those local variables that need separate instances for
3542 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3543 needs to "recurse". Keeping track of where control must go is done by means
3544 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3545 hide most of the details from the main code, and operates only if
3546 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3547 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3548 operating.
3549
3550 To make it easier for callers to provide specially tailored get/free
3551 functions for this usage, two new functions, pcre_stack_malloc, and
3552 pcre_stack_free, are used. They are always called in strict stacking order,
3553 and the size of block requested is always the same.
3554
3555 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3556 PCRE has been compiled to use the stack or the heap for recursion. The
3557 -C option of pcretest uses this to show which version is compiled.
3558
3559 A new data escape \S, is added to pcretest; it causes the amounts of store
3560 obtained and freed by both kinds of malloc/free at match time to be added
3561 to the output.
3562
3563 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3564 what's available on my current Linux desktop machine.
3565
3566 3. When matching a UTF-8 string, the test for a valid string at the start has
3567 been extended. If start_offset is not zero, PCRE now checks that it points
3568 to a byte that is the start of a UTF-8 character. If not, it returns
3569 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3570 this is necessary because there may be backward assertions in the pattern.
3571 When matching the same subject several times, it may save resources to use
3572 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3573
3574 4. The code for checking the validity of UTF-8 strings has been tightened so
3575 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3576 containing "overlong sequences".
3577
3578 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3579 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3580 should have been "&", but it just so happened that all the cases this let
3581 through by mistake were picked up later in the function.
3582
3583 6. I had used a variable called "isblank" - this is a C99 function, causing
3584 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3585
3586 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3587 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3588 so the version is shown; (c) stop "make test" echoing "./RunTest".
3589
3590 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3591
3592 9. The prototype for memmove() for systems that don't have it was using
3593 size_t, but the inclusion of the header that defines size_t was later. I've
3594 moved the #includes for the C headers earlier to avoid this.
3595
3596 10. Added some adjustments to the code to make it easier to compiler on certain
3597 special systems:
3598
3599 (a) Some "const" qualifiers were missing.
3600 (b) Added the macro EXPORT before all exported functions; by default this
3601 is defined to be empty.
3602 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3603 that it reads its output file name as an argument instead of writing
3604 to the standard output and assuming this can be redirected.
3605
3606 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3607 class containing characters with values greater than 255, PCRE compilation
3608 went into a loop.
3609
3610 12. A recursive reference to a subpattern that was within another subpattern
3611 that had a minimum quantifier of zero caused PCRE to crash. For example,
3612 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3613 recursion. If the recursively-called subpattern itself had a zero repeat,
3614 that was OK.
3615
3616 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3617 buffer into which it was copied (for escape processing) was still set at
3618 1024, so long lines caused crashes.
3619
3620 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3621 "internal error: code overflow...". This applied to any character class
3622 that was followed by a possessive quantifier.
3623
3624 15. Modified the Makefile to add libpcre.la as a prerequisite for
3625 libpcreposix.la because I was told this is needed for a parallel build to
3626 work.
3627
3628 16. If a pattern that contained .* following optional items at the start was
3629 studied, the wrong optimizing data was generated, leading to matching
3630 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3631 matching string must start with a or b or c. The correct conclusion for
3632 this pattern is that a match can start with any character.
3633
3634
3635 Version 4.4 13-Aug-03
3636 ---------------------
3637
3638 1. In UTF-8 mode, a character class containing characters with values between
3639 127 and 255 was not handled correctly if the compiled pattern was studied.
3640 In fixing this, I have also improved the studying algorithm for such
3641 classes (slightly).
3642
3643 2. Three internal functions had redundant arguments passed to them. Removal
3644 might give a very teeny performance improvement.
3645
3646 3. Documentation bug: the value of the capture_top field in a callout is *one
3647 more than* the number of the hightest numbered captured substring.
3648
3649 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3650 in incorrectly linking with a previously installed version. They now link
3651 explicitly with libpcre.la.
3652
3653 5. configure.in no longer needs to recognize Cygwin specially.
3654
3655 6. A problem in pcre.in for Windows platforms is fixed.
3656
3657 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3658 pcretest, it used to include the size of the study block as part of its
3659 output. Unfortunately, the structure contains a field that has a different
3660 size on different hardware architectures. This meant that the tests that
3661 showed this size failed. As the block is currently always of a fixed size,
3662 this information isn't actually particularly useful in pcretest output, so
3663 I have just removed it.
3664
3665 8. Three pre-processor statements accidentally did not start in column 1.
3666 Sadly, there are *still* compilers around that complain, even though
3667 standard C has not required this for well over a decade. Sigh.
3668
3669 9. In pcretest, the code for checking callouts passed small integers in the
3670 callout_data field, which is a void * field. However, some picky compilers
3671 complained about the casts involved for this on 64-bit systems. Now
3672 pcretest passes the address of the small integer instead, which should get
3673 rid of the warnings.
3674
3675 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3676 both compile and run time, and gives an error if an invalid UTF-8 sequence
3677 is found. There is a option for disabling this check in cases where the
3678 string is known to be correct and/or the maximum performance is wanted.
3679
3680 11. In response to a bug report, I changed one line in Makefile.in from
3681
3682 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3683 to
3684 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3685
3686 to look similar to other lines, but I have no way of telling whether this
3687 is the right thing to do, as I do not use Windows. No doubt I'll get told
3688 if it's wrong...
3689
3690
3691 Version 4.3 21-May-03
3692 ---------------------
3693
3694 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3695 Makefile.
3696
3697 2. Some refactoring to improve the quality of the code:
3698
3699 (i) The utf8_table... variables are now declared "const".
3700
3701 (ii) The code for \cx, which used the "case flipping" table to upper case
3702 lower case letters, now just substracts 32. This is ASCII-specific,
3703 but the whole concept of \cx is ASCII-specific, so it seems
3704 reasonable.
3705
3706 (iii) PCRE was using its character types table to recognize decimal and
3707 hexadecimal digits in the pattern. This is silly, because it handles
3708 only 0-9, a-f, and A-F, but the character types table is locale-
3709 specific, which means strange things might happen. A private
3710 table is now used for this - though it costs 256 bytes, a table is
3711 much faster than multiple explicit tests. Of course, the standard
3712 character types table is still used for matching digits in subject
3713 strings against \d.
3714
3715 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3716 ending in _t are). So I've renamed it as ESC_tee.
3717
3718 3. The first argument for regexec() in the POSIX wrapper should have been
3719 defined as "const".
3720
3721 4. Changed pcretest to use malloc() for its buffers so that they can be
3722 Electric Fenced for debugging.
3723
3724 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3725 to read one or more bytes before the start of the subject string. Often this
3726 had no effect on PCRE's behaviour, but in some circumstances it could
3727 provoke a segmentation fault.
3728
3729 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3730 to try to read one or more bytes before the start of the subject string.
3731
3732 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3733 UTF-8 support could misbehave in various ways if the subject string
3734 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3735 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3736 back over UTF-8 characters.)
3737
3738
3739 Version 4.2 14-Apr-03
3740 ---------------------
3741
3742 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3743
3744 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3745 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3746 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3747 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3748 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3749 and BUILD_EXEEXT
3750 Note: automatic setting of the BUILD variables is not yet working
3751 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3752 compile-time but not at link-time
3753 [LINK]: use for linking executables only
3754 make different versions for Windows and non-Windows
3755 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3756 libraries
3757 [LINK_FOR_BUILD]: new variable
3758 [OBJEXT]: use throughout
3759 [EXEEXT]: use throughout
3760 <winshared>: new target
3761 <wininstall>: new target
3762 <dftables.o>: use native compiler
3763 <dftables>: use native linker
3764 <install>: handle Windows platform correctly
3765 <clean>: ditto
3766 <check>: ditto
3767 copy DLL to top builddir before testing
3768
3769 As part of these changes, -no-undefined was removed again. This was reported
3770 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3771 in any case.
3772
3773 3. Some tidies to get rid of compiler warnings:
3774
3775 . In the match_data structure, match_limit was an unsigned long int, whereas
3776 match_call_count was an int. I've made them both unsigned long ints.
3777
3778 . In pcretest the fact that a const uschar * doesn't automatically cast to
3779 a void * provoked a warning.
3780
3781 . Turning on some more compiler warnings threw up some "shadow" variables
3782 and a few more missing casts.
3783
3784 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3785 option, a class that contained a single character with a value between 128
3786 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3787
3788 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3789 option, a class that contained several characters, but with at least one
3790 whose value was between 128 and 255 caused PCRE to crash.
3791
3792
3793 Version 4.1 12-Mar-03
3794 ---------------------
3795
3796 1. Compiling with gcc -pedantic found a couple of places where casts were
3797 needed, and a string in dftables.c that was longer than standard compilers are
3798 required to support.
3799
3800 2. Compiling with Sun's compiler found a few more places where the code could
3801 be tidied up in order to avoid warnings.
3802
3803 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3804 first of these names is deprecated in the latest Autoconf in favour of the name
3805 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3806 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3807 analogy I have changed it to CFLAGS_FOR_BUILD.
3808
3809 4. Added -no-undefined to the linking command in the Makefile, because this is
3810 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3811 linking step for the pcreposix library.
3812
3813 5. PCRE was failing to diagnose the case of two named groups with the same
3814 name.
3815
3816 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3817 literal character that is needed in the subject for a match, and scans along to
3818 ensure that it is present before embarking on the full matching process. This
3819 saves time in cases of nested unlimited repeats that are never going to match.
3820 Problem: the scan can take a lot of time if the subject is very long (e.g.
3821 megabytes), thus penalizing straightforward matches. It is now done only if the
3822 amount of subject to be scanned is less than 1000 bytes.
3823
3824 7. A lesser problem with the same optimization is that it was recording the
3825 first character of an anchored pattern as "needed", thus provoking a search
3826 right along the subject, even when the first match of the pattern was going to
3827 fail. The "needed" character is now not set for anchored patterns, unless it
3828 follows something in the pattern that is of non-fixed length. Thus, it still
3829 fulfils its original purpose of finding quick non-matches in cases of nested
3830 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3831
3832
3833 Version 4.0 17-Feb-03
3834 ---------------------
3835
3836 1. If a comment in an extended regex that started immediately after a meta-item
3837 extended to the end of string, PCRE compiled incorrect data. This could lead to
3838 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3839
3840 2. Moved to autoconf 2.53 and libtool 1.4.2.
3841
3842 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3843 the special perltest8 script is no longer needed - all the tests can be run
3844 from a single perltest script.
3845
3846 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3847 by \s. It has now been removed in PCRE. This means it isn't recognized as
3848 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3849 class [:space:] *does* include VT, thereby creating a mess.
3850
3851 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3852 space and tab.
3853
3854 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3855 its new features into the main test script, reducing the number of scripts.
3856
3857 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3858 were backward compatible, and made the (?i) apply to the whole pattern, as if
3859 /i were given. Now it behaves more logically, and applies the option setting
3860 only to what follows. PCRE has been changed to follow suit. However, if it
3861 finds options settings right at the start of the pattern, it extracts them into
3862 the global options, as before. Thus, they show up in the info data.
3863
3864 8. Added support for the \Q...\E escape sequence. Characters in between are
3865 treated as literals. This is slightly different from Perl in that $ and @ are
3866 also handled as literals inside the quotes. In Perl, they will cause variable
3867 interpolation. Note the following examples:
3868
3869 Pattern PCRE matches Perl matches
3870
3871 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3872 \Qabc\$xyz\E abc\$xyz abc\$xyz
3873 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3874
3875 For compatibility with Perl, \Q...\E sequences are recognized inside character
3876 classes as well as outside them.
3877
3878 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3879 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3880 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3881 signed/unsigned warnings.
3882
3883 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3884 option for pcretest, so I've replaced it by a simple function that does just
3885 that job.
3886
3887 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3888 "pcregrep -".
3889
3890 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3891 Java package. This provides some syntactic sugar for simple cases of what my
3892 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3893 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3894 item, you can use this simplified notation. Note that only makes sense with
3895 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3896 greediness, whatever the setting of the PCRE_UNGREEDY option.
3897
3898 13. A change of greediness default within a pattern was not taking effect at
3899 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3900 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3901 was abstracted outside.
3902
3903 14. PCRE now supports the \G assertion. It is true when the current matching
3904 position is at the start point of the match. This differs from \A when the
3905 starting offset is non-zero. Used with the /g option of pcretest (or similar
3906 code), it works in the same way as it does for Perl's /g option. If all
3907 alternatives of a regex begin with \G, the expression is anchored to the start
3908 match position, and the "anchored" flag is set in the compiled expression.
3909
3910 15. Some bugs concerning the handling of certain option changes within patterns
3911 have been fixed. These applied to options other than (?ims). For example,
3912 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3913 been the other way round. Some of this was related to change 7 above.
3914
3915 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3916 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3917 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3918 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3919
3920 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3921 mode. Unlike ".", it always matches newline, whatever the setting of
3922 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3923 assertions. Perl allows it, but it doesn't (in general) work because it can't
3924 calculate the length of the lookbehind. At least, that's the case for Perl
3925 5.8.0 - I've been told they are going to document that it doesn't work in
3926 future.
3927
3928 18. Added an error diagnosis for escapes that PCRE does not support: these are
3929 \L, \l, \N, \P, \p, \U, \u, and \X.
3930
3931 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3932 reading past the end of the pattern in cases such as /[abcd/.
3933
3934 20. PCRE was getting more memory than necessary for patterns with classes that
3935 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3936
3937 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3938 compiling PCRE for use with Virtual Pascal.
3939
3940 22. Small fix to the Makefile to make it work properly if the build is done
3941 outside the source tree.
3942
3943 23. Added a new extension: a condition to go with recursion. If a conditional
3944 subpattern starts with (?(R) the "true" branch is used if recursion has
3945 happened, whereas the "false" branch is used only at the top level.
3946
3947 24. When there was a very long string of literal characters (over 255 bytes
3948 without UTF support, over 250 bytes with UTF support), the computation of how
3949 much memory was required could be incorrect, leading to segfaults or other
3950 strange effects.
3951
3952 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3953 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3954 there was a subsequent back reference to those brackets. This meant that, for
3955 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3956 possible to check for precisely this case. All we can do is abandon the
3957 optimization if .* occurs inside capturing brackets when there are any back
3958 references whatsoever. (See below for a better fix that came later.)
3959
3960 26. The handling of the optimization for finding the first character of a
3961 non-anchored pattern, and for finding a character that is required later in the
3962 match were failing in some cases. This didn't break the matching; it just
3963 failed to optimize when it could. The way this is done has been re-implemented.
3964
3965 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3966
3967 28. Added a new feature that provides some of the functionality that Perl
3968 provides with (?{...}). The facility is termed a "callout". The way it is done
3969 in PCRE is for the caller to provide an optional function, by setting
3970 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3971 global variable. By default it is unset, which disables all calling out. To get
3972 the function called, the regex must include (?C) at appropriate points. This
3973 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3974 This provides a means of identifying different callout points. When PCRE
3975 reaches such a point in the regex, if pcre_callout has been set, the external
3976 function is called. It is provided with data in a structure called
3977 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3978 matching continues; if it returns a non-zero value, the match at the current
3979 point fails. However, backtracking will occur if possible. [This was changed
3980 later and other features added - see item 49 below.]
3981
3982 29. pcretest is upgraded to test the callout functionality. It provides a
3983 callout function that displays information. By default, it shows the start of
3984 the match and the current position in the text. There are some new data escapes
3985 to vary what happens:
3986
3987 \C+ in addition, show current contents of captured substrings
3988 \C- do not supply a callout function
3989 \C!n return 1 when callout number n is reached
3990 \C!n!m return 1 when callout number n is reached for the mth time
3991
3992 30. If pcregrep was called with the -l option and just a single file name, it
3993 output "<stdin>" if a match was found, instead of the file name.
3994
3995 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3996 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3997 pcre_exec(). This saves a malloc/free per call. The default value of
3998 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3999 when configuring.
4000
4001 32. The default maximum size of a compiled pattern is 64K. There have been a
4002 few cases of people hitting this limit. The code now uses macros to handle the
4003 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4004 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4005 configuring. Tests 2 and 5 work only with 2-byte links because they output
4006 debugging information about compiled patterns.
4007
4008 33. Internal code re-arrangements:
4009
4010 (a) Moved the debugging function for printing out a compiled regex into
4011 its own source file (printint.c) and used #include to pull it into
4012 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4013 separate copies.
4014
4015 (b) Defined the list of op-code names for debugging as a macro in
4016 internal.h so that it is next to the definition of the opcodes.
4017
4018 (c) Defined a table of op-code lengths for simpler skipping along compiled
4019 code. This is again a macro in internal.h so that it is next to the
4020 definition of the opcodes.
4021
4022 34. Added support for recursive calls to individual subpatterns, along the
4023 lines of Robin Houston's patch (but implemented somewhat differently).
4024
4025 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4026 allow it to read and process whole directories in Win32. This code was
4027 contributed by Lionel Fourquaux; it has not been tested by me.
4028
4029 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4030 used to name a group. Names consist of alphanumerics and underscores, and must
4031 be unique. Back references use the syntax (?P=name) and recursive calls use
4032 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4033 numbers. The function pcre_fullinfo() can be used after compilation to extract
4034 a name/number map. There are three relevant calls:
4035
4036 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4037 PCRE_INFO_NAMECOUNT yields the number of entries
4038 PCRE_INFO_NAMETABLE yields a pointer to the map.
4039
4040 The map is a vector of fixed-size entries. The size of each entry depends on
4041 the length of the longest name used. The first two bytes of each entry are the
4042 group number, most significant byte first. There follows the corresponding
4043 name, zero terminated. The names are in alphabetical order.
4044
4045 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4046 case instead of 255. Making it the same both with and without UTF-8 support
4047 means that the same test output works with both.
4048
4049 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4050 calling malloc() with a zero argument.
4051
4052 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4053 optimization. I've improved things by keeping a bitmap of backreferences with
4054 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4055 fact referenced, the optimization can be applied. It is unlikely that a
4056 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4057 the match to follow \n) will appear inside brackets with a number greater than
4058 31, but if it does, any back reference > 31 suppresses the optimization.
4059
4060 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4061 of disabling numbered capturing parentheses. Any opening parenthesis that is
4062 not followed by ? behaves as if it were followed by ?: but named parentheses
4063 can still be used for capturing (and they will acquire numbers in the usual
4064 way).
4065
4066 41. Redesigned the return codes from the match() function into yes/no/error so
4067 that errors can be passed back from deep inside the nested calls. A malloc
4068 failure while inside a recursive subpattern call now causes the
4069 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4070
4071 42. It is now possible to set a limit on the number of times the match()
4072 function is called in a call to pcre_exec(). This facility makes it possible to
4073 limit the amount of recursion and backtracking, though not in a directly
4074 obvious way, because the match() function is used in a number of different
4075 circumstances. The count starts from zero for each position in the subject
4076 string (for non-anchored patterns). The default limit is, for compatibility, a
4077 large number, namely 10 000 000. You can change this in two ways:
4078
4079 (a) When configuring PCRE before making, you can use --with-match-limit=n
4080 to set a default value for the compiled library.
4081
4082 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4083 a different value is set. See 45 below.
4084
4085 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4086
4087 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4088 of things that can be changed at compile time. The first argument specifies
4089 what is wanted and the second points to where the information is to be placed.
4090 The current list of available information is:
4091
4092 PCRE_CONFIG_UTF8
4093
4094 The output is an integer that is set to one if UTF-8 support is available;
4095 otherwise it is set to zero.
4096
4097 PCRE_CONFIG_NEWLINE
4098
4099 The output is an integer that it set to the value of the code that is used for
4100 newline. It is either LF (10) or CR (13).
4101
4102 PCRE_CONFIG_LINK_SIZE
4103
4104 The output is an integer that contains the number of bytes used for internal
4105 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4106
4107 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4108
4109 The output is an integer that contains the threshold above which the POSIX
4110 interface uses malloc() for output vectors. See item 31 above.
4111
4112 PCRE_CONFIG_MATCH_LIMIT
4113
4114 The output is an unsigned integer that contains the default limit of the number
4115 of match() calls in a pcre_exec() execution. See 42 above.
4116
4117 44. pcretest has been upgraded by the addition of the -C option. This causes it
4118 to extract all the available output from the new pcre_config() function, and to
4119 output it. The program then exits immediately.
4120
4121 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4122 order to support additional features. One way would have been to define
4123 pcre_exec2() (for example) with extra arguments, but this would not have been
4124 extensible, and would also have required all calls to the original function to
4125 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4126 is used for passing in "extra" data from pcre_study().
4127
4128 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4129 contains the following fields:
4130
4131 flags a bitmap indicating which of the following fields are set
4132 study_data opaque data from pcre_study()
4133 match_limit a way of specifying a limit on match() calls for a specific
4134 call to pcre_exec()
4135 callout_data data for callouts (see 49 below)
4136
4137 The flag bits are also defined in pcre.h, and are
4138
4139 PCRE_EXTRA_STUDY_DATA
4140 PCRE_EXTRA_MATCH_LIMIT
4141 PCRE_EXTRA_CALLOUT_DATA
4142
4143 The pcre_study() function now returns one of these new pcre_extra blocks, with
4144 the actual study data pointed to by the study_data field, and the
4145 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4146 before. That is, this change is entirely upwards-compatible and requires no
4147 change to existing code.
4148
4149 If you want to pass in additional data to pcre_exec(), you can either place it
4150 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4151 block.
4152
4153 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4154 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4155 times with different match limits, until it finds the minimum value needed for
4156 pcre_exec() to complete. The value is then output. This can be instructive; for
4157 most simple matches the number is quite small, but for pathological cases it
4158 gets very large very quickly.
4159
4160 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4161 returns the size of the data block pointed to by the study_data field in a
4162 pcre_extra block, that is, the value that was passed as the argument to
4163 pcre_malloc() when PCRE was getting memory in which to place the information
4164 created by pcre_study(). The fourth argument should point to a size_t variable.
4165 pcretest has been extended so that this information is shown after a successful
4166 pcre_study() call when information about the compiled regex is being displayed.
4167
4168 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4169 because what follows is always an absolute path. (Later: it turns out that this
4170 is more than cosmetic for MinGW, because it doesn't like empty path
4171 components.)
4172
4173 49. Some changes have been made to the callout feature (see 28 above):
4174
4175 (i) A callout function now has three choices for what it returns:
4176
4177 0 => success, carry on matching
4178 > 0 => failure at this point, but backtrack if possible
4179 < 0 => serious error, return this value from pcre_exec()
4180
4181 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4182 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4183 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4184 use by callout functions. It will never be used by PCRE itself.
4185
4186 (ii) The pcre_extra structure (see 45 above) has a void * field called
4187 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4188 pcre_callout_block structure has a field of the same name. The contents of
4189 the field passed in the pcre_extra structure are passed to the callout
4190 function in the corresponding field in the callout block. This makes it
4191 easier to use the same callout-containing regex from multiple threads. For
4192 testing, the pcretest program has a new data escape
4193
4194 \C*n pass the number n (may be negative) as callout_data
4195
4196 If the callout function in pcretest receives a non-zero value as
4197 callout_data, it returns that value.
4198
4199 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4200 there were some redundant $(CFLAGS) in commands that are now specified as
4201 $(LINK), which already includes $(CFLAGS).
4202
4203 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4204 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4205 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4206 one-byte characters throughout. Note that case-insensitive matching applies
4207 only to characters whose values are less than 256. PCRE doesn't support the
4208 notion of cases for higher-valued characters.
4209
4210 (i) A character class whose characters are all within 0-255 is handled as
4211 a bit map, and the map is inverted for negative classes. Previously, a
4212 character > 255 always failed to match such a class; however it should
4213 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4214
4215 (ii) A negated character class with a single character < 255 is coded as
4216 "not this character" (OP_NOT). This wasn't working properly when the test
4217 character was multibyte, either singly or repeated.
4218
4219 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4220 mode, for example: \x{100}{2,3}.
4221
4222 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4223 singly or repeated) now correctly test multibyte characters. However,
4224 PCRE doesn't recognize any characters with values greater than 255 as
4225 digits, spaces, or word characters. Such characters always match \D, \S,
4226 and \W, and never match \d, \s, or \w.
4227
4228 (v) Classes may now contain characters and character ranges with values
4229 greater than 255. For example: [ab\x{100}-\x{400}].
4230
4231 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4232 PCRE in UTF-8 mode.
4233
4234 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4235 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4236 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4237 value.)
4238
4239 53. The single man page has become too large. I have therefore split it up into
4240 a number of separate man pages. These also give rise to individual HTML pages;
4241 these are now put in a separate directory, and there is an index.html page that
4242 lists them all. Some hyperlinking between the pages has been installed.
4243
4244 54. Added convenience functions for handling named capturing parentheses.
4245
4246 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4247 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4248 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4249 are faulted.
4250
4251 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4252 calling configure. These values are used when compiling the dftables.c program
4253 which is run to generate the source of the default character tables. They
4254 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4255 you will need to set these values.
4256
4257 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4258
4259
4260 Version 3.9 02-Jan-02
4261 ---------------------
4262
4263 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4264
4265 2. If --disable-static was given, the building process failed when trying to
4266 build pcretest and pcregrep. (For some reason it was using libtool to compile
4267 them, which is not right, as they aren't part of the library.)
4268
4269
4270 Version 3.8 18-Dec-01
4271 ---------------------
4272
4273 1. The experimental UTF-8 code was completely screwed up. It was packing the
4274 bytes in the wrong order. How dumb can you get?
4275
4276
4277 Version 3.7 29-Oct-01
4278 ---------------------
4279
4280 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4281 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4282 this didn't happen under Solaris 8, where I normally test things.
4283
4284 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4285 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4286 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4287 this fix an hour or so after the initial 3.7 release.)
4288
4289
4290 Version 3.6 23-Oct-01
4291 ---------------------
4292
4293 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4294 offsets passed as NULL with zero offset count.
4295
4296 2. The config.guess and config.sub files had not been updated when I moved to
4297 the latest autoconf.
4298
4299
4300 Version 3.5 15-Aug-01
4301 ---------------------
4302
4303 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4304 had been forgotten.
4305
4306 2. By using declared but undefined structures, we can avoid using "void"
4307 definitions in pcre.h while keeping the internal definitions of the structures
4308 private.
4309
4310 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4311 user point of view, this means that both static and shared libraries are built
4312 by default, but this can be individually controlled. More of the work of
4313 handling this static/shared cases is now inside libtool instead of PCRE's make
4314 file.
4315
4316 4. The pcretest utility is now installed along with pcregrep because it is
4317 useful for users (to test regexs) and by doing this, it automatically gets
4318 relinked by libtool. The documentation has been turned into a man page, so
4319 there are now .1, .txt, and .html versions in /doc.
4320
4321 5. Upgrades to pcregrep:
4322 (i) Added long-form option names like gnu grep.
4323 (ii) Added --help to list all options with an explanatory phrase.
4324 (iii) Added -r, --recursive to recurse into sub-directories.
4325 (iv) Added -f, --file to read patterns from a file.
4326
4327 6. pcre_exec() was referring to its "code" argument before testing that
4328 argument for NULL (and giving an error if it was NULL).
4329
4330 7. Upgraded Makefile.in to allow for compiling in a different directory from
4331 the source directory.
4332
4333 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4334 options bits, the pointer it was passed was to an int instead of to an unsigned
4335 long int. This mattered only on 64-bit systems.
4336
4337 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4338 generated) instead of pcre.in, which it its source. Also made the same change
4339 in several of the .c files.
4340
4341 10. A new release of gcc defines printf() as a macro, which broke pcretest
4342 because it had an ifdef in the middle of a string argument for printf(). Fixed
4343 by using separate calls to printf().
4344
4345 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4346 script, to force use of CR or LF instead of \n in the source. On non-Unix
4347 systems, the value can be set in config.h.
4348
4349 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4350 absolute limit. Changed the text of the error message to make this clear, and
4351 likewise updated the man page.
4352
4353 13. The limit of 99 on the number of capturing subpatterns has been removed.
4354 The new limit is 65535, which I hope will not be a "real" limit.
4355
4356
4357 Version 3.4 22-Aug-00
4358 ---------------------
4359
4360 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4361
4362 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4363
4364
4365 Version 3.3 01-Aug-00
4366 ---------------------
4367
4368 1. If an octal character was given, but the value was greater than \377, it
4369 was not getting masked to the least significant bits, as documented. This could
4370 lead to crashes in some systems.
4371
4372 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4373 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4374
4375 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4376 These just pass their arguments on to (pcre_free)(), but they are provided
4377 because some uses of PCRE bind it to non-C systems that can call its functions,
4378 but cannot call free() or pcre_free() directly.
4379
4380 4. Add "make test" as a synonym for "make check". Corrected some comments in
4381 the Makefile.
4382
4383 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4384 Makefile.
4385
4386 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4387 command called pgrep for grepping around the active processes.
4388
4389 7. Added the beginnings of support for UTF-8 character strings.
4390
4391 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4392 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4393 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4394 out for the ar command.)
4395
4396
4397 Version 3.2 12-May-00
4398 ---------------------
4399
4400 This is purely a bug fixing release.
4401
4402 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4403 of ZA. This was just one example of several cases that could provoke this bug,
4404 which was introduced by change 9 of version 2.00. The code for breaking
4405 infinite loops after an iteration that matches an empty string was't working
4406 correctly.
4407
4408 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4409 when matched against abbab (for example). After matching an empty string, it
4410 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4411 caused it to match further down the string than it should.
4412
4413 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4414 was there because it doesn't seem to be needed, and it causes trouble on some
4415 systems, as it is not a Standard C header. It has been removed.
4416
4417 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4418 were reported on the Macintosh. The changes were from
4419
4420 while ((c = *(++ptr)) != 0 && c != '\n');
4421 to
4422 while ((c = *(++ptr)) != 0 && c != '\n') ;
4423
4424 Totally extraordinary, but if that's what it takes...
4425
4426 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4427 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4428 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4429 assumes the way PCRE uses memmove() (always moving upwards).
4430
4431 6. PCRE is being used in one environment where strchr() is not available. There
4432 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4433 faster code anyway.
4434
4435
4436 Version 3.1 09-Feb-00
4437 ---------------------
4438
4439 The only change in this release is the fixing of some bugs in Makefile.in for
4440 the "install" target:
4441
4442 (1) It was failing to install pcreposix.h.
4443
4444 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4445
4446
4447 Version 3.0 01-Feb-00
4448 ---------------------
4449
4450 1. Add support for the /+ modifier to perltest (to output $` like it does in
4451 pcretest).
4452
4453 2. Add support for the /g modifier to perltest.
4454
4455 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4456 matches null strings.
4457
4458 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4459 pattern. Perl treats empty patterns specially - it reuses the most recent
4460 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4461 effect.
4462
4463 5. The POSIX interface was broken in that it was just handing over the POSIX
4464 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4465 required a bigger vector, with some working space on the end. This means that
4466 the POSIX wrapper now has to get and free some memory, and copy the results.
4467
4468 6. Added some simple autoconf support, placing the test data and the
4469 documentation in separate directories, re-organizing some of the
4470 information files, and making it build pcre-config (a GNU standard). Also added
4471 libtool support for building PCRE as a shared library, which is now the
4472 default.
4473
4474 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4475 09 are not valid octal constants. Single digits will be used for minor values
4476 less than 10.
4477
4478 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4479 existing programs that set these in the POSIX interface can use PCRE without
4480 modification.
4481
4482 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4483 return all that pcre_info() returns, plus additional data. The pcre_info()
4484 function is retained for compatibility, but is considered to be obsolete.
4485
4486 10. Added experimental recursion feature (?R) to handle one common case that
4487 Perl 5.6 will be able to do with (?p{...}).
4488
4489 11. Added support for POSIX character classes like [:alpha:], which Perl is
4490 adopting.
4491
4492
4493 Version 2.08 31-Aug-99
4494 ----------------------
4495
4496 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4497 trying to match at the startoffset position, but instead was moving forward to
4498 the next newline as if a previous match had failed.
4499
4500 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4501 and could get into a loop if a null string was matched other than at the start
4502 of the subject.
4503
4504 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4505 be distinguished at compile time, and for completeness also added PCRE_DATE.
4506
4507 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4508 in GnuWin32 environments.
4509
4510
4511 Version 2.07 29-Jul-99
4512 ----------------------
4513
4514 1. The documentation is now supplied in plain text form and HTML as well as in
4515 the form of man page sources.
4516
4517 2. C++ compilers don't like assigning (void *) values to other pointer types.
4518 In particular this affects malloc(). Although there is no problem in Standard
4519 C, I've put in casts to keep C++ compilers happy.
4520
4521 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4522 should be (const char *).
4523
4524 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4525 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4526 However, I haven't made this a standard facility. The documentation doesn't
4527 mention it, and the Makefile doesn't support it.
4528
4529 5. The Makefile now contains an "install" target, with editable destinations at
4530 the top of the file. The pcretest program is not installed.
4531
4532 6. pgrep -V now gives the PCRE version number and date.
4533
4534 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4535 causing the entire string to be ignored, instead of just the last character.
4536
4537 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4538 non-matching string, it can take a very, very long time, even for strings of
4539 quite modest length, because of the nested recursion. PCRE now does better in
4540 some of these cases. It does this by remembering the last required literal
4541 character in the pattern, and pre-searching the subject to ensure it is present
4542 before running the real match. In other words, it applies a heuristic to detect
4543 some types of certain failure quickly, and in the above example, if presented
4544 with a string that has no trailing " it gives "no match" very quickly.
4545
4546 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4547 other alternatives are tried instead.
4548
4549
4550 Version 2.06 09-Jun-99
4551 ----------------------
4552
4553 1. Change pcretest's output for amount of store used to show just the code
4554 space, because the remainder (the data block) varies in size between 32-bit and
4555 64-bit systems.
4556
4557 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4558 start matching at. This allows lookbehinds to work when searching for multiple
4559 occurrences in a string.
4560
4561 3. Added additional options to pcretest for testing multiple occurrences:
4562
4563 /+ outputs the rest of the string that follows a match
4564 /g loops for multiple occurrences, using the new startoffset argument
4565 /G loops for multiple occurrences by passing an incremented pointer
4566
4567 4. PCRE wasn't doing the "first character" optimization for patterns starting
4568 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4569 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4570 the letter 'x'. On long subject strings, this gives a significant speed-up.
4571
4572
4573 Version 2.05 21-Apr-99
4574 ----------------------
4575
4576 1. Changed the type of magic_number from int to long int so that it works
4577 properly on 16-bit systems.
4578
4579 2. Fixed a bug which caused patterns starting with .* not to work correctly
4580 when the subject string contained newline characters. PCRE was assuming
4581 anchoring for such patterns in all cases, which is not correct because .* will
4582 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4583 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4584 must be retried after every newline in the subject.
4585
4586
4587 Version 2.04 18-Feb-99
4588 ----------------------
4589
4590 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4591 computation of the store needed to hold the pattern was incorrect (too large).
4592 If such patterns were nested a few deep, this could multiply and become a real
4593 problem.
4594
4595 2. Added /M option to pcretest to show the memory requirement of a specific
4596 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4597
4598 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4599 compiled in such a way that the backtracking after subsequent failure was
4600 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4601 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4602
4603
4604 Version 2.03 02-Feb-99
4605 ----------------------
4606
4607 1. Fixed typo and small mistake in man page.
4608
4609 2. Added 4th condition (GPL supersedes if conflict) and created separate
4610 LICENCE file containing the conditions.
4611
4612 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4613 Perl, that is the internal \ allows the delimiter to be included in the
4614 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4615 the final delimiter, add \ to the end of the pattern (to test the error).
4616
4617 4. Added the convenience functions for extracting substrings after a successful
4618 match. Updated pcretest to make it able to test these functions.
4619
4620
4621 Version 2.02 14-Jan-99
4622 ----------------------
4623
4624 1. Initialized the working variables associated with each extraction so that
4625 their saving and restoring doesn't refer to uninitialized store.
4626
4627 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4628 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4629 fix the problem.
4630
4631 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4632 calls, and wasn't printing the correct value for compiling calls. Increased the
4633 default value of LOOPREPEAT, and the number of significant figures in the
4634 times.
4635
4636 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4637
4638 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4639 a building problem on Windows NT with a FAT file system.
4640
4641
4642 Version 2.01 21-Oct-98
4643 ----------------------
4644
4645 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4646 to character tables built by pcre_maketables() in the current locale. If NULL
4647 is passed, the default tables are used.
4648
4649
4650 Version 2.00 24-Sep-98
4651 ----------------------
4652
4653 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4654 it any more.
4655
4656 2. Allow quantification of (?>) groups, and make it work correctly.
4657
4658 3. The first character computation wasn't working for (?>) groups.
4659
4660 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4661 end of the subject) and add 5.005's \z, which really does match only at the
4662 very end of the subject.
4663
4664 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4665
4666 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4667 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4668 localized options. All options to pcre_study() were also removed.
4669
4670 7. Add other new features from 5.005:
4671
4672 $(?<= positive lookbehind
4673 $(?<! negative lookbehind
4674 (?imsx-imsx) added the unsetting capability
4675 such a setting is global if at outer level; local otherwise
4676 (?imsx-imsx:) non-capturing groups with option setting
4677 (?(cond)re|re) conditional pattern matching
4678
4679 A backreference to itself in a repeated group matches the previous
4680 captured string.
4681
4682 8. General tidying up of studying (both automatic and via "study")
4683 consequential on the addition of new assertions.
4684
4685 9. As in 5.005, unlimited repeated groups that could match an empty substring
4686 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4687 runtime if any iteration does actually match an empty substring.
4688
4689 10. Include the RunTest script in the distribution.
4690
4691 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4692 discrepancies, some of which were old and were also with respect to 5.004. They
4693 have now been fixed.
4694
4695
4696 Version 1.09 28-Apr-98
4697 ----------------------
4698
4699 1. A negated single character class followed by a quantifier with a minimum
4700 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4701 program crashes, or just wrong answers. This did not apply to negated classes
4702 containing more than one character, or to minima other than one.
4703
4704
4705 Version 1.08 27-Mar-98
4706 ----------------------
4707
4708 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4709
4710 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4711 latter must appear before anything that relies on it in the pattern.
4712
4713
4714 Version 1.07 16-Feb-98
4715 ----------------------
4716
4717 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4718 repeat of a potentially empty string).
4719
4720
4721 Version 1.06 23-Jan-98
4722 ----------------------
4723
4724 1. Added Markus Oberhumer's little patches for C++.
4725
4726 2. Literal strings longer than 255 characters were broken.
4727
4728
4729 Version 1.05 23-Dec-97
4730 ----------------------
4731
4732 1. Negated character classes containing more than one character were failing if
4733 PCRE_CASELESS was set at run time.
4734
4735
4736 Version 1.04 19-Dec-97
4737 ----------------------
4738
4739 1. Corrected the man page, where some "const" qualifiers had been omitted.
4740
4741 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4742 input syntax.
4743
4744 3. Fixed memory leak which occurred when a regex with back references was
4745 matched with an offsets vector that wasn't big enough. The temporary memory
4746 that is used in this case wasn't being freed if the match failed.
4747
4748 4. Tidied pcretest to ensure it frees memory that it gets.
4749
4750 5. Temporary memory was being obtained in the case where the passed offsets
4751 vector was exactly big enough.
4752
4753 6. Corrected definition of offsetof() from change 5 below.
4754
4755 7. I had screwed up change 6 below and broken the rules for the use of
4756 setjmp(). Now fixed.
4757
4758
4759 Version 1.03 18-Dec-97
4760 ----------------------
4761
4762 1. A erroneous regex with a missing opening parenthesis was correctly
4763 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4764 on some systems.
4765
4766 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4767 it was reported that one broken compiler failed on the former because "code" is
4768 also an independent variable.
4769
4770 3. The erroneous regex a[]b caused an array overrun reference.
4771
4772 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4773 fail on data ending with that character. (It was going on too far, and checking
4774 the next character, typically a binary zero.) This was specific to the
4775 optimized code for single-character negative classes.
4776
4777 5. Added a contributed patch from the TIN world which does the following:
4778
4779 + Add an undef for memmove, in case the the system defines a macro for it.
4780
4781 + Add a definition of offsetof(), in case there isn't one. (I don't know
4782 the reason behind this - offsetof() is part of the ANSI standard - but
4783 it does no harm).
4784
4785 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4786 most of the places where whitespace preceded '#'. I have given up and
4787 allowed the remaining 2 cases to be at the margin.
4788
4789 + Rename some variables in pcre to eliminate shadowing. This seems very
4790 pedantic, but does no harm, of course.
4791
4792 6. Moved the call to setjmp() into its own function, to get rid of warnings
4793 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4794
4795 7. Constructs such as \d{8,} were compiling into the equivalent of
4796 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4797 outcome, but in this particular case used more store than had been allocated,
4798 which caused the bug to be discovered because it threw up an internal error.
4799
4800 8. The debugging code in both pcre and pcretest for outputting the compiled
4801 form of a regex was going wrong in the case of back references followed by
4802 curly-bracketed repeats.
4803
4804
4805 Version 1.02 12-Dec-97
4806 ----------------------
4807
4808 1. Typos in pcre.3 and comments in the source fixed.
4809
4810 2. Applied a contributed patch to get rid of places where it used to remove
4811 'const' from variables, and fixed some signed/unsigned and uninitialized
4812 variable warnings.
4813
4814 3. Added the "runtest" target to Makefile.
4815
4816 4. Set default compiler flag to -O2 rather than just -O.
4817
4818
4819 Version 1.01 19-Nov-97
4820 ----------------------
4821
4822 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4823 like /([ab]*)*/, that is, for classes with more than one character in them.
4824
4825 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4826 as /((?>a*))*/ (a PCRE_EXTRA facility).
4827
4828
4829 Version 1.00 18-Nov-97
4830 ----------------------
4831
4832 1. Added compile-time macros to support systems such as SunOS4 which don't have
4833 memmove() or strerror() but have other things that can be used instead.
4834
4835 2. Arranged that "make clean" removes the executables.
4836
4837
4838 Version 0.99 27-Oct-97
4839 ----------------------
4840
4841 1. Fixed bug in code for optimizing classes with only one character. It was
4842 initializing a 32-byte map regardless, which could cause it to run off the end
4843 of the memory it had got.
4844
4845 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4846
4847
4848 Version 0.98 22-Oct-97
4849 ----------------------
4850
4851 1. Fixed bug in code for handling temporary memory usage when there are more
4852 back references than supplied space in the ovector. This could cause segfaults.
4853
4854
4855 Version 0.97 21-Oct-97
4856 ----------------------
4857
4858 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4859
4860 2. Optimized negated single characters not to use a bit map.
4861
4862 3. Brought error texts together as macro definitions; clarified some of them;
4863 fixed one that was wrong - it said "range out of order" when it meant "invalid
4864 escape sequence".
4865
4866 4. Changed some char * arguments to const char *.
4867
4868 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4869
4870 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4871 pcretest.
4872
4873
4874 Version 0.96 16-Oct-97
4875 ----------------------
4876
4877 1. Added a simple "pgrep" utility to the distribution.
4878
4879 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4880 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4881 where "ddd" means "one or more decimal digits".
4882
4883 3. Fixed serious bug. If a pattern had a back reference, but the call to
4884 pcre_exec() didn't supply a large enough ovector to record the related
4885 identifying subpattern, the match always failed. PCRE now remembers the number
4886 of the largest back reference, and gets some temporary memory in which to save
4887 the offsets during matching if necessary, in order to ensure that
4888 backreferences always work.
4889
4890 4. Increased the compatibility with Perl in a number of ways:
4891
4892 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4893 to request this handling. The option can be set at compile or exec time.
4894
4895 (b) $ matches before a terminating newline by default; an option
4896 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4897 mode). The option can be set at compile or exec time.
4898
4899 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4900 the same as Perl's. If the decimal number it represents is less than 10
4901 or there aren't that many previous left capturing parentheses, an octal
4902 escape is read. Inside a character class, it's always an octal escape,
4903 even if it is a single digit.
4904
4905 (d) An escaped but undefined alphabetic character is taken as a literal,
4906 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4907 escapes.
4908
4909 (e) {0} is now permitted. (The previous item is removed from the compiled
4910 pattern).
4911
4912 5. Changed all the names of code files so that the basic parts are no longer
4913 than 10 characters, and abolished the teeny "globals.c" file.
4914
4915 6. Changed the handling of character classes; they are now done with a 32-byte
4916 bit map always.
4917
4918 7. Added the -d and /D options to pcretest to make it possible to look at the
4919 internals of compilation without having to recompile pcre.
4920
4921
4922 Version 0.95 23-Sep-97
4923 ----------------------
4924
4925 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4926 \x20 at the start of a run of normal characters. These were being treated as
4927 real characters, instead of the source characters being re-checked.
4928
4929
4930 Version 0.94 18-Sep-97
4931 ----------------------
4932
4933 1. The functions are now thread-safe, with the caveat that the global variables
4934 containing pointers to malloc() and free() or alternative functions are the
4935 same for all threads.
4936
4937 2. Get pcre_study() to generate a bitmap of initial characters for non-
4938 anchored patterns when this is possible, and use it if passed to pcre_exec().
4939
4940
4941 Version 0.93 15-Sep-97
4942 ----------------------
4943
4944 1. /(b)|(:+)/ was computing an incorrect first character.
4945
4946 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4947 but not actually doing anything yet.
4948
4949 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4950 as Perl does (e.g. [-az] or [az-]).
4951
4952 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4953 all possible positions.
4954
4955 5. Split up into different modules to avoid including unneeded functions in a
4956 compiled binary. However, compile and exec are still in one module. The "study"
4957 function is split off.
4958
4959 6. The character tables are now in a separate module whose source is generated
4960 by an auxiliary program - but can then be edited by hand if required. There are
4961 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4962 toupper() in the code.
4963
4964 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4965 make them global. Abolish the function for setting them, as the caller can now
4966 set them directly.
4967
4968
4969 Version 0.92 11-Sep-97
4970 ----------------------
4971
4972 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4973 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4974
4975 2. Caseless matching was not working in character classes if the characters in
4976 the pattern were in upper case.
4977
4978 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4979
4980 4. Make PCRE_ANCHORED public and accept as a compile option.
4981
4982 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4983 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4984 pass them.
4985
4986 6. Give an error if bad option bits passed at compile or run time.
4987
4988 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4989 pcretest to cause it to pass that flag.
4990
4991 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4992 options, and the first character, if set.
4993
4994 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4995
4996
4997 Version 0.91 10-Sep-97
4998 ----------------------
4999
5000 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5001 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5002
5003 2. PCRE was looping on encountering an indefinitely repeated back reference to
5004 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5005 Perl does - treats the match as successful.
5006
5007 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5