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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1284 - (show annotations)
Fri Mar 15 11:54:58 2013 UTC (6 years, 5 months ago) by ph10
File size: 245178 byte(s)
Change backtracking behaviour to "first verb encountered".
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80 be expressed in certain programming languages using Universal Character
81 Names.
82
83 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84 which points out that "non characters" are not "characters that may not
85 appear in Unicode strings" but rather "characters that are reserved for
86 internal use and have only local meaning".
87
88 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89 there was a conditional group that depended on an assertion, if the
90 assertion was false, the callout that immediately followed the alternation
91 in the condition was skipped when pcre_exec() was used for matching.
92
93 23. Allow an explicit callout to be inserted before an assertion that is the
94 condition for a conditional group, for compatibility with automatic
95 callouts, which always insert a callout at this point.
96
97 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99
100 25. (*PRUNE) is now supported by the JIT compiler.
101
102 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103
104 27. Fix the case where there are two or more SKIPs with arguments that may be
105 ignored.
106
107 28. Experimental support of (*SKIP) backtracking verb in the JIT compiler.
108
109 29. Experimental support of (*THEN) backtracking verb in the JIT compiler.
110
111 30. Update RunTest with additional test selector options.
112
113 31. PCRE has been changed to be more compatible with Perl when there is more
114 than one backtracking verb present. Previously, in something like
115 (*COMMIT)(*SKIP), COMMIT would override SKIP. Apart from one anomaly (which
116 has been reported), Perl seems to act on whichever backtracking verb is
117 reached first, so PCRE has been changed to follow this behaviour.
118
119
120 Version 8.32 30-November-2012
121 -----------------------------
122
123 1. Improved JIT compiler optimizations for first character search and single
124 character iterators.
125
126 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
127 Patch by Daniel Richard G.
128
129 3. Single character iterator optimizations in the JIT compiler.
130
131 4. Improved JIT compiler optimizations for character ranges.
132
133 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
134 Reported by Giuseppe D'Angelo.
135
136 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
137 of a line, was being set incorrectly in cases where .* appeared inside
138 atomic brackets at the start of a pattern, or where there was a subsequent
139 *PRUNE or *SKIP.
140
141 7. Improved instruction cache flush for POWER/PowerPC.
142 Patch by Daniel Richard G.
143
144 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
145 grep:
146
147 (a) There is now no limit to the number of patterns to be matched.
148
149 (b) An error is given if a pattern is too long.
150
151 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
152 are now supported.
153
154 (d) --exclude-from and --include-from (multiple use) have been added.
155
156 (e) Exclusions and inclusions now apply to all files and directories, not
157 just to those obtained from scanning a directory recursively.
158
159 (f) Multiple uses of -f and --file-list are now supported.
160
161 (g) In a Windows environment, the default for -d has been changed from
162 "read" (the GNU grep default) to "skip", because otherwise the presence
163 of a directory in the file list provokes an error.
164
165 (h) The documentation has been revised and clarified in places.
166
167 9. Improve the matching speed of capturing brackets.
168
169 10. Changed the meaning of \X so that it now matches a Unicode extended
170 grapheme cluster.
171
172 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
173 out POSIX threads when JIT support is configured.
174
175 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
176
177 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
178 structure could go wrong in environments where size_t is not the same size
179 as int.
180
181 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
182 set.
183
184 15. The EBCDIC support had decayed; later updates to the code had included
185 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
186 general tidy up of EBCDIC-related issues, and the documentation was also
187 not quite right. There is now a test that can be run on ASCII systems to
188 check some of the EBCDIC-related things (but is it not a full test).
189
190 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
191 in a small tidy to the code.
192
193 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
194
195 18. If the --only-matching (-o) option in pcregrep is specified multiple
196 times, each one causes appropriate output. For example, -o1 -o2 outputs the
197 substrings matched by the 1st and 2nd capturing parentheses. A separating
198 string can be specified by --om-separator (default empty).
199
200 19. Improving the first n character searches.
201
202 20. Turn case lists for horizontal and vertical white space into macros so that
203 they are defined only once.
204
205 21. This set of changes together give more compatible Unicode case-folding
206 behaviour for characters that have more than one other case when UCP
207 support is available.
208
209 (a) The Unicode property table now has offsets into a new table of sets of
210 three or more characters that are case-equivalent. The MultiStage2.py
211 script that generates these tables (the pcre_ucd.c file) now scans
212 CaseFolding.txt instead of UnicodeData.txt for character case
213 information.
214
215 (b) The code for adding characters or ranges of characters to a character
216 class has been abstracted into a generalized function that also handles
217 case-independence. In UTF-mode with UCP support, this uses the new data
218 to handle characters with more than one other case.
219
220 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
221 whose other case is greater than 256 are now correctly matched
222 caselessly. Previously, the high codepoint matched the low one, but not
223 vice versa.
224
225 (d) The processing of \h, \H, \v, and \ in character classes now makes use
226 of the new class addition function, using character lists defined as
227 macros alongside the case definitions of 20 above.
228
229 (e) Caseless back references now work with characters that have more than
230 one other case.
231
232 (f) General caseless matching of characters with more than one other case
233 is supported.
234
235 22. Unicode character properties were updated from Unicode 6.2.0
236
237 23. Improved CMake support under Windows. Patch by Daniel Richard G.
238
239 24. Add support for 32-bit character strings, and UTF-32
240
241 25. Major JIT compiler update (code refactoring and bugfixing).
242 Experimental Sparc 32 support is added.
243
244 26. Applied a modified version of Daniel Richard G's patch to create
245 pcre.h.generic and config.h.generic by "make" instead of in the
246 PrepareRelease script.
247
248 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
249 pcre_compile.c when checking for a zero character.
250
251 28. Introducing a native interface for JIT. Through this interface, the compiled
252 machine code can be directly executed. The purpose of this interface is to
253 provide fast pattern matching, so several sanity checks are not performed.
254 However, feature tests are still performed. The new interface provides
255 1.4x speedup compared to the old one.
256
257 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
258 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
259 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
260
261 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
262 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
263 the "old" RFC 2279). Instead, it ended up passing a negative length to
264 pcre_exec().
265
266 31. Add support for GCC's visibility feature to hide internal functions.
267
268 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
269 "unknown -C option" after outputting 0 or 1.
270
271 33. There is now support for generating a code coverage report for the test
272 suite in environments where gcc is the compiler and lcov is installed. This
273 is mainly for the benefit of the developers.
274
275 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
276 unaddressable using valgrind annotations, allowing valgrind to detect
277 invalid memory accesses. This is mainly for the benefit of the developers.
278
279 25. (*UTF) can now be used to start a pattern in any of the three libraries.
280
281 26. Give configure error if --enable-cpp but no C++ compiler found.
282
283
284 Version 8.31 06-July-2012
285 -------------------------
286
287 1. Fixing a wrong JIT test case and some compiler warnings.
288
289 2. Removed a bashism from the RunTest script.
290
291 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
292 to unsigned type, result still unsigned" that was given by an MS compiler
293 on encountering the code "-sizeof(xxx)".
294
295 4. Partial matching support is added to the JIT compiler.
296
297 5. Fixed several bugs concerned with partial matching of items that consist
298 of more than one character:
299
300 (a) /^(..)\1/ did not partially match "aba" because checking references was
301 done on an "all or nothing" basis. This also applied to repeated
302 references.
303
304 (b) \R did not give a hard partial match if \r was found at the end of the
305 subject.
306
307 (c) \X did not give a hard partial match after matching one or more
308 characters at the end of the subject.
309
310 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
311 a partial match for the string "\r".
312
313 (e) When newline was set to CRLF, the metacharacter "." did not recognize
314 a partial match for a CR character at the end of the subject string.
315
316 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
317 running pcretest, the text "(JIT)" added to the output whenever JIT is
318 actually used to run the match.
319
320 7. Individual JIT compile options can be set in pcretest by following -s+[+]
321 or /S+[+] with a digit between 1 and 7.
322
323 8. OP_NOT now supports any UTF character not just single-byte ones.
324
325 9. (*MARK) control verb is now supported by the JIT compiler.
326
327 10. The command "./RunTest list" lists the available tests without actually
328 running any of them. (Because I keep forgetting what they all are.)
329
330 11. Add PCRE_INFO_MAXLOOKBEHIND.
331
332 12. Applied a (slightly modified) user-supplied patch that improves performance
333 when the heap is used for recursion (compiled with --disable-stack-for-
334 recursion). Instead of malloc and free for each heap frame each time a
335 logical recursion happens, frames are retained on a chain and re-used where
336 possible. This sometimes gives as much as 30% improvement.
337
338 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
339 call.
340
341 14. As documented, (*COMMIT) is now confined to within a positive assertion.
342
343 15. It is now possible to link pcretest with libedit as an alternative to
344 libreadline.
345
346 16. (*COMMIT) control verb is now supported by the JIT compiler.
347
348 17. The Unicode data tables have been updated to Unicode 6.1.0.
349
350 18. Added --file-list option to pcregrep.
351
352 19. Added binary file support to pcregrep, including the -a, --binary-files,
353 -I, and --text options.
354
355 20. The madvise function is renamed for posix_madvise for QNX compatibility
356 reasons. Fixed by Giuseppe D'Angelo.
357
358 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
359
360 22. Changed the option for creating symbolic links for 16-bit man pages from
361 -s to -sf so that re-installing does not cause issues.
362
363 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
364
365 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
366 matches in certain environments (the workspace was not being correctly
367 retained). Also added to pcre_dfa_exec() a simple plausibility check on
368 some of the workspace data at the beginning of a restart.
369
370 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
371 was not doing so when it should - probably a typo introduced by SVN 528
372 (change 8.10/14).
373
374 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
375 \w+ when the character tables indicated that \x{c4} was a word character.
376 There were several related cases, all because the tests for doing a table
377 lookup were testing for characters less than 127 instead of 255.
378
379 27. If a pattern contains capturing parentheses that are not used in a match,
380 their slots in the ovector are set to -1. For those that are higher than
381 any matched groups, this happens at the end of processing. In the case when
382 there were back references that the ovector was too small to contain
383 (causing temporary malloc'd memory to be used during matching), and the
384 highest capturing number was not used, memory off the end of the ovector
385 was incorrectly being set to -1. (It was using the size of the temporary
386 memory instead of the true size.)
387
388 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
389 ovector size, it uses memory at the end of the block that it has got.
390
391 29. Check for an overlong MARK name and give an error at compile time. The
392 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
393
394 30. JIT compiler update.
395
396 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
397 Rill for the patch.
398
399 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
400
401 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
402
403 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
404 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
405 was enabled.
406
407 35. Improve JIT code generation for greedy plus quantifier.
408
409 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
410 1 to "aa" instead of to an empty string. The bug affected repeated groups
411 that could potentially match an empty string.
412
413 37. Optimizing single character iterators in JIT.
414
415 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
416 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
417 codepoints that are too big for the mode are faulted, and in a UTF mode,
418 disallowed codepoints are also faulted.
419
420 39. If PCRE was compiled with UTF support, in three places in the DFA
421 matcher there was code that should only have been obeyed in UTF mode, but
422 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
423 processing when bytes with values greater than 127 were present. In 16-bit
424 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
425 both cases the values are those that cannot be the first data item in a UTF
426 character. The three items that might have provoked this were recursions,
427 possessively repeated groups, and atomic groups.
428
429 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
430 and pcregrep, because some OS require shared objects to be explicitly
431 passed to ld, causing the link step to fail if they are not.
432
433 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
434 mode, patterns that started with \h* or \R* might be incorrectly matched.
435
436
437 Version 8.30 04-February-2012
438 -----------------------------
439
440 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
441 name is defined in ctype.h.
442
443 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
444 only in quite long subpatterns.
445
446 3. Removed the function pcre_info(), which has been obsolete and deprecated
447 since it was replaced by pcre_fullinfo() in February 2000.
448
449 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
450 match a (*MARK), and the match failed at the start of the subject, a
451 reference to memory before the start of the subject could occur. This bug
452 was introduced by fix 17 of release 8.21.
453
454 5. A reference to an unset group with zero minimum repetition was giving
455 totally wrong answers (in non-JavaScript-compatibility mode). For example,
456 /(another)?(\1?)test/ matched against "hello world test". This bug was
457 introduced in release 8.13.
458
459 6. Add support for 16-bit character strings (a large amount of work involving
460 many changes and refactorings).
461
462 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
463 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
464 from a file.
465
466 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
467 rounding is not applied in this particular case).
468
469 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
470 if they appear, or are escaped, in patterns.
471
472 10. Get rid of a number of -Wunused-but-set-variable warnings.
473
474 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
475 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
476 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
477 also returns the mark "x". This bug applied to capturing parentheses,
478 non-capturing parentheses, and atomic parentheses. It also applied to some
479 assertions.
480
481 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
482 information out of configure.ac instead of relying on pcre.h.generic, which
483 is not stored in the repository.
484
485 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
486 -lreadline.
487
488 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
489
490 15. Applied Graycode's patch to put the top-level frame on the stack rather
491 than the heap when not using the stack for recursion. This gives a
492 performance improvement in many cases when recursion is not deep.
493
494 16. Experimental code added to "pcretest -C" to output the stack frame size.
495
496
497 Version 8.21 12-Dec-2011
498 ------------------------
499
500 1. Updating the JIT compiler.
501
502 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
503 are added as well.
504
505 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
506 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
507 calling _pcre_jit_exec. Some extra comments are added.
508
509 4. (*MARK) settings inside atomic groups that do not contain any capturing
510 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
511 was introduced by change 18 for 8.20.
512
513 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
514 ECMA-262 standard.
515
516 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
517 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
518 This bug was probably introduced by change 9 of 8.13.
519
520 7. While fixing 6 above, I noticed that a number of other items were being
521 incorrectly rejected as "not fixed length". This arose partly because newer
522 opcodes had not been added to the fixed-length checking code. I have (a)
523 corrected the bug and added tests for these items, and (b) arranged for an
524 error to occur if an unknown opcode is encountered while checking for fixed
525 length instead of just assuming "not fixed length". The items that were
526 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
527 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
528 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
529
530 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
531 being incorrectly compiled and would have given unpredicatble results.
532
533 9. A possessively repeated subpattern with minimum repeat count greater than
534 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
535 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
536 the first (A) could occur when it should not.
537
538 10. Add a cast and remove a redundant test from the code.
539
540 11. JIT should use pcre_malloc/pcre_free for allocation.
541
542 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
543 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
544 is anything other than /usr, -L is still shown).
545
546 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
547
548 14. Perl does not support \N without a following name in a [] class; PCRE now
549 also gives an error.
550
551 15. If a forward reference was repeated with an upper limit of around 2000,
552 it caused the error "internal error: overran compiling workspace". The
553 maximum number of forward references (including repeats) was limited by the
554 internal workspace, and dependent on the LINK_SIZE. The code has been
555 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
556 the default depends on LINK_SIZE. There is a new upper limit (for safety)
557 of around 200,000 forward references. While doing this, I also speeded up
558 the filling in of repeated forward references.
559
560 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
561 incorrectly expecting the subject to contain another "a" after the start.
562
563 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
564 in the match, the SKIP should be ignored. This was not happening; instead
565 the SKIP was being treated as NOMATCH. For patterns such as
566 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
567 tested.
568
569 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
570 now much more compatible with Perl, in particular in cases where the result
571 is a non-match for a non-anchored pattern. For example, if
572 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
573 "m", where previously it did not return a name. A side effect of this
574 change is that for partial matches, the last encountered mark name is
575 returned, as for non matches. A number of tests that were previously not
576 Perl-compatible have been moved into the Perl-compatible test files. The
577 refactoring has had the pleasing side effect of removing one argument from
578 the match() function, thus reducing its stack requirements.
579
580 19. If the /S+ option was used in pcretest to study a pattern using JIT,
581 subsequent uses of /S (without +) incorrectly behaved like /S+.
582
583 21. Retrieve executable code size support for the JIT compiler and fixing
584 some warnings.
585
586 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
587 not work when the shorter character appeared right at the end of the
588 subject string.
589
590 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
591 systems.
592
593 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
594 output it when the /M option is used in pcretest.
595
596 25. The CheckMan script was not being included in the distribution. Also, added
597 an explicit "perl" to run Perl scripts from the PrepareRelease script
598 because this is reportedly needed in Windows.
599
600 26. If study data was being save in a file and studying had not found a set of
601 "starts with" bytes for the pattern, the data written to the file (though
602 never used) was taken from uninitialized memory and so caused valgrind to
603 complain.
604
605 27. Updated RunTest.bat as provided by Sheri Pierce.
606
607 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
608
609 29. Computation of memory usage for the table of capturing group names was
610 giving an unnecessarily large value.
611
612
613 Version 8.20 21-Oct-2011
614 ------------------------
615
616 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
617 a POSIX class. After further experiments with Perl, which convinced me that
618 Perl has bugs and confusions, a closing square bracket is no longer allowed
619 in a POSIX name. This bug also affected patterns with classes that started
620 with full stops.
621
622 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
623 captured substring, but while checking the failing first alternative,
624 substring 1 is temporarily captured. If the output vector supplied to
625 pcre_exec() was not big enough for this capture, the yield of the function
626 was still zero ("insufficient space for captured substrings"). This cannot
627 be totally fixed without adding another stack variable, which seems a lot
628 of expense for a edge case. However, I have improved the situation in cases
629 such as /(a)(b)x|abc/ matched against "abc", where the return code
630 indicates that fewer than the maximum number of slots in the ovector have
631 been set.
632
633 3. Related to (2) above: when there are more back references in a pattern than
634 slots in the output vector, pcre_exec() uses temporary memory during
635 matching, and copies in the captures as far as possible afterwards. It was
636 using the entire output vector, but this conflicts with the specification
637 that only 2/3 is used for passing back captured substrings. Now it uses
638 only the first 2/3, for compatibility. This is, of course, another edge
639 case.
640
641 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
642 main code base, and can be used by building with --enable-jit. When this is
643 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
644 runtime --no-jit option is given.
645
646 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
647 ovector, the return from the function was zero, implying that there were
648 other matches that did not fit. The correct "exactly full" value is now
649 returned.
650
651 6. If a subpattern that was called recursively or as a subroutine contained
652 (*PRUNE) or any other control that caused it to give a non-standard return,
653 invalid errors such as "Error -26 (nested recursion at the same subject
654 position)" or even infinite loops could occur.
655
656 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
657 computing the minimum length on reaching *ACCEPT, and so ended up with the
658 wrong value of 1 rather than 0. Further investigation indicates that
659 computing a minimum subject length in the presence of *ACCEPT is difficult
660 (think back references, subroutine calls), and so I have changed the code
661 so that no minimum is registered for a pattern that contains *ACCEPT.
662
663 8. If (*THEN) was present in the first (true) branch of a conditional group,
664 it was not handled as intended. [But see 16 below.]
665
666 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
667 Sheri Pierce.
668
669 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
670 the first byte in a match must be "a".
671
672 11. Change 17 for 8.13 increased the recursion depth for patterns like
673 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
674 pattern contains any instances of (*THEN). If it does not, the old
675 optimizations are restored. It would be nice to do this on a per-group
676 basis, but at the moment that is not feasible.
677
678 12. In some environments, the output of pcretest -C is CRLF terminated. This
679 broke RunTest's code that checks for the link size. A single white space
680 character after the value is now allowed for.
681
682 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
683 For "fr", it uses the Windows-specific input and output files.
684
685 14. If (*THEN) appeared in a group that was called recursively or as a
686 subroutine, it did not work as intended. [But see next item.]
687
688 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
689 pattern fragments (but not containing any | characters). If A and B are
690 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
691 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
692 D. In other words, Perl considers parentheses that do not contain any |
693 characters to be part of a surrounding alternative, whereas PCRE was
694 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
695 differently. PCRE now behaves in the same way as Perl, except in the case
696 of subroutine/recursion calls such as (?1) which have in any case always
697 been different (but PCRE had them first :-).
698
699 16. Related to 15 above: Perl does not treat the | in a conditional group as
700 creating alternatives. Such a group is treated in the same way as an
701 ordinary group without any | characters when processing (*THEN). PCRE has
702 been changed to match Perl's behaviour.
703
704 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
705 RunGrepTest script failed.
706
707 18. Change 22 for version 13 caused atomic groups to use more stack. This is
708 inevitable for groups that contain captures, but it can lead to a lot of
709 stack use in large patterns. The old behaviour has been restored for atomic
710 groups that do not contain any capturing parentheses.
711
712 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
713 suppress the check for a minimum subject length at run time. (If it was
714 given to pcre_exec() or pcre_dfa_exec() it did work.)
715
716 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
717 fail to work when decoding hex characters in data strings in EBCDIC
718 environments.
719
720 21. It appears that in at least one Mac OS environment, the isxdigit() function
721 is implemented as a macro that evaluates to its argument more than once,
722 contravening the C 90 Standard (I haven't checked a later standard). There
723 was an instance in pcretest which caused it to go wrong when processing
724 \x{...} escapes in subject strings. The has been rewritten to avoid using
725 things like p++ in the argument of isxdigit().
726
727
728 Version 8.13 16-Aug-2011
729 ------------------------
730
731 1. The Unicode data tables have been updated to Unicode 6.0.0.
732
733 2. Two minor typos in pcre_internal.h have been fixed.
734
735 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
736 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
737 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
738
739 4. There were a number of related bugs in the code for matching backrefences
740 caselessly in UTF-8 mode when codes for the characters concerned were
741 different numbers of bytes. For example, U+023A and U+2C65 are an upper
742 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
743 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
744 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
745 2-byte code at the end of the subject (it thought there wasn't enough data
746 left).
747
748 5. Comprehensive information about what went wrong is now returned by
749 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
750 as the output vector has at least 2 elements. The offset of the start of
751 the failing character and a reason code are placed in the vector.
752
753 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
754 now returned is for the first byte of the failing character, instead of the
755 last byte inspected. This is an incompatible change, but I hope it is small
756 enough not to be a problem. It makes the returned offset consistent with
757 pcre_exec() and pcre_dfa_exec().
758
759 7. pcretest now gives a text phrase as well as the error number when
760 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
761 failure, the offset and reason code are output.
762
763 8. When \R was used with a maximizing quantifier it failed to skip backwards
764 over a \r\n pair if the subsequent match failed. Instead, it just skipped
765 back over a single character (\n). This seems wrong (because it treated the
766 two characters as a single entity when going forwards), conflicts with the
767 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
768 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
769 has been changed.
770
771 9. Some internal refactoring has changed the processing so that the handling
772 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
773 time (the PCRE_DOTALL option was changed this way some time ago: version
774 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
775 which was always a bit of a fudge. It also means that there is one less
776 argument for the match() function, which reduces its stack requirements
777 slightly. This change also fixes an incompatibility with Perl: the pattern
778 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
779
780 10. More internal refactoring has drastically reduced the number of recursive
781 calls to match() for possessively repeated groups such as (abc)++ when
782 using pcre_exec().
783
784 11. While implementing 10, a number of bugs in the handling of groups were
785 discovered and fixed:
786
787 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
788 (a|)*(?1) gave a compile-time internal error.
789 ((a|)+)+ did not notice that the outer group could match an empty string.
790 (^a|^)+ was not marked as anchored.
791 (.*a|.*)+ was not marked as matching at start or after a newline.
792
793 12. Yet more internal refactoring has removed another argument from the match()
794 function. Special calls to this function are now indicated by setting a
795 value in a variable in the "match data" data block.
796
797 13. Be more explicit in pcre_study() instead of relying on "default" for
798 opcodes that mean there is no starting character; this means that when new
799 ones are added and accidentally left out of pcre_study(), testing should
800 pick them up.
801
802 14. The -s option of pcretest has been documented for ages as being an old
803 synonym of -m (show memory usage). I have changed it to mean "force study
804 for every regex", that is, assume /S for every regex. This is similar to -i
805 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
806 using it. It makes it easier to run collections of tests with and without
807 study enabled, and thereby test pcre_study() more easily. All the standard
808 tests are now run with and without -s (but some patterns can be marked as
809 "never study" - see 20 below).
810
811 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
812 restoration of the capturing data to the outer values was not happening
813 correctly.
814
815 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
816 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
817 pattern had matched an empty string, and so incorrectly returned a no
818 match.
819
820 17. There was optimizing code for the last branch of non-capturing parentheses,
821 and also for the obeyed branch of a conditional subexpression, which used
822 tail recursion to cut down on stack usage. Unfortunately, now that there is
823 the possibility of (*THEN) occurring in these branches, tail recursion is
824 no longer possible because the return has to be checked for (*THEN). These
825 two optimizations have therefore been removed. [But see 8.20/11 above.]
826
827 18. If a pattern containing \R was studied, it was assumed that \R always
828 matched two bytes, thus causing the minimum subject length to be
829 incorrectly computed because \R can also match just one byte.
830
831 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
832 was incorrectly computed.
833
834 20. If /S is present twice on a test pattern in pcretest input, it now
835 *disables* studying, thereby overriding the use of -s on the command line
836 (see 14 above). This is necessary for one or two tests to keep the output
837 identical in both cases.
838
839 21. When (*ACCEPT) was used in an assertion that matched an empty string and
840 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
841
842 22. When an atomic group that contained a capturing parenthesis was
843 successfully matched, but the branch in which it appeared failed, the
844 capturing was not being forgotten if a higher numbered group was later
845 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
846 group 1 to "a", when in fact it should be unset. This applied to multi-
847 branched capturing and non-capturing groups, repeated or not, and also to
848 positive assertions (capturing in negative assertions does not happen
849 in PCRE) and also to nested atomic groups.
850
851 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
852 subject after a captured substring, to make it easier to tell which of a
853 number of identical substrings has been captured.
854
855 24. The way atomic groups are processed by pcre_exec() has been changed so that
856 if they are repeated, backtracking one repetition now resets captured
857 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
858 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
859 "aaa". Previously, it would have been "a". As part of this code
860 refactoring, the way recursive calls are handled has also been changed.
861
862 25. If an assertion condition captured any substrings, they were not passed
863 back unless some other capturing happened later. For example, if
864 (?(?=(a))a) was matched against "a", no capturing was returned.
865
866 26. When studying a pattern that contained subroutine calls or assertions,
867 the code for finding the minimum length of a possible match was handling
868 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
869 group 1 called group 2 while simultaneously a separate group 2 called group
870 1). A stack overflow occurred in this case. I have fixed this by limiting
871 the recursion depth to 10.
872
873 27. Updated RunTest.bat in the distribution to the version supplied by Tom
874 Fortmann. This supports explicit test numbers on the command line, and has
875 argument validation and error reporting.
876
877 28. An instance of \X with an unlimited repeat could fail if at any point the
878 first character it looked at was a mark character.
879
880 29. Some minor code refactoring concerning Unicode properties and scripts
881 should reduce the stack requirement of match() slightly.
882
883 30. Added the '=' option to pcretest to check the setting of unused capturing
884 slots at the end of the pattern, which are documented as being -1, but are
885 not included in the return count.
886
887 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
888 compiled something random. Now it gives a compile-time error (as does
889 Perl).
890
891 32. A *MARK encountered during the processing of a positive assertion is now
892 recorded and passed back (compatible with Perl).
893
894 33. If --only-matching or --colour was set on a pcregrep call whose pattern
895 had alternative anchored branches, the search for a second match in a line
896 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
897 matched the line "0102" twice. The same bug affected patterns that started
898 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
899 twice.
900
901 34. Previously, PCRE did not allow quantification of assertions. However, Perl
902 does, and because of capturing effects, quantifying parenthesized
903 assertions may at times be useful. Quantifiers are now allowed for
904 parenthesized assertions.
905
906 35. A minor code tidy in pcre_compile() when checking options for \R usage.
907
908 36. \g was being checked for fancy things in a character class, when it should
909 just be a literal "g".
910
911 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
912 appearance of a nested POSIX class supersedes an apparent external class.
913 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
914 unescaped square brackets may also appear as part of class names. For
915 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
916 more like Perl. (But see 8.20/1 above.)
917
918 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
919 was because it thought it was \N{name}, which is not supported).
920
921 39. Add minix to OS list not supporting the -S option in pcretest.
922
923 40. PCRE tries to detect cases of infinite recursion at compile time, but it
924 cannot analyze patterns in sufficient detail to catch mutual recursions
925 such as ((?1))((?2)). There is now a runtime test that gives an error if a
926 subgroup is called recursively as a subpattern for a second time at the
927 same position in the subject string. In previous releases this might have
928 been caught by the recursion limit, or it might have run out of stack.
929
930 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
931 happen only once. PCRE was, however incorrectly giving a compile time error
932 "recursive call could loop indefinitely" because it cannot analyze the
933 pattern in sufficient detail. The compile time test no longer happens when
934 PCRE is compiling a conditional subpattern, but actual runaway loops are
935 now caught at runtime (see 40 above).
936
937 42. It seems that Perl allows any characters other than a closing parenthesis
938 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
939 has been changed to be the same.
940
941 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
942 as not to get warnings when autogen.sh is called. Also changed
943 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
944
945 44. To help people who use pcregrep to scan files containing exceedingly long
946 lines, the following changes have been made:
947
948 (a) The default value of the buffer size parameter has been increased from
949 8K to 20K. (The actual buffer used is three times this size.)
950
951 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
952 PCRE is built.
953
954 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
955 to be set at run time.
956
957 (d) Numerical values in pcregrep options can be followed by K or M, for
958 example --buffer-size=50K.
959
960 (e) If a line being scanned overflows pcregrep's buffer, an error is now
961 given and the return code is set to 2.
962
963 45. Add a pointer to the latest mark to the callout data block.
964
965 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
966 partial match of an empty string instead of no match. This was specific to
967 the use of ".".
968
969 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
970 complete match instead of a partial match. This bug was dependent on both
971 the PCRE_UTF8 and PCRE_DOTALL options being set.
972
973 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
974 starting byte set, because \b was not being ignored.
975
976
977 Version 8.12 15-Jan-2011
978 ------------------------
979
980 1. Fixed some typos in the markup of the man pages, and wrote a script that
981 checks for such things as part of the documentation building process.
982
983 2. On a big-endian 64-bit system, pcregrep did not correctly process the
984 --match-limit and --recursion-limit options (added for 8.11). In
985 particular, this made one of the standard tests fail. (The integer value
986 went into the wrong half of a long int.)
987
988 3. If the --colour option was given to pcregrep with -v (invert match), it
989 did strange things, either producing crazy output, or crashing. It should,
990 of course, ignore a request for colour when reporting lines that do not
991 match.
992
993 4. Another pcregrep bug caused similar problems if --colour was specified with
994 -M (multiline) and the pattern match finished with a line ending.
995
996 5. In pcregrep, when a pattern that ended with a literal newline sequence was
997 matched in multiline mode, the following line was shown as part of the
998 match. This seems wrong, so I have changed it.
999
1000 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1001 the check for further matches in the same line (so they could be coloured)
1002 to overrun the end of the current line. If another match was found, it was
1003 incorrectly shown (and then shown again when found in the next line).
1004
1005 7. If pcregrep was compiled under Windows, there was a reference to the
1006 function pcregrep_exit() before it was defined. I am assuming this was
1007 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1008 reported by a user. I've moved the definition above the reference.
1009
1010
1011 Version 8.11 10-Dec-2010
1012 ------------------------
1013
1014 1. (*THEN) was not working properly if there were untried alternatives prior
1015 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1016 backtracked to try for "b" instead of moving to the next alternative branch
1017 at the same level (in this case, to look for "c"). The Perl documentation
1018 is clear that when (*THEN) is backtracked onto, it goes to the "next
1019 alternative in the innermost enclosing group".
1020
1021 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1022 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1023 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1024 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1025 (*THEN).
1026
1027 3. If \s appeared in a character class, it removed the VT character from
1028 the class, even if it had been included by some previous item, for example
1029 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1030 of \s, but is part of the POSIX "space" class.)
1031
1032 4. A partial match never returns an empty string (because you can always
1033 match an empty string at the end of the subject); however the checking for
1034 an empty string was starting at the "start of match" point. This has been
1035 changed to the "earliest inspected character" point, because the returned
1036 data for a partial match starts at this character. This means that, for
1037 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1038 (previously it gave "no match").
1039
1040 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1041 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1042 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1043 has an implication that the given string is incomplete (because a partial
1044 match is preferred over a full match). For this reason, these items now
1045 give a partial match in this situation. [Aside: previously, the one case
1046 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1047 match rather than a full match, which was wrong by the old rules, but is
1048 now correct.]
1049
1050 6. There was a bug in the handling of #-introduced comments, recognized when
1051 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1052 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1053 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1054 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1055 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1056 places in pcre_compile().
1057
1058 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1059 comments when looking ahead for named forward references to subpatterns,
1060 the only newline sequence it recognized was NL. It now handles newlines
1061 according to the set newline convention.
1062
1063 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1064 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1065 cater for a lack of strerror(). These oversights have been fixed.
1066
1067 9. Added --match-limit and --recursion-limit to pcregrep.
1068
1069 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1070
1071 11. When the -o option was used, pcregrep was setting a return code of 1, even
1072 when matches were found, and --line-buffered was not being honoured.
1073
1074 12. Added an optional parentheses number to the -o and --only-matching options
1075 of pcregrep.
1076
1077 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1078 can match an empty string. The code to do it in pcretest and pcredemo
1079 needed fixing:
1080
1081 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1082 only one byte after an empty string match just before CRLF (this case
1083 just got forgotten; "any" and "anycrlf" were OK).
1084
1085 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1086 mode when an empty string match preceded an ASCII character followed by
1087 a non-ASCII character. (The code for advancing by one character rather
1088 than one byte was nonsense.)
1089
1090 (c) The pcredemo.c sample program did not have any code at all to handle
1091 the cases when CRLF is a valid newline sequence.
1092
1093 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1094 as a starting offset was within the subject string. There is now a new
1095 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1096 negative or greater than the length of the string. In order to test this,
1097 pcretest is extended to allow the setting of negative starting offsets.
1098
1099 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1100 starting offset points to the beginning of a UTF-8 character was
1101 unnecessarily clumsy. I tidied it up.
1102
1103 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1104 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1105
1106 17. Nobody had reported that the --include_dir option, which was added in
1107 release 7.7 should have been called --include-dir (hyphen, not underscore)
1108 for compatibility with GNU grep. I have changed it to --include-dir, but
1109 left --include_dir as an undocumented synonym, and the same for
1110 --exclude-dir, though that is not available in GNU grep, at least as of
1111 release 2.5.4.
1112
1113 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1114 characters from a string of bytes) have been redefined so as not to use
1115 loops, in order to improve performance in some environments. At the same
1116 time, I abstracted some of the common code into auxiliary macros to save
1117 repetition (this should not affect the compiled code).
1118
1119 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1120 compile-time error is now given if \c is not followed by an ASCII
1121 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1122 different, and any byte value is allowed.)
1123
1124 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1125 START_OPTIMIZE option, which is now allowed at compile time - but just
1126 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1127 to pcregrep and other applications that have no direct access to PCRE
1128 options. The new /Y option in pcretest sets this option when calling
1129 pcre_compile().
1130
1131 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1132 back references. Groups containing recursive back references were forced to
1133 be atomic by that change, but in the case of named groups, the amount of
1134 memory required was incorrectly computed, leading to "Failed: internal
1135 error: code overflow". This has been fixed.
1136
1137 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1138 pcretest.c, to avoid build problems in some Borland environments.
1139
1140
1141 Version 8.10 25-Jun-2010
1142 ------------------------
1143
1144 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1145 THEN.
1146
1147 2. (*ACCEPT) was not working when inside an atomic group.
1148
1149 3. Inside a character class, \B is treated as a literal by default, but
1150 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1151 causes the error). The code is unchanged, but I tidied the documentation.
1152
1153 4. Inside a character class, PCRE always treated \R and \X as literals,
1154 whereas Perl faults them if its -w option is set. I have changed PCRE so
1155 that it faults them when PCRE_EXTRA is set.
1156
1157 5. Added support for \N, which always matches any character other than
1158 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1159
1160 6. When compiling pcregrep with newer versions of gcc which may have
1161 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1162 declared with attribute warn_unused_result" were given. Just casting the
1163 result to (void) does not stop the warnings; a more elaborate fudge is
1164 needed. I've used a macro to implement this.
1165
1166 7. Minor change to pcretest.c to avoid a compiler warning.
1167
1168 8. Added four artifical Unicode properties to help with an option to make
1169 \s etc use properties (see next item). The new properties are: Xan
1170 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1171
1172 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1173 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1174 this option. Modified pcretest to add /W to test this facility. Added
1175 REG_UCP to make it available via the POSIX interface.
1176
1177 10. Added --line-buffered to pcregrep.
1178
1179 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1180 studied, and the match started with a letter with a code point greater than
1181 127 whose first byte was different to the first byte of the other case of
1182 the letter, the other case of this starting letter was not recognized
1183 (#976).
1184
1185 12. If a pattern that was studied started with a repeated Unicode property
1186 test, for example, \p{Nd}+, there was the theoretical possibility of
1187 setting up an incorrect bitmap of starting bytes, but fortunately it could
1188 not have actually happened in practice until change 8 above was made (it
1189 added property types that matched character-matching opcodes).
1190
1191 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1192 possible starting bytes for non-anchored patterns.
1193
1194 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1195 \R, and also a number of cases that involve Unicode properties, both
1196 explicit and implicit when PCRE_UCP is set.
1197
1198 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1199 input, it could crash or give wrong results if characters with values
1200 greater than 0xc0 were present in the subject string. (Detail: it assumed
1201 UTF-8 input when processing these items.)
1202
1203 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1204 size_t is 64-bit (#991).
1205
1206 17. Added a check for running out of memory when PCRE is compiled with
1207 --disable-stack-for-recursion (#990).
1208
1209 18. If the last data line in a file for pcretest does not have a newline on
1210 the end, a newline was missing in the output.
1211
1212 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1213 less than 128) in its various bitmaps. However, there is a facility for
1214 generating tables according to the current locale when PCRE is compiled. It
1215 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1216 characters, are recognized by isspace() and therefore were getting set in
1217 these tables, and indeed these tables seem to approximate to ISO 8859. This
1218 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1219 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1220 which of course cannot start UTF-8 characters. I have changed the code so
1221 that only real ASCII characters (less than 128) and the correct starting
1222 bytes for UTF-8 encodings are set for characters greater than 127 when in
1223 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1224 altogether.)
1225
1226 20. Added the /T option to pcretest so as to be able to run tests with non-
1227 standard character tables, thus making it possible to include the tests
1228 used for 19 above in the standard set of tests.
1229
1230 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1231 reference to a subpattern the other side of a comment that contains an
1232 opening parenthesis caused either an internal compiling error, or a
1233 reference to the wrong subpattern.
1234
1235
1236 Version 8.02 19-Mar-2010
1237 ------------------------
1238
1239 1. The Unicode data tables have been updated to Unicode 5.2.0.
1240
1241 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1242 configured.
1243
1244 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1245 original author of that file, following a query about its status.
1246
1247 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1248 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1249
1250 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1251 quantifier applied to a forward-referencing subroutine call, could compile
1252 incorrect code or give the error "internal error: previously-checked
1253 referenced subpattern not found".
1254
1255 6. Both MS Visual Studio and Symbian OS have problems with initializing
1256 variables to point to external functions. For these systems, therefore,
1257 pcre_malloc etc. are now initialized to local functions that call the
1258 relevant global functions.
1259
1260 7. There were two entries missing in the vectors called coptable and poptable
1261 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1262 I've fixed the data, and added a kludgy way of testing at compile time that
1263 the lengths are correct (equal to the number of opcodes).
1264
1265 8. Following on from 7, I added a similar kludge to check the length of the
1266 eint vector in pcreposix.c.
1267
1268 9. Error texts for pcre_compile() are held as one long string to avoid too
1269 much relocation at load time. To find a text, the string is searched,
1270 counting zeros. There was no check for running off the end of the string,
1271 which could happen if a new error number was added without updating the
1272 string.
1273
1274 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1275
1276 11. \K was not working if it appeared in an atomic group or in a group that
1277 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1278 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1279 the assertion is positive, but not if it is negative.
1280
1281 12. Change 11 fortuitously reduced the size of the stack frame used in the
1282 "match()" function of pcre_exec.c by one pointer. Forthcoming
1283 implementation of support for (*MARK) will need an extra pointer on the
1284 stack; I have reserved it now, so that the stack frame size does not
1285 decrease.
1286
1287 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1288 item in branch that calls a recursion is a subroutine call - as in the
1289 second branch in the above example - was incorrectly given the compile-
1290 time error "recursive call could loop indefinitely" because pcre_compile()
1291 was not correctly checking the subroutine for matching a non-empty string.
1292
1293 14. The checks for overrunning compiling workspace could trigger after an
1294 overrun had occurred. This is a "should never occur" error, but it can be
1295 triggered by pathological patterns such as hundreds of nested parentheses.
1296 The checks now trigger 100 bytes before the end of the workspace.
1297
1298 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1299
1300
1301 Version 8.01 19-Jan-2010
1302 ------------------------
1303
1304 1. If a pattern contained a conditional subpattern with only one branch (in
1305 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1306 computed the wrong minimum data length (which is of course zero for such
1307 subpatterns). This could cause incorrect "no match" results.
1308
1309 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1310 the pattern is reset in the first branch, pcre_compile() failed with
1311 "internal error: code overflow at offset...". This happened only when
1312 the reset was to the original external option setting. (An optimization
1313 abstracts leading options settings into an external setting, which was the
1314 cause of this.)
1315
1316 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1317 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1318 assertion pattern did not match (meaning that the assertion was true), it
1319 was incorrectly treated as false if the SKIP had been reached during the
1320 matching. This also applied to assertions used as conditions.
1321
1322 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1323 assertion subpattern, including such a pattern used as a condition,
1324 unpredictable results occurred, instead of the error return
1325 PCRE_ERROR_DFA_UITEM.
1326
1327 5. The C++ GlobalReplace function was not working like Perl for the special
1328 situation when an empty string is matched. It now does the fancy magic
1329 stuff that is necessary.
1330
1331 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1332 removed. (These were left over from very, very early versions of PCRE.)
1333
1334 7. Some cosmetic changes to the code to make life easier when compiling it
1335 as part of something else:
1336
1337 (a) Change DEBUG to PCRE_DEBUG.
1338
1339 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1340 called "current" as "current_branch", to prevent a collision with the
1341 Linux macro when compiled as a kernel module.
1342
1343 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1344 prevent a collision with the Linux macro when compiled as a kernel
1345 module.
1346
1347 8. In pcre_compile() there are some checks for integer overflows that used to
1348 cast potentially large values to (double). This has been changed to that
1349 when building, a check for int64_t is made, and if it is found, it is used
1350 instead, thus avoiding the use of floating point arithmetic. (There is no
1351 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1352 double.
1353
1354 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1355 2005 (difference between two addresses compared to an unsigned value).
1356
1357 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1358 custom one, because of the following reported problem in Windows:
1359
1360 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1361 under Win32.
1362 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1363 therefore missing the function definition.
1364 - The compiler thus generates a "C" signature for the test function.
1365 - The linker fails to find the "C" function.
1366 - PCRE fails to configure if asked to do so against libbz2.
1367
1368 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1369 messages were output:
1370
1371 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1372 rerunning libtoolize, to keep the correct libtool macros in-tree.
1373 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1374
1375 I have done both of these things.
1376
1377 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1378 most of the time, it *can* run out if it is given a pattern that contains a
1379 runaway infinite recursion. I updated the discussion in the pcrestack man
1380 page.
1381
1382 13. Now that we have gone to the x.xx style of version numbers, the minor
1383 version may start with zero. Using 08 or 09 is a bad idea because users
1384 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1385 interpreted as invalid octal numbers. I've updated the previous comment in
1386 configure.ac, and also added a check that gives an error if 08 or 09 are
1387 used.
1388
1389 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1390 causing partial matching to fail when the end of the subject matched \W
1391 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1392
1393 15. There were some discrepancies between the declarations in pcre_internal.h
1394 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1395 their definitions. The declarations used "const uschar *" and the
1396 definitions used USPTR. Even though USPTR is normally defined as "const
1397 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1398 reported that: "This difference in casting confuses some C++ compilers, for
1399 example, SunCC recognizes above declarations as different functions and
1400 generates broken code for hbpcre." I have changed the declarations to use
1401 USPTR.
1402
1403 16. GNU libtool is named differently on some systems. The autogen.sh script now
1404 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1405 (FreeBSD).
1406
1407 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1408 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1409 comment: "Figure out how to create a longlong from a string: strtoll and
1410 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1411 instance, but it only takes 2 args instead of 3!"
1412
1413 18. A subtle bug concerned with back references has been fixed by a change of
1414 specification, with a corresponding code fix. A pattern such as
1415 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1416 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1417 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1418 same bug. Such groups have to be quantified to be useful, or contained
1419 inside another quantified group. (If there's no repetition, the reference
1420 can never match.) The problem arises because, having left the group and
1421 moved on to the rest of the pattern, a later failure that backtracks into
1422 the group uses the captured value from the final iteration of the group
1423 rather than the correct earlier one. I have fixed this in PCRE by forcing
1424 any group that contains a reference to itself to be an atomic group; that
1425 is, there cannot be any backtracking into it once it has completed. This is
1426 similar to recursive and subroutine calls.
1427
1428
1429 Version 8.00 19-Oct-09
1430 ----------------------
1431
1432 1. The table for translating pcre_compile() error codes into POSIX error codes
1433 was out-of-date, and there was no check on the pcre_compile() error code
1434 being within the table. This could lead to an OK return being given in
1435 error.
1436
1437 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1438 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1439 in a Windows environment.
1440
1441 3. The pcregrep --count option prints the count for each file even when it is
1442 zero, as does GNU grep. However, pcregrep was also printing all files when
1443 --files-with-matches was added. Now, when both options are given, it prints
1444 counts only for those files that have at least one match. (GNU grep just
1445 prints the file name in this circumstance, but including the count seems
1446 more useful - otherwise, why use --count?) Also ensured that the
1447 combination -clh just lists non-zero counts, with no names.
1448
1449 4. The long form of the pcregrep -F option was incorrectly implemented as
1450 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1451 but it seems right to fix it, and I didn't think it was worth preserving
1452 the old behaviour.
1453
1454 5. The command line items --regex=pattern and --regexp=pattern were not
1455 recognized by pcregrep, which required --regex pattern or --regexp pattern
1456 (with a space rather than an '='). The man page documented the '=' forms,
1457 which are compatible with GNU grep; these now work.
1458
1459 6. No libpcreposix.pc file was created for pkg-config; there was just
1460 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1461
1462 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1463 when UCP support is not needed, by modifying the Python script that
1464 generates it from Unicode data files. This should not matter if the module
1465 is correctly used as a library, but I received one complaint about 50K of
1466 unwanted data. My guess is that the person linked everything into his
1467 program rather than using a library. Anyway, it does no harm.
1468
1469 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1470 was a minimum greater than 1 for a wide character in a possessive
1471 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1472 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1473 character. Chaos in the form of incorrect output or a compiling loop could
1474 result.
1475
1476 9. The restrictions on what a pattern can contain when partial matching is
1477 requested for pcre_exec() have been removed. All patterns can now be
1478 partially matched by this function. In addition, if there are at least two
1479 slots in the offset vector, the offset of the earliest inspected character
1480 for the match and the offset of the end of the subject are set in them when
1481 PCRE_ERROR_PARTIAL is returned.
1482
1483 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1484 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1485 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1486 and may be more useful for multi-segment matching.
1487
1488 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1489 used to be given if ever the end of the subject was reached; now it is
1490 given only if matching could not proceed because another character was
1491 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1492 string "Z", which now yields "no match" instead of "partial match". In the
1493 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1494 final character ended with (*FAIL).
1495
1496 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1497 if the pattern had a "must contain" character that was already found in the
1498 earlier partial match, unless partial matching was again requested. For
1499 example, with the pattern /dog.(body)?/, the "must contain" character is
1500 "g". If the first part-match was for the string "dog", restarting with
1501 "sbody" failed. This bug has been fixed.
1502
1503 13. The string returned by pcre_dfa_exec() after a partial match has been
1504 changed so that it starts at the first inspected character rather than the
1505 first character of the match. This makes a difference only if the pattern
1506 starts with a lookbehind assertion or \b or \B (\K is not supported by
1507 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1508 matching functions compatible, and I think it's the right thing to do.
1509
1510 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1511 so that the demonstration program is easily available in environments where
1512 PCRE has not been installed from source.
1513
1514 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1515 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1516 library.
1517
1518 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1519 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1520 is not the first non-POSIX option to be added. Clearly some people find
1521 these options useful.
1522
1523 17. If a caller to the POSIX matching function regexec() passes a non-zero
1524 value for nmatch with a NULL value for pmatch, the value of
1525 nmatch is forced to zero.
1526
1527 18. RunGrepTest did not have a test for the availability of the -u option of
1528 the diff command, as RunTest does. It now checks in the same way as
1529 RunTest, and also checks for the -b option.
1530
1531 19. If an odd number of negated classes containing just a single character
1532 interposed, within parentheses, between a forward reference to a named
1533 subpattern and the definition of the subpattern, compilation crashed with
1534 an internal error, complaining that it could not find the referenced
1535 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1536 [The bug was that it was starting one character too far in when skipping
1537 over the character class, thus treating the ] as data rather than
1538 terminating the class. This meant it could skip too much.]
1539
1540 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1541 /g option in pcretest when the pattern contains \K, which makes it possible
1542 to have an empty string match not at the start, even when the pattern is
1543 anchored. Updated pcretest and pcredemo to use this option.
1544
1545 21. If the maximum number of capturing subpatterns in a recursion was greater
1546 than the maximum at the outer level, the higher number was returned, but
1547 with unset values at the outer level. The correct (outer level) value is
1548 now given.
1549
1550 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1551 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1552 make it do so. The string so far is captured, making this feature
1553 compatible with Perl.
1554
1555 23. The tests have been re-organized, adding tests 11 and 12, to make it
1556 possible to check the Perl 5.10 features against Perl 5.10.
1557
1558 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1559 pattern matches a fixed length string. PCRE did not allow this; now it
1560 does. Neither allows recursion.
1561
1562 25. I finally figured out how to implement a request to provide the minimum
1563 length of subject string that was needed in order to match a given pattern.
1564 (It was back references and recursion that I had previously got hung up
1565 on.) This code has now been added to pcre_study(); it finds a lower bound
1566 to the length of subject needed. It is not necessarily the greatest lower
1567 bound, but using it to avoid searching strings that are too short does give
1568 some useful speed-ups. The value is available to calling programs via
1569 pcre_fullinfo().
1570
1571 26. While implementing 25, I discovered to my embarrassment that pcretest had
1572 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1573 study optimizations had never been tested with that matching function.
1574 Oops. What is worse, even when it was passed study data, there was a bug in
1575 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1576 were also very few tests of studied patterns with pcre_dfa_exec().
1577
1578 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1579 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1580 on the other side of the coin, they are no longer allowed to have different
1581 names, because these cannot be distinguished in PCRE, and this has caused
1582 confusion. (This is a difference from Perl.)
1583
1584 28. When duplicate subpattern names are present (necessarily with different
1585 numbers, as required by 27 above), and a test is made by name in a
1586 conditional pattern, either for a subpattern having been matched, or for
1587 recursion in such a pattern, all the associated numbered subpatterns are
1588 tested, and the overall condition is true if the condition is true for any
1589 one of them. This is the way Perl works, and is also more like the way
1590 testing by number works.
1591
1592
1593 Version 7.9 11-Apr-09
1594 ---------------------
1595
1596 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1597 (pcretest), all targets were linked against these libraries. This included
1598 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1599 libraries. This caused unwanted dependencies to be created. This problem
1600 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1601 pcretest is linked with readline.
1602
1603 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1604 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1605 moved outside it again, because FALSE and TRUE are already defined in AIX,
1606 but BOOL is not.
1607
1608 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1609 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1610
1611 4. The pcregrep documentation said spaces were inserted as well as colons (or
1612 hyphens) following file names and line numbers when outputting matching
1613 lines. This is not true; no spaces are inserted. I have also clarified the
1614 wording for the --colour (or --color) option.
1615
1616 5. In pcregrep, when --colour was used with -o, the list of matching strings
1617 was not coloured; this is different to GNU grep, so I have changed it to be
1618 the same.
1619
1620 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1621 each matching line was coloured. Now it goes on to look for further matches
1622 of any of the test patterns, which is the same behaviour as GNU grep.
1623
1624 7. A pattern that could match an empty string could cause pcregrep to loop; it
1625 doesn't make sense to accept an empty string match in pcregrep, so I have
1626 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1627 seems to be how GNU grep behaves.
1628
1629 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1630 start or after a newline", because the conditional assertion was not being
1631 correctly handled. The rule now is that both the assertion and what follows
1632 in the first alternative must satisfy the test.
1633
1634 9. If auto-callout was enabled in a pattern with a conditional group whose
1635 condition was an assertion, PCRE could crash during matching, both with
1636 pcre_exec() and pcre_dfa_exec().
1637
1638 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1639 used for matching.
1640
1641 11. Unicode property support in character classes was not working for
1642 characters (bytes) greater than 127 when not in UTF-8 mode.
1643
1644 12. Added the -M command line option to pcretest.
1645
1646 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1647
1648 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1649
1650 16. Added comments and documentation about mis-use of no_arg in the C++
1651 wrapper.
1652
1653 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1654 from Martin Jerabek that uses macro names for all relevant character and
1655 string constants.
1656
1657 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1658 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1659 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1660 these, but not everybody uses configure.
1661
1662 19. A conditional group that had only one branch was not being correctly
1663 recognized as an item that could match an empty string. This meant that an
1664 enclosing group might also not be so recognized, causing infinite looping
1665 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1666 with the subject "ab", where knowledge that the repeated group can match
1667 nothing is needed in order to break the loop.
1668
1669 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1670 exec(), but without supplying a callout function, matching went wrong.
1671
1672 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1673 leak if the size of the offset vector was greater than 30. When the vector
1674 is smaller, the saved offsets during recursion go onto a local stack
1675 vector, but for larger vectors malloc() is used. It was failing to free
1676 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1677 error, in fact).
1678
1679 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1680 heapframe that is used only when UTF-8 support is enabled. This caused no
1681 problem, but was untidy.
1682
1683 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1684 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1685 included within another project.
1686
1687 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1688 slightly modified by me:
1689
1690 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1691 not building pcregrep.
1692
1693 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1694 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1695
1696 25. Forward references, both numeric and by name, in patterns that made use of
1697 duplicate group numbers, could behave incorrectly or give incorrect errors,
1698 because when scanning forward to find the reference group, PCRE was not
1699 taking into account the duplicate group numbers. A pattern such as
1700 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1701
1702 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1703 the feature of a custom pointer more persuasive (as requested by a user).
1704
1705 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1706 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1707 pre-defined.
1708
1709 28. Added support for (*UTF8) at the start of a pattern.
1710
1711 29. Arrange for flags added by the "release type" setting in CMake to be shown
1712 in the configuration summary.
1713
1714
1715 Version 7.8 05-Sep-08
1716 ---------------------
1717
1718 1. Replaced UCP searching code with optimized version as implemented for Ad
1719 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1720 stage table and inline lookup instead of a function, giving speed ups of 2
1721 to 5 times on some simple patterns that I tested. Permission was given to
1722 distribute the MultiStage2.py script that generates the tables (it's not in
1723 the tarball, but is in the Subversion repository).
1724
1725 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1726 scripts.
1727
1728 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1729 a group with a zero qualifier. The result of the study could be incorrect,
1730 or the function might crash, depending on the pattern.
1731
1732 4. Caseless matching was not working for non-ASCII characters in back
1733 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1734 It now works when Unicode Property Support is available.
1735
1736 5. In pcretest, an escape such as \x{de} in the data was always generating
1737 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1738 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1739 truncation.
1740
1741 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1742
1743 7. Added two (int) casts to pcregrep when printing the difference of two
1744 pointers, in case they are 64-bit values.
1745
1746 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1747 test 2 if it fails.
1748
1749 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1750 and a #define of that name to empty if it is not externally set. This is to
1751 allow users of MSVC to set it if necessary.
1752
1753 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1754 the convenience functions in the pcre_get.c source file.
1755
1756 11. An option change at the start of a pattern that had top-level alternatives
1757 could cause overwriting and/or a crash. This command provoked a crash in
1758 some environments:
1759
1760 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1761
1762 This potential security problem was recorded as CVE-2008-2371.
1763
1764 12. For a pattern where the match had to start at the beginning or immediately
1765 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1766 pcre_dfa_exec() could read past the end of the passed subject if there was
1767 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1768 pcretest so that it places the subject at the end of its malloc-ed buffer.
1769
1770 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1771 exec() might read past the end of the data buffer in UTF-8 mode.
1772
1773 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1774 the data contained the byte 0x85 as part of a UTF-8 character within its
1775 first line. This applied both to normal and DFA matching.
1776
1777 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1778 /^[^d]*?$/8 failed to match "abc".
1779
1780 16. Added a missing copyright notice to pcrecpp_internal.h.
1781
1782 17. Make it more clear in the documentation that values returned from
1783 pcre_exec() in ovector are byte offsets, not character counts.
1784
1785 18. Tidied a few places to stop certain compilers from issuing warnings.
1786
1787 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1788 supplied by Stefan Weber. I made a further small update for 7.8 because
1789 there is a change of source arrangements: the pcre_searchfuncs.c module is
1790 replaced by pcre_ucd.c.
1791
1792
1793 Version 7.7 07-May-08
1794 ---------------------
1795
1796 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1797 a string to a long long, pretend we don't even have a long long." This is
1798 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1799
1800 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1801 pre-7.6 versions, which defined a global no_arg variable instead of putting
1802 it in the RE class. (See also #8 below.)
1803
1804 3. Remove a line of dead code, identified by coverity and reported by Nuno
1805 Lopes.
1806
1807 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1808
1809 (1) The include/exclude patterns were being applied to the whole pathnames
1810 of files, instead of just to the final components.
1811
1812 (2) If there was more than one level of directory, the subdirectories were
1813 skipped unless they satisfied the include/exclude conditions. This is
1814 inconsistent with GNU grep (and could even be seen as contrary to the
1815 pcregrep specification - which I improved to make it absolutely clear).
1816 The action now is always to scan all levels of directory, and just
1817 apply the include/exclude patterns to regular files.
1818
1819 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1820 --exclude_dir in the tests to avoid scanning .svn directories.
1821
1822 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1823 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1824 doesn't support NULs in patterns.
1825
1826 7. Added some missing "const"s to declarations of static tables in
1827 pcre_compile.c and pcre_dfa_exec.c.
1828
1829 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1830 caused by fix #2 above. (Subsequently also a second patch to fix the
1831 first patch. And a third patch - this was a messy problem.)
1832
1833 9. Applied Craig's patch to remove the use of push_back().
1834
1835 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1836 matching function regexec().
1837
1838 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1839 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1840 references. PCRE supports relative numbers with this syntax (I don't think
1841 Oniguruma does).
1842
1843 12. Previously, a group with a zero repeat such as (...){0} was completely
1844 omitted from the compiled regex. However, this means that if the group
1845 was called as a subroutine from elsewhere in the pattern, things went wrong
1846 (an internal error was given). Such groups are now left in the compiled
1847 pattern, with a new opcode that causes them to be skipped at execution
1848 time.
1849
1850 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1851 to the way PCRE behaves:
1852
1853 (a) A lone ] character is dis-allowed (Perl treats it as data).
1854
1855 (b) A back reference to an unmatched subpattern matches an empty string
1856 (Perl fails the current match path).
1857
1858 (c) A data ] in a character class must be notated as \] because if the
1859 first data character in a class is ], it defines an empty class. (In
1860 Perl it is not possible to have an empty class.) The empty class []
1861 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1862 The negative empty class [^] matches any one character, independently
1863 of the DOTALL setting.
1864
1865 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1866 non-existent subpattern following a character class starting with ']' and
1867 containing () gave an internal compiling error instead of "reference to
1868 non-existent subpattern". Fortunately, when the pattern did exist, the
1869 compiled code was correct. (When scanning forwards to check for the
1870 existencd of the subpattern, it was treating the data ']' as terminating
1871 the class, so got the count wrong. When actually compiling, the reference
1872 was subsequently set up correctly.)
1873
1874 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1875 it was being rejected as not supported by pcre_dfa_exec(), even though
1876 other assertions are supported. I have made pcre_dfa_exec() support
1877 (*FAIL).
1878
1879 16. The implementation of 13c above involved the invention of a new opcode,
1880 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1881 cannot be changed at match time, I realized I could make a small
1882 improvement to matching performance by compiling OP_ALLANY instead of
1883 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1884 on the OP_ANY path.
1885
1886 17. Compiling pcretest on Windows with readline support failed without the
1887 following two fixes: (1) Make the unistd.h include conditional on
1888 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1889
1890 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1891 ncurses library to be included for pcretest when ReadLine support is
1892 requested, but also to allow for it to be overridden. This patch came from
1893 Daniel Bergström.
1894
1895 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1896 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1897 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1898 spotting this.
1899
1900
1901 Version 7.6 28-Jan-08
1902 ---------------------
1903
1904 1. A character class containing a very large number of characters with
1905 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1906 overflow.
1907
1908 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1909 HAVE_LONG_LONG is not defined.
1910
1911 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1912 bring them up to date and include new features. This patch includes:
1913
1914 - Fixed PH's badly added libz and libbz2 support.
1915 - Fixed a problem with static linking.
1916 - Added pcredemo. [But later removed - see 7 below.]
1917 - Fixed dftables problem and added an option.
1918 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1919 HAVE_LONG_LONG.
1920 - Added readline support for pcretest.
1921 - Added an listing of the option settings after cmake has run.
1922
1923 4. A user submitted a patch to Makefile that makes it easy to create
1924 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1925 Makefile.am that cause it to include this special target, without
1926 affecting anything else. Note that the same mingw target plus all
1927 the other distribution libraries and programs are now supported
1928 when configuring with CMake (see 6 below) instead of with
1929 Configure/Make.
1930
1931 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1932 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1933 exported in the Windows port". It has not yet been confirmed that the patch
1934 solves the problem, but it does no harm.
1935
1936 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1937 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1938 with CMake, and also correct the comment about stack recursion.
1939
1940 7. Remove the automatic building of pcredemo from the ./configure system and
1941 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1942 of a program that users should build themselves after PCRE is installed, so
1943 building it automatically is not really right. What is more, it gave
1944 trouble in some build environments.
1945
1946 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1947
1948
1949 Version 7.5 10-Jan-08
1950 ---------------------
1951
1952 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1953 values in parens when parsing an RE using the C++ wrapper."
1954
1955 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1956 Characters greater than 255 were excluded from the class instead of being
1957 included.
1958
1959 3. The same bug as (2) above applied to negated POSIX classes such as
1960 [:^space:].
1961
1962 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1963 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1964 I have changed it.
1965
1966 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1967 first named subpattern) and a construct such as (?&a) would reference the
1968 first named subpattern whose name started with "a" (in other words, the
1969 length check was missing). Both these problems are fixed. "Subpattern name
1970 expected" is now given for (?&) (a zero-length name), and this patch also
1971 makes it give the same error for \k'' (previously it complained that that
1972 was a reference to a non-existent subpattern).
1973
1974 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1975 this is right because (?- can be followed by option settings as well as by
1976 digits. I have, however, made the messages clearer.
1977
1978 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1979 than the number used in the conditional) now cause a compile-time error.
1980 This is actually not compatible with Perl, which accepts such patterns, but
1981 treats the conditional as always being FALSE (as PCRE used to), but it
1982 seems to me that giving a diagnostic is better.
1983
1984 8. Change "alphameric" to the more common word "alphanumeric" in comments
1985 and messages.
1986
1987 9. Fix two occurrences of "backslash" in comments that should have been
1988 "backspace".
1989
1990 10. Remove two redundant lines of code that can never be obeyed (their function
1991 was moved elsewhere).
1992
1993 11. The program that makes PCRE's Unicode character property table had a bug
1994 which caused it to generate incorrect table entries for sequences of
1995 characters that have the same character type, but are in different scripts.
1996 It amalgamated them into a single range, with the script of the first of
1997 them. In other words, some characters were in the wrong script. There were
1998 thirteen such cases, affecting characters in the following ranges:
1999
2000 U+002b0 - U+002c1
2001 U+0060c - U+0060d
2002 U+0061e - U+00612
2003 U+0064b - U+0065e
2004 U+0074d - U+0076d
2005 U+01800 - U+01805
2006 U+01d00 - U+01d77
2007 U+01d9b - U+01dbf
2008 U+0200b - U+0200f
2009 U+030fc - U+030fe
2010 U+03260 - U+0327f
2011 U+0fb46 - U+0fbb1
2012 U+10450 - U+1049d
2013
2014 12. The -o option (show only the matching part of a line) for pcregrep was not
2015 compatible with GNU grep in that, if there was more than one match in a
2016 line, it showed only the first of them. It now behaves in the same way as
2017 GNU grep.
2018
2019 13. If the -o and -v options were combined for pcregrep, it printed a blank
2020 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2021 does the same. The return code can be used to tell if there were any
2022 non-matching lines.
2023
2024 14. Added --file-offsets and --line-offsets to pcregrep.
2025
2026 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2027 infinitely looping recursion. The bug was that positive lookaheads were not
2028 being skipped when checking for a possible empty match (negative lookaheads
2029 and both kinds of lookbehind were skipped).
2030
2031 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2032 inclusion of <windows.h> to before rather than after the definition of
2033 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2034
2035 17. Specifying a possessive quantifier with a specific limit for a Unicode
2036 character property caused pcre_compile() to compile bad code, which led at
2037 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2038 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2039 caused the error; without that there was no problem.
2040
2041 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2042
2043 19. Added --enable-pcretest-libreadline.
2044
2045 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2046 RE::GlobalReplace(). As a result, the number of replacements returned was
2047 double what it should be. I removed one of the increments, but Craig sent a
2048 later patch that removed the other one (the right fix) and added unit tests
2049 that check the return values (which was not done before).
2050
2051 21. Several CMake things:
2052
2053 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2054 the names libpcre and libpcreposix, not just pcre and pcreposix.
2055
2056 (2) The above change means that pcretest and pcregrep are now correctly
2057 linked with the newly-built libraries, not previously installed ones.
2058
2059 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2060
2061 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2062 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2063 UTF-8 newline character). The key issue is that the pattern starts .*;
2064 this means that the match must be either at the beginning, or after a
2065 newline. The bug was in the code for advancing after a failed match and
2066 checking that the new position followed a newline. It was not taking
2067 account of UTF-8 characters correctly.
2068
2069 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2070 character classes. PCRE was not treating the sequence [:...:] as a
2071 character class unless the ... were all letters. Perl, however, seems to
2072 allow any characters between [: and :], though of course it rejects as
2073 unknown any "names" that contain non-letters, because all the known class
2074 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2075 for example, whereas PCRE did not - it did not recognize a POSIX character
2076 class. This seemed a bit dangerous, so the code has been changed to be
2077 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2078 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2079 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2080 Perl does, and where it didn't before.
2081
2082 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2083 Windows environments %n is disabled by default.
2084
2085
2086 Version 7.4 21-Sep-07
2087 ---------------------
2088
2089 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2090 means that a class such as [\s] counted as "explicit reference to CR or
2091 LF". That isn't really right - the whole point of the change was to try to
2092 help when there was an actual mention of one of the two characters. So now
2093 the change happens only if \r or \n (or a literal CR or LF) character is
2094 encountered.
2095
2096 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2097 of both had grown to the point where there were only 3 bits left.
2098 Fortunately, there was spare space in the data structure, and so I have
2099 moved the internal flags into a new 16-bit field to free up more option
2100 bits.
2101
2102 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2103 but did not set the internal JCHANGED flag - either of these is enough to
2104 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2105 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2106 start sets both bits.
2107
2108 4. Added options (at build time, compile time, exec time) to change \R from
2109 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2110
2111 5. doc/pcresyntax.html was missing from the distribution.
2112
2113 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2114 compatibility, even though it is no longer used.
2115
2116 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2117 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2118 windows.h file is present (where different names are used). [This was
2119 reversed later after testing - see 16 below.]
2120
2121 8. Changed all #include <config.h> to #include "config.h". There were also
2122 some further <pcre.h> cases that I changed to "pcre.h".
2123
2124 9. When pcregrep was used with the --colour option, it missed the line ending
2125 sequence off the lines that it output.
2126
2127 10. It was pointed out to me that arrays of string pointers cause lots of
2128 relocations when a shared library is dynamically loaded. A technique of
2129 using a single long string with a table of offsets can drastically reduce
2130 these. I have refactored PCRE in four places to do this. The result is
2131 dramatic:
2132
2133 Originally: 290
2134 After changing UCP table: 187
2135 After changing error message table: 43
2136 After changing table of "verbs" 36
2137 After changing table of Posix names 22
2138
2139 Thanks to the folks working on Gregex for glib for this insight.
2140
2141 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2142 unicode-properties was also set.
2143
2144 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2145
2146 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2147 checked only for CRLF.
2148
2149 14. Added casts to pcretest.c to avoid compiler warnings.
2150
2151 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2152
2153 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2154 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2155 entirely. This removes changes made in 7 above.
2156
2157 17. The CMake files have been updated, and there is now more information about
2158 building with CMake in the NON-UNIX-USE document.
2159
2160
2161 Version 7.3 28-Aug-07
2162 ---------------------
2163
2164 1. In the rejigging of the build system that eventually resulted in 7.1, the
2165 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2166 brackets there is not right, since it causes compilers to look for an
2167 installed pcre.h, not the version that is in the source that is being
2168 compiled (which of course may be different). I have changed it back to:
2169
2170 #include "pcre.h"
2171
2172 I have a vague recollection that the change was concerned with compiling in
2173 different directories, but in the new build system, that is taken care of
2174 by the VPATH setting the Makefile.
2175
2176 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2177 when the subject happened to end in the byte 0x85 (e.g. if the last
2178 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2179 characters but of course it shouldn't be taken as a newline when it is part
2180 of another character. The bug was that, for an unlimited repeat of . in
2181 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2182 characters when looking for a newline.
2183
2184 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2185
2186 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2187 in debug output.
2188
2189 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2190 long printing in the pcrecpp unittest when running under MinGW.
2191
2192 6. ESC_K was left out of the EBCDIC table.
2193
2194 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2195 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2196 limit also applies to "virtual nesting" when a pattern is recursive, and in
2197 this case 1000 isn't so big. I have been able to remove this limit at the
2198 expense of backing off one optimization in certain circumstances. Normally,
2199 when pcre_exec() would call its internal match() function recursively and
2200 immediately return the result unconditionally, it uses a "tail recursion"
2201 feature to save stack. However, when a subpattern that can match an empty
2202 string has an unlimited repetition quantifier, it no longer makes this
2203 optimization. That gives it a stack frame in which to save the data for
2204 checking that an empty string has been matched. Previously this was taken
2205 from the 1000-entry workspace that had been reserved. So now there is no
2206 explicit limit, but more stack is used.
2207
2208 8. Applied Daniel's patches to solve problems with the import/export magic
2209 syntax that is required for Windows, and which was going wrong for the
2210 pcreposix and pcrecpp parts of the library. These were overlooked when this
2211 problem was solved for the main library.
2212
2213 9. There were some crude static tests to avoid integer overflow when computing
2214 the size of patterns that contain repeated groups with explicit upper
2215 limits. As the maximum quantifier is 65535, the maximum group length was
2216 set at 30,000 so that the product of these two numbers did not overflow a
2217 32-bit integer. However, it turns out that people want to use groups that
2218 are longer than 30,000 bytes (though not repeat them that many times).
2219 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2220 made it possible to implement the integer overflow checks in a much more
2221 dynamic way, which I have now done. The artificial limitation on group
2222 length has been removed - we now have only the limit on the total length of
2223 the compiled pattern, which depends on the LINK_SIZE setting.
2224
2225 10. Fixed a bug in the documentation for get/copy named substring when
2226 duplicate names are permitted. If none of the named substrings are set, the
2227 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2228 empty string.
2229
2230 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2231 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2232 because the ] is interpreted as the first data character and the
2233 terminating ] is not found. PCRE has been made compatible with Perl in this
2234 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2235 cause memory overwriting.
2236
2237 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2238 string has been matched (to stop an infinite loop). It was not recognizing
2239 a conditional subpattern that could match an empty string if that
2240 subpattern was within another subpattern. For example, it looped when
2241 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2242 condition was not nested. This bug has been fixed.
2243
2244 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2245 past the start of the subject in the presence of bytes with the top bit
2246 set, for example "\x8aBCD".
2247
2248 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2249 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2250
2251 14. Optimized (?!) to (*FAIL).
2252
2253 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2254 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2255 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2256 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2257 does: it's just the validity check that is more restrictive.
2258
2259 16. Inserted checks for integer overflows during escape sequence (backslash)
2260 processing, and also fixed erroneous offset values for syntax errors during
2261 backslash processing.
2262
2263 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2264 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2265
2266 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2267 caused an overrun.
2268
2269 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2270 something other than just ASCII characters) inside a group that had an
2271 unlimited repeat caused a loop at compile time (while checking to see
2272 whether the group could match an empty string).
2273
2274 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2275 [\P{Any}] did so. (Error in the code for printing property names.)
2276
2277 21. An orphan \E inside a character class could cause a crash.
2278
2279 22. A repeated capturing bracket such as (A)? could cause a wild memory
2280 reference during compilation.
2281
2282 23. There are several functions in pcre_compile() that scan along a compiled
2283 expression for various reasons (e.g. to see if it's fixed length for look
2284 behind). There were bugs in these functions when a repeated \p or \P was
2285 present in the pattern. These operators have additional parameters compared
2286 with \d, etc, and these were not being taken into account when moving along
2287 the compiled data. Specifically:
2288
2289 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2290 length.
2291
2292 (b) An item such as \pL+ within a repeated group could cause crashes or
2293 loops.
2294
2295 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2296 "reference to non-existent subpattern" error.
2297
2298 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2299
2300 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2301 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2302
2303 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2304
2305 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2306 character were causing crashes (broken optimization).
2307
2308 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2309 \p or \P) caused a compile-time loop.
2310
2311 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2312 break. For example, the unstudied pattern [\r\n]A does not match the string
2313 "\r\nA" because change 7.0/46 below moves the current point on by two
2314 characters after failing to match at the start. However, the pattern \nA
2315 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2316 the same is true. There doesn't seem any very clean way out of this, but
2317 what I have chosen to do makes the common cases work: PCRE now takes note
2318 of whether there can be an explicit match for \r or \n anywhere in the
2319 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2320 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2321 pattern has explicit CR or LF references.
2322
2323 29. Added (*CR) etc for changing newline setting at start of pattern.
2324
2325
2326 Version 7.2 19-Jun-07
2327 ---------------------
2328
2329 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2330 which is apparently normally available under Windows.
2331
2332 2. Re-jig the pcregrep tests with different newline settings in an attempt
2333 to make them independent of the local environment's newline setting.
2334
2335 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2336
2337 4. Some of the "internals" tests were previously cut out when the link size
2338 was not 2, because the output contained actual offsets. The recent new
2339 "Z" feature of pcretest means that these can be cut out, making the tests
2340 usable with all link sizes.
2341
2342 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2343 stack recursion. This gives a massive performance boost under BSD, but just
2344 a small improvement under Linux. However, it saves one field in the frame
2345 in all cases.
2346
2347 6. Added more features from the forthcoming Perl 5.10:
2348
2349 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2350 recursion call. It refers to the nth most recently opened parentheses.
2351
2352 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2353 to be opened parentheses.
2354
2355 (c) Conditions that refer to capturing parentheses can be specified
2356 relatively, for example, (?(-2)... or (?(+3)...
2357
2358 (d) \K resets the start of the current match so that everything before
2359 is not part of it.
2360
2361 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2362
2363 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2364 reference syntax.
2365
2366 (g) (?| introduces a group in which the numbering of parentheses in each
2367 alternative starts with the same number.
2368
2369 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2370
2371 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2372 PCRE_INFO_JCHANGED.
2373
2374 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2375 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2376 for detecting groups that can match an empty string.
2377
2378 9. A pattern with a very large number of alternatives (more than several
2379 hundred) was running out of internal workspace during the pre-compile
2380 phase, where pcre_compile() figures out how much memory will be needed. A
2381 bit of new cunning has reduced the workspace needed for groups with
2382 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2383 workspace instead of running out of the 4096 that are available.
2384
2385 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2386
2387 11. Applied patch from Google to remove an optimization that didn't quite work.
2388 The report of the bug said:
2389
2390 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2391 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2392 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2393
2394 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2395 it matched the wrong number of bytes.
2396
2397
2398 Version 7.1 24-Apr-07
2399 ---------------------
2400
2401 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2402 that is more "standard", making use of automake and other Autotools. There
2403 is some re-arrangement of the files and adjustment of comments consequent
2404 on this.
2405
2406 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2407 for recursive directory scanning broke on some systems because the files
2408 are not scanned in any specific order and on different systems the order
2409 was different. A call to "sort" has been inserted into RunGrepTest for the
2410 approprate test as a short-term fix. In the longer term there may be an
2411 alternative.
2412
2413 3. I had an email from Eric Raymond about problems translating some of PCRE's
2414 man pages to HTML (despite the fact that I distribute HTML pages, some
2415 people do their own conversions for various reasons). The problems
2416 concerned the use of low-level troff macros .br and .in. I have therefore
2417 removed all such uses from the man pages (some were redundant, some could
2418 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2419 HTML has been updated to handle .nf/.fi and to complain if it encounters
2420 .br or .in.
2421
2422 4. Updated comments in configure.ac that get placed in config.h.in and also
2423 arranged for config.h to be included in the distribution, with the name
2424 config.h.generic, for the benefit of those who have to compile without
2425 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2426
2427 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2428 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2429 makevp.bat for the current PCRE, using the additional files
2430 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2431
2432 6. A Windows user reported a minor discrepancy with test 2, which turned out
2433 to be caused by a trailing space on an input line that had got lost in his
2434 copy. The trailing space was an accident, so I've just removed it.
2435
2436 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2437 that is needed.
2438
2439 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2440 as "const" (a) because they are and (b) because it helps the PHP
2441 maintainers who have recently made a script to detect big data structures
2442 in the php code that should be moved to the .rodata section. I remembered
2443 to update Builducptable as well, so it won't revert if ucptable.h is ever
2444 re-created.
2445
2446 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2447 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2448 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2449 support is not required. This saves 1.5-2K of code, which is important in
2450 some applications.
2451
2452 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2453 so as not to refer to the tables, even though these functions will never be
2454 called when UTF-8 support is disabled. Otherwise there are problems with a
2455 shared library.
2456
2457 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2458
2459 (a) It was defining its arguments as char * instead of void *.
2460
2461 (b) It was assuming that all moves were upwards in memory; this was true
2462 a long time ago when I wrote it, but is no longer the case.
2463
2464 The emulated memove() is provided for those environments that have neither
2465 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2466 is clearly not the case, as these two bugs were recently reported.
2467
2468 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2469 and Detrail to create the HTML documentation, the .txt form of the man
2470 pages, and it removes trailing spaces from listed files. It also creates
2471 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2472 case, it wraps all the #defines with #ifndefs. This script should be run
2473 before "make dist".
2474
2475 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2476 with Unicode property support.
2477
2478 (a) For a maximizing quantifier, if the two different cases of the
2479 character were of different lengths in their UTF-8 codings (there are
2480 some cases like this - I found 11), and the matching function had to
2481 back up over a mixture of the two cases, it incorrectly assumed they
2482 were both the same length.
2483
2484 (b) When PCRE was configured to use the heap rather than the stack for
2485 recursion during matching, it was not correctly preserving the data for
2486 the other case of a UTF-8 character when checking ahead for a match
2487 while processing a minimizing repeat. If the check also involved
2488 matching a wide character, but failed, corruption could cause an
2489 erroneous result when trying to check for a repeat of the original
2490 character.
2491
2492 13. Some tidying changes to the testing mechanism:
2493
2494 (a) The RunTest script now detects the internal link size and whether there
2495 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2496 values substituted by "configure". (The RunGrepTest script already did
2497 this for UTF-8.) The configure.ac script no longer substitutes the
2498 relevant variables.
2499
2500 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2501 with length and offset values. This means that the output is different
2502 for different internal link sizes. Test 2 is skipped for link sizes
2503 other than 2 because of this, bypassing the problem. Unfortunately,
2504 there was also a test in test 3 (the locale tests) that used /B and
2505 failed for link sizes other than 2. Rather than cut the whole test out,
2506 I have added a new /Z option to pcretest that replaces the length and
2507 offset values with spaces. This is now used to make test 3 independent
2508 of link size. (Test 2 will be tidied up later.)
2509
2510 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2511 segmentation fault instead of returning the appropriate error message.
2512
2513 15. In multiline mode when the newline sequence was set to "any", the pattern
2514 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2515 This doesn't seem right; it now treats the CRLF combination as the line
2516 ending, and so does not match in that case. It's only a pattern such as ^$
2517 that would hit this one: something like ^ABC$ would have failed after \r
2518 and then tried again after \r\n.
2519
2520 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2521 in an attempt to make files that differ only in their line terminators
2522 compare equal. This works on Linux.
2523
2524 17. Under certain error circumstances pcregrep might try to free random memory
2525 as it exited. This is now fixed, thanks to valgrind.
2526
2527 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2528 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2529 was because its rules for how to advance for /g after matching an empty
2530 string at the end of a line did not allow for this case. They now check for
2531 it specially.
2532
2533 20. pcretest is supposed to handle patterns and data of any length, by
2534 extending its buffers when necessary. It was getting this wrong when the
2535 buffer for a data line had to be extended.
2536
2537 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2538 CRLF as a newline sequence.
2539
2540 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2541 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2542 I have nevertheless tidied it up.
2543
2544 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2545
2546 24. Added a man page for pcre-config.
2547
2548
2549 Version 7.0 19-Dec-06
2550 ---------------------
2551
2552 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2553 moving to gcc 4.1.1.
2554
2555 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2556 sys/time.h, which is documented as needed for this function. It doesn't
2557 seem to matter on Linux, but it showed up on some releases of OS X.
2558
2559 3. It seems that there are systems where bytes whose values are greater than
2560 127 match isprint() in the "C" locale. The "C" locale should be the
2561 default when a C program starts up. In most systems, only ASCII printing
2562 characters match isprint(). This difference caused the output from pcretest
2563 to vary, making some of the tests fail. I have changed pcretest so that:
2564
2565 (a) When it is outputting text in the compiled version of a pattern, bytes
2566 other than 32-126 are always shown as hex escapes.
2567
2568 (b) When it is outputting text that is a matched part of a subject string,
2569 it does the same, unless a different locale has been set for the match
2570 (using the /L modifier). In this case, it uses isprint() to decide.
2571
2572 4. Fixed a major bug that caused incorrect computation of the amount of memory
2573 required for a compiled pattern when options that changed within the
2574 pattern affected the logic of the preliminary scan that determines the
2575 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2576 that the computed length was too small. The symptoms of this bug were
2577 either the PCRE error "internal error: code overflow" from pcre_compile(),
2578 or a glibc crash with a message such as "pcretest: free(): invalid next
2579 size (fast)". Examples of patterns that provoked this bug (shown in
2580 pcretest format) are:
2581
2582 /(?-x: )/x
2583 /(?x)(?-x: \s*#\s*)/
2584 /((?i)[\x{c0}])/8
2585 /(?i:[\x{c0}])/8
2586
2587 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2588 is now done differently.
2589
2590 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2591 wrapper classes; (b) implement a new function in the C++ scanner that is
2592 more efficient than the old way of doing things because it avoids levels of
2593 recursion in the regex matching; (c) add a paragraph to the documentation
2594 for the FullMatch() function.
2595
2596 6. The escape sequence \n was being treated as whatever was defined as
2597 "newline". Not only was this contrary to the documentation, which states
2598 that \n is character 10 (hex 0A), but it also went horribly wrong when
2599 "newline" was defined as CRLF. This has been fixed.
2600
2601 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2602 was being set to -1 for the "end of line" case (supposedly a value that no
2603 character can have). Though this value is never used (the check for end of
2604 line is "zero bytes in current character"), it caused compiler complaints.
2605 I've changed it to 0xffffffff.
2606
2607 8. In pcre_version.c, the version string was being built by a sequence of
2608 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2609 string (as it is for production releases) called a macro with an empty
2610 argument. The C standard says the result of this is undefined. The gcc
2611 compiler treats it as an empty string (which was what was wanted) but it is
2612 reported that Visual C gives an error. The source has been hacked around to
2613 avoid this problem.
2614
2615 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2616 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2617 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2618 of them did).
2619
2620 10. Originally, pcretest opened its input and output without "b"; then I was
2621 told that "b" was needed in some environments, so it was added for release
2622 5.0 to both the input and output. (It makes no difference on Unix-like
2623 systems.) Later I was told that it is wrong for the input on Windows. I've
2624 now abstracted the modes into two macros, to make it easier to fiddle with
2625 them, and removed "b" from the input mode under Windows.
2626
2627 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2628
2629 12. Added -help and --help to pcretest as an official way of being reminded
2630 of the options.
2631
2632 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2633 and pcrecpp.cc because they annoy compilers at high warning levels.
2634
2635 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2636
2637 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2638 scripts are not C programs :-) and which was not noticed because it works
2639 on Linux.
2640
2641 16. pcretest is supposed to handle any length of pattern and data line (as one
2642 line or as a continued sequence of lines) by extending its input buffer if
2643 necessary. This feature was broken for very long pattern lines, leading to
2644 a string of junk being passed to pcre_compile() if the pattern was longer
2645 than about 50K.
2646
2647 17. I have done a major re-factoring of the way pcre_compile() computes the
2648 amount of memory needed for a compiled pattern. Previously, there was code
2649 that made a preliminary scan of the pattern in order to do this. That was
2650 OK when PCRE was new, but as the facilities have expanded, it has become
2651 harder and harder to keep it in step with the real compile phase, and there
2652 have been a number of bugs (see for example, 4 above). I have now found a
2653 cunning way of running the real compile function in a "fake" mode that
2654 enables it to compute how much memory it would need, while actually only
2655 ever using a few hundred bytes of working memory and without too many
2656 tests of the mode. This should make future maintenance and development
2657 easier. A side effect of this work is that the limit of 200 on the nesting
2658 depth of parentheses has been removed (though this was never a serious
2659 limitation, I suspect). However, there is a downside: pcre_compile() now
2660 runs more slowly than before (30% or more, depending on the pattern). I
2661 hope this isn't a big issue. There is no effect on runtime performance.
2662
2663 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2664 newline (only possible for the last line of a file) and it was a
2665 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2666
2667 19. Added additional timing features to pcretest. (1) The -tm option now times
2668 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2669 separate command line item, by a number that specifies the number of
2670 repeats to use when timing. The default is 50000; this gives better
2671 precision, but takes uncomfortably long for very large patterns.
2672
2673 20. Extended pcre_study() to be more clever in cases where a branch of a
2674 subpattern has no definite first character. For example, (a*|b*)[cd] would
2675 previously give no result from pcre_study(). Now it recognizes that the
2676 first character must be a, b, c, or d.
2677
2678 21. There was an incorrect error "recursive call could loop indefinitely" if
2679 a subpattern (or the entire pattern) that was being tested for matching an
2680 empty string contained only one non-empty item after a nested subpattern.
2681 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2682 incorrectly, because the \d was being skipped in the check.
2683
2684 22. The pcretest program now has a new pattern option /B and a command line
2685 option -b, which is equivalent to adding /B to every pattern. This causes
2686 it to show the compiled bytecode, without the additional information that
2687 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2688 is the same as /B/I).
2689
2690 23. A new optimization is now able automatically to treat some sequences such
2691 as a*b as a*+b. More specifically, if something simple (such as a character
2692 or a simple class like \d) has an unlimited quantifier, and is followed by
2693 something that cannot possibly match the quantified thing, the quantifier
2694 is automatically "possessified".
2695
2696 24. A recursive reference to a subpattern whose number was greater than 39
2697 went wrong under certain circumstances in UTF-8 mode. This bug could also
2698 have affected the operation of pcre_study().
2699
2700 25. Realized that a little bit of performance could be had by replacing
2701 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2702
2703 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2704
2705 27. Possessive quantifiers such as a++ were previously implemented by turning
2706 them into atomic groups such as ($>a+). Now they have their own opcodes,
2707 which improves performance. This includes the automatically created ones
2708 from 23 above.
2709
2710 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2711 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2712 the first matched character to be a colon. This applied both to named and
2713 numbered groups.
2714
2715 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2716
2717 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2718 building PCRE on Windows easier, so I have included it in the distribution.
2719
2720 31. There is now a check in pcretest against a ridiculously large number being
2721 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2722 loop, the loop is abandoned.
2723
2724 32. Forward references to subpatterns in conditions such as (?(2)...) where
2725 subpattern 2 is defined later cause pcre_compile() to search forwards in
2726 the pattern for the relevant set of parentheses. This search went wrong
2727 when there were unescaped parentheses in a character class, parentheses
2728 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2729
2730 33. "Subroutine" calls and backreferences were previously restricted to
2731 referencing subpatterns earlier in the regex. This restriction has now
2732 been removed.
2733
2734 34. Added a number of extra features that are going to be in Perl 5.10. On the
2735 whole, these are just syntactic alternatives for features that PCRE had
2736 previously implemented using the Python syntax or my own invention. The
2737 other formats are all retained for compatibility.
2738
2739 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2740 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2741 also .NET compatible.
2742
2743 (b) A recursion or subroutine call to a named group can now be defined as
2744 (?&name) as well as (?P>name).
2745
2746 (c) A backreference to a named group can now be defined as \k<name> or
2747 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2748 5.10, are also .NET compatible.
2749
2750 (d) A conditional reference to a named group can now use the syntax
2751 (?(<name>) or (?('name') as well as (?(name).
2752
2753 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2754 groups (named and numbered) that are never evaluated inline, but can be
2755 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2756 is always false. There may be only one alternative in such a group.
2757
2758 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2759 as the simple (?(R). The condition is true only if the most recent
2760 recursion is that of the given number or name. It does not search out
2761 through the entire recursion stack.
2762
2763 (g) The escape \gN or \g{N} has been added, where N is a positive or
2764 negative number, specifying an absolute or relative reference.
2765
2766 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2767 some "unreachable code" warnings.
2768
2769 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2770 things, this adds five new scripts.
2771
2772 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2773 There were also incompatibilities regarding the handling of \Q..\E inside
2774 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2775 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2776
2777 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2778 matches an empty string, and forcibly breaks the loop. There were bugs in
2779 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2780 against aaaa the result was just "a" rather than "aaaa", for example. Two
2781 separate and independent bugs (that affected different cases) have been
2782 fixed.
2783
2784 39. Refactored the code to abolish the use of different opcodes for small
2785 capturing bracket numbers. This is a tidy that I avoided doing when I
2786 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2787 The new approach is not only tidier, it makes it possible to reduce the
2788 memory needed to fix the previous bug (38).
2789
2790 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2791 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2792 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2793 mode.
2794
2795 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2796 report.
2797
2798 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2799 copy construction and assignment in the C++ wrapper.
2800
2801 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2802 couple of bugs that could have given wrong results in the "--newline=crlf"
2803 case.
2804
2805 44. Added a number of casts and did some reorganization of signed/unsigned int
2806 variables following suggestions from Dair Grant. Also renamed the variable
2807 "this" as "item" because it is a C++ keyword.
2808
2809 45. Arranged for dftables to add
2810
2811 #include "pcre_internal.h"
2812
2813 to pcre_chartables.c because without it, gcc 4.x may remove the array
2814 definition from the final binary if PCRE is built into a static library and
2815 dead code stripping is activated.
2816
2817 46. For an unanchored pattern, if a match attempt fails at the start of a
2818 newline sequence, and the newline setting is CRLF or ANY, and the next two
2819 characters are CRLF, advance by two characters instead of one.
2820
2821
2822 Version 6.7 04-Jul-06
2823 ---------------------
2824
2825 1. In order to handle tests when input lines are enormously long, pcretest has
2826 been re-factored so that it automatically extends its buffers when
2827 necessary. The code is crude, but this _is_ just a test program. The
2828 default size has been increased from 32K to 50K.
2829
2830 2. The code in pcre_study() was using the value of the re argument before
2831 testing it for NULL. (Of course, in any sensible call of the function, it
2832 won't be NULL.)
2833
2834 3. The memmove() emulation function in pcre_internal.h, which is used on
2835 systems that lack both memmove() and bcopy() - that is, hardly ever -
2836 was missing a "static" storage class specifier.
2837
2838 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2839 containing an extended class (one that cannot be represented by a bitmap
2840 because it contains high-valued characters or Unicode property items, e.g.
2841 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2842 pattern, but PCRE should not loop if you do not (it no longer does).
2843 [Detail: two cases were found: (a) a repeated subpattern containing an
2844 extended class; (b) a recursive reference to a subpattern that followed a
2845 previous extended class. It wasn't skipping over the extended class
2846 correctly when UTF-8 mode was not set.]
2847
2848 5. A negated single-character class was not being recognized as fixed-length
2849 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2850 compile error "lookbehind assertion is not fixed length".
2851
2852 6. The RunPerlTest auxiliary script was showing an unexpected difference
2853 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2854 write a Perl script that can interpret lines of an input file either as
2855 byte characters or as UTF-8, which is what "perltest" was being required to
2856 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2857 can't do is switch easily at run time between having the "use utf8;" pragma
2858 or not. In the end, I fudged it by using the RunPerlTest script to insert
2859 "use utf8;" explicitly for the UTF-8 tests.
2860
2861 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2862 the end of the subject string, contrary to the documentation and to what
2863 Perl does. This was true of both matching functions. Now it matches only at
2864 the start of the subject and immediately after *internal* newlines.
2865
2866 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2867 a pointer to an int instead of a pointer to an unsigned long int. This
2868 caused problems on 64-bit systems.
2869
2870 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2871 instance of the 'standard' template library not being so standard".
2872
2873 10. There was no check on the number of named subpatterns nor the maximum
2874 length of a subpattern name. The product of these values is used to compute
2875 the size of the memory block for a compiled pattern. By supplying a very
2876 long subpattern name and a large number of named subpatterns, the size
2877 computation could be caused to overflow. This is now prevented by limiting
2878 the length of names to 32 characters, and the number of named subpatterns
2879 to 10,000.
2880
2881 11. Subpatterns that are repeated with specific counts have to be replicated in
2882 the compiled pattern. The size of memory for this was computed from the
2883 length of the subpattern and the repeat count. The latter is limited to
2884 65535, but there was no limit on the former, meaning that integer overflow
2885 could in principle occur. The compiled length of a repeated subpattern is
2886 now limited to 30,000 bytes in order to prevent this.
2887
2888 12. Added the optional facility to have named substrings with the same name.
2889
2890 13. Added the ability to use a named substring as a condition, using the
2891 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2892 are numbers (not recommended). Forward references are permitted.
2893
2894 14. Added forward references in named backreferences (if you see what I mean).
2895
2896 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2897 pattern could run off the end of the subject. For example, the pattern
2898 "(?s)(.{1,5})"8 did this with the subject "ab".
2899
2900 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2901 PCRE_CASELESS was set when matching characters that were quantified with ?
2902 or *.
2903
2904 17. A character class other than a single negated character that had a minimum
2905 but no maximum quantifier - for example [ab]{6,} - was not handled
2906 correctly by pce_dfa_exec(). It would match only one character.
2907
2908 18. A valid (though odd) pattern that looked like a POSIX character
2909 class but used an invalid character after [ (for example [[,abc,]]) caused
2910 pcre_compile() to give the error "Failed: internal error: code overflow" or
2911 in some cases to crash with a glibc free() error. This could even happen if
2912 the pattern terminated after [[ but there just happened to be a sequence of
2913 letters, a binary zero, and a closing ] in the memory that followed.
2914
2915 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2916 over the years. Originally (before any Unicode support), just the bottom 8
2917 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2918 output from "man perlunicode" includes this:
2919
2920 The regular expression compiler produces polymorphic opcodes. That
2921 is, the pattern adapts to the data and automatically switches to
2922 the Unicode character scheme when presented with Unicode data--or
2923 instead uses a traditional byte scheme when presented with byte
2924 data.
2925
2926 Sadly, a wide octal escape does not cause a switch, and in a string with
2927 no other multibyte characters, these octal escapes are treated as before.
2928 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2929 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2930 Unicode string.
2931
2932 I have not perpetrated such confusion in PCRE. Up till now, it took just
2933 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2934 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2935 translate to the appropriate multibyte character.
2936
2937 29. Applied some refactoring to reduce the number of warnings from Microsoft
2938 and Borland compilers. This has included removing the fudge introduced
2939 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2940 a warning about an unused variable.
2941
2942 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2943 characters since release 4.0, because Perl (from release 5.004) does not.
2944 [Or at least, is documented not to: some releases seem to be in conflict
2945 with the documentation.] However, when a pattern was studied with
2946 pcre_study() and all its branches started with \s, PCRE still included VT
2947 as a possible starting character. Of course, this did no harm; it just
2948 caused an unnecessary match attempt.
2949
2950 22. Removed a now-redundant internal flag bit that recorded the fact that case
2951 dependency changed within the pattern. This was once needed for "required
2952 byte" processing, but is no longer used. This recovers a now-scarce options
2953 bit. Also moved the least significant internal flag bit to the most-
2954 significant bit of the word, which was not previously used (hangover from
2955 the days when it was an int rather than a uint) to free up another bit for
2956 the future.
2957
2958 23. Added support for CRLF line endings as well as CR and LF. As well as the
2959 default being selectable at build time, it can now be changed at runtime
2960 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2961 specify that it is scanning data with non-default line endings.
2962
2963 24. Changed the definition of CXXLINK to make it agree with the definition of
2964 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2965
2966 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2967 recursions. This makes a big different to stack usage for some patterns.
2968
2969 26. If a subpattern containing a named recursion or subroutine reference such
2970 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2971 the space required for the compiled pattern went wrong and gave too small a
2972 value. Depending on the environment, this could lead to "Failed: internal
2973 error: code overflow at offset 49" or "glibc detected double free or
2974 corruption" errors.
2975
2976 27. Applied patches from Google (a) to support the new newline modes and (b) to
2977 advance over multibyte UTF-8 characters in GlobalReplace.
2978
2979 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2980 difference for some implementation of PCRE in some Windows version.
2981
2982 29. Added some extra testing facilities to pcretest:
2983
2984 \q<number> in a data line sets the "match limit" value
2985 \Q<number> in a data line sets the "match recursion limt" value
2986 -S <number> sets the stack size, where <number> is in megabytes
2987
2988 The -S option isn't available for Windows.
2989
2990
2991 Version 6.6 06-Feb-06
2992 ---------------------
2993
2994 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2995 in pcreposix.h. I have copied the definition from pcre.h.
2996
2997 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2998 because pcre.h is no longer a built file.
2999
3000 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3001 not normally included in the compiled code.
3002
3003
3004 Version 6.5 01-Feb-06
3005 ---------------------
3006
3007 1. When using the partial match feature with pcre_dfa_exec(), it was not
3008 anchoring the second and subsequent partial matches at the new starting
3009 point. This could lead to incorrect results. For example, with the pattern
3010 /1234/, partially matching against "123" and then "a4" gave a match.
3011
3012 2. Changes to pcregrep:
3013
3014 (a) All non-match returns from pcre_exec() were being treated as failures
3015 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3016 error message is output. Some extra information is given for the
3017 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3018 probably the only errors that are likely to be caused by users (by
3019 specifying a regex that has nested indefinite repeats, for instance).
3020 If there are more than 20 of these errors, pcregrep is abandoned.
3021
3022 (b) A binary zero was treated as data while matching, but terminated the
3023 output line if it was written out. This has been fixed: binary zeroes
3024 are now no different to any other data bytes.
3025
3026 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3027 used to set a locale for matching. The --locale=xxxx long option has
3028 been added (no short equivalent) to specify a locale explicitly on the
3029 pcregrep command, overriding the environment variables.
3030
3031 (d) When -B was used with -n, some line numbers in the output were one less
3032 than they should have been.
3033
3034 (e) Added the -o (--only-matching) option.
3035
3036 (f) If -A or -C was used with -c (count only), some lines of context were
3037 accidentally printed for the final match.
3038
3039 (g) Added the -H (--with-filename) option.
3040
3041 (h) The combination of options -rh failed to suppress file names for files
3042 that were found from directory arguments.
3043
3044 (i) Added the -D (--devices) and -d (--directories) options.
3045
3046 (j) Added the -F (--fixed-strings) option.
3047
3048 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3049
3050 (l) Added the --colo(u)r option.
3051
3052 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3053 is not present by default.
3054
3055 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3056 items such as (?R) or (?1), when the recursion could match a number of
3057 alternatives. If it matched one of the alternatives, but subsequently,
3058 outside the recursion, there was a failure, the code tried to back up into
3059 the recursion. However, because of the way PCRE is implemented, this is not
3060 possible, and the result was an incorrect result from the match.
3061
3062 In order to prevent this happening, the specification of recursion has
3063 been changed so that all such subpatterns are automatically treated as
3064 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3065
3066 4. I had overlooked the fact that, in some locales, there are characters for
3067 which isalpha() is true but neither isupper() nor islower() are true. In
3068 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3069 and ordfeminine) are like this. This affected the treatment of \w and \W
3070 when they appeared in character classes, but not when they appeared outside
3071 a character class. The bit map for "word" characters is now created
3072 separately from the results of isalnum() instead of just taking it from the
3073 upper, lower, and digit maps. (Plus the underscore character, of course.)
3074
3075 5. The above bug also affected the handling of POSIX character classes such as
3076 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3077 permanent tables. Instead, the bit maps for such a class were previously
3078 created as the appropriate unions of the upper, lower, and digit bitmaps.
3079 Now they are created by subtraction from the [[:word:]] class, which has
3080 its own bitmap.
3081
3082 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3083 It is created by subtracting the vertical space characters (\x09, \x0a,
3084 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3085 subtraction was done in the overall bitmap for a character class, meaning
3086 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3087 be recognized. This bug has been fixed.
3088
3089 7. Patches from the folks at Google:
3090
3091 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3092 real life, but is still worth protecting against".
3093
3094 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3095 regular expressions".
3096
3097 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3098 have it.
3099
3100 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3101 "configure" and the latter not, in order to fix a problem somebody had
3102 with compiling the Arg class on HP-UX.
3103
3104 (e) Improve the error-handling of the C++ wrapper a little bit.
3105
3106 (f) New tests for checking recursion limiting.
3107
3108 8. The pcre_memmove() function, which is used only if the environment does not
3109 have a standard memmove() function (and is therefore rarely compiled),
3110 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3111 returning a result (though PCRE never actually uses the result).
3112
3113 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3114 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3115 returned instead of calling malloc() with an overflowing number that would
3116 most likely cause subsequent chaos.
3117
3118 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3119
3120 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3121 with this option is matched, the nmatch and pmatch options of regexec() are
3122 ignored.
3123
3124 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3125 provided in case anyone wants to the the POSIX interface with UTF-8
3126 strings.
3127
3128 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3129 C++ linking (needed for some HP-UX environments).
3130
3131 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3132 (unused parameter) and in the pcre_printint() function (omitted "default"
3133 switch label when the default is to do nothing).
3134
3135 15. Added some code to make it possible, when PCRE is compiled as a C++
3136 library, to replace subject pointers for pcre_exec() with a smart pointer
3137 class, thus making it possible to process discontinuous strings.
3138
3139 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3140 much the same function. They were added by different people who were trying
3141 to make PCRE easy to compile on non-Unix systems. It has been suggested
3142 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3143 for compiling on Windows systems. I have therefore replaced it with
3144 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3145 defaults to "extern" for C or "extern C" for C++, which works fine on
3146 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3147 SCOPE with something explicit in config.h. In addition:
3148
3149 (a) pcreposix.h still had just "extern" instead of either of these macros;
3150 I have replaced it with PCRE_DATA_SCOPE.
3151
3152 (b) Functions such as _pcre_xclass(), which are internal to the library,
3153 but external in the C sense, all had PCRE_EXPORT in their definitions.
3154 This is apparently wrong for the Windows case, so I have removed it.
3155 (It makes no difference on Unix-like systems.)
3156
3157 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3158 of recursive calls to match(). This is different to MATCH_LIMIT because
3159 that limits the total number of calls to match(), not all of which increase
3160 the depth of recursion. Limiting the recursion depth limits the amount of
3161 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3162 when PCRE is compiled, and changed at run time. A patch from Google adds
3163 this functionality to the C++ interface.
3164
3165 18. Changes to the handling of Unicode character properties:
3166
3167 (a) Updated the table to Unicode 4.1.0.
3168
3169 (b) Recognize characters that are not in the table as "Cn" (undefined).
3170
3171 (c) I revised the way the table is implemented to a much improved format
3172 which includes recognition of ranges. It now supports the ranges that
3173 are defined in UnicodeData.txt, and it also amalgamates other
3174 characters into ranges. This has reduced the number of entries in the
3175 table from around 16,000 to around 3,000, thus reducing its size
3176 considerably. I realized I did not need to use a tree structure after
3177 all - a binary chop search is just as efficient. Having reduced the
3178 number of entries, I extended their size from 6 bytes to 8 bytes to
3179 allow for more data.
3180
3181 (d) Added support for Unicode script names via properties such as \p{Han}.
3182
3183 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3184 matching that character.
3185
3186 20. When matching a repeated Unicode property with a minimum greater than zero,
3187 (for example \pL{2,}), PCRE could look past the end of the subject if it
3188 reached it while seeking the minimum number of characters. This could
3189 happen only if some of the characters were more than one byte long, because
3190 there is a check for at least the minimum number of bytes.
3191
3192 21. Refactored the implementation of \p and \P so as to be more general, to
3193 allow for more different types of property in future. This has changed the
3194 compiled form incompatibly. Anybody with saved compiled patterns that use
3195 \p or \P will have to recompile them.
3196
3197 22. Added "Any" and "L&" to the supported property types.
3198
3199 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3200 but give a compile time error if the value is greater than 0xff.
3201
3202 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3203 accidentally not being installed or uninstalled.
3204
3205 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3206 made were to insert the current release number. This seemed silly, because
3207 it made things harder for people building PCRE on systems that don't run
3208 "configure". I have turned pcre.h into a distributed file, no longer built
3209 by "configure", with the version identification directly included. There is
3210 no longer a pcre.h.in file.
3211
3212 However, this change necessitated a change to the pcre-config script as
3213 well. It is built from pcre-config.in, and one of the substitutions was the
3214 release number. I have updated configure.ac so that ./configure now finds
3215 the release number by grepping pcre.h.
3216
3217 26. Added the ability to run the tests under valgrind.
3218
3219
3220 Version 6.4 05-Sep-05
3221 ---------------------
3222
3223 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3224 "--" to be printed when multiple files were scanned, even when none of the
3225 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3226 consider it to be a bug, and have restored the previous behaviour.
3227
3228 2. A couple of code tidies to get rid of compiler warnings.
3229
3230 3. The pcretest program used to cheat by referring to symbols in the library
3231 whose names begin with _pcre_. These are internal symbols that are not
3232 really supposed to be visible externally, and in some environments it is
3233 possible to suppress them. The cheating is now confined to including
3234 certain files from the library's source, which is a bit cleaner.
3235
3236 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3237 file's purpose clearer.
3238
3239 5. Reorganized pcre_ucp_findchar().
3240
3241
3242 Version 6.3 15-Aug-05
3243 ---------------------
3244
3245 1. The file libpcre.pc.in did not have general read permission in the tarball.
3246
3247 2. There were some problems when building without C++ support:
3248
3249 (a) If C++ support was not built, "make install" and "make test" still
3250 tried to test it.
3251
3252 (b) There were problems when the value of CXX was explicitly set. Some
3253 changes have been made to try to fix these, and ...
3254
3255 (c) --disable-cpp can now be used to explicitly disable C++ support.
3256
3257 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3258 backslash in a target when C++ was disabled. This confuses some
3259 versions of "make", apparently. Using an intermediate variable solves
3260 this. (Same for CPP_LOBJ.)
3261
3262 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3263 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3264 necessary on certain architectures.
3265
3266 4. Added a setting of -export-symbols-regex to the link command to remove
3267 those symbols that are exported in the C sense, but actually are local
3268 within the library, and not documented. Their names all begin with
3269 "_pcre_". This is not a perfect job, because (a) we have to except some
3270 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3271 available (and never for static libraries). I have made a note to try to
3272 find a way round (a) in the future.
3273
3274
3275 Version 6.2 01-Aug-05
3276 ---------------------
3277
3278 1. There was no test for integer overflow of quantifier values. A construction
3279 such as {1111111111111111} would give undefined results. What is worse, if
3280 a minimum quantifier for a parenthesized subpattern overflowed and became
3281 negative, the calculation of the memory size went wrong. This could have
3282 led to memory overwriting.
3283
3284 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3285
3286 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3287 operating environments where this matters.
3288
3289 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3290 PCRE options from within the C++ wrapper.
3291
3292 5. Named capturing subpatterns were not being correctly counted when a pattern
3293 was compiled. This caused two problems: (a) If there were more than 100
3294 such subpatterns, the calculation of the memory needed for the whole
3295 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3296 back references of the form \12, where the number was greater than 9, were
3297 not recognized as back references, even though there were sufficient
3298 previous subpatterns.
3299
3300 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3301 versions of gcc, e.g. 2.95.4.
3302
3303
3304 Version 6.1 21-Jun-05
3305 ---------------------
3306
3307 1. There was one reference to the variable "posix" in pcretest.c that was not
3308 surrounded by "#if !defined NOPOSIX".
3309
3310 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3311 the cross-check on the old pcre_info() function, for the benefit of the
3312 cut-down version of PCRE that is currently imported into Exim.
3313
3314 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3315 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3316 patterns that start (?i) but I don't think that matters. The use of (?i) is
3317 just an example; this all applies to the other options as well.
3318
3319 4. Since libtool seems to echo the compile commands it is issuing, the output
3320 from "make" can be reduced a bit by putting "@" in front of each libtool
3321 compile command.
3322
3323 5. Patch from the folks at Google for configure.in to be a bit more thorough
3324 in checking for a suitable C++ installation before trying to compile the
3325 C++ stuff. This should fix a reported problem when a compiler was present,
3326 but no suitable headers.
3327
3328 6. The man pages all had just "PCRE" as their title. I have changed them to
3329 be the relevant file name. I have also arranged that these names are
3330 retained in the file doc/pcre.txt, which is a concatenation in text format
3331 of all the man pages except the little individual ones for each function.
3332
3333 7. The NON-UNIX-USE file had not been updated for the different set of source
3334 files that come with release 6. I also added a few comments about the C++
3335 wrapper.
3336
3337
3338 Version 6.0 07-Jun-05
3339 ---------------------
3340
3341 1. Some minor internal re-organization to help with my DFA experiments.
3342
3343 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3344 didn't matter for the library itself when fully configured, but did matter
3345 when compiling without UCP support, or within Exim, where the ucp files are
3346 not imported.
3347
3348 3. Refactoring of the library code to split up the various functions into
3349 different source modules. The addition of the new DFA matching code (see
3350 below) to a single monolithic source would have made it really too
3351 unwieldy, quite apart from causing all the code to be include in a
3352 statically linked application, when only some functions are used. This is
3353 relevant even without the DFA addition now that patterns can be compiled in
3354 one application and matched in another.
3355
3356 The downside of splitting up is that there have to be some external
3357 functions and data tables that are used internally in different modules of
3358 the library but which are not part of the API. These have all had their
3359 names changed to start with "_pcre_" so that they are unlikely to clash
3360 with other external names.
3361
3362 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3363 a different (DFA) algorithm. Although it is slower than the original
3364 function, it does have some advantages for certain types of matching
3365 problem.
3366
3367 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3368 including restarting after a partial match.
3369
3370 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3371 defined when compiling for Windows was sent to me. I have put it into the
3372 code, though I have no means of testing or verifying it.
3373
3374 7. Added the pcre_refcount() auxiliary function.
3375
3376 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3377 match before or at the first newline in the subject string. In pcretest,
3378 the /f option on a pattern can be used to set this.
3379
3380 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3381 would behave wrongly. This has been present in PCRE since release 4.0.
3382
3383 10. A number of changes to the pcregrep command:
3384
3385 (a) Refactored how -x works; insert ^(...)$ instead of setting
3386 PCRE_ANCHORED and checking the length, in preparation for adding
3387 something similar for -w.
3388
3389 (b) Added the -w (match as a word) option.
3390
3391 (c) Refactored the way lines are read and buffered so as to have more
3392 than one at a time available.
3393
3394 (d) Implemented a pcregrep test script.
3395
3396 (e) Added the -M (multiline match) option. This allows patterns to match
3397 over several lines of the subject. The buffering ensures that at least
3398 8K, or the rest of the document (whichever is the shorter) is available
3399 for matching (and similarly the previous 8K for lookbehind assertions).
3400
3401 (f) Changed the --help output so that it now says
3402
3403 -w, --word-regex(p)
3404
3405 instead of two lines, one with "regex" and the other with "regexp"
3406 because that confused at least one person since the short forms are the
3407 same. (This required a bit of code, as the output is generated
3408 automatically from a table. It wasn't just a text change.)
3409
3410 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3411 option but starts with a hyphen. Could be a pattern or a path name
3412 starting with a hyphen, for instance.
3413
3414 (h) "-" can be given as a file name to represent stdin.
3415
3416 (i) When file names are being printed, "(standard input)" is used for
3417 the standard input, for compatibility with GNU grep. Previously
3418 "<stdin>" was used.
3419
3420 (j) The option --label=xxx can be used to supply a name to be used for
3421 stdin when file names are being printed. There is no short form.
3422
3423 (k) Re-factored the options decoding logic because we are going to add
3424 two more options that take data. Such options can now be given in four
3425 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3426
3427 (l) Added the -A, -B, and -C options for requesting that lines of context
3428 around matches be printed.
3429
3430 (m) Added the -L option to print the names of files that do not contain
3431 any matching lines, that is, the complement of -l.
3432
3433 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3434 continue to scan other files.
3435
3436 (o) The -s option was incorrectly implemented. For compatibility with other
3437 greps, it now suppresses the error message for a non-existent or non-
3438 accessible file (but not the return code). There is a new option called
3439 -q that suppresses the output of matching lines, which was what -s was
3440 previously doing.
3441
3442 (p) Added --include and --exclude options to specify files for inclusion
3443 and exclusion when recursing.
3444
3445 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3446 Hopefully, it now does.
3447
3448 12. Missing cast in pcre_study().
3449
3450 13. Added an "uninstall" target to the makefile.
3451
3452 14. Replaced "extern" in the function prototypes in Makefile.in with
3453 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3454 world, but is set differently for Windows.
3455
3456 15. Added a second compiling function called pcre_compile2(). The only
3457 difference is that it has an extra argument, which is a pointer to an
3458 integer error code. When there is a compile-time failure, this is set
3459 non-zero, in addition to the error test pointer being set to point to an
3460 error message. The new argument may be NULL if no error number is required
3461 (but then you may as well call pcre_compile(), which is now just a
3462 wrapper). This facility is provided because some applications need a
3463 numeric error indication, but it has also enabled me to tidy up the way
3464 compile-time errors are handled in the POSIX wrapper.
3465
3466 16. Added VPATH=.libs to the makefile; this should help when building with one
3467 prefix path and installing with another. (Or so I'm told by someone who
3468 knows more about this stuff than I do.)
3469
3470 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3471 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3472 match everything, including newlines. This is not POSIX-compatible, but
3473 somebody wanted the feature. From pcretest it can be activated by using
3474 both the P and the s flags.
3475
3476 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3477
3478 19. libpcre.pc was being incorrectly installed as executable.
3479
3480 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3481 it now also looks for '\r' so that it will work unmodified on Windows.
3482
3483 21. Added Google's contributed C++ wrapper to the distribution.
3484
3485 22. Added some untidy missing memory free() calls in pcretest, to keep
3486 Electric Fence happy when testing.
3487
3488
3489
3490 Version 5.0 13-Sep-04
3491 ---------------------
3492
3493 1. Internal change: literal characters are no longer packed up into items
3494 containing multiple characters in a single byte-string. Each character
3495 is now matched using a separate opcode. However, there may be more than one
3496 byte in the character in UTF-8 mode.
3497
3498 2. The pcre_callout_block structure has two new fields: pattern_position and
3499 next_item_length. These contain the offset in the pattern to the next match
3500 item, and its length, respectively.
3501
3502 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3503 insertion of callouts before each pattern item. Added the /C option to
3504 pcretest to make use of this.
3505
3506 4. On the advice of a Windows user, the lines
3507
3508 #if defined(_WIN32) || defined(WIN32)
3509 _setmode( _fileno( stdout ), 0x8000 );
3510 #endif /* defined(_WIN32) || defined(WIN32) */
3511
3512 have been added to the source of pcretest. This apparently does useful
3513 magic in relation to line terminators.
3514
3515 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3516 for the benefit of those environments where the "b" makes a difference.
3517
3518 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3519 to know about it. I have put a hack into configure.in that adds in code
3520 to set GCC=yes if CC=icc. This seems to end up at a point in the
3521 generated configure script that is early enough to affect the setting of
3522 compiler options, which is what is needed, but I have no means of testing
3523 whether it really works. (The user who reported this had patched the
3524 generated configure script, which of course I cannot do.)
3525
3526 LATER: After change 22 below (new libtool files), the configure script
3527 seems to know about icc (and also ecc). Therefore, I have commented out
3528 this hack in configure.in.
3529
3530 7. Added support for pkg-config (2 patches were sent in).
3531
3532 8. Negated POSIX character classes that used a combination of internal tables
3533 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3534 [[:^ascii]]. Typically, they would match almost any characters. The other
3535 POSIX classes were not broken in this way.
3536
3537 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3538 to find the match, as PCRE was deluded into thinking that the match had to
3539 start at the start point or following a newline. The same bug applied to
3540 patterns with negative forward assertions or any backward assertions
3541 preceding ".*" at the start, unless the pattern required a fixed first
3542 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3543
3544 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3545 starting at the last subject character, bytes beyond the end of the subject
3546 string were read.
3547
3548 11. Renamed the variable "class" as "classbits" to make life easier for C++
3549 users. (Previously there was a macro definition, but it apparently wasn't
3550 enough.)
3551
3552 12. Added the new field "tables" to the extra data so that tables can be passed
3553 in at exec time, or the internal tables can be re-selected. This allows
3554 a compiled regex to be saved and re-used at a later time by a different
3555 program that might have everything at different addresses.
3556
3557 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3558 -R library as well as a -L library.
3559
3560 14. The debugging options of pcretest (-d on the command line or D on a
3561 pattern) showed incorrect output for anything following an extended class
3562 that contained multibyte characters and which was followed by a quantifier.
3563
3564 15. Added optional support for general category Unicode character properties
3565 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3566 support. It adds about 90K to the size of the library. The meanings of the
3567 inbuilt class escapes such as \d and \s have NOT been changed.
3568
3569 16. Updated pcredemo.c to include calls to free() to release the memory for the
3570 compiled pattern.
3571
3572 17. The generated file chartables.c was being created in the source directory
3573 instead of in the building directory. This caused the build to fail if the
3574 source directory was different from the building directory, and was
3575 read-only.
3576
3577 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3578 file. No doubt somebody will tell me if they don't make sense... Also added
3579 Dan Mooney's comments about building on OpenVMS.
3580
3581 19. Added support for partial matching via the PCRE_PARTIAL option for
3582 pcre_exec() and the \P data escape in pcretest.
3583
3584 20. Extended pcretest with 3 new pattern features:
3585
3586 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3587 write the compiled pattern to the file whose name is "rest-of-line".
3588 This is a straight binary dump of the data, with the saved pointer to
3589 the character tables forced to be NULL. The study data, if any, is
3590 written too. After writing, pcretest reads a new pattern.
3591
3592 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3593 compiled pattern from the given file. There must not be any
3594 occurrences of "<" in the file name (pretty unlikely); if there are,
3595 pcretest will instead treat the initial "<" as a pattern delimiter.
3596 After reading in the pattern, pcretest goes on to read data lines as
3597 usual.
3598
3599 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3600 and 16-bit fields in a compiled pattern, to simulate a pattern that
3601 was compiled on a host of opposite endianness.
3602
3603 21. The pcre-exec() function can now cope with patterns that were compiled on
3604 hosts of opposite endianness, with this restriction:
3605
3606 As for any compiled expression that is saved and used later, the tables
3607 pointer field cannot be preserved; the extra_data field in the arguments
3608 to pcre_exec() should be used to pass in a tables address if a value
3609 other than the default internal tables were used at compile time.
3610
3611 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3612 now diagnosed as an error. Previously, most of the time, a negative number
3613 would have been treated as zero, but if in addition "ovector" was passed as
3614 NULL, a crash could occur.
3615
3616 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3617 new versions from the libtool 1.5 distribution (the last one is a copy of
3618 a file called libtool.m4). This seems to have fixed the need to patch
3619 "configure" to support Darwin 1.3 (which I used to do). However, I still
3620 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3621 workstation).
3622
3623 24. Changed the PCRE licence to be the more standard "BSD" licence.
3624
3625
3626 Version 4.5 01-Dec-03
3627 ---------------------
3628
3629 1. There has been some re-arrangement of the code for the match() function so
3630 that it can be compiled in a version that does not call itself recursively.
3631 Instead, it keeps those local variables that need separate instances for
3632 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3633 needs to "recurse". Keeping track of where control must go is done by means
3634 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3635 hide most of the details from the main code, and operates only if
3636 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3637 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3638 operating.
3639
3640 To make it easier for callers to provide specially tailored get/free
3641 functions for this usage, two new functions, pcre_stack_malloc, and
3642 pcre_stack_free, are used. They are always called in strict stacking order,
3643 and the size of block requested is always the same.
3644
3645 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3646 PCRE has been compiled to use the stack or the heap for recursion. The
3647 -C option of pcretest uses this to show which version is compiled.
3648
3649 A new data escape \S, is added to pcretest; it causes the amounts of store
3650 obtained and freed by both kinds of malloc/free at match time to be added
3651 to the output.
3652
3653 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3654 what's available on my current Linux desktop machine.
3655
3656 3. When matching a UTF-8 string, the test for a valid string at the start has
3657 been extended. If start_offset is not zero, PCRE now checks that it points
3658 to a byte that is the start of a UTF-8 character. If not, it returns
3659 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3660 this is necessary because there may be backward assertions in the pattern.
3661 When matching the same subject several times, it may save resources to use
3662 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3663
3664 4. The code for checking the validity of UTF-8 strings has been tightened so
3665 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3666 containing "overlong sequences".
3667
3668 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3669 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3670 should have been "&", but it just so happened that all the cases this let
3671 through by mistake were picked up later in the function.
3672
3673 6. I had used a variable called "isblank" - this is a C99 function, causing
3674 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3675
3676 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3677 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3678 so the version is shown; (c) stop "make test" echoing "./RunTest".
3679
3680 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3681
3682 9. The prototype for memmove() for systems that don't have it was using
3683 size_t, but the inclusion of the header that defines size_t was later. I've
3684 moved the #includes for the C headers earlier to avoid this.
3685
3686 10. Added some adjustments to the code to make it easier to compiler on certain
3687 special systems:
3688
3689 (a) Some "const" qualifiers were missing.
3690 (b) Added the macro EXPORT before all exported functions; by default this
3691 is defined to be empty.
3692 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3693 that it reads its output file name as an argument instead of writing
3694 to the standard output and assuming this can be redirected.
3695
3696 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3697 class containing characters with values greater than 255, PCRE compilation
3698 went into a loop.
3699
3700 12. A recursive reference to a subpattern that was within another subpattern
3701 that had a minimum quantifier of zero caused PCRE to crash. For example,
3702 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3703 recursion. If the recursively-called subpattern itself had a zero repeat,
3704 that was OK.
3705
3706 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3707 buffer into which it was copied (for escape processing) was still set at
3708 1024, so long lines caused crashes.
3709
3710 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3711 "internal error: code overflow...". This applied to any character class
3712 that was followed by a possessive quantifier.
3713
3714 15. Modified the Makefile to add libpcre.la as a prerequisite for
3715 libpcreposix.la because I was told this is needed for a parallel build to
3716 work.
3717
3718 16. If a pattern that contained .* following optional items at the start was
3719 studied, the wrong optimizing data was generated, leading to matching
3720 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3721 matching string must start with a or b or c. The correct conclusion for
3722 this pattern is that a match can start with any character.
3723
3724
3725 Version 4.4 13-Aug-03
3726 ---------------------
3727
3728 1. In UTF-8 mode, a character class containing characters with values between
3729 127 and 255 was not handled correctly if the compiled pattern was studied.
3730 In fixing this, I have also improved the studying algorithm for such
3731 classes (slightly).
3732
3733 2. Three internal functions had redundant arguments passed to them. Removal
3734 might give a very teeny performance improvement.
3735
3736 3. Documentation bug: the value of the capture_top field in a callout is *one
3737 more than* the number of the hightest numbered captured substring.
3738
3739 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3740 in incorrectly linking with a previously installed version. They now link
3741 explicitly with libpcre.la.
3742
3743 5. configure.in no longer needs to recognize Cygwin specially.
3744
3745 6. A problem in pcre.in for Windows platforms is fixed.
3746
3747 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3748 pcretest, it used to include the size of the study block as part of its
3749 output. Unfortunately, the structure contains a field that has a different
3750 size on different hardware architectures. This meant that the tests that
3751 showed this size failed. As the block is currently always of a fixed size,
3752 this information isn't actually particularly useful in pcretest output, so
3753 I have just removed it.
3754
3755 8. Three pre-processor statements accidentally did not start in column 1.
3756 Sadly, there are *still* compilers around that complain, even though
3757 standard C has not required this for well over a decade. Sigh.
3758
3759 9. In pcretest, the code for checking callouts passed small integers in the
3760 callout_data field, which is a void * field. However, some picky compilers
3761 complained about the casts involved for this on 64-bit systems. Now
3762 pcretest passes the address of the small integer instead, which should get
3763 rid of the warnings.
3764
3765 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3766 both compile and run time, and gives an error if an invalid UTF-8 sequence
3767 is found. There is a option for disabling this check in cases where the
3768 string is known to be correct and/or the maximum performance is wanted.
3769
3770 11. In response to a bug report, I changed one line in Makefile.in from
3771
3772 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3773 to
3774 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3775
3776 to look similar to other lines, but I have no way of telling whether this
3777 is the right thing to do, as I do not use Windows. No doubt I'll get told
3778 if it's wrong...
3779
3780
3781 Version 4.3 21-May-03
3782 ---------------------
3783
3784 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3785 Makefile.
3786
3787 2. Some refactoring to improve the quality of the code:
3788
3789 (i) The utf8_table... variables are now declared "const".
3790
3791 (ii) The code for \cx, which used the "case flipping" table to upper case
3792 lower case letters, now just substracts 32. This is ASCII-specific,
3793 but the whole concept of \cx is ASCII-specific, so it seems
3794 reasonable.
3795
3796 (iii) PCRE was using its character types table to recognize decimal and
3797 hexadecimal digits in the pattern. This is silly, because it handles
3798 only 0-9, a-f, and A-F, but the character types table is locale-
3799 specific, which means strange things might happen. A private
3800 table is now used for this - though it costs 256 bytes, a table is
3801 much faster than multiple explicit tests. Of course, the standard
3802 character types table is still used for matching digits in subject
3803 strings against \d.
3804
3805 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3806 ending in _t are). So I've renamed it as ESC_tee.
3807
3808 3. The first argument for regexec() in the POSIX wrapper should have been
3809 defined as "const".
3810
3811 4. Changed pcretest to use malloc() for its buffers so that they can be
3812 Electric Fenced for debugging.
3813
3814 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3815 to read one or more bytes before the start of the subject string. Often this
3816 had no effect on PCRE's behaviour, but in some circumstances it could
3817 provoke a segmentation fault.
3818
3819 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3820 to try to read one or more bytes before the start of the subject string.
3821
3822 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3823 UTF-8 support could misbehave in various ways if the subject string
3824 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3825 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3826 back over UTF-8 characters.)
3827
3828
3829 Version 4.2 14-Apr-03
3830 ---------------------
3831
3832 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3833
3834 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3835 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3836 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3837 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3838 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3839 and BUILD_EXEEXT
3840 Note: automatic setting of the BUILD variables is not yet working
3841 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3842 compile-time but not at link-time
3843 [LINK]: use for linking executables only
3844 make different versions for Windows and non-Windows
3845 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3846 libraries
3847 [LINK_FOR_BUILD]: new variable
3848 [OBJEXT]: use throughout
3849 [EXEEXT]: use throughout
3850 <winshared>: new target
3851 <wininstall>: new target
3852 <dftables.o>: use native compiler
3853 <dftables>: use native linker
3854 <install>: handle Windows platform correctly
3855 <clean>: ditto
3856 <check>: ditto
3857 copy DLL to top builddir before testing
3858
3859 As part of these changes, -no-undefined was removed again. This was reported
3860 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3861 in any case.
3862
3863 3. Some tidies to get rid of compiler warnings:
3864
3865 . In the match_data structure, match_limit was an unsigned long int, whereas
3866 match_call_count was an int. I've made them both unsigned long ints.
3867
3868 . In pcretest the fact that a const uschar * doesn't automatically cast to
3869 a void * provoked a warning.
3870
3871 . Turning on some more compiler warnings threw up some "shadow" variables
3872 and a few more missing casts.
3873
3874 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3875 option, a class that contained a single character with a value between 128
3876 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3877
3878 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3879 option, a class that contained several characters, but with at least one
3880 whose value was between 128 and 255 caused PCRE to crash.
3881
3882
3883 Version 4.1 12-Mar-03
3884 ---------------------
3885
3886 1. Compiling with gcc -pedantic found a couple of places where casts were
3887 needed, and a string in dftables.c that was longer than standard compilers are
3888 required to support.
3889
3890 2. Compiling with Sun's compiler found a few more places where the code could
3891 be tidied up in order to avoid warnings.
3892
3893 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3894 first of these names is deprecated in the latest Autoconf in favour of the name
3895 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3896 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3897 analogy I have changed it to CFLAGS_FOR_BUILD.
3898
3899 4. Added -no-undefined to the linking command in the Makefile, because this is
3900 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3901 linking step for the pcreposix library.
3902
3903 5. PCRE was failing to diagnose the case of two named groups with the same
3904 name.
3905
3906 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3907 literal character that is needed in the subject for a match, and scans along to
3908 ensure that it is present before embarking on the full matching process. This
3909 saves time in cases of nested unlimited repeats that are never going to match.
3910 Problem: the scan can take a lot of time if the subject is very long (e.g.
3911 megabytes), thus penalizing straightforward matches. It is now done only if the
3912 amount of subject to be scanned is less than 1000 bytes.
3913
3914 7. A lesser problem with the same optimization is that it was recording the
3915 first character of an anchored pattern as "needed", thus provoking a search
3916 right along the subject, even when the first match of the pattern was going to
3917 fail. The "needed" character is now not set for anchored patterns, unless it
3918 follows something in the pattern that is of non-fixed length. Thus, it still
3919 fulfils its original purpose of finding quick non-matches in cases of nested
3920 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3921
3922
3923 Version 4.0 17-Feb-03
3924 ---------------------
3925
3926 1. If a comment in an extended regex that started immediately after a meta-item
3927 extended to the end of string, PCRE compiled incorrect data. This could lead to
3928 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3929
3930 2. Moved to autoconf 2.53 and libtool 1.4.2.
3931
3932 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3933 the special perltest8 script is no longer needed - all the tests can be run
3934 from a single perltest script.
3935
3936 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3937 by \s. It has now been removed in PCRE. This means it isn't recognized as
3938 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3939 class [:space:] *does* include VT, thereby creating a mess.
3940
3941 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3942 space and tab.
3943
3944 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3945 its new features into the main test script, reducing the number of scripts.
3946
3947 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3948 were backward compatible, and made the (?i) apply to the whole pattern, as if
3949 /i were given. Now it behaves more logically, and applies the option setting
3950 only to what follows. PCRE has been changed to follow suit. However, if it
3951 finds options settings right at the start of the pattern, it extracts them into
3952 the global options, as before. Thus, they show up in the info data.
3953
3954 8. Added support for the \Q...\E escape sequence. Characters in between are
3955 treated as literals. This is slightly different from Perl in that $ and @ are
3956 also handled as literals inside the quotes. In Perl, they will cause variable
3957 interpolation. Note the following examples:
3958
3959 Pattern PCRE matches Perl matches
3960
3961 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3962 \Qabc\$xyz\E abc\$xyz abc\$xyz
3963 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3964
3965 For compatibility with Perl, \Q...\E sequences are recognized inside character
3966 classes as well as outside them.
3967
3968 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3969 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3970 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3971 signed/unsigned warnings.
3972
3973 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3974 option for pcretest, so I've replaced it by a simple function that does just
3975 that job.
3976
3977 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3978 "pcregrep -".
3979
3980 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3981 Java package. This provides some syntactic sugar for simple cases of what my
3982 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3983 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3984 item, you can use this simplified notation. Note that only makes sense with
3985 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3986 greediness, whatever the setting of the PCRE_UNGREEDY option.
3987
3988 13. A change of greediness default within a pattern was not taking effect at
3989 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3990 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3991 was abstracted outside.
3992
3993 14. PCRE now supports the \G assertion. It is true when the current matching
3994 position is at the start point of the match. This differs from \A when the
3995 starting offset is non-zero. Used with the /g option of pcretest (or similar
3996 code), it works in the same way as it does for Perl's /g option. If all
3997 alternatives of a regex begin with \G, the expression is anchored to the start
3998 match position, and the "anchored" flag is set in the compiled expression.
3999
4000 15. Some bugs concerning the handling of certain option changes within patterns
4001 have been fixed. These applied to options other than (?ims). For example,
4002 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4003 been the other way round. Some of this was related to change 7 above.
4004
4005 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4006 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4007 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4008 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4009
4010 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4011 mode. Unlike ".", it always matches newline, whatever the setting of
4012 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4013 assertions. Perl allows it, but it doesn't (in general) work because it can't
4014 calculate the length of the lookbehind. At least, that's the case for Perl
4015 5.8.0 - I've been told they are going to document that it doesn't work in
4016 future.
4017
4018 18. Added an error diagnosis for escapes that PCRE does not support: these are
4019 \L, \l, \N, \P, \p, \U, \u, and \X.
4020
4021 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4022 reading past the end of the pattern in cases such as /[abcd/.
4023
4024 20. PCRE was getting more memory than necessary for patterns with classes that
4025 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4026
4027 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4028 compiling PCRE for use with Virtual Pascal.
4029
4030 22. Small fix to the Makefile to make it work properly if the build is done
4031 outside the source tree.
4032
4033 23. Added a new extension: a condition to go with recursion. If a conditional
4034 subpattern starts with (?(R) the "true" branch is used if recursion has
4035 happened, whereas the "false" branch is used only at the top level.
4036
4037 24. When there was a very long string of literal characters (over 255 bytes
4038 without UTF support, over 250 bytes with UTF support), the computation of how
4039 much memory was required could be incorrect, leading to segfaults or other
4040 strange effects.
4041
4042 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4043 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4044 there was a subsequent back reference to those brackets. This meant that, for
4045 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4046 possible to check for precisely this case. All we can do is abandon the
4047 optimization if .* occurs inside capturing brackets when there are any back
4048 references whatsoever. (See below for a better fix that came later.)
4049
4050 26. The handling of the optimization for finding the first character of a
4051 non-anchored pattern, and for finding a character that is required later in the
4052 match were failing in some cases. This didn't break the matching; it just
4053 failed to optimize when it could. The way this is done has been re-implemented.
4054
4055 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4056
4057 28. Added a new feature that provides some of the functionality that Perl
4058 provides with (?{...}). The facility is termed a "callout". The way it is done
4059 in PCRE is for the caller to provide an optional function, by setting
4060 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4061 global variable. By default it is unset, which disables all calling out. To get
4062 the function called, the regex must include (?C) at appropriate points. This
4063 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4064 This provides a means of identifying different callout points. When PCRE
4065 reaches such a point in the regex, if pcre_callout has been set, the external
4066 function is called. It is provided with data in a structure called
4067 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4068 matching continues; if it returns a non-zero value, the match at the current
4069 point fails. However, backtracking will occur if possible. [This was changed
4070 later and other features added - see item 49 below.]
4071
4072 29. pcretest is upgraded to test the callout functionality. It provides a
4073 callout function that displays information. By default, it shows the start of
4074 the match and the current position in the text. There are some new data escapes
4075 to vary what happens:
4076
4077 \C+ in addition, show current contents of captured substrings
4078 \C- do not supply a callout function
4079 \C!n return 1 when callout number n is reached
4080 \C!n!m return 1 when callout number n is reached for the mth time
4081
4082 30. If pcregrep was called with the -l option and just a single file name, it
4083 output "<stdin>" if a match was found, instead of the file name.
4084
4085 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4086 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4087 pcre_exec(). This saves a malloc/free per call. The default value of
4088 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4089 when configuring.
4090
4091 32. The default maximum size of a compiled pattern is 64K. There have been a
4092 few cases of people hitting this limit. The code now uses macros to handle the
4093 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4094 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4095 configuring. Tests 2 and 5 work only with 2-byte links because they output
4096 debugging information about compiled patterns.
4097
4098 33. Internal code re-arrangements:
4099
4100 (a) Moved the debugging function for printing out a compiled regex into
4101 its own source file (printint.c) and used #include to pull it into
4102 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4103 separate copies.
4104
4105 (b) Defined the list of op-code names for debugging as a macro in
4106 internal.h so that it is next to the definition of the opcodes.
4107
4108 (c) Defined a table of op-code lengths for simpler skipping along compiled
4109 code. This is again a macro in internal.h so that it is next to the
4110 definition of the opcodes.
4111
4112 34. Added support for recursive calls to individual subpatterns, along the
4113 lines of Robin Houston's patch (but implemented somewhat differently).
4114
4115 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4116 allow it to read and process whole directories in Win32. This code was
4117 contributed by Lionel Fourquaux; it has not been tested by me.
4118
4119 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4120 used to name a group. Names consist of alphanumerics and underscores, and must
4121 be unique. Back references use the syntax (?P=name) and recursive calls use
4122 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4123 numbers. The function pcre_fullinfo() can be used after compilation to extract
4124 a name/number map. There are three relevant calls:
4125
4126 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4127 PCRE_INFO_NAMECOUNT yields the number of entries
4128 PCRE_INFO_NAMETABLE yields a pointer to the map.
4129
4130 The map is a vector of fixed-size entries. The size of each entry depends on
4131 the length of the longest name used. The first two bytes of each entry are the
4132 group number, most significant byte first. There follows the corresponding
4133 name, zero terminated. The names are in alphabetical order.
4134
4135 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4136 case instead of 255. Making it the same both with and without UTF-8 support
4137 means that the same test output works with both.
4138
4139 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4140 calling malloc() with a zero argument.
4141
4142 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4143 optimization. I've improved things by keeping a bitmap of backreferences with
4144 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4145 fact referenced, the optimization can be applied. It is unlikely that a
4146 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4147 the match to follow \n) will appear inside brackets with a number greater than
4148 31, but if it does, any back reference > 31 suppresses the optimization.
4149
4150 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4151 of disabling numbered capturing parentheses. Any opening parenthesis that is
4152 not followed by ? behaves as if it were followed by ?: but named parentheses
4153 can still be used for capturing (and they will acquire numbers in the usual
4154 way).
4155
4156 41. Redesigned the return codes from the match() function into yes/no/error so
4157 that errors can be passed back from deep inside the nested calls. A malloc
4158 failure while inside a recursive subpattern call now causes the
4159 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4160
4161 42. It is now possible to set a limit on the number of times the match()
4162 function is called in a call to pcre_exec(). This facility makes it possible to
4163 limit the amount of recursion and backtracking, though not in a directly
4164 obvious way, because the match() function is used in a number of different
4165 circumstances. The count starts from zero for each position in the subject
4166 string (for non-anchored patterns). The default limit is, for compatibility, a
4167 large number, namely 10 000 000. You can change this in two ways:
4168
4169 (a) When configuring PCRE before making, you can use --with-match-limit=n
4170 to set a default value for the compiled library.
4171
4172 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4173 a different value is set. See 45 below.
4174
4175 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4176
4177 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4178 of things that can be changed at compile time. The first argument specifies
4179 what is wanted and the second points to where the information is to be placed.
4180 The current list of available information is:
4181
4182 PCRE_CONFIG_UTF8
4183
4184 The output is an integer that is set to one if UTF-8 support is available;
4185 otherwise it is set to zero.
4186
4187 PCRE_CONFIG_NEWLINE
4188
4189 The output is an integer that it set to the value of the code that is used for
4190 newline. It is either LF (10) or CR (13).
4191
4192 PCRE_CONFIG_LINK_SIZE
4193
4194 The output is an integer that contains the number of bytes used for internal
4195 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4196
4197 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4198
4199 The output is an integer that contains the threshold above which the POSIX
4200 interface uses malloc() for output vectors. See item 31 above.
4201
4202 PCRE_CONFIG_MATCH_LIMIT
4203
4204 The output is an unsigned integer that contains the default limit of the number
4205 of match() calls in a pcre_exec() execution. See 42 above.
4206
4207 44. pcretest has been upgraded by the addition of the -C option. This causes it
4208 to extract all the available output from the new pcre_config() function, and to
4209 output it. The program then exits immediately.
4210
4211 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4212 order to support additional features. One way would have been to define
4213 pcre_exec2() (for example) with extra arguments, but this would not have been
4214 extensible, and would also have required all calls to the original function to
4215 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4216 is used for passing in "extra" data from pcre_study().
4217
4218 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4219 contains the following fields:
4220
4221 flags a bitmap indicating which of the following fields are set
4222 study_data opaque data from pcre_study()
4223 match_limit a way of specifying a limit on match() calls for a specific
4224 call to pcre_exec()
4225 callout_data data for callouts (see 49 below)
4226
4227 The flag bits are also defined in pcre.h, and are
4228
4229 PCRE_EXTRA_STUDY_DATA
4230 PCRE_EXTRA_MATCH_LIMIT
4231 PCRE_EXTRA_CALLOUT_DATA
4232
4233 The pcre_study() function now returns one of these new pcre_extra blocks, with
4234 the actual study data pointed to by the study_data field, and the
4235 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4236 before. That is, this change is entirely upwards-compatible and requires no
4237 change to existing code.
4238
4239 If you want to pass in additional data to pcre_exec(), you can either place it
4240 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4241 block.
4242
4243 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4244 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4245 times with different match limits, until it finds the minimum value needed for
4246 pcre_exec() to complete. The value is then output. This can be instructive; for
4247 most simple matches the number is quite small, but for pathological cases it
4248 gets very large very quickly.
4249
4250 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4251 returns the size of the data block pointed to by the study_data field in a
4252 pcre_extra block, that is, the value that was passed as the argument to
4253 pcre_malloc() when PCRE was getting memory in which to place the information
4254 created by pcre_study(). The fourth argument should point to a size_t variable.
4255 pcretest has been extended so that this information is shown after a successful
4256 pcre_study() call when information about the compiled regex is being displayed.
4257
4258 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4259 because what follows is always an absolute path. (Later: it turns out that this
4260 is more than cosmetic for MinGW, because it doesn't like empty path
4261 components.)
4262
4263 49. Some changes have been made to the callout feature (see 28 above):
4264
4265 (i) A callout function now has three choices for what it returns:
4266
4267 0 => success, carry on matching
4268 > 0 => failure at this point, but backtrack if possible
4269 < 0 => serious error, return this value from pcre_exec()
4270
4271 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4272 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4273 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4274 use by callout functions. It will never be used by PCRE itself.
4275
4276 (ii) The pcre_extra structure (see 45 above) has a void * field called
4277 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4278 pcre_callout_block structure has a field of the same name. The contents of
4279 the field passed in the pcre_extra structure are passed to the callout
4280 function in the corresponding field in the callout block. This makes it
4281 easier to use the same callout-containing regex from multiple threads. For
4282 testing, the pcretest program has a new data escape
4283
4284 \C*n pass the number n (may be negative) as callout_data
4285
4286 If the callout function in pcretest receives a non-zero value as
4287 callout_data, it returns that value.
4288
4289 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4290 there were some redundant $(CFLAGS) in commands that are now specified as
4291 $(LINK), which already includes $(CFLAGS).
4292
4293 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4294 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4295 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4296 one-byte characters throughout. Note that case-insensitive matching applies
4297 only to characters whose values are less than 256. PCRE doesn't support the
4298 notion of cases for higher-valued characters.
4299
4300 (i) A character class whose characters are all within 0-255 is handled as
4301 a bit map, and the map is inverted for negative classes. Previously, a
4302 character > 255 always failed to match such a class; however it should
4303 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4304
4305 (ii) A negated character class with a single character < 255 is coded as
4306 "not this character" (OP_NOT). This wasn't working properly when the test
4307 character was multibyte, either singly or repeated.
4308
4309 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4310 mode, for example: \x{100}{2,3}.
4311
4312 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4313 singly or repeated) now correctly test multibyte characters. However,
4314 PCRE doesn't recognize any characters with values greater than 255 as
4315 digits, spaces, or word characters. Such characters always match \D, \S,
4316 and \W, and never match \d, \s, or \w.
4317
4318 (v) Classes may now contain characters and character ranges with values
4319 greater than 255. For example: [ab\x{100}-\x{400}].
4320
4321 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4322 PCRE in UTF-8 mode.
4323
4324 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4325 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4326 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4327 value.)
4328
4329 53. The single man page has become too large. I have therefore split it up into
4330 a number of separate man pages. These also give rise to individual HTML pages;
4331 these are now put in a separate directory, and there is an index.html page that
4332 lists them all. Some hyperlinking between the pages has been installed.
4333
4334 54. Added convenience functions for handling named capturing parentheses.
4335
4336 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4337 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4338 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4339 are faulted.
4340
4341 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4342 calling configure. These values are used when compiling the dftables.c program
4343 which is run to generate the source of the default character tables. They
4344 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4345 you will need to set these values.
4346
4347 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4348
4349
4350 Version 3.9 02-Jan-02
4351 ---------------------
4352
4353 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4354
4355 2. If --disable-static was given, the building process failed when trying to
4356 build pcretest and pcregrep. (For some reason it was using libtool to compile
4357 them, which is not right, as they aren't part of the library.)
4358
4359
4360 Version 3.8 18-Dec-01
4361 ---------------------
4362
4363 1. The experimental UTF-8 code was completely screwed up. It was packing the
4364 bytes in the wrong order. How dumb can you get?
4365
4366
4367 Version 3.7 29-Oct-01
4368 ---------------------
4369
4370 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4371 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4372 this didn't happen under Solaris 8, where I normally test things.
4373
4374 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4375 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4376 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4377 this fix an hour or so after the initial 3.7 release.)
4378
4379
4380 Version 3.6 23-Oct-01
4381 ---------------------
4382
4383 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4384 offsets passed as NULL with zero offset count.
4385
4386 2. The config.guess and config.sub files had not been updated when I moved to
4387 the latest autoconf.
4388
4389
4390 Version 3.5 15-Aug-01
4391 ---------------------
4392
4393 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4394 had been forgotten.
4395
4396 2. By using declared but undefined structures, we can avoid using "void"
4397 definitions in pcre.h while keeping the internal definitions of the structures
4398 private.
4399
4400 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4401 user point of view, this means that both static and shared libraries are built
4402 by default, but this can be individually controlled. More of the work of
4403 handling this static/shared cases is now inside libtool instead of PCRE's make
4404 file.
4405
4406 4. The pcretest utility is now installed along with pcregrep because it is
4407 useful for users (to test regexs) and by doing this, it automatically gets
4408 relinked by libtool. The documentation has been turned into a man page, so
4409 there are now .1, .txt, and .html versions in /doc.
4410
4411 5. Upgrades to pcregrep:
4412 (i) Added long-form option names like gnu grep.
4413 (ii) Added --help to list all options with an explanatory phrase.
4414 (iii) Added -r, --recursive to recurse into sub-directories.
4415 (iv) Added -f, --file to read patterns from a file.
4416
4417 6. pcre_exec() was referring to its "code" argument before testing that
4418 argument for NULL (and giving an error if it was NULL).
4419
4420 7. Upgraded Makefile.in to allow for compiling in a different directory from
4421 the source directory.
4422
4423 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4424 options bits, the pointer it was passed was to an int instead of to an unsigned
4425 long int. This mattered only on 64-bit systems.
4426
4427 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4428 generated) instead of pcre.in, which it its source. Also made the same change
4429 in several of the .c files.
4430
4431 10. A new release of gcc defines printf() as a macro, which broke pcretest
4432 because it had an ifdef in the middle of a string argument for printf(). Fixed
4433 by using separate calls to printf().
4434
4435 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4436 script, to force use of CR or LF instead of \n in the source. On non-Unix
4437 systems, the value can be set in config.h.
4438
4439 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4440 absolute limit. Changed the text of the error message to make this clear, and
4441 likewise updated the man page.
4442
4443 13. The limit of 99 on the number of capturing subpatterns has been removed.
4444 The new limit is 65535, which I hope will not be a "real" limit.
4445
4446
4447 Version 3.4 22-Aug-00
4448 ---------------------
4449
4450 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4451
4452 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4453
4454
4455 Version 3.3 01-Aug-00
4456 ---------------------
4457
4458 1. If an octal character was given, but the value was greater than \377, it
4459 was not getting masked to the least significant bits, as documented. This could
4460 lead to crashes in some systems.
4461
4462 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4463 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4464
4465 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4466 These just pass their arguments on to (pcre_free)(), but they are provided
4467 because some uses of PCRE bind it to non-C systems that can call its functions,
4468 but cannot call free() or pcre_free() directly.
4469
4470 4. Add "make test" as a synonym for "make check". Corrected some comments in
4471 the Makefile.
4472
4473 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4474 Makefile.
4475
4476 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4477 command called pgrep for grepping around the active processes.
4478
4479 7. Added the beginnings of support for UTF-8 character strings.
4480
4481 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4482 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4483 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4484 out for the ar command.)
4485
4486
4487 Version 3.2 12-May-00
4488 ---------------------
4489
4490 This is purely a bug fixing release.
4491
4492 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4493 of ZA. This was just one example of several cases that could provoke this bug,
4494 which was introduced by change 9 of version 2.00. The code for breaking
4495 infinite loops after an iteration that matches an empty string was't working
4496 correctly.
4497
4498 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4499 when matched against abbab (for example). After matching an empty string, it
4500 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4501 caused it to match further down the string than it should.
4502
4503 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4504 was there because it doesn't seem to be needed, and it causes trouble on some
4505 systems, as it is not a Standard C header. It has been removed.
4506
4507 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4508 were reported on the Macintosh. The changes were from
4509
4510 while ((c = *(++ptr)) != 0 && c != '\n');
4511 to
4512 while ((c = *(++ptr)) != 0 && c != '\n') ;
4513
4514 Totally extraordinary, but if that's what it takes...
4515
4516 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4517 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4518 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4519 assumes the way PCRE uses memmove() (always moving upwards).
4520
4521 6. PCRE is being used in one environment where strchr() is not available. There
4522 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4523 faster code anyway.
4524
4525
4526 Version 3.1 09-Feb-00
4527 ---------------------
4528
4529 The only change in this release is the fixing of some bugs in Makefile.in for
4530 the "install" target:
4531
4532 (1) It was failing to install pcreposix.h.
4533
4534 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4535
4536
4537 Version 3.0 01-Feb-00
4538 ---------------------
4539
4540 1. Add support for the /+ modifier to perltest (to output $` like it does in
4541 pcretest).
4542
4543 2. Add support for the /g modifier to perltest.
4544
4545 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4546 matches null strings.
4547
4548 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4549 pattern. Perl treats empty patterns specially - it reuses the most recent
4550 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4551 effect.
4552
4553 5. The POSIX interface was broken in that it was just handing over the POSIX
4554 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4555 required a bigger vector, with some working space on the end. This means that
4556 the POSIX wrapper now has to get and free some memory, and copy the results.
4557
4558 6. Added some simple autoconf support, placing the test data and the
4559 documentation in separate directories, re-organizing some of the
4560 information files, and making it build pcre-config (a GNU standard). Also added
4561 libtool support for building PCRE as a shared library, which is now the
4562 default.
4563
4564 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4565 09 are not valid octal constants. Single digits will be used for minor values
4566 less than 10.
4567
4568 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4569 existing programs that set these in the POSIX interface can use PCRE without
4570 modification.
4571
4572 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4573 return all that pcre_info() returns, plus additional data. The pcre_info()
4574 function is retained for compatibility, but is considered to be obsolete.
4575
4576 10. Added experimental recursion feature (?R) to handle one common case that
4577 Perl 5.6 will be able to do with (?p{...}).
4578
4579 11. Added support for POSIX character classes like [:alpha:], which Perl is
4580 adopting.
4581
4582
4583 Version 2.08 31-Aug-99
4584 ----------------------
4585
4586 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4587 trying to match at the startoffset position, but instead was moving forward to
4588 the next newline as if a previous match had failed.
4589
4590 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4591 and could get into a loop if a null string was matched other than at the start
4592 of the subject.
4593
4594 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4595 be distinguished at compile time, and for completeness also added PCRE_DATE.
4596
4597 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4598 in GnuWin32 environments.
4599
4600
4601 Version 2.07 29-Jul-99
4602 ----------------------
4603
4604 1. The documentation is now supplied in plain text form and HTML as well as in
4605 the form of man page sources.
4606
4607 2. C++ compilers don't like assigning (void *) values to other pointer types.
4608 In particular this affects malloc(). Although there is no problem in Standard
4609 C, I've put in casts to keep C++ compilers happy.
4610
4611 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4612 should be (const char *).
4613
4614 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4615 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4616 However, I haven't made this a standard facility. The documentation doesn't
4617 mention it, and the Makefile doesn't support it.
4618
4619 5. The Makefile now contains an "install" target, with editable destinations at
4620 the top of the file. The pcretest program is not installed.
4621
4622 6. pgrep -V now gives the PCRE version number and date.
4623
4624 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4625 causing the entire string to be ignored, instead of just the last character.
4626
4627 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4628 non-matching string, it can take a very, very long time, even for strings of
4629 quite modest length, because of the nested recursion. PCRE now does better in
4630 some of these cases. It does this by remembering the last required literal
4631 character in the pattern, and pre-searching the subject to ensure it is present
4632 before running the real match. In other words, it applies a heuristic to detect
4633 some types of certain failure quickly, and in the above example, if presented
4634 with a string that has no trailing " it gives "no match" very quickly.
4635
4636 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4637 other alternatives are tried instead.
4638
4639
4640 Version 2.06 09-Jun-99
4641 ----------------------
4642
4643 1. Change pcretest's output for amount of store used to show just the code
4644 space, because the remainder (the data block) varies in size between 32-bit and
4645 64-bit systems.
4646
4647 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4648 start matching at. This allows lookbehinds to work when searching for multiple
4649 occurrences in a string.
4650
4651 3. Added additional options to pcretest for testing multiple occurrences:
4652
4653 /+ outputs the rest of the string that follows a match
4654 /g loops for multiple occurrences, using the new startoffset argument
4655 /G loops for multiple occurrences by passing an incremented pointer
4656
4657 4. PCRE wasn't doing the "first character" optimization for patterns starting
4658 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4659 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4660 the letter 'x'. On long subject strings, this gives a significant speed-up.
4661
4662
4663 Version 2.05 21-Apr-99
4664 ----------------------
4665
4666 1. Changed the type of magic_number from int to long int so that it works
4667 properly on 16-bit systems.
4668
4669 2. Fixed a bug which caused patterns starting with .* not to work correctly
4670 when the subject string contained newline characters. PCRE was assuming
4671 anchoring for such patterns in all cases, which is not correct because .* will
4672 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4673 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4674 must be retried after every newline in the subject.
4675
4676
4677 Version 2.04 18-Feb-99
4678 ----------------------
4679
4680 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4681 computation of the store needed to hold the pattern was incorrect (too large).
4682 If such patterns were nested a few deep, this could multiply and become a real
4683 problem.
4684
4685 2. Added /M option to pcretest to show the memory requirement of a specific
4686 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4687
4688 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4689 compiled in such a way that the backtracking after subsequent failure was
4690 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4691 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4692
4693
4694 Version 2.03 02-Feb-99
4695 ----------------------
4696
4697 1. Fixed typo and small mistake in man page.
4698
4699 2. Added 4th condition (GPL supersedes if conflict) and created separate
4700 LICENCE file containing the conditions.
4701
4702 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4703 Perl, that is the internal \ allows the delimiter to be included in the
4704 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4705 the final delimiter, add \ to the end of the pattern (to test the error).
4706
4707 4. Added the convenience functions for extracting substrings after a successful
4708 match. Updated pcretest to make it able to test these functions.
4709
4710
4711 Version 2.02 14-Jan-99
4712 ----------------------
4713
4714 1. Initialized the working variables associated with each extraction so that
4715 their saving and restoring doesn't refer to uninitialized store.
4716
4717 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4718 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4719 fix the problem.
4720
4721 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4722 calls, and wasn't printing the correct value for compiling calls. Increased the
4723 default value of LOOPREPEAT, and the number of significant figures in the
4724 times.
4725
4726 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4727
4728 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4729 a building problem on Windows NT with a FAT file system.
4730
4731
4732 Version 2.01 21-Oct-98
4733 ----------------------
4734
4735 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4736 to character tables built by pcre_maketables() in the current locale. If NULL
4737 is passed, the default tables are used.
4738
4739
4740 Version 2.00 24-Sep-98
4741 ----------------------
4742
4743 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4744 it any more.
4745
4746 2. Allow quantification of (?>) groups, and make it work correctly.
4747
4748 3. The first character computation wasn't working for (?>) groups.
4749
4750 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4751 end of the subject) and add 5.005's \z, which really does match only at the
4752 very end of the subject.
4753
4754 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4755
4756 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4757 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4758 localized options. All options to pcre_study() were also removed.
4759
4760 7. Add other new features from 5.005:
4761
4762 $(?<= positive lookbehind
4763 $(?<! negative lookbehind
4764 (?imsx-imsx) added the unsetting capability
4765 such a setting is global if at outer level; local otherwise
4766 (?imsx-imsx:) non-capturing groups with option setting
4767 (?(cond)re|re) conditional pattern matching
4768
4769 A backreference to itself in a repeated group matches the previous
4770 captured string.
4771
4772 8. General tidying up of studying (both automatic and via "study")
4773 consequential on the addition of new assertions.
4774
4775 9. As in 5.005, unlimited repeated groups that could match an empty substring
4776 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4777 runtime if any iteration does actually match an empty substring.
4778
4779 10. Include the RunTest script in the distribution.
4780
4781 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4782 discrepancies, some of which were old and were also with respect to 5.004. They
4783 have now been fixed.
4784
4785
4786 Version 1.09 28-Apr-98
4787 ----------------------
4788
4789 1. A negated single character class followed by a quantifier with a minimum
4790 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4791 program crashes, or just wrong answers. This did not apply to negated classes
4792 containing more than one character, or to minima other than one.
4793
4794
4795 Version 1.08 27-Mar-98
4796 ----------------------
4797
4798 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4799
4800 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4801 latter must appear before anything that relies on it in the pattern.
4802
4803
4804 Version 1.07 16-Feb-98
4805 ----------------------
4806
4807 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4808 repeat of a potentially empty string).
4809
4810
4811 Version 1.06 23-Jan-98
4812 ----------------------
4813
4814 1. Added Markus Oberhumer's little patches for C++.
4815
4816 2. Literal strings longer than 255 characters were broken.
4817
4818
4819 Version 1.05 23-Dec-97
4820 ----------------------
4821
4822 1. Negated character classes containing more than one character were failing if
4823 PCRE_CASELESS was set at run time.
4824
4825
4826 Version 1.04 19-Dec-97
4827 ----------------------
4828
4829 1. Corrected the man page, where some "const" qualifiers had been omitted.
4830
4831 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4832 input syntax.
4833
4834 3. Fixed memory leak which occurred when a regex with back references was
4835 matched with an offsets vector that wasn't big enough. The temporary memory
4836 that is used in this case wasn't being freed if the match failed.
4837
4838 4. Tidied pcretest to ensure it frees memory that it gets.
4839
4840 5. Temporary memory was being obtained in the case where the passed offsets
4841 vector was exactly big enough.
4842
4843 6. Corrected definition of offsetof() from change 5 below.
4844
4845 7. I had screwed up change 6 below and broken the rules for the use of
4846 setjmp(). Now fixed.
4847
4848
4849 Version 1.03 18-Dec-97
4850 ----------------------
4851
4852 1. A erroneous regex with a missing opening parenthesis was correctly
4853 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4854 on some systems.
4855
4856 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4857 it was reported that one broken compiler failed on the former because "code" is
4858 also an independent variable.
4859
4860 3. The erroneous regex a[]b caused an array overrun reference.
4861
4862 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4863 fail on data ending with that character. (It was going on too far, and checking
4864 the next character, typically a binary zero.) This was specific to the
4865 optimized code for single-character negative classes.
4866
4867 5. Added a contributed patch from the TIN world which does the following:
4868
4869 + Add an undef for memmove, in case the the system defines a macro for it.
4870
4871 + Add a definition of offsetof(), in case there isn't one. (I don't know
4872 the reason behind this - offsetof() is part of the ANSI standard - but
4873 it does no harm).
4874
4875 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4876 most of the places where whitespace preceded '#'. I have given up and
4877 allowed the remaining 2 cases to be at the margin.
4878
4879 + Rename some variables in pcre to eliminate shadowing. This seems very
4880 pedantic, but does no harm, of course.
4881
4882 6. Moved the call to setjmp() into its own function, to get rid of warnings
4883 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4884
4885 7. Constructs such as \d{8,} were compiling into the equivalent of
4886 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4887 outcome, but in this particular case used more store than had been allocated,
4888 which caused the bug to be discovered because it threw up an internal error.
4889
4890 8. The debugging code in both pcre and pcretest for outputting the compiled
4891 form of a regex was going wrong in the case of back references followed by
4892 curly-bracketed repeats.
4893
4894
4895 Version 1.02 12-Dec-97
4896 ----------------------
4897
4898 1. Typos in pcre.3 and comments in the source fixed.
4899
4900 2. Applied a contributed patch to get rid of places where it used to remove
4901 'const' from variables, and fixed some signed/unsigned and uninitialized
4902 variable warnings.
4903
4904 3. Added the "runtest" target to Makefile.
4905
4906 4. Set default compiler flag to -O2 rather than just -O.
4907
4908
4909 Version 1.01 19-Nov-97
4910 ----------------------
4911
4912 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4913 like /([ab]*)*/, that is, for classes with more than one character in them.
4914
4915 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4916 as /((?>a*))*/ (a PCRE_EXTRA facility).
4917
4918
4919 Version 1.00 18-Nov-97
4920 ----------------------
4921
4922 1. Added compile-time macros to support systems such as SunOS4 which don't have
4923 memmove() or strerror() but have other things that can be used instead.
4924
4925 2. Arranged that "make clean" removes the executables.
4926
4927
4928 Version 0.99 27-Oct-97
4929 ----------------------
4930
4931 1. Fixed bug in code for optimizing classes with only one character. It was
4932 initializing a 32-byte map regardless, which could cause it to run off the end
4933 of the memory it had got.
4934
4935 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4936
4937
4938 Version 0.98 22-Oct-97
4939 ----------------------
4940
4941 1. Fixed bug in code for handling temporary memory usage when there are more
4942 back references than supplied space in the ovector. This could cause segfaults.
4943
4944
4945 Version 0.97 21-Oct-97
4946 ----------------------
4947
4948 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4949
4950 2. Optimized negated single characters not to use a bit map.
4951
4952 3. Brought error texts together as macro definitions; clarified some of them;
4953 fixed one that was wrong - it said "range out of order" when it meant "invalid
4954 escape sequence".
4955
4956 4. Changed some char * arguments to const char *.
4957
4958 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4959
4960 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4961 pcretest.
4962
4963
4964 Version 0.96 16-Oct-97
4965 ----------------------
4966
4967 1. Added a simple "pgrep" utility to the distribution.
4968
4969 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4970 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4971 where "ddd" means "one or more decimal digits".
4972
4973 3. Fixed serious bug. If a pattern had a back reference, but the call to
4974 pcre_exec() didn't supply a large enough ovector to record the related
4975 identifying subpattern, the match always failed. PCRE now remembers the number
4976 of the largest back reference, and gets some temporary memory in which to save
4977 the offsets during matching if necessary, in order to ensure that
4978 backreferences always work.
4979
4980 4. Increased the compatibility with Perl in a number of ways:
4981
4982 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4983 to request this handling. The option can be set at compile or exec time.
4984
4985 (b) $ matches before a terminating newline by default; an option
4986 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4987 mode). The option can be set at compile or exec time.
4988
4989 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4990 the same as Perl's. If the decimal number it represents is less than 10
4991 or there aren't that many previous left capturing parentheses, an octal
4992 escape is read. Inside a character class, it's always an octal escape,
4993 even if it is a single digit.
4994
4995 (d) An escaped but undefined alphabetic character is taken as a literal,
4996 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4997 escapes.
4998
4999 (e) {0} is now permitted. (The previous item is removed from the compiled
5000 pattern).
5001
5002 5. Changed all the names of code files so that the basic parts are no longer
5003 than 10 characters, and abolished the teeny "globals.c" file.
5004
5005 6. Changed the handling of character classes; they are now done with a 32-byte
5006 bit map always.
5007
5008 7. Added the -d and /D options to pcretest to make it possible to look at the
5009 internals of compilation without having to recompile pcre.
5010
5011
5012 Version 0.95 23-Sep-97
5013 ----------------------
5014
5015 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5016 \x20 at the start of a run of normal characters. These were being treated as
5017 real characters, instead of the source characters being re-checked.
5018
5019
5020 Version 0.94 18-Sep-97
5021 ----------------------
5022
5023 1. The functions are now thread-safe, with the caveat that the global variables
5024 containing pointers to malloc() and free() or alternative functions are the
5025 same for all threads.
5026
5027 2. Get pcre_study() to generate a bitmap of initial characters for non-
5028 anchored patterns when this is possible, and use it if passed to pcre_exec().
5029
5030
5031 Version 0.93 15-Sep-97
5032 ----------------------
5033
5034 1. /(b)|(:+)/ was computing an incorrect first character.
5035
5036 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5037 but not actually doing anything yet.
5038
5039 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5040 as Perl does (e.g. [-az] or [az-]).
5041
5042 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5043 all possible positions.
5044
5045 5. Split up into different modules to avoid including unneeded functions in a
5046 compiled binary. However, compile and exec are still in one module. The "study"
5047 function is split off.
5048
5049 6. The character tables are now in a separate module whose source is generated
5050 by an auxiliary program - but can then be edited by hand if required. There are
5051 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5052 toupper() in the code.
5053
5054 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5055 make them global. Abolish the function for setting them, as the caller can now
5056 set them directly.
5057
5058
5059 Version 0.92 11-Sep-97
5060 ----------------------
5061
5062 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5063 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5064
5065 2. Caseless matching was not working in character classes if the characters in
5066 the pattern were in upper case.
5067
5068 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5069
5070 4. Make PCRE_ANCHORED public and accept as a compile option.
5071
5072 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5073 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5074 pass them.
5075
5076 6. Give an error if bad option bits passed at compile or run time.
5077
5078 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5079 pcretest to cause it to pass that flag.
5080
5081 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5082 options, and the first character, if set.
5083
5084 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5085
5086
5087 Version 0.91 10-Sep-97
5088 ----------------------
5089
5090 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5091 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5092
5093 2. PCRE was looping on encountering an indefinitely repeated back reference to
5094 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5095 Perl does - treats the match as successful.
5096
5097 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5