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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1298 - (show annotations)
Fri Mar 22 16:13:13 2013 UTC (6 years, 8 months ago) by ph10
File size: 245378 byte(s)
Error occurred while calculating annotation data.
Fix COMMIT in recursion; document backtracking verbs in assertions and 
subroutines.
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. The way PCRE handles backtracking verbs has been changed in to ways.
114
115 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
116 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
117 backtracking. In some cases this makes it more Perl-compatible, but Perl's
118 rather obscure rules do not always do the same thing.
119
120 (2) Previously, backtracking verbs were confined within assertions. This is
121 no longer the case, except for (*ACCEPT). Again, this sometimes improves
122 Perl compatibility, and sometimes does not.
123
124
125 Version 8.32 30-November-2012
126 -----------------------------
127
128 1. Improved JIT compiler optimizations for first character search and single
129 character iterators.
130
131 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
132 Patch by Daniel Richard G.
133
134 3. Single character iterator optimizations in the JIT compiler.
135
136 4. Improved JIT compiler optimizations for character ranges.
137
138 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
139 Reported by Giuseppe D'Angelo.
140
141 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
142 of a line, was being set incorrectly in cases where .* appeared inside
143 atomic brackets at the start of a pattern, or where there was a subsequent
144 *PRUNE or *SKIP.
145
146 7. Improved instruction cache flush for POWER/PowerPC.
147 Patch by Daniel Richard G.
148
149 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
150 grep:
151
152 (a) There is now no limit to the number of patterns to be matched.
153
154 (b) An error is given if a pattern is too long.
155
156 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
157 are now supported.
158
159 (d) --exclude-from and --include-from (multiple use) have been added.
160
161 (e) Exclusions and inclusions now apply to all files and directories, not
162 just to those obtained from scanning a directory recursively.
163
164 (f) Multiple uses of -f and --file-list are now supported.
165
166 (g) In a Windows environment, the default for -d has been changed from
167 "read" (the GNU grep default) to "skip", because otherwise the presence
168 of a directory in the file list provokes an error.
169
170 (h) The documentation has been revised and clarified in places.
171
172 9. Improve the matching speed of capturing brackets.
173
174 10. Changed the meaning of \X so that it now matches a Unicode extended
175 grapheme cluster.
176
177 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
178 out POSIX threads when JIT support is configured.
179
180 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
181
182 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
183 structure could go wrong in environments where size_t is not the same size
184 as int.
185
186 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
187 set.
188
189 15. The EBCDIC support had decayed; later updates to the code had included
190 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
191 general tidy up of EBCDIC-related issues, and the documentation was also
192 not quite right. There is now a test that can be run on ASCII systems to
193 check some of the EBCDIC-related things (but is it not a full test).
194
195 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
196 in a small tidy to the code.
197
198 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
199
200 18. If the --only-matching (-o) option in pcregrep is specified multiple
201 times, each one causes appropriate output. For example, -o1 -o2 outputs the
202 substrings matched by the 1st and 2nd capturing parentheses. A separating
203 string can be specified by --om-separator (default empty).
204
205 19. Improving the first n character searches.
206
207 20. Turn case lists for horizontal and vertical white space into macros so that
208 they are defined only once.
209
210 21. This set of changes together give more compatible Unicode case-folding
211 behaviour for characters that have more than one other case when UCP
212 support is available.
213
214 (a) The Unicode property table now has offsets into a new table of sets of
215 three or more characters that are case-equivalent. The MultiStage2.py
216 script that generates these tables (the pcre_ucd.c file) now scans
217 CaseFolding.txt instead of UnicodeData.txt for character case
218 information.
219
220 (b) The code for adding characters or ranges of characters to a character
221 class has been abstracted into a generalized function that also handles
222 case-independence. In UTF-mode with UCP support, this uses the new data
223 to handle characters with more than one other case.
224
225 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
226 whose other case is greater than 256 are now correctly matched
227 caselessly. Previously, the high codepoint matched the low one, but not
228 vice versa.
229
230 (d) The processing of \h, \H, \v, and \ in character classes now makes use
231 of the new class addition function, using character lists defined as
232 macros alongside the case definitions of 20 above.
233
234 (e) Caseless back references now work with characters that have more than
235 one other case.
236
237 (f) General caseless matching of characters with more than one other case
238 is supported.
239
240 22. Unicode character properties were updated from Unicode 6.2.0
241
242 23. Improved CMake support under Windows. Patch by Daniel Richard G.
243
244 24. Add support for 32-bit character strings, and UTF-32
245
246 25. Major JIT compiler update (code refactoring and bugfixing).
247 Experimental Sparc 32 support is added.
248
249 26. Applied a modified version of Daniel Richard G's patch to create
250 pcre.h.generic and config.h.generic by "make" instead of in the
251 PrepareRelease script.
252
253 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
254 pcre_compile.c when checking for a zero character.
255
256 28. Introducing a native interface for JIT. Through this interface, the compiled
257 machine code can be directly executed. The purpose of this interface is to
258 provide fast pattern matching, so several sanity checks are not performed.
259 However, feature tests are still performed. The new interface provides
260 1.4x speedup compared to the old one.
261
262 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
263 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
264 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
265
266 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
267 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
268 the "old" RFC 2279). Instead, it ended up passing a negative length to
269 pcre_exec().
270
271 31. Add support for GCC's visibility feature to hide internal functions.
272
273 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
274 "unknown -C option" after outputting 0 or 1.
275
276 33. There is now support for generating a code coverage report for the test
277 suite in environments where gcc is the compiler and lcov is installed. This
278 is mainly for the benefit of the developers.
279
280 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
281 unaddressable using valgrind annotations, allowing valgrind to detect
282 invalid memory accesses. This is mainly for the benefit of the developers.
283
284 25. (*UTF) can now be used to start a pattern in any of the three libraries.
285
286 26. Give configure error if --enable-cpp but no C++ compiler found.
287
288
289 Version 8.31 06-July-2012
290 -------------------------
291
292 1. Fixing a wrong JIT test case and some compiler warnings.
293
294 2. Removed a bashism from the RunTest script.
295
296 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
297 to unsigned type, result still unsigned" that was given by an MS compiler
298 on encountering the code "-sizeof(xxx)".
299
300 4. Partial matching support is added to the JIT compiler.
301
302 5. Fixed several bugs concerned with partial matching of items that consist
303 of more than one character:
304
305 (a) /^(..)\1/ did not partially match "aba" because checking references was
306 done on an "all or nothing" basis. This also applied to repeated
307 references.
308
309 (b) \R did not give a hard partial match if \r was found at the end of the
310 subject.
311
312 (c) \X did not give a hard partial match after matching one or more
313 characters at the end of the subject.
314
315 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
316 a partial match for the string "\r".
317
318 (e) When newline was set to CRLF, the metacharacter "." did not recognize
319 a partial match for a CR character at the end of the subject string.
320
321 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
322 running pcretest, the text "(JIT)" added to the output whenever JIT is
323 actually used to run the match.
324
325 7. Individual JIT compile options can be set in pcretest by following -s+[+]
326 or /S+[+] with a digit between 1 and 7.
327
328 8. OP_NOT now supports any UTF character not just single-byte ones.
329
330 9. (*MARK) control verb is now supported by the JIT compiler.
331
332 10. The command "./RunTest list" lists the available tests without actually
333 running any of them. (Because I keep forgetting what they all are.)
334
335 11. Add PCRE_INFO_MAXLOOKBEHIND.
336
337 12. Applied a (slightly modified) user-supplied patch that improves performance
338 when the heap is used for recursion (compiled with --disable-stack-for-
339 recursion). Instead of malloc and free for each heap frame each time a
340 logical recursion happens, frames are retained on a chain and re-used where
341 possible. This sometimes gives as much as 30% improvement.
342
343 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
344 call.
345
346 14. As documented, (*COMMIT) is now confined to within a positive assertion.
347
348 15. It is now possible to link pcretest with libedit as an alternative to
349 libreadline.
350
351 16. (*COMMIT) control verb is now supported by the JIT compiler.
352
353 17. The Unicode data tables have been updated to Unicode 6.1.0.
354
355 18. Added --file-list option to pcregrep.
356
357 19. Added binary file support to pcregrep, including the -a, --binary-files,
358 -I, and --text options.
359
360 20. The madvise function is renamed for posix_madvise for QNX compatibility
361 reasons. Fixed by Giuseppe D'Angelo.
362
363 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
364
365 22. Changed the option for creating symbolic links for 16-bit man pages from
366 -s to -sf so that re-installing does not cause issues.
367
368 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
369
370 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
371 matches in certain environments (the workspace was not being correctly
372 retained). Also added to pcre_dfa_exec() a simple plausibility check on
373 some of the workspace data at the beginning of a restart.
374
375 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
376 was not doing so when it should - probably a typo introduced by SVN 528
377 (change 8.10/14).
378
379 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
380 \w+ when the character tables indicated that \x{c4} was a word character.
381 There were several related cases, all because the tests for doing a table
382 lookup were testing for characters less than 127 instead of 255.
383
384 27. If a pattern contains capturing parentheses that are not used in a match,
385 their slots in the ovector are set to -1. For those that are higher than
386 any matched groups, this happens at the end of processing. In the case when
387 there were back references that the ovector was too small to contain
388 (causing temporary malloc'd memory to be used during matching), and the
389 highest capturing number was not used, memory off the end of the ovector
390 was incorrectly being set to -1. (It was using the size of the temporary
391 memory instead of the true size.)
392
393 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
394 ovector size, it uses memory at the end of the block that it has got.
395
396 29. Check for an overlong MARK name and give an error at compile time. The
397 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
398
399 30. JIT compiler update.
400
401 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
402 Rill for the patch.
403
404 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
405
406 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
407
408 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
409 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
410 was enabled.
411
412 35. Improve JIT code generation for greedy plus quantifier.
413
414 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
415 1 to "aa" instead of to an empty string. The bug affected repeated groups
416 that could potentially match an empty string.
417
418 37. Optimizing single character iterators in JIT.
419
420 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
421 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
422 codepoints that are too big for the mode are faulted, and in a UTF mode,
423 disallowed codepoints are also faulted.
424
425 39. If PCRE was compiled with UTF support, in three places in the DFA
426 matcher there was code that should only have been obeyed in UTF mode, but
427 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
428 processing when bytes with values greater than 127 were present. In 16-bit
429 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
430 both cases the values are those that cannot be the first data item in a UTF
431 character. The three items that might have provoked this were recursions,
432 possessively repeated groups, and atomic groups.
433
434 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
435 and pcregrep, because some OS require shared objects to be explicitly
436 passed to ld, causing the link step to fail if they are not.
437
438 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
439 mode, patterns that started with \h* or \R* might be incorrectly matched.
440
441
442 Version 8.30 04-February-2012
443 -----------------------------
444
445 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
446 name is defined in ctype.h.
447
448 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
449 only in quite long subpatterns.
450
451 3. Removed the function pcre_info(), which has been obsolete and deprecated
452 since it was replaced by pcre_fullinfo() in February 2000.
453
454 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
455 match a (*MARK), and the match failed at the start of the subject, a
456 reference to memory before the start of the subject could occur. This bug
457 was introduced by fix 17 of release 8.21.
458
459 5. A reference to an unset group with zero minimum repetition was giving
460 totally wrong answers (in non-JavaScript-compatibility mode). For example,
461 /(another)?(\1?)test/ matched against "hello world test". This bug was
462 introduced in release 8.13.
463
464 6. Add support for 16-bit character strings (a large amount of work involving
465 many changes and refactorings).
466
467 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
468 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
469 from a file.
470
471 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
472 rounding is not applied in this particular case).
473
474 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
475 if they appear, or are escaped, in patterns.
476
477 10. Get rid of a number of -Wunused-but-set-variable warnings.
478
479 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
480 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
481 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
482 also returns the mark "x". This bug applied to capturing parentheses,
483 non-capturing parentheses, and atomic parentheses. It also applied to some
484 assertions.
485
486 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
487 information out of configure.ac instead of relying on pcre.h.generic, which
488 is not stored in the repository.
489
490 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
491 -lreadline.
492
493 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
494
495 15. Applied Graycode's patch to put the top-level frame on the stack rather
496 than the heap when not using the stack for recursion. This gives a
497 performance improvement in many cases when recursion is not deep.
498
499 16. Experimental code added to "pcretest -C" to output the stack frame size.
500
501
502 Version 8.21 12-Dec-2011
503 ------------------------
504
505 1. Updating the JIT compiler.
506
507 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
508 are added as well.
509
510 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
511 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
512 calling _pcre_jit_exec. Some extra comments are added.
513
514 4. (*MARK) settings inside atomic groups that do not contain any capturing
515 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
516 was introduced by change 18 for 8.20.
517
518 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
519 ECMA-262 standard.
520
521 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
522 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
523 This bug was probably introduced by change 9 of 8.13.
524
525 7. While fixing 6 above, I noticed that a number of other items were being
526 incorrectly rejected as "not fixed length". This arose partly because newer
527 opcodes had not been added to the fixed-length checking code. I have (a)
528 corrected the bug and added tests for these items, and (b) arranged for an
529 error to occur if an unknown opcode is encountered while checking for fixed
530 length instead of just assuming "not fixed length". The items that were
531 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
532 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
533 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
534
535 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
536 being incorrectly compiled and would have given unpredicatble results.
537
538 9. A possessively repeated subpattern with minimum repeat count greater than
539 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
540 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
541 the first (A) could occur when it should not.
542
543 10. Add a cast and remove a redundant test from the code.
544
545 11. JIT should use pcre_malloc/pcre_free for allocation.
546
547 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
548 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
549 is anything other than /usr, -L is still shown).
550
551 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
552
553 14. Perl does not support \N without a following name in a [] class; PCRE now
554 also gives an error.
555
556 15. If a forward reference was repeated with an upper limit of around 2000,
557 it caused the error "internal error: overran compiling workspace". The
558 maximum number of forward references (including repeats) was limited by the
559 internal workspace, and dependent on the LINK_SIZE. The code has been
560 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
561 the default depends on LINK_SIZE. There is a new upper limit (for safety)
562 of around 200,000 forward references. While doing this, I also speeded up
563 the filling in of repeated forward references.
564
565 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
566 incorrectly expecting the subject to contain another "a" after the start.
567
568 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
569 in the match, the SKIP should be ignored. This was not happening; instead
570 the SKIP was being treated as NOMATCH. For patterns such as
571 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
572 tested.
573
574 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
575 now much more compatible with Perl, in particular in cases where the result
576 is a non-match for a non-anchored pattern. For example, if
577 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
578 "m", where previously it did not return a name. A side effect of this
579 change is that for partial matches, the last encountered mark name is
580 returned, as for non matches. A number of tests that were previously not
581 Perl-compatible have been moved into the Perl-compatible test files. The
582 refactoring has had the pleasing side effect of removing one argument from
583 the match() function, thus reducing its stack requirements.
584
585 19. If the /S+ option was used in pcretest to study a pattern using JIT,
586 subsequent uses of /S (without +) incorrectly behaved like /S+.
587
588 21. Retrieve executable code size support for the JIT compiler and fixing
589 some warnings.
590
591 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
592 not work when the shorter character appeared right at the end of the
593 subject string.
594
595 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
596 systems.
597
598 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
599 output it when the /M option is used in pcretest.
600
601 25. The CheckMan script was not being included in the distribution. Also, added
602 an explicit "perl" to run Perl scripts from the PrepareRelease script
603 because this is reportedly needed in Windows.
604
605 26. If study data was being save in a file and studying had not found a set of
606 "starts with" bytes for the pattern, the data written to the file (though
607 never used) was taken from uninitialized memory and so caused valgrind to
608 complain.
609
610 27. Updated RunTest.bat as provided by Sheri Pierce.
611
612 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
613
614 29. Computation of memory usage for the table of capturing group names was
615 giving an unnecessarily large value.
616
617
618 Version 8.20 21-Oct-2011
619 ------------------------
620
621 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
622 a POSIX class. After further experiments with Perl, which convinced me that
623 Perl has bugs and confusions, a closing square bracket is no longer allowed
624 in a POSIX name. This bug also affected patterns with classes that started
625 with full stops.
626
627 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
628 captured substring, but while checking the failing first alternative,
629 substring 1 is temporarily captured. If the output vector supplied to
630 pcre_exec() was not big enough for this capture, the yield of the function
631 was still zero ("insufficient space for captured substrings"). This cannot
632 be totally fixed without adding another stack variable, which seems a lot
633 of expense for a edge case. However, I have improved the situation in cases
634 such as /(a)(b)x|abc/ matched against "abc", where the return code
635 indicates that fewer than the maximum number of slots in the ovector have
636 been set.
637
638 3. Related to (2) above: when there are more back references in a pattern than
639 slots in the output vector, pcre_exec() uses temporary memory during
640 matching, and copies in the captures as far as possible afterwards. It was
641 using the entire output vector, but this conflicts with the specification
642 that only 2/3 is used for passing back captured substrings. Now it uses
643 only the first 2/3, for compatibility. This is, of course, another edge
644 case.
645
646 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
647 main code base, and can be used by building with --enable-jit. When this is
648 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
649 runtime --no-jit option is given.
650
651 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
652 ovector, the return from the function was zero, implying that there were
653 other matches that did not fit. The correct "exactly full" value is now
654 returned.
655
656 6. If a subpattern that was called recursively or as a subroutine contained
657 (*PRUNE) or any other control that caused it to give a non-standard return,
658 invalid errors such as "Error -26 (nested recursion at the same subject
659 position)" or even infinite loops could occur.
660
661 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
662 computing the minimum length on reaching *ACCEPT, and so ended up with the
663 wrong value of 1 rather than 0. Further investigation indicates that
664 computing a minimum subject length in the presence of *ACCEPT is difficult
665 (think back references, subroutine calls), and so I have changed the code
666 so that no minimum is registered for a pattern that contains *ACCEPT.
667
668 8. If (*THEN) was present in the first (true) branch of a conditional group,
669 it was not handled as intended. [But see 16 below.]
670
671 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
672 Sheri Pierce.
673
674 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
675 the first byte in a match must be "a".
676
677 11. Change 17 for 8.13 increased the recursion depth for patterns like
678 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
679 pattern contains any instances of (*THEN). If it does not, the old
680 optimizations are restored. It would be nice to do this on a per-group
681 basis, but at the moment that is not feasible.
682
683 12. In some environments, the output of pcretest -C is CRLF terminated. This
684 broke RunTest's code that checks for the link size. A single white space
685 character after the value is now allowed for.
686
687 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
688 For "fr", it uses the Windows-specific input and output files.
689
690 14. If (*THEN) appeared in a group that was called recursively or as a
691 subroutine, it did not work as intended. [But see next item.]
692
693 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
694 pattern fragments (but not containing any | characters). If A and B are
695 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
696 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
697 D. In other words, Perl considers parentheses that do not contain any |
698 characters to be part of a surrounding alternative, whereas PCRE was
699 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
700 differently. PCRE now behaves in the same way as Perl, except in the case
701 of subroutine/recursion calls such as (?1) which have in any case always
702 been different (but PCRE had them first :-).
703
704 16. Related to 15 above: Perl does not treat the | in a conditional group as
705 creating alternatives. Such a group is treated in the same way as an
706 ordinary group without any | characters when processing (*THEN). PCRE has
707 been changed to match Perl's behaviour.
708
709 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
710 RunGrepTest script failed.
711
712 18. Change 22 for version 13 caused atomic groups to use more stack. This is
713 inevitable for groups that contain captures, but it can lead to a lot of
714 stack use in large patterns. The old behaviour has been restored for atomic
715 groups that do not contain any capturing parentheses.
716
717 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
718 suppress the check for a minimum subject length at run time. (If it was
719 given to pcre_exec() or pcre_dfa_exec() it did work.)
720
721 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
722 fail to work when decoding hex characters in data strings in EBCDIC
723 environments.
724
725 21. It appears that in at least one Mac OS environment, the isxdigit() function
726 is implemented as a macro that evaluates to its argument more than once,
727 contravening the C 90 Standard (I haven't checked a later standard). There
728 was an instance in pcretest which caused it to go wrong when processing
729 \x{...} escapes in subject strings. The has been rewritten to avoid using
730 things like p++ in the argument of isxdigit().
731
732
733 Version 8.13 16-Aug-2011
734 ------------------------
735
736 1. The Unicode data tables have been updated to Unicode 6.0.0.
737
738 2. Two minor typos in pcre_internal.h have been fixed.
739
740 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
741 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
742 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
743
744 4. There were a number of related bugs in the code for matching backrefences
745 caselessly in UTF-8 mode when codes for the characters concerned were
746 different numbers of bytes. For example, U+023A and U+2C65 are an upper
747 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
748 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
749 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
750 2-byte code at the end of the subject (it thought there wasn't enough data
751 left).
752
753 5. Comprehensive information about what went wrong is now returned by
754 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
755 as the output vector has at least 2 elements. The offset of the start of
756 the failing character and a reason code are placed in the vector.
757
758 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
759 now returned is for the first byte of the failing character, instead of the
760 last byte inspected. This is an incompatible change, but I hope it is small
761 enough not to be a problem. It makes the returned offset consistent with
762 pcre_exec() and pcre_dfa_exec().
763
764 7. pcretest now gives a text phrase as well as the error number when
765 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
766 failure, the offset and reason code are output.
767
768 8. When \R was used with a maximizing quantifier it failed to skip backwards
769 over a \r\n pair if the subsequent match failed. Instead, it just skipped
770 back over a single character (\n). This seems wrong (because it treated the
771 two characters as a single entity when going forwards), conflicts with the
772 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
773 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
774 has been changed.
775
776 9. Some internal refactoring has changed the processing so that the handling
777 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
778 time (the PCRE_DOTALL option was changed this way some time ago: version
779 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
780 which was always a bit of a fudge. It also means that there is one less
781 argument for the match() function, which reduces its stack requirements
782 slightly. This change also fixes an incompatibility with Perl: the pattern
783 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
784
785 10. More internal refactoring has drastically reduced the number of recursive
786 calls to match() for possessively repeated groups such as (abc)++ when
787 using pcre_exec().
788
789 11. While implementing 10, a number of bugs in the handling of groups were
790 discovered and fixed:
791
792 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
793 (a|)*(?1) gave a compile-time internal error.
794 ((a|)+)+ did not notice that the outer group could match an empty string.
795 (^a|^)+ was not marked as anchored.
796 (.*a|.*)+ was not marked as matching at start or after a newline.
797
798 12. Yet more internal refactoring has removed another argument from the match()
799 function. Special calls to this function are now indicated by setting a
800 value in a variable in the "match data" data block.
801
802 13. Be more explicit in pcre_study() instead of relying on "default" for
803 opcodes that mean there is no starting character; this means that when new
804 ones are added and accidentally left out of pcre_study(), testing should
805 pick them up.
806
807 14. The -s option of pcretest has been documented for ages as being an old
808 synonym of -m (show memory usage). I have changed it to mean "force study
809 for every regex", that is, assume /S for every regex. This is similar to -i
810 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
811 using it. It makes it easier to run collections of tests with and without
812 study enabled, and thereby test pcre_study() more easily. All the standard
813 tests are now run with and without -s (but some patterns can be marked as
814 "never study" - see 20 below).
815
816 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
817 restoration of the capturing data to the outer values was not happening
818 correctly.
819
820 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
821 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
822 pattern had matched an empty string, and so incorrectly returned a no
823 match.
824
825 17. There was optimizing code for the last branch of non-capturing parentheses,
826 and also for the obeyed branch of a conditional subexpression, which used
827 tail recursion to cut down on stack usage. Unfortunately, now that there is
828 the possibility of (*THEN) occurring in these branches, tail recursion is
829 no longer possible because the return has to be checked for (*THEN). These
830 two optimizations have therefore been removed. [But see 8.20/11 above.]
831
832 18. If a pattern containing \R was studied, it was assumed that \R always
833 matched two bytes, thus causing the minimum subject length to be
834 incorrectly computed because \R can also match just one byte.
835
836 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
837 was incorrectly computed.
838
839 20. If /S is present twice on a test pattern in pcretest input, it now
840 *disables* studying, thereby overriding the use of -s on the command line
841 (see 14 above). This is necessary for one or two tests to keep the output
842 identical in both cases.
843
844 21. When (*ACCEPT) was used in an assertion that matched an empty string and
845 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
846
847 22. When an atomic group that contained a capturing parenthesis was
848 successfully matched, but the branch in which it appeared failed, the
849 capturing was not being forgotten if a higher numbered group was later
850 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
851 group 1 to "a", when in fact it should be unset. This applied to multi-
852 branched capturing and non-capturing groups, repeated or not, and also to
853 positive assertions (capturing in negative assertions does not happen
854 in PCRE) and also to nested atomic groups.
855
856 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
857 subject after a captured substring, to make it easier to tell which of a
858 number of identical substrings has been captured.
859
860 24. The way atomic groups are processed by pcre_exec() has been changed so that
861 if they are repeated, backtracking one repetition now resets captured
862 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
863 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
864 "aaa". Previously, it would have been "a". As part of this code
865 refactoring, the way recursive calls are handled has also been changed.
866
867 25. If an assertion condition captured any substrings, they were not passed
868 back unless some other capturing happened later. For example, if
869 (?(?=(a))a) was matched against "a", no capturing was returned.
870
871 26. When studying a pattern that contained subroutine calls or assertions,
872 the code for finding the minimum length of a possible match was handling
873 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
874 group 1 called group 2 while simultaneously a separate group 2 called group
875 1). A stack overflow occurred in this case. I have fixed this by limiting
876 the recursion depth to 10.
877
878 27. Updated RunTest.bat in the distribution to the version supplied by Tom
879 Fortmann. This supports explicit test numbers on the command line, and has
880 argument validation and error reporting.
881
882 28. An instance of \X with an unlimited repeat could fail if at any point the
883 first character it looked at was a mark character.
884
885 29. Some minor code refactoring concerning Unicode properties and scripts
886 should reduce the stack requirement of match() slightly.
887
888 30. Added the '=' option to pcretest to check the setting of unused capturing
889 slots at the end of the pattern, which are documented as being -1, but are
890 not included in the return count.
891
892 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
893 compiled something random. Now it gives a compile-time error (as does
894 Perl).
895
896 32. A *MARK encountered during the processing of a positive assertion is now
897 recorded and passed back (compatible with Perl).
898
899 33. If --only-matching or --colour was set on a pcregrep call whose pattern
900 had alternative anchored branches, the search for a second match in a line
901 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
902 matched the line "0102" twice. The same bug affected patterns that started
903 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
904 twice.
905
906 34. Previously, PCRE did not allow quantification of assertions. However, Perl
907 does, and because of capturing effects, quantifying parenthesized
908 assertions may at times be useful. Quantifiers are now allowed for
909 parenthesized assertions.
910
911 35. A minor code tidy in pcre_compile() when checking options for \R usage.
912
913 36. \g was being checked for fancy things in a character class, when it should
914 just be a literal "g".
915
916 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
917 appearance of a nested POSIX class supersedes an apparent external class.
918 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
919 unescaped square brackets may also appear as part of class names. For
920 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
921 more like Perl. (But see 8.20/1 above.)
922
923 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
924 was because it thought it was \N{name}, which is not supported).
925
926 39. Add minix to OS list not supporting the -S option in pcretest.
927
928 40. PCRE tries to detect cases of infinite recursion at compile time, but it
929 cannot analyze patterns in sufficient detail to catch mutual recursions
930 such as ((?1))((?2)). There is now a runtime test that gives an error if a
931 subgroup is called recursively as a subpattern for a second time at the
932 same position in the subject string. In previous releases this might have
933 been caught by the recursion limit, or it might have run out of stack.
934
935 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
936 happen only once. PCRE was, however incorrectly giving a compile time error
937 "recursive call could loop indefinitely" because it cannot analyze the
938 pattern in sufficient detail. The compile time test no longer happens when
939 PCRE is compiling a conditional subpattern, but actual runaway loops are
940 now caught at runtime (see 40 above).
941
942 42. It seems that Perl allows any characters other than a closing parenthesis
943 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
944 has been changed to be the same.
945
946 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
947 as not to get warnings when autogen.sh is called. Also changed
948 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
949
950 44. To help people who use pcregrep to scan files containing exceedingly long
951 lines, the following changes have been made:
952
953 (a) The default value of the buffer size parameter has been increased from
954 8K to 20K. (The actual buffer used is three times this size.)
955
956 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
957 PCRE is built.
958
959 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
960 to be set at run time.
961
962 (d) Numerical values in pcregrep options can be followed by K or M, for
963 example --buffer-size=50K.
964
965 (e) If a line being scanned overflows pcregrep's buffer, an error is now
966 given and the return code is set to 2.
967
968 45. Add a pointer to the latest mark to the callout data block.
969
970 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
971 partial match of an empty string instead of no match. This was specific to
972 the use of ".".
973
974 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
975 complete match instead of a partial match. This bug was dependent on both
976 the PCRE_UTF8 and PCRE_DOTALL options being set.
977
978 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
979 starting byte set, because \b was not being ignored.
980
981
982 Version 8.12 15-Jan-2011
983 ------------------------
984
985 1. Fixed some typos in the markup of the man pages, and wrote a script that
986 checks for such things as part of the documentation building process.
987
988 2. On a big-endian 64-bit system, pcregrep did not correctly process the
989 --match-limit and --recursion-limit options (added for 8.11). In
990 particular, this made one of the standard tests fail. (The integer value
991 went into the wrong half of a long int.)
992
993 3. If the --colour option was given to pcregrep with -v (invert match), it
994 did strange things, either producing crazy output, or crashing. It should,
995 of course, ignore a request for colour when reporting lines that do not
996 match.
997
998 4. Another pcregrep bug caused similar problems if --colour was specified with
999 -M (multiline) and the pattern match finished with a line ending.
1000
1001 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1002 matched in multiline mode, the following line was shown as part of the
1003 match. This seems wrong, so I have changed it.
1004
1005 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1006 the check for further matches in the same line (so they could be coloured)
1007 to overrun the end of the current line. If another match was found, it was
1008 incorrectly shown (and then shown again when found in the next line).
1009
1010 7. If pcregrep was compiled under Windows, there was a reference to the
1011 function pcregrep_exit() before it was defined. I am assuming this was
1012 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1013 reported by a user. I've moved the definition above the reference.
1014
1015
1016 Version 8.11 10-Dec-2010
1017 ------------------------
1018
1019 1. (*THEN) was not working properly if there were untried alternatives prior
1020 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1021 backtracked to try for "b" instead of moving to the next alternative branch
1022 at the same level (in this case, to look for "c"). The Perl documentation
1023 is clear that when (*THEN) is backtracked onto, it goes to the "next
1024 alternative in the innermost enclosing group".
1025
1026 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1027 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1028 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1029 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1030 (*THEN).
1031
1032 3. If \s appeared in a character class, it removed the VT character from
1033 the class, even if it had been included by some previous item, for example
1034 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1035 of \s, but is part of the POSIX "space" class.)
1036
1037 4. A partial match never returns an empty string (because you can always
1038 match an empty string at the end of the subject); however the checking for
1039 an empty string was starting at the "start of match" point. This has been
1040 changed to the "earliest inspected character" point, because the returned
1041 data for a partial match starts at this character. This means that, for
1042 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1043 (previously it gave "no match").
1044
1045 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1046 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1047 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1048 has an implication that the given string is incomplete (because a partial
1049 match is preferred over a full match). For this reason, these items now
1050 give a partial match in this situation. [Aside: previously, the one case
1051 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1052 match rather than a full match, which was wrong by the old rules, but is
1053 now correct.]
1054
1055 6. There was a bug in the handling of #-introduced comments, recognized when
1056 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1057 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1058 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1059 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1060 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1061 places in pcre_compile().
1062
1063 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1064 comments when looking ahead for named forward references to subpatterns,
1065 the only newline sequence it recognized was NL. It now handles newlines
1066 according to the set newline convention.
1067
1068 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1069 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1070 cater for a lack of strerror(). These oversights have been fixed.
1071
1072 9. Added --match-limit and --recursion-limit to pcregrep.
1073
1074 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1075
1076 11. When the -o option was used, pcregrep was setting a return code of 1, even
1077 when matches were found, and --line-buffered was not being honoured.
1078
1079 12. Added an optional parentheses number to the -o and --only-matching options
1080 of pcregrep.
1081
1082 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1083 can match an empty string. The code to do it in pcretest and pcredemo
1084 needed fixing:
1085
1086 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1087 only one byte after an empty string match just before CRLF (this case
1088 just got forgotten; "any" and "anycrlf" were OK).
1089
1090 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1091 mode when an empty string match preceded an ASCII character followed by
1092 a non-ASCII character. (The code for advancing by one character rather
1093 than one byte was nonsense.)
1094
1095 (c) The pcredemo.c sample program did not have any code at all to handle
1096 the cases when CRLF is a valid newline sequence.
1097
1098 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1099 as a starting offset was within the subject string. There is now a new
1100 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1101 negative or greater than the length of the string. In order to test this,
1102 pcretest is extended to allow the setting of negative starting offsets.
1103
1104 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1105 starting offset points to the beginning of a UTF-8 character was
1106 unnecessarily clumsy. I tidied it up.
1107
1108 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1109 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1110
1111 17. Nobody had reported that the --include_dir option, which was added in
1112 release 7.7 should have been called --include-dir (hyphen, not underscore)
1113 for compatibility with GNU grep. I have changed it to --include-dir, but
1114 left --include_dir as an undocumented synonym, and the same for
1115 --exclude-dir, though that is not available in GNU grep, at least as of
1116 release 2.5.4.
1117
1118 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1119 characters from a string of bytes) have been redefined so as not to use
1120 loops, in order to improve performance in some environments. At the same
1121 time, I abstracted some of the common code into auxiliary macros to save
1122 repetition (this should not affect the compiled code).
1123
1124 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1125 compile-time error is now given if \c is not followed by an ASCII
1126 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1127 different, and any byte value is allowed.)
1128
1129 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1130 START_OPTIMIZE option, which is now allowed at compile time - but just
1131 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1132 to pcregrep and other applications that have no direct access to PCRE
1133 options. The new /Y option in pcretest sets this option when calling
1134 pcre_compile().
1135
1136 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1137 back references. Groups containing recursive back references were forced to
1138 be atomic by that change, but in the case of named groups, the amount of
1139 memory required was incorrectly computed, leading to "Failed: internal
1140 error: code overflow". This has been fixed.
1141
1142 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1143 pcretest.c, to avoid build problems in some Borland environments.
1144
1145
1146 Version 8.10 25-Jun-2010
1147 ------------------------
1148
1149 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1150 THEN.
1151
1152 2. (*ACCEPT) was not working when inside an atomic group.
1153
1154 3. Inside a character class, \B is treated as a literal by default, but
1155 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1156 causes the error). The code is unchanged, but I tidied the documentation.
1157
1158 4. Inside a character class, PCRE always treated \R and \X as literals,
1159 whereas Perl faults them if its -w option is set. I have changed PCRE so
1160 that it faults them when PCRE_EXTRA is set.
1161
1162 5. Added support for \N, which always matches any character other than
1163 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1164
1165 6. When compiling pcregrep with newer versions of gcc which may have
1166 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1167 declared with attribute warn_unused_result" were given. Just casting the
1168 result to (void) does not stop the warnings; a more elaborate fudge is
1169 needed. I've used a macro to implement this.
1170
1171 7. Minor change to pcretest.c to avoid a compiler warning.
1172
1173 8. Added four artifical Unicode properties to help with an option to make
1174 \s etc use properties (see next item). The new properties are: Xan
1175 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1176
1177 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1178 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1179 this option. Modified pcretest to add /W to test this facility. Added
1180 REG_UCP to make it available via the POSIX interface.
1181
1182 10. Added --line-buffered to pcregrep.
1183
1184 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1185 studied, and the match started with a letter with a code point greater than
1186 127 whose first byte was different to the first byte of the other case of
1187 the letter, the other case of this starting letter was not recognized
1188 (#976).
1189
1190 12. If a pattern that was studied started with a repeated Unicode property
1191 test, for example, \p{Nd}+, there was the theoretical possibility of
1192 setting up an incorrect bitmap of starting bytes, but fortunately it could
1193 not have actually happened in practice until change 8 above was made (it
1194 added property types that matched character-matching opcodes).
1195
1196 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1197 possible starting bytes for non-anchored patterns.
1198
1199 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1200 \R, and also a number of cases that involve Unicode properties, both
1201 explicit and implicit when PCRE_UCP is set.
1202
1203 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1204 input, it could crash or give wrong results if characters with values
1205 greater than 0xc0 were present in the subject string. (Detail: it assumed
1206 UTF-8 input when processing these items.)
1207
1208 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1209 size_t is 64-bit (#991).
1210
1211 17. Added a check for running out of memory when PCRE is compiled with
1212 --disable-stack-for-recursion (#990).
1213
1214 18. If the last data line in a file for pcretest does not have a newline on
1215 the end, a newline was missing in the output.
1216
1217 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1218 less than 128) in its various bitmaps. However, there is a facility for
1219 generating tables according to the current locale when PCRE is compiled. It
1220 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1221 characters, are recognized by isspace() and therefore were getting set in
1222 these tables, and indeed these tables seem to approximate to ISO 8859. This
1223 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1224 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1225 which of course cannot start UTF-8 characters. I have changed the code so
1226 that only real ASCII characters (less than 128) and the correct starting
1227 bytes for UTF-8 encodings are set for characters greater than 127 when in
1228 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1229 altogether.)
1230
1231 20. Added the /T option to pcretest so as to be able to run tests with non-
1232 standard character tables, thus making it possible to include the tests
1233 used for 19 above in the standard set of tests.
1234
1235 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1236 reference to a subpattern the other side of a comment that contains an
1237 opening parenthesis caused either an internal compiling error, or a
1238 reference to the wrong subpattern.
1239
1240
1241 Version 8.02 19-Mar-2010
1242 ------------------------
1243
1244 1. The Unicode data tables have been updated to Unicode 5.2.0.
1245
1246 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1247 configured.
1248
1249 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1250 original author of that file, following a query about its status.
1251
1252 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1253 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1254
1255 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1256 quantifier applied to a forward-referencing subroutine call, could compile
1257 incorrect code or give the error "internal error: previously-checked
1258 referenced subpattern not found".
1259
1260 6. Both MS Visual Studio and Symbian OS have problems with initializing
1261 variables to point to external functions. For these systems, therefore,
1262 pcre_malloc etc. are now initialized to local functions that call the
1263 relevant global functions.
1264
1265 7. There were two entries missing in the vectors called coptable and poptable
1266 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1267 I've fixed the data, and added a kludgy way of testing at compile time that
1268 the lengths are correct (equal to the number of opcodes).
1269
1270 8. Following on from 7, I added a similar kludge to check the length of the
1271 eint vector in pcreposix.c.
1272
1273 9. Error texts for pcre_compile() are held as one long string to avoid too
1274 much relocation at load time. To find a text, the string is searched,
1275 counting zeros. There was no check for running off the end of the string,
1276 which could happen if a new error number was added without updating the
1277 string.
1278
1279 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1280
1281 11. \K was not working if it appeared in an atomic group or in a group that
1282 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1283 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1284 the assertion is positive, but not if it is negative.
1285
1286 12. Change 11 fortuitously reduced the size of the stack frame used in the
1287 "match()" function of pcre_exec.c by one pointer. Forthcoming
1288 implementation of support for (*MARK) will need an extra pointer on the
1289 stack; I have reserved it now, so that the stack frame size does not
1290 decrease.
1291
1292 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1293 item in branch that calls a recursion is a subroutine call - as in the
1294 second branch in the above example - was incorrectly given the compile-
1295 time error "recursive call could loop indefinitely" because pcre_compile()
1296 was not correctly checking the subroutine for matching a non-empty string.
1297
1298 14. The checks for overrunning compiling workspace could trigger after an
1299 overrun had occurred. This is a "should never occur" error, but it can be
1300 triggered by pathological patterns such as hundreds of nested parentheses.
1301 The checks now trigger 100 bytes before the end of the workspace.
1302
1303 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1304
1305
1306 Version 8.01 19-Jan-2010
1307 ------------------------
1308
1309 1. If a pattern contained a conditional subpattern with only one branch (in
1310 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1311 computed the wrong minimum data length (which is of course zero for such
1312 subpatterns). This could cause incorrect "no match" results.
1313
1314 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1315 the pattern is reset in the first branch, pcre_compile() failed with
1316 "internal error: code overflow at offset...". This happened only when
1317 the reset was to the original external option setting. (An optimization
1318 abstracts leading options settings into an external setting, which was the
1319 cause of this.)
1320
1321 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1322 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1323 assertion pattern did not match (meaning that the assertion was true), it
1324 was incorrectly treated as false if the SKIP had been reached during the
1325 matching. This also applied to assertions used as conditions.
1326
1327 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1328 assertion subpattern, including such a pattern used as a condition,
1329 unpredictable results occurred, instead of the error return
1330 PCRE_ERROR_DFA_UITEM.
1331
1332 5. The C++ GlobalReplace function was not working like Perl for the special
1333 situation when an empty string is matched. It now does the fancy magic
1334 stuff that is necessary.
1335
1336 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1337 removed. (These were left over from very, very early versions of PCRE.)
1338
1339 7. Some cosmetic changes to the code to make life easier when compiling it
1340 as part of something else:
1341
1342 (a) Change DEBUG to PCRE_DEBUG.
1343
1344 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1345 called "current" as "current_branch", to prevent a collision with the
1346 Linux macro when compiled as a kernel module.
1347
1348 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1349 prevent a collision with the Linux macro when compiled as a kernel
1350 module.
1351
1352 8. In pcre_compile() there are some checks for integer overflows that used to
1353 cast potentially large values to (double). This has been changed to that
1354 when building, a check for int64_t is made, and if it is found, it is used
1355 instead, thus avoiding the use of floating point arithmetic. (There is no
1356 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1357 double.
1358
1359 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1360 2005 (difference between two addresses compared to an unsigned value).
1361
1362 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1363 custom one, because of the following reported problem in Windows:
1364
1365 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1366 under Win32.
1367 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1368 therefore missing the function definition.
1369 - The compiler thus generates a "C" signature for the test function.
1370 - The linker fails to find the "C" function.
1371 - PCRE fails to configure if asked to do so against libbz2.
1372
1373 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1374 messages were output:
1375
1376 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1377 rerunning libtoolize, to keep the correct libtool macros in-tree.
1378 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1379
1380 I have done both of these things.
1381
1382 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1383 most of the time, it *can* run out if it is given a pattern that contains a
1384 runaway infinite recursion. I updated the discussion in the pcrestack man
1385 page.
1386
1387 13. Now that we have gone to the x.xx style of version numbers, the minor
1388 version may start with zero. Using 08 or 09 is a bad idea because users
1389 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1390 interpreted as invalid octal numbers. I've updated the previous comment in
1391 configure.ac, and also added a check that gives an error if 08 or 09 are
1392 used.
1393
1394 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1395 causing partial matching to fail when the end of the subject matched \W
1396 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1397
1398 15. There were some discrepancies between the declarations in pcre_internal.h
1399 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1400 their definitions. The declarations used "const uschar *" and the
1401 definitions used USPTR. Even though USPTR is normally defined as "const
1402 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1403 reported that: "This difference in casting confuses some C++ compilers, for
1404 example, SunCC recognizes above declarations as different functions and
1405 generates broken code for hbpcre." I have changed the declarations to use
1406 USPTR.
1407
1408 16. GNU libtool is named differently on some systems. The autogen.sh script now
1409 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1410 (FreeBSD).
1411
1412 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1413 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1414 comment: "Figure out how to create a longlong from a string: strtoll and
1415 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1416 instance, but it only takes 2 args instead of 3!"
1417
1418 18. A subtle bug concerned with back references has been fixed by a change of
1419 specification, with a corresponding code fix. A pattern such as
1420 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1421 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1422 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1423 same bug. Such groups have to be quantified to be useful, or contained
1424 inside another quantified group. (If there's no repetition, the reference
1425 can never match.) The problem arises because, having left the group and
1426 moved on to the rest of the pattern, a later failure that backtracks into
1427 the group uses the captured value from the final iteration of the group
1428 rather than the correct earlier one. I have fixed this in PCRE by forcing
1429 any group that contains a reference to itself to be an atomic group; that
1430 is, there cannot be any backtracking into it once it has completed. This is
1431 similar to recursive and subroutine calls.
1432
1433
1434 Version 8.00 19-Oct-09
1435 ----------------------
1436
1437 1. The table for translating pcre_compile() error codes into POSIX error codes
1438 was out-of-date, and there was no check on the pcre_compile() error code
1439 being within the table. This could lead to an OK return being given in
1440 error.
1441
1442 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1443 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1444 in a Windows environment.
1445
1446 3. The pcregrep --count option prints the count for each file even when it is
1447 zero, as does GNU grep. However, pcregrep was also printing all files when
1448 --files-with-matches was added. Now, when both options are given, it prints
1449 counts only for those files that have at least one match. (GNU grep just
1450 prints the file name in this circumstance, but including the count seems
1451 more useful - otherwise, why use --count?) Also ensured that the
1452 combination -clh just lists non-zero counts, with no names.
1453
1454 4. The long form of the pcregrep -F option was incorrectly implemented as
1455 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1456 but it seems right to fix it, and I didn't think it was worth preserving
1457 the old behaviour.
1458
1459 5. The command line items --regex=pattern and --regexp=pattern were not
1460 recognized by pcregrep, which required --regex pattern or --regexp pattern
1461 (with a space rather than an '='). The man page documented the '=' forms,
1462 which are compatible with GNU grep; these now work.
1463
1464 6. No libpcreposix.pc file was created for pkg-config; there was just
1465 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1466
1467 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1468 when UCP support is not needed, by modifying the Python script that
1469 generates it from Unicode data files. This should not matter if the module
1470 is correctly used as a library, but I received one complaint about 50K of
1471 unwanted data. My guess is that the person linked everything into his
1472 program rather than using a library. Anyway, it does no harm.
1473
1474 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1475 was a minimum greater than 1 for a wide character in a possessive
1476 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1477 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1478 character. Chaos in the form of incorrect output or a compiling loop could
1479 result.
1480
1481 9. The restrictions on what a pattern can contain when partial matching is
1482 requested for pcre_exec() have been removed. All patterns can now be
1483 partially matched by this function. In addition, if there are at least two
1484 slots in the offset vector, the offset of the earliest inspected character
1485 for the match and the offset of the end of the subject are set in them when
1486 PCRE_ERROR_PARTIAL is returned.
1487
1488 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1489 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1490 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1491 and may be more useful for multi-segment matching.
1492
1493 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1494 used to be given if ever the end of the subject was reached; now it is
1495 given only if matching could not proceed because another character was
1496 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1497 string "Z", which now yields "no match" instead of "partial match". In the
1498 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1499 final character ended with (*FAIL).
1500
1501 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1502 if the pattern had a "must contain" character that was already found in the
1503 earlier partial match, unless partial matching was again requested. For
1504 example, with the pattern /dog.(body)?/, the "must contain" character is
1505 "g". If the first part-match was for the string "dog", restarting with
1506 "sbody" failed. This bug has been fixed.
1507
1508 13. The string returned by pcre_dfa_exec() after a partial match has been
1509 changed so that it starts at the first inspected character rather than the
1510 first character of the match. This makes a difference only if the pattern
1511 starts with a lookbehind assertion or \b or \B (\K is not supported by
1512 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1513 matching functions compatible, and I think it's the right thing to do.
1514
1515 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1516 so that the demonstration program is easily available in environments where
1517 PCRE has not been installed from source.
1518
1519 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1520 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1521 library.
1522
1523 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1524 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1525 is not the first non-POSIX option to be added. Clearly some people find
1526 these options useful.
1527
1528 17. If a caller to the POSIX matching function regexec() passes a non-zero
1529 value for nmatch with a NULL value for pmatch, the value of
1530 nmatch is forced to zero.
1531
1532 18. RunGrepTest did not have a test for the availability of the -u option of
1533 the diff command, as RunTest does. It now checks in the same way as
1534 RunTest, and also checks for the -b option.
1535
1536 19. If an odd number of negated classes containing just a single character
1537 interposed, within parentheses, between a forward reference to a named
1538 subpattern and the definition of the subpattern, compilation crashed with
1539 an internal error, complaining that it could not find the referenced
1540 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1541 [The bug was that it was starting one character too far in when skipping
1542 over the character class, thus treating the ] as data rather than
1543 terminating the class. This meant it could skip too much.]
1544
1545 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1546 /g option in pcretest when the pattern contains \K, which makes it possible
1547 to have an empty string match not at the start, even when the pattern is
1548 anchored. Updated pcretest and pcredemo to use this option.
1549
1550 21. If the maximum number of capturing subpatterns in a recursion was greater
1551 than the maximum at the outer level, the higher number was returned, but
1552 with unset values at the outer level. The correct (outer level) value is
1553 now given.
1554
1555 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1556 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1557 make it do so. The string so far is captured, making this feature
1558 compatible with Perl.
1559
1560 23. The tests have been re-organized, adding tests 11 and 12, to make it
1561 possible to check the Perl 5.10 features against Perl 5.10.
1562
1563 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1564 pattern matches a fixed length string. PCRE did not allow this; now it
1565 does. Neither allows recursion.
1566
1567 25. I finally figured out how to implement a request to provide the minimum
1568 length of subject string that was needed in order to match a given pattern.
1569 (It was back references and recursion that I had previously got hung up
1570 on.) This code has now been added to pcre_study(); it finds a lower bound
1571 to the length of subject needed. It is not necessarily the greatest lower
1572 bound, but using it to avoid searching strings that are too short does give
1573 some useful speed-ups. The value is available to calling programs via
1574 pcre_fullinfo().
1575
1576 26. While implementing 25, I discovered to my embarrassment that pcretest had
1577 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1578 study optimizations had never been tested with that matching function.
1579 Oops. What is worse, even when it was passed study data, there was a bug in
1580 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1581 were also very few tests of studied patterns with pcre_dfa_exec().
1582
1583 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1584 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1585 on the other side of the coin, they are no longer allowed to have different
1586 names, because these cannot be distinguished in PCRE, and this has caused
1587 confusion. (This is a difference from Perl.)
1588
1589 28. When duplicate subpattern names are present (necessarily with different
1590 numbers, as required by 27 above), and a test is made by name in a
1591 conditional pattern, either for a subpattern having been matched, or for
1592 recursion in such a pattern, all the associated numbered subpatterns are
1593 tested, and the overall condition is true if the condition is true for any
1594 one of them. This is the way Perl works, and is also more like the way
1595 testing by number works.
1596
1597
1598 Version 7.9 11-Apr-09
1599 ---------------------
1600
1601 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1602 (pcretest), all targets were linked against these libraries. This included
1603 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1604 libraries. This caused unwanted dependencies to be created. This problem
1605 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1606 pcretest is linked with readline.
1607
1608 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1609 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1610 moved outside it again, because FALSE and TRUE are already defined in AIX,
1611 but BOOL is not.
1612
1613 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1614 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1615
1616 4. The pcregrep documentation said spaces were inserted as well as colons (or
1617 hyphens) following file names and line numbers when outputting matching
1618 lines. This is not true; no spaces are inserted. I have also clarified the
1619 wording for the --colour (or --color) option.
1620
1621 5. In pcregrep, when --colour was used with -o, the list of matching strings
1622 was not coloured; this is different to GNU grep, so I have changed it to be
1623 the same.
1624
1625 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1626 each matching line was coloured. Now it goes on to look for further matches
1627 of any of the test patterns, which is the same behaviour as GNU grep.
1628
1629 7. A pattern that could match an empty string could cause pcregrep to loop; it
1630 doesn't make sense to accept an empty string match in pcregrep, so I have
1631 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1632 seems to be how GNU grep behaves.
1633
1634 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1635 start or after a newline", because the conditional assertion was not being
1636 correctly handled. The rule now is that both the assertion and what follows
1637 in the first alternative must satisfy the test.
1638
1639 9. If auto-callout was enabled in a pattern with a conditional group whose
1640 condition was an assertion, PCRE could crash during matching, both with
1641 pcre_exec() and pcre_dfa_exec().
1642
1643 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1644 used for matching.
1645
1646 11. Unicode property support in character classes was not working for
1647 characters (bytes) greater than 127 when not in UTF-8 mode.
1648
1649 12. Added the -M command line option to pcretest.
1650
1651 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1652
1653 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1654
1655 16. Added comments and documentation about mis-use of no_arg in the C++
1656 wrapper.
1657
1658 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1659 from Martin Jerabek that uses macro names for all relevant character and
1660 string constants.
1661
1662 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1663 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1664 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1665 these, but not everybody uses configure.
1666
1667 19. A conditional group that had only one branch was not being correctly
1668 recognized as an item that could match an empty string. This meant that an
1669 enclosing group might also not be so recognized, causing infinite looping
1670 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1671 with the subject "ab", where knowledge that the repeated group can match
1672 nothing is needed in order to break the loop.
1673
1674 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1675 exec(), but without supplying a callout function, matching went wrong.
1676
1677 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1678 leak if the size of the offset vector was greater than 30. When the vector
1679 is smaller, the saved offsets during recursion go onto a local stack
1680 vector, but for larger vectors malloc() is used. It was failing to free
1681 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1682 error, in fact).
1683
1684 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1685 heapframe that is used only when UTF-8 support is enabled. This caused no
1686 problem, but was untidy.
1687
1688 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1689 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1690 included within another project.
1691
1692 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1693 slightly modified by me:
1694
1695 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1696 not building pcregrep.
1697
1698 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1699 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1700
1701 25. Forward references, both numeric and by name, in patterns that made use of
1702 duplicate group numbers, could behave incorrectly or give incorrect errors,
1703 because when scanning forward to find the reference group, PCRE was not
1704 taking into account the duplicate group numbers. A pattern such as
1705 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1706
1707 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1708 the feature of a custom pointer more persuasive (as requested by a user).
1709
1710 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1711 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1712 pre-defined.
1713
1714 28. Added support for (*UTF8) at the start of a pattern.
1715
1716 29. Arrange for flags added by the "release type" setting in CMake to be shown
1717 in the configuration summary.
1718
1719
1720 Version 7.8 05-Sep-08
1721 ---------------------
1722
1723 1. Replaced UCP searching code with optimized version as implemented for Ad
1724 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1725 stage table and inline lookup instead of a function, giving speed ups of 2
1726 to 5 times on some simple patterns that I tested. Permission was given to
1727 distribute the MultiStage2.py script that generates the tables (it's not in
1728 the tarball, but is in the Subversion repository).
1729
1730 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1731 scripts.
1732
1733 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1734 a group with a zero qualifier. The result of the study could be incorrect,
1735 or the function might crash, depending on the pattern.
1736
1737 4. Caseless matching was not working for non-ASCII characters in back
1738 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1739 It now works when Unicode Property Support is available.
1740
1741 5. In pcretest, an escape such as \x{de} in the data was always generating
1742 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1743 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1744 truncation.
1745
1746 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1747
1748 7. Added two (int) casts to pcregrep when printing the difference of two
1749 pointers, in case they are 64-bit values.
1750
1751 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1752 test 2 if it fails.
1753
1754 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1755 and a #define of that name to empty if it is not externally set. This is to
1756 allow users of MSVC to set it if necessary.
1757
1758 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1759 the convenience functions in the pcre_get.c source file.
1760
1761 11. An option change at the start of a pattern that had top-level alternatives
1762 could cause overwriting and/or a crash. This command provoked a crash in
1763 some environments:
1764
1765 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1766
1767 This potential security problem was recorded as CVE-2008-2371.
1768
1769 12. For a pattern where the match had to start at the beginning or immediately
1770 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1771 pcre_dfa_exec() could read past the end of the passed subject if there was
1772 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1773 pcretest so that it places the subject at the end of its malloc-ed buffer.
1774
1775 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1776 exec() might read past the end of the data buffer in UTF-8 mode.
1777
1778 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1779 the data contained the byte 0x85 as part of a UTF-8 character within its
1780 first line. This applied both to normal and DFA matching.
1781
1782 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1783 /^[^d]*?$/8 failed to match "abc".
1784
1785 16. Added a missing copyright notice to pcrecpp_internal.h.
1786
1787 17. Make it more clear in the documentation that values returned from
1788 pcre_exec() in ovector are byte offsets, not character counts.
1789
1790 18. Tidied a few places to stop certain compilers from issuing warnings.
1791
1792 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1793 supplied by Stefan Weber. I made a further small update for 7.8 because
1794 there is a change of source arrangements: the pcre_searchfuncs.c module is
1795 replaced by pcre_ucd.c.
1796
1797
1798 Version 7.7 07-May-08
1799 ---------------------
1800
1801 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1802 a string to a long long, pretend we don't even have a long long." This is
1803 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1804
1805 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1806 pre-7.6 versions, which defined a global no_arg variable instead of putting
1807 it in the RE class. (See also #8 below.)
1808
1809 3. Remove a line of dead code, identified by coverity and reported by Nuno
1810 Lopes.
1811
1812 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1813
1814 (1) The include/exclude patterns were being applied to the whole pathnames
1815 of files, instead of just to the final components.
1816
1817 (2) If there was more than one level of directory, the subdirectories were
1818 skipped unless they satisfied the include/exclude conditions. This is
1819 inconsistent with GNU grep (and could even be seen as contrary to the
1820 pcregrep specification - which I improved to make it absolutely clear).
1821 The action now is always to scan all levels of directory, and just
1822 apply the include/exclude patterns to regular files.
1823
1824 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1825 --exclude_dir in the tests to avoid scanning .svn directories.
1826
1827 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1828 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1829 doesn't support NULs in patterns.
1830
1831 7. Added some missing "const"s to declarations of static tables in
1832 pcre_compile.c and pcre_dfa_exec.c.
1833
1834 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1835 caused by fix #2 above. (Subsequently also a second patch to fix the
1836 first patch. And a third patch - this was a messy problem.)
1837
1838 9. Applied Craig's patch to remove the use of push_back().
1839
1840 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1841 matching function regexec().
1842
1843 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1844 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1845 references. PCRE supports relative numbers with this syntax (I don't think
1846 Oniguruma does).
1847
1848 12. Previously, a group with a zero repeat such as (...){0} was completely
1849 omitted from the compiled regex. However, this means that if the group
1850 was called as a subroutine from elsewhere in the pattern, things went wrong
1851 (an internal error was given). Such groups are now left in the compiled
1852 pattern, with a new opcode that causes them to be skipped at execution
1853 time.
1854
1855 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1856 to the way PCRE behaves:
1857
1858 (a) A lone ] character is dis-allowed (Perl treats it as data).
1859
1860 (b) A back reference to an unmatched subpattern matches an empty string
1861 (Perl fails the current match path).
1862
1863 (c) A data ] in a character class must be notated as \] because if the
1864 first data character in a class is ], it defines an empty class. (In
1865 Perl it is not possible to have an empty class.) The empty class []
1866 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1867 The negative empty class [^] matches any one character, independently
1868 of the DOTALL setting.
1869
1870 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1871 non-existent subpattern following a character class starting with ']' and
1872 containing () gave an internal compiling error instead of "reference to
1873 non-existent subpattern". Fortunately, when the pattern did exist, the
1874 compiled code was correct. (When scanning forwards to check for the
1875 existencd of the subpattern, it was treating the data ']' as terminating
1876 the class, so got the count wrong. When actually compiling, the reference
1877 was subsequently set up correctly.)
1878
1879 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1880 it was being rejected as not supported by pcre_dfa_exec(), even though
1881 other assertions are supported. I have made pcre_dfa_exec() support
1882 (*FAIL).
1883
1884 16. The implementation of 13c above involved the invention of a new opcode,
1885 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1886 cannot be changed at match time, I realized I could make a small
1887 improvement to matching performance by compiling OP_ALLANY instead of
1888 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1889 on the OP_ANY path.
1890
1891 17. Compiling pcretest on Windows with readline support failed without the
1892 following two fixes: (1) Make the unistd.h include conditional on
1893 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1894
1895 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1896 ncurses library to be included for pcretest when ReadLine support is
1897 requested, but also to allow for it to be overridden. This patch came from
1898 Daniel Bergström.
1899
1900 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1901 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1902 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1903 spotting this.
1904
1905
1906 Version 7.6 28-Jan-08
1907 ---------------------
1908
1909 1. A character class containing a very large number of characters with
1910 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1911 overflow.
1912
1913 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1914 HAVE_LONG_LONG is not defined.
1915
1916 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1917 bring them up to date and include new features. This patch includes:
1918
1919 - Fixed PH's badly added libz and libbz2 support.
1920 - Fixed a problem with static linking.
1921 - Added pcredemo. [But later removed - see 7 below.]
1922 - Fixed dftables problem and added an option.
1923 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1924 HAVE_LONG_LONG.
1925 - Added readline support for pcretest.
1926 - Added an listing of the option settings after cmake has run.
1927
1928 4. A user submitted a patch to Makefile that makes it easy to create
1929 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1930 Makefile.am that cause it to include this special target, without
1931 affecting anything else. Note that the same mingw target plus all
1932 the other distribution libraries and programs are now supported
1933 when configuring with CMake (see 6 below) instead of with
1934 Configure/Make.
1935
1936 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1937 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1938 exported in the Windows port". It has not yet been confirmed that the patch
1939 solves the problem, but it does no harm.
1940
1941 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1942 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1943 with CMake, and also correct the comment about stack recursion.
1944
1945 7. Remove the automatic building of pcredemo from the ./configure system and
1946 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1947 of a program that users should build themselves after PCRE is installed, so
1948 building it automatically is not really right. What is more, it gave
1949 trouble in some build environments.
1950
1951 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1952
1953
1954 Version 7.5 10-Jan-08
1955 ---------------------
1956
1957 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1958 values in parens when parsing an RE using the C++ wrapper."
1959
1960 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1961 Characters greater than 255 were excluded from the class instead of being
1962 included.
1963
1964 3. The same bug as (2) above applied to negated POSIX classes such as
1965 [:^space:].
1966
1967 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1968 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1969 I have changed it.
1970
1971 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1972 first named subpattern) and a construct such as (?&a) would reference the
1973 first named subpattern whose name started with "a" (in other words, the
1974 length check was missing). Both these problems are fixed. "Subpattern name
1975 expected" is now given for (?&) (a zero-length name), and this patch also
1976 makes it give the same error for \k'' (previously it complained that that
1977 was a reference to a non-existent subpattern).
1978
1979 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1980 this is right because (?- can be followed by option settings as well as by
1981 digits. I have, however, made the messages clearer.
1982
1983 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1984 than the number used in the conditional) now cause a compile-time error.
1985 This is actually not compatible with Perl, which accepts such patterns, but
1986 treats the conditional as always being FALSE (as PCRE used to), but it
1987 seems to me that giving a diagnostic is better.
1988
1989 8. Change "alphameric" to the more common word "alphanumeric" in comments
1990 and messages.
1991
1992 9. Fix two occurrences of "backslash" in comments that should have been
1993 "backspace".
1994
1995 10. Remove two redundant lines of code that can never be obeyed (their function
1996 was moved elsewhere).
1997
1998 11. The program that makes PCRE's Unicode character property table had a bug
1999 which caused it to generate incorrect table entries for sequences of
2000 characters that have the same character type, but are in different scripts.
2001 It amalgamated them into a single range, with the script of the first of
2002 them. In other words, some characters were in the wrong script. There were
2003 thirteen such cases, affecting characters in the following ranges:
2004
2005 U+002b0 - U+002c1
2006 U+0060c - U+0060d
2007 U+0061e - U+00612
2008 U+0064b - U+0065e
2009 U+0074d - U+0076d
2010 U+01800 - U+01805
2011 U+01d00 - U+01d77
2012 U+01d9b - U+01dbf
2013 U+0200b - U+0200f
2014 U+030fc - U+030fe
2015 U+03260 - U+0327f
2016 U+0fb46 - U+0fbb1
2017 U+10450 - U+1049d
2018
2019 12. The -o option (show only the matching part of a line) for pcregrep was not
2020 compatible with GNU grep in that, if there was more than one match in a
2021 line, it showed only the first of them. It now behaves in the same way as
2022 GNU grep.
2023
2024 13. If the -o and -v options were combined for pcregrep, it printed a blank
2025 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2026 does the same. The return code can be used to tell if there were any
2027 non-matching lines.
2028
2029 14. Added --file-offsets and --line-offsets to pcregrep.
2030
2031 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2032 infinitely looping recursion. The bug was that positive lookaheads were not
2033 being skipped when checking for a possible empty match (negative lookaheads
2034 and both kinds of lookbehind were skipped).
2035
2036 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2037 inclusion of <windows.h> to before rather than after the definition of
2038 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2039
2040 17. Specifying a possessive quantifier with a specific limit for a Unicode
2041 character property caused pcre_compile() to compile bad code, which led at
2042 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2043 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2044 caused the error; without that there was no problem.
2045
2046 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2047
2048 19. Added --enable-pcretest-libreadline.
2049
2050 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2051 RE::GlobalReplace(). As a result, the number of replacements returned was
2052 double what it should be. I removed one of the increments, but Craig sent a
2053 later patch that removed the other one (the right fix) and added unit tests
2054 that check the return values (which was not done before).
2055
2056 21. Several CMake things:
2057
2058 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2059 the names libpcre and libpcreposix, not just pcre and pcreposix.
2060
2061 (2) The above change means that pcretest and pcregrep are now correctly
2062 linked with the newly-built libraries, not previously installed ones.
2063
2064 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2065
2066 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2067 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2068 UTF-8 newline character). The key issue is that the pattern starts .*;
2069 this means that the match must be either at the beginning, or after a
2070 newline. The bug was in the code for advancing after a failed match and
2071 checking that the new position followed a newline. It was not taking
2072 account of UTF-8 characters correctly.
2073
2074 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2075 character classes. PCRE was not treating the sequence [:...:] as a
2076 character class unless the ... were all letters. Perl, however, seems to
2077 allow any characters between [: and :], though of course it rejects as
2078 unknown any "names" that contain non-letters, because all the known class
2079 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2080 for example, whereas PCRE did not - it did not recognize a POSIX character
2081 class. This seemed a bit dangerous, so the code has been changed to be
2082 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2083 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2084 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2085 Perl does, and where it didn't before.
2086
2087 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2088 Windows environments %n is disabled by default.
2089
2090
2091 Version 7.4 21-Sep-07
2092 ---------------------
2093
2094 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2095 means that a class such as [\s] counted as "explicit reference to CR or
2096 LF". That isn't really right - the whole point of the change was to try to
2097 help when there was an actual mention of one of the two characters. So now
2098 the change happens only if \r or \n (or a literal CR or LF) character is
2099 encountered.
2100
2101 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2102 of both had grown to the point where there were only 3 bits left.
2103 Fortunately, there was spare space in the data structure, and so I have
2104 moved the internal flags into a new 16-bit field to free up more option
2105 bits.
2106
2107 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2108 but did not set the internal JCHANGED flag - either of these is enough to
2109 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2110 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2111 start sets both bits.
2112
2113 4. Added options (at build time, compile time, exec time) to change \R from
2114 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2115
2116 5. doc/pcresyntax.html was missing from the distribution.
2117
2118 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2119 compatibility, even though it is no longer used.
2120
2121 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2122 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2123 windows.h file is present (where different names are used). [This was
2124 reversed later after testing - see 16 below.]
2125
2126 8. Changed all #include <config.h> to #include "config.h". There were also
2127 some further <pcre.h> cases that I changed to "pcre.h".
2128
2129 9. When pcregrep was used with the --colour option, it missed the line ending
2130 sequence off the lines that it output.
2131
2132 10. It was pointed out to me that arrays of string pointers cause lots of
2133 relocations when a shared library is dynamically loaded. A technique of
2134 using a single long string with a table of offsets can drastically reduce
2135 these. I have refactored PCRE in four places to do this. The result is
2136 dramatic:
2137
2138 Originally: 290
2139 After changing UCP table: 187
2140 After changing error message table: 43
2141 After changing table of "verbs" 36
2142 After changing table of Posix names 22
2143
2144 Thanks to the folks working on Gregex for glib for this insight.
2145
2146 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2147 unicode-properties was also set.
2148
2149 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2150
2151 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2152 checked only for CRLF.
2153
2154 14. Added casts to pcretest.c to avoid compiler warnings.
2155
2156 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2157
2158 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2159 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2160 entirely. This removes changes made in 7 above.
2161
2162 17. The CMake files have been updated, and there is now more information about
2163 building with CMake in the NON-UNIX-USE document.
2164
2165
2166 Version 7.3 28-Aug-07
2167 ---------------------
2168
2169 1. In the rejigging of the build system that eventually resulted in 7.1, the
2170 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2171 brackets there is not right, since it causes compilers to look for an
2172 installed pcre.h, not the version that is in the source that is being
2173 compiled (which of course may be different). I have changed it back to:
2174
2175 #include "pcre.h"
2176
2177 I have a vague recollection that the change was concerned with compiling in
2178 different directories, but in the new build system, that is taken care of
2179 by the VPATH setting the Makefile.
2180
2181 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2182 when the subject happened to end in the byte 0x85 (e.g. if the last
2183 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2184 characters but of course it shouldn't be taken as a newline when it is part
2185 of another character. The bug was that, for an unlimited repeat of . in
2186 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2187 characters when looking for a newline.
2188
2189 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2190
2191 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2192 in debug output.
2193
2194 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2195 long printing in the pcrecpp unittest when running under MinGW.
2196
2197 6. ESC_K was left out of the EBCDIC table.
2198
2199 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2200 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2201 limit also applies to "virtual nesting" when a pattern is recursive, and in
2202 this case 1000 isn't so big. I have been able to remove this limit at the
2203 expense of backing off one optimization in certain circumstances. Normally,
2204 when pcre_exec() would call its internal match() function recursively and
2205 immediately return the result unconditionally, it uses a "tail recursion"
2206 feature to save stack. However, when a subpattern that can match an empty
2207 string has an unlimited repetition quantifier, it no longer makes this
2208 optimization. That gives it a stack frame in which to save the data for
2209 checking that an empty string has been matched. Previously this was taken
2210 from the 1000-entry workspace that had been reserved. So now there is no
2211 explicit limit, but more stack is used.
2212
2213 8. Applied Daniel's patches to solve problems with the import/export magic
2214 syntax that is required for Windows, and which was going wrong for the
2215 pcreposix and pcrecpp parts of the library. These were overlooked when this
2216 problem was solved for the main library.
2217
2218 9. There were some crude static tests to avoid integer overflow when computing
2219 the size of patterns that contain repeated groups with explicit upper
2220 limits. As the maximum quantifier is 65535, the maximum group length was
2221 set at 30,000 so that the product of these two numbers did not overflow a
2222 32-bit integer. However, it turns out that people want to use groups that
2223 are longer than 30,000 bytes (though not repeat them that many times).
2224 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2225 made it possible to implement the integer overflow checks in a much more
2226 dynamic way, which I have now done. The artificial limitation on group
2227 length has been removed - we now have only the limit on the total length of
2228 the compiled pattern, which depends on the LINK_SIZE setting.
2229
2230 10. Fixed a bug in the documentation for get/copy named substring when
2231 duplicate names are permitted. If none of the named substrings are set, the
2232 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2233 empty string.
2234
2235 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2236 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2237 because the ] is interpreted as the first data character and the
2238 terminating ] is not found. PCRE has been made compatible with Perl in this
2239 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2240 cause memory overwriting.
2241
2242 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2243 string has been matched (to stop an infinite loop). It was not recognizing
2244 a conditional subpattern that could match an empty string if that
2245 subpattern was within another subpattern. For example, it looped when
2246 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2247 condition was not nested. This bug has been fixed.
2248
2249 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2250 past the start of the subject in the presence of bytes with the top bit
2251 set, for example "\x8aBCD".
2252
2253 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2254 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2255
2256 14. Optimized (?!) to (*FAIL).
2257
2258 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2259 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2260 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2261 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2262 does: it's just the validity check that is more restrictive.
2263
2264 16. Inserted checks for integer overflows during escape sequence (backslash)
2265 processing, and also fixed erroneous offset values for syntax errors during
2266 backslash processing.
2267
2268 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2269 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2270
2271 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2272 caused an overrun.
2273
2274 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2275 something other than just ASCII characters) inside a group that had an
2276 unlimited repeat caused a loop at compile time (while checking to see
2277 whether the group could match an empty string).
2278
2279 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2280 [\P{Any}] did so. (Error in the code for printing property names.)
2281
2282 21. An orphan \E inside a character class could cause a crash.
2283
2284 22. A repeated capturing bracket such as (A)? could cause a wild memory
2285 reference during compilation.
2286
2287 23. There are several functions in pcre_compile() that scan along a compiled
2288 expression for various reasons (e.g. to see if it's fixed length for look
2289 behind). There were bugs in these functions when a repeated \p or \P was
2290 present in the pattern. These operators have additional parameters compared
2291 with \d, etc, and these were not being taken into account when moving along
2292 the compiled data. Specifically:
2293
2294 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2295 length.
2296
2297 (b) An item such as \pL+ within a repeated group could cause crashes or
2298 loops.
2299
2300 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2301 "reference to non-existent subpattern" error.
2302
2303 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2304
2305 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2306 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2307
2308 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2309
2310 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2311 character were causing crashes (broken optimization).
2312
2313 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2314 \p or \P) caused a compile-time loop.
2315
2316 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2317 break. For example, the unstudied pattern [\r\n]A does not match the string
2318 "\r\nA" because change 7.0/46 below moves the current point on by two
2319 characters after failing to match at the start. However, the pattern \nA
2320 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2321 the same is true. There doesn't seem any very clean way out of this, but
2322 what I have chosen to do makes the common cases work: PCRE now takes note
2323 of whether there can be an explicit match for \r or \n anywhere in the
2324 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2325 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2326 pattern has explicit CR or LF references.
2327
2328 29. Added (*CR) etc for changing newline setting at start of pattern.
2329
2330
2331 Version 7.2 19-Jun-07
2332 ---------------------
2333
2334 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2335 which is apparently normally available under Windows.
2336
2337 2. Re-jig the pcregrep tests with different newline settings in an attempt
2338 to make them independent of the local environment's newline setting.
2339
2340 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2341
2342 4. Some of the "internals" tests were previously cut out when the link size
2343 was not 2, because the output contained actual offsets. The recent new
2344 "Z" feature of pcretest means that these can be cut out, making the tests
2345 usable with all link sizes.
2346
2347 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2348 stack recursion. This gives a massive performance boost under BSD, but just
2349 a small improvement under Linux. However, it saves one field in the frame
2350 in all cases.
2351
2352 6. Added more features from the forthcoming Perl 5.10:
2353
2354 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2355 recursion call. It refers to the nth most recently opened parentheses.
2356
2357 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2358 to be opened parentheses.
2359
2360 (c) Conditions that refer to capturing parentheses can be specified
2361 relatively, for example, (?(-2)... or (?(+3)...
2362
2363 (d) \K resets the start of the current match so that everything before
2364 is not part of it.
2365
2366 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2367
2368 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2369 reference syntax.
2370
2371 (g) (?| introduces a group in which the numbering of parentheses in each
2372 alternative starts with the same number.
2373
2374 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2375
2376 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2377 PCRE_INFO_JCHANGED.
2378
2379 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2380 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2381 for detecting groups that can match an empty string.
2382
2383 9. A pattern with a very large number of alternatives (more than several
2384 hundred) was running out of internal workspace during the pre-compile
2385 phase, where pcre_compile() figures out how much memory will be needed. A
2386 bit of new cunning has reduced the workspace needed for groups with
2387 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2388 workspace instead of running out of the 4096 that are available.
2389
2390 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2391
2392 11. Applied patch from Google to remove an optimization that didn't quite work.
2393 The report of the bug said:
2394
2395 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2396 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2397 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2398
2399 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2400 it matched the wrong number of bytes.
2401
2402
2403 Version 7.1 24-Apr-07
2404 ---------------------
2405
2406 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2407 that is more "standard", making use of automake and other Autotools. There
2408 is some re-arrangement of the files and adjustment of comments consequent
2409 on this.
2410
2411 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2412 for recursive directory scanning broke on some systems because the files
2413 are not scanned in any specific order and on different systems the order
2414 was different. A call to "sort" has been inserted into RunGrepTest for the
2415 approprate test as a short-term fix. In the longer term there may be an
2416 alternative.
2417
2418 3. I had an email from Eric Raymond about problems translating some of PCRE's
2419 man pages to HTML (despite the fact that I distribute HTML pages, some
2420 people do their own conversions for various reasons). The problems
2421 concerned the use of low-level troff macros .br and .in. I have therefore
2422 removed all such uses from the man pages (some were redundant, some could
2423 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2424 HTML has been updated to handle .nf/.fi and to complain if it encounters
2425 .br or .in.
2426
2427 4. Updated comments in configure.ac that get placed in config.h.in and also
2428 arranged for config.h to be included in the distribution, with the name
2429 config.h.generic, for the benefit of those who have to compile without
2430 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2431
2432 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2433 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2434 makevp.bat for the current PCRE, using the additional files
2435 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2436
2437 6. A Windows user reported a minor discrepancy with test 2, which turned out
2438 to be caused by a trailing space on an input line that had got lost in his
2439 copy. The trailing space was an accident, so I've just removed it.
2440
2441 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2442 that is needed.
2443
2444 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2445 as "const" (a) because they are and (b) because it helps the PHP
2446 maintainers who have recently made a script to detect big data structures
2447 in the php code that should be moved to the .rodata section. I remembered
2448 to update Builducptable as well, so it won't revert if ucptable.h is ever
2449 re-created.
2450
2451 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2452 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2453 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2454 support is not required. This saves 1.5-2K of code, which is important in
2455 some applications.
2456
2457 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2458 so as not to refer to the tables, even though these functions will never be
2459 called when UTF-8 support is disabled. Otherwise there are problems with a
2460 shared library.
2461
2462 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2463
2464 (a) It was defining its arguments as char * instead of void *.
2465
2466 (b) It was assuming that all moves were upwards in memory; this was true
2467 a long time ago when I wrote it, but is no longer the case.
2468
2469 The emulated memove() is provided for those environments that have neither
2470 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2471 is clearly not the case, as these two bugs were recently reported.
2472
2473 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2474 and Detrail to create the HTML documentation, the .txt form of the man
2475 pages, and it removes trailing spaces from listed files. It also creates
2476 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2477 case, it wraps all the #defines with #ifndefs. This script should be run
2478 before "make dist".
2479
2480 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2481 with Unicode property support.
2482
2483 (a) For a maximizing quantifier, if the two different cases of the
2484 character were of different lengths in their UTF-8 codings (there are
2485 some cases like this - I found 11), and the matching function had to
2486 back up over a mixture of the two cases, it incorrectly assumed they
2487 were both the same length.
2488
2489 (b) When PCRE was configured to use the heap rather than the stack for
2490 recursion during matching, it was not correctly preserving the data for
2491 the other case of a UTF-8 character when checking ahead for a match
2492 while processing a minimizing repeat. If the check also involved
2493 matching a wide character, but failed, corruption could cause an
2494 erroneous result when trying to check for a repeat of the original
2495 character.
2496
2497 13. Some tidying changes to the testing mechanism:
2498
2499 (a) The RunTest script now detects the internal link size and whether there
2500 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2501 values substituted by "configure". (The RunGrepTest script already did
2502 this for UTF-8.) The configure.ac script no longer substitutes the
2503 relevant variables.
2504
2505 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2506 with length and offset values. This means that the output is different
2507 for different internal link sizes. Test 2 is skipped for link sizes
2508 other than 2 because of this, bypassing the problem. Unfortunately,
2509 there was also a test in test 3 (the locale tests) that used /B and
2510 failed for link sizes other than 2. Rather than cut the whole test out,
2511 I have added a new /Z option to pcretest that replaces the length and
2512 offset values with spaces. This is now used to make test 3 independent
2513 of link size. (Test 2 will be tidied up later.)
2514
2515 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2516 segmentation fault instead of returning the appropriate error message.
2517
2518 15. In multiline mode when the newline sequence was set to "any", the pattern
2519 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2520 This doesn't seem right; it now treats the CRLF combination as the line
2521 ending, and so does not match in that case. It's only a pattern such as ^$
2522 that would hit this one: something like ^ABC$ would have failed after \r
2523 and then tried again after \r\n.
2524
2525 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2526 in an attempt to make files that differ only in their line terminators
2527 compare equal. This works on Linux.
2528
2529 17. Under certain error circumstances pcregrep might try to free random memory
2530 as it exited. This is now fixed, thanks to valgrind.
2531
2532 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2533 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2534 was because its rules for how to advance for /g after matching an empty
2535 string at the end of a line did not allow for this case. They now check for
2536 it specially.
2537
2538 20. pcretest is supposed to handle patterns and data of any length, by
2539 extending its buffers when necessary. It was getting this wrong when the
2540 buffer for a data line had to be extended.
2541
2542 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2543 CRLF as a newline sequence.
2544
2545 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2546 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2547 I have nevertheless tidied it up.
2548
2549 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2550
2551 24. Added a man page for pcre-config.
2552
2553
2554 Version 7.0 19-Dec-06
2555 ---------------------
2556
2557 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2558 moving to gcc 4.1.1.
2559
2560 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2561 sys/time.h, which is documented as needed for this function. It doesn't
2562 seem to matter on Linux, but it showed up on some releases of OS X.
2563
2564 3. It seems that there are systems where bytes whose values are greater than
2565 127 match isprint() in the "C" locale. The "C" locale should be the
2566 default when a C program starts up. In most systems, only ASCII printing
2567 characters match isprint(). This difference caused the output from pcretest
2568 to vary, making some of the tests fail. I have changed pcretest so that:
2569
2570 (a) When it is outputting text in the compiled version of a pattern, bytes
2571 other than 32-126 are always shown as hex escapes.
2572
2573 (b) When it is outputting text that is a matched part of a subject string,
2574 it does the same, unless a different locale has been set for the match
2575 (using the /L modifier). In this case, it uses isprint() to decide.
2576
2577 4. Fixed a major bug that caused incorrect computation of the amount of memory
2578 required for a compiled pattern when options that changed within the
2579 pattern affected the logic of the preliminary scan that determines the
2580 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2581 that the computed length was too small. The symptoms of this bug were
2582 either the PCRE error "internal error: code overflow" from pcre_compile(),
2583 or a glibc crash with a message such as "pcretest: free(): invalid next
2584 size (fast)". Examples of patterns that provoked this bug (shown in
2585 pcretest format) are:
2586
2587 /(?-x: )/x
2588 /(?x)(?-x: \s*#\s*)/
2589 /((?i)[\x{c0}])/8
2590 /(?i:[\x{c0}])/8
2591
2592 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2593 is now done differently.
2594
2595 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2596 wrapper classes; (b) implement a new function in the C++ scanner that is
2597 more efficient than the old way of doing things because it avoids levels of
2598 recursion in the regex matching; (c) add a paragraph to the documentation
2599 for the FullMatch() function.
2600
2601 6. The escape sequence \n was being treated as whatever was defined as
2602 "newline". Not only was this contrary to the documentation, which states
2603 that \n is character 10 (hex 0A), but it also went horribly wrong when
2604 "newline" was defined as CRLF. This has been fixed.
2605
2606 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2607 was being set to -1 for the "end of line" case (supposedly a value that no
2608 character can have). Though this value is never used (the check for end of
2609 line is "zero bytes in current character"), it caused compiler complaints.
2610 I've changed it to 0xffffffff.
2611
2612 8. In pcre_version.c, the version string was being built by a sequence of
2613 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2614 string (as it is for production releases) called a macro with an empty
2615 argument. The C standard says the result of this is undefined. The gcc
2616 compiler treats it as an empty string (which was what was wanted) but it is
2617 reported that Visual C gives an error. The source has been hacked around to
2618 avoid this problem.
2619
2620 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2621 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2622 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2623 of them did).
2624
2625 10. Originally, pcretest opened its input and output without "b"; then I was
2626 told that "b" was needed in some environments, so it was added for release
2627 5.0 to both the input and output. (It makes no difference on Unix-like
2628 systems.) Later I was told that it is wrong for the input on Windows. I've
2629 now abstracted the modes into two macros, to make it easier to fiddle with
2630 them, and removed "b" from the input mode under Windows.
2631
2632 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2633
2634 12. Added -help and --help to pcretest as an official way of being reminded
2635 of the options.
2636
2637 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2638 and pcrecpp.cc because they annoy compilers at high warning levels.
2639
2640 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2641
2642 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2643 scripts are not C programs :-) and which was not noticed because it works
2644 on Linux.
2645
2646 16. pcretest is supposed to handle any length of pattern and data line (as one
2647 line or as a continued sequence of lines) by extending its input buffer if
2648 necessary. This feature was broken for very long pattern lines, leading to
2649 a string of junk being passed to pcre_compile() if the pattern was longer
2650 than about 50K.
2651
2652 17. I have done a major re-factoring of the way pcre_compile() computes the
2653 amount of memory needed for a compiled pattern. Previously, there was code
2654 that made a preliminary scan of the pattern in order to do this. That was
2655 OK when PCRE was new, but as the facilities have expanded, it has become
2656 harder and harder to keep it in step with the real compile phase, and there
2657 have been a number of bugs (see for example, 4 above). I have now found a
2658 cunning way of running the real compile function in a "fake" mode that
2659 enables it to compute how much memory it would need, while actually only
2660 ever using a few hundred bytes of working memory and without too many
2661 tests of the mode. This should make future maintenance and development
2662 easier. A side effect of this work is that the limit of 200 on the nesting
2663 depth of parentheses has been removed (though this was never a serious
2664 limitation, I suspect). However, there is a downside: pcre_compile() now
2665 runs more slowly than before (30% or more, depending on the pattern). I
2666 hope this isn't a big issue. There is no effect on runtime performance.
2667
2668 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2669 newline (only possible for the last line of a file) and it was a
2670 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2671
2672 19. Added additional timing features to pcretest. (1) The -tm option now times
2673 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2674 separate command line item, by a number that specifies the number of
2675 repeats to use when timing. The default is 50000; this gives better
2676 precision, but takes uncomfortably long for very large patterns.
2677
2678 20. Extended pcre_study() to be more clever in cases where a branch of a
2679 subpattern has no definite first character. For example, (a*|b*)[cd] would
2680 previously give no result from pcre_study(). Now it recognizes that the
2681 first character must be a, b, c, or d.
2682
2683 21. There was an incorrect error "recursive call could loop indefinitely" if
2684 a subpattern (or the entire pattern) that was being tested for matching an
2685 empty string contained only one non-empty item after a nested subpattern.
2686 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2687 incorrectly, because the \d was being skipped in the check.
2688
2689 22. The pcretest program now has a new pattern option /B and a command line
2690 option -b, which is equivalent to adding /B to every pattern. This causes
2691 it to show the compiled bytecode, without the additional information that
2692 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2693 is the same as /B/I).
2694
2695 23. A new optimization is now able automatically to treat some sequences such
2696 as a*b as a*+b. More specifically, if something simple (such as a character
2697 or a simple class like \d) has an unlimited quantifier, and is followed by
2698 something that cannot possibly match the quantified thing, the quantifier
2699 is automatically "possessified".
2700
2701 24. A recursive reference to a subpattern whose number was greater than 39
2702 went wrong under certain circumstances in UTF-8 mode. This bug could also
2703 have affected the operation of pcre_study().
2704
2705 25. Realized that a little bit of performance could be had by replacing
2706 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2707
2708 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2709
2710 27. Possessive quantifiers such as a++ were previously implemented by turning
2711 them into atomic groups such as ($>a+). Now they have their own opcodes,
2712 which improves performance. This includes the automatically created ones
2713 from 23 above.
2714
2715 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2716 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2717 the first matched character to be a colon. This applied both to named and
2718 numbered groups.
2719
2720 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2721
2722 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2723 building PCRE on Windows easier, so I have included it in the distribution.
2724
2725 31. There is now a check in pcretest against a ridiculously large number being
2726 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2727 loop, the loop is abandoned.
2728
2729 32. Forward references to subpatterns in conditions such as (?(2)...) where
2730 subpattern 2 is defined later cause pcre_compile() to search forwards in
2731 the pattern for the relevant set of parentheses. This search went wrong
2732 when there were unescaped parentheses in a character class, parentheses
2733 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2734
2735 33. "Subroutine" calls and backreferences were previously restricted to
2736 referencing subpatterns earlier in the regex. This restriction has now
2737 been removed.
2738
2739 34. Added a number of extra features that are going to be in Perl 5.10. On the
2740 whole, these are just syntactic alternatives for features that PCRE had
2741 previously implemented using the Python syntax or my own invention. The
2742 other formats are all retained for compatibility.
2743
2744 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2745 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2746 also .NET compatible.
2747
2748 (b) A recursion or subroutine call to a named group can now be defined as
2749 (?&name) as well as (?P>name).
2750
2751 (c) A backreference to a named group can now be defined as \k<name> or
2752 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2753 5.10, are also .NET compatible.
2754
2755 (d) A conditional reference to a named group can now use the syntax
2756 (?(<name>) or (?('name') as well as (?(name).
2757
2758 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2759 groups (named and numbered) that are never evaluated inline, but can be
2760 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2761 is always false. There may be only one alternative in such a group.
2762
2763 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2764 as the simple (?(R). The condition is true only if the most recent
2765 recursion is that of the given number or name. It does not search out
2766 through the entire recursion stack.
2767
2768 (g) The escape \gN or \g{N} has been added, where N is a positive or
2769 negative number, specifying an absolute or relative reference.
2770
2771 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2772 some "unreachable code" warnings.
2773
2774 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2775 things, this adds five new scripts.
2776
2777 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2778 There were also incompatibilities regarding the handling of \Q..\E inside
2779 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2780 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2781
2782 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2783 matches an empty string, and forcibly breaks the loop. There were bugs in
2784 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2785 against aaaa the result was just "a" rather than "aaaa", for example. Two
2786 separate and independent bugs (that affected different cases) have been
2787 fixed.
2788
2789 39. Refactored the code to abolish the use of different opcodes for small
2790 capturing bracket numbers. This is a tidy that I avoided doing when I
2791 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2792 The new approach is not only tidier, it makes it possible to reduce the
2793 memory needed to fix the previous bug (38).
2794
2795 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2796 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2797 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2798 mode.
2799
2800 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2801 report.
2802
2803 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2804 copy construction and assignment in the C++ wrapper.
2805
2806 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2807 couple of bugs that could have given wrong results in the "--newline=crlf"
2808 case.
2809
2810 44. Added a number of casts and did some reorganization of signed/unsigned int
2811 variables following suggestions from Dair Grant. Also renamed the variable
2812 "this" as "item" because it is a C++ keyword.
2813
2814 45. Arranged for dftables to add
2815
2816 #include "pcre_internal.h"
2817
2818 to pcre_chartables.c because without it, gcc 4.x may remove the array
2819 definition from the final binary if PCRE is built into a static library and
2820 dead code stripping is activated.
2821
2822 46. For an unanchored pattern, if a match attempt fails at the start of a
2823 newline sequence, and the newline setting is CRLF or ANY, and the next two
2824 characters are CRLF, advance by two characters instead of one.
2825
2826
2827 Version 6.7 04-Jul-06
2828 ---------------------
2829
2830 1. In order to handle tests when input lines are enormously long, pcretest has
2831 been re-factored so that it automatically extends its buffers when
2832 necessary. The code is crude, but this _is_ just a test program. The
2833 default size has been increased from 32K to 50K.
2834
2835 2. The code in pcre_study() was using the value of the re argument before
2836 testing it for NULL. (Of course, in any sensible call of the function, it
2837 won't be NULL.)
2838
2839 3. The memmove() emulation function in pcre_internal.h, which is used on
2840 systems that lack both memmove() and bcopy() - that is, hardly ever -
2841 was missing a "static" storage class specifier.
2842
2843 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2844 containing an extended class (one that cannot be represented by a bitmap
2845 because it contains high-valued characters or Unicode property items, e.g.
2846 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2847 pattern, but PCRE should not loop if you do not (it no longer does).
2848 [Detail: two cases were found: (a) a repeated subpattern containing an
2849 extended class; (b) a recursive reference to a subpattern that followed a
2850 previous extended class. It wasn't skipping over the extended class
2851 correctly when UTF-8 mode was not set.]
2852
2853 5. A negated single-character class was not being recognized as fixed-length
2854 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2855 compile error "lookbehind assertion is not fixed length".
2856
2857 6. The RunPerlTest auxiliary script was showing an unexpected difference
2858 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2859 write a Perl script that can interpret lines of an input file either as
2860 byte characters or as UTF-8, which is what "perltest" was being required to
2861 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2862 can't do is switch easily at run time between having the "use utf8;" pragma
2863 or not. In the end, I fudged it by using the RunPerlTest script to insert
2864 "use utf8;" explicitly for the UTF-8 tests.
2865
2866 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2867 the end of the subject string, contrary to the documentation and to what
2868 Perl does. This was true of both matching functions. Now it matches only at
2869 the start of the subject and immediately after *internal* newlines.
2870
2871 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2872 a pointer to an int instead of a pointer to an unsigned long int. This
2873 caused problems on 64-bit systems.
2874
2875 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2876 instance of the 'standard' template library not being so standard".
2877
2878 10. There was no check on the number of named subpatterns nor the maximum
2879 length of a subpattern name. The product of these values is used to compute
2880 the size of the memory block for a compiled pattern. By supplying a very
2881 long subpattern name and a large number of named subpatterns, the size
2882 computation could be caused to overflow. This is now prevented by limiting
2883 the length of names to 32 characters, and the number of named subpatterns
2884 to 10,000.
2885
2886 11. Subpatterns that are repeated with specific counts have to be replicated in
2887 the compiled pattern. The size of memory for this was computed from the
2888 length of the subpattern and the repeat count. The latter is limited to
2889 65535, but there was no limit on the former, meaning that integer overflow
2890 could in principle occur. The compiled length of a repeated subpattern is
2891 now limited to 30,000 bytes in order to prevent this.
2892
2893 12. Added the optional facility to have named substrings with the same name.
2894
2895 13. Added the ability to use a named substring as a condition, using the
2896 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2897 are numbers (not recommended). Forward references are permitted.
2898
2899 14. Added forward references in named backreferences (if you see what I mean).
2900
2901 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2902 pattern could run off the end of the subject. For example, the pattern
2903 "(?s)(.{1,5})"8 did this with the subject "ab".
2904
2905 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2906 PCRE_CASELESS was set when matching characters that were quantified with ?
2907 or *.
2908
2909 17. A character class other than a single negated character that had a minimum
2910 but no maximum quantifier - for example [ab]{6,} - was not handled
2911 correctly by pce_dfa_exec(). It would match only one character.
2912
2913 18. A valid (though odd) pattern that looked like a POSIX character
2914 class but used an invalid character after [ (for example [[,abc,]]) caused
2915 pcre_compile() to give the error "Failed: internal error: code overflow" or
2916 in some cases to crash with a glibc free() error. This could even happen if
2917 the pattern terminated after [[ but there just happened to be a sequence of
2918 letters, a binary zero, and a closing ] in the memory that followed.
2919
2920 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2921 over the years. Originally (before any Unicode support), just the bottom 8
2922 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2923 output from "man perlunicode" includes this:
2924
2925 The regular expression compiler produces polymorphic opcodes. That
2926 is, the pattern adapts to the data and automatically switches to
2927 the Unicode character scheme when presented with Unicode data--or
2928 instead uses a traditional byte scheme when presented with byte
2929 data.
2930
2931 Sadly, a wide octal escape does not cause a switch, and in a string with
2932 no other multibyte characters, these octal escapes are treated as before.
2933 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2934 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2935 Unicode string.
2936
2937 I have not perpetrated such confusion in PCRE. Up till now, it took just
2938 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2939 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2940 translate to the appropriate multibyte character.
2941
2942 29. Applied some refactoring to reduce the number of warnings from Microsoft
2943 and Borland compilers. This has included removing the fudge introduced
2944 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2945 a warning about an unused variable.
2946
2947 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2948 characters since release 4.0, because Perl (from release 5.004) does not.
2949 [Or at least, is documented not to: some releases seem to be in conflict
2950 with the documentation.] However, when a pattern was studied with
2951 pcre_study() and all its branches started with \s, PCRE still included VT
2952 as a possible starting character. Of course, this did no harm; it just
2953 caused an unnecessary match attempt.
2954
2955 22. Removed a now-redundant internal flag bit that recorded the fact that case
2956 dependency changed within the pattern. This was once needed for "required
2957 byte" processing, but is no longer used. This recovers a now-scarce options
2958 bit. Also moved the least significant internal flag bit to the most-
2959 significant bit of the word, which was not previously used (hangover from
2960 the days when it was an int rather than a uint) to free up another bit for
2961 the future.
2962
2963 23. Added support for CRLF line endings as well as CR and LF. As well as the
2964 default being selectable at build time, it can now be changed at runtime
2965 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2966 specify that it is scanning data with non-default line endings.
2967
2968 24. Changed the definition of CXXLINK to make it agree with the definition of
2969 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2970
2971 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2972 recursions. This makes a big different to stack usage for some patterns.
2973
2974 26. If a subpattern containing a named recursion or subroutine reference such
2975 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2976 the space required for the compiled pattern went wrong and gave too small a
2977 value. Depending on the environment, this could lead to "Failed: internal
2978 error: code overflow at offset 49" or "glibc detected double free or
2979 corruption" errors.
2980
2981 27. Applied patches from Google (a) to support the new newline modes and (b) to
2982 advance over multibyte UTF-8 characters in GlobalReplace.
2983
2984 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2985 difference for some implementation of PCRE in some Windows version.
2986
2987 29. Added some extra testing facilities to pcretest:
2988
2989 \q<number> in a data line sets the "match limit" value
2990 \Q<number> in a data line sets the "match recursion limt" value
2991 -S <number> sets the stack size, where <number> is in megabytes
2992
2993 The -S option isn't available for Windows.
2994
2995
2996 Version 6.6 06-Feb-06
2997 ---------------------
2998
2999 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3000 in pcreposix.h. I have copied the definition from pcre.h.
3001
3002 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3003 because pcre.h is no longer a built file.
3004
3005 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3006 not normally included in the compiled code.
3007
3008
3009 Version 6.5 01-Feb-06
3010 ---------------------
3011
3012 1. When using the partial match feature with pcre_dfa_exec(), it was not
3013 anchoring the second and subsequent partial matches at the new starting
3014 point. This could lead to incorrect results. For example, with the pattern
3015 /1234/, partially matching against "123" and then "a4" gave a match.
3016
3017 2. Changes to pcregrep:
3018
3019 (a) All non-match returns from pcre_exec() were being treated as failures
3020 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3021 error message is output. Some extra information is given for the
3022 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3023 probably the only errors that are likely to be caused by users (by
3024 specifying a regex that has nested indefinite repeats, for instance).
3025 If there are more than 20 of these errors, pcregrep is abandoned.
3026
3027 (b) A binary zero was treated as data while matching, but terminated the
3028 output line if it was written out. This has been fixed: binary zeroes
3029 are now no different to any other data bytes.
3030
3031 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3032 used to set a locale for matching. The --locale=xxxx long option has
3033 been added (no short equivalent) to specify a locale explicitly on the
3034 pcregrep command, overriding the environment variables.
3035
3036 (d) When -B was used with -n, some line numbers in the output were one less
3037 than they should have been.
3038
3039 (e) Added the -o (--only-matching) option.
3040
3041 (f) If -A or -C was used with -c (count only), some lines of context were
3042 accidentally printed for the final match.
3043
3044 (g) Added the -H (--with-filename) option.
3045
3046 (h) The combination of options -rh failed to suppress file names for files
3047 that were found from directory arguments.
3048
3049 (i) Added the -D (--devices) and -d (--directories) options.
3050
3051 (j) Added the -F (--fixed-strings) option.
3052
3053 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3054
3055 (l) Added the --colo(u)r option.
3056
3057 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3058 is not present by default.
3059
3060 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3061 items such as (?R) or (?1), when the recursion could match a number of
3062 alternatives. If it matched one of the alternatives, but subsequently,
3063 outside the recursion, there was a failure, the code tried to back up into
3064 the recursion. However, because of the way PCRE is implemented, this is not
3065 possible, and the result was an incorrect result from the match.
3066
3067 In order to prevent this happening, the specification of recursion has
3068 been changed so that all such subpatterns are automatically treated as
3069 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3070
3071 4. I had overlooked the fact that, in some locales, there are characters for
3072 which isalpha() is true but neither isupper() nor islower() are true. In
3073 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3074 and ordfeminine) are like this. This affected the treatment of \w and \W
3075 when they appeared in character classes, but not when they appeared outside
3076 a character class. The bit map for "word" characters is now created
3077 separately from the results of isalnum() instead of just taking it from the
3078 upper, lower, and digit maps. (Plus the underscore character, of course.)
3079
3080 5. The above bug also affected the handling of POSIX character classes such as
3081 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3082 permanent tables. Instead, the bit maps for such a class were previously
3083 created as the appropriate unions of the upper, lower, and digit bitmaps.
3084 Now they are created by subtraction from the [[:word:]] class, which has
3085 its own bitmap.
3086
3087 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3088 It is created by subtracting the vertical space characters (\x09, \x0a,
3089 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3090 subtraction was done in the overall bitmap for a character class, meaning
3091 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3092 be recognized. This bug has been fixed.
3093
3094 7. Patches from the folks at Google:
3095
3096 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3097 real life, but is still worth protecting against".
3098
3099 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3100 regular expressions".
3101
3102 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3103 have it.
3104
3105 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3106 "configure" and the latter not, in order to fix a problem somebody had
3107 with compiling the Arg class on HP-UX.
3108
3109 (e) Improve the error-handling of the C++ wrapper a little bit.
3110
3111 (f) New tests for checking recursion limiting.
3112
3113 8. The pcre_memmove() function, which is used only if the environment does not
3114 have a standard memmove() function (and is therefore rarely compiled),
3115 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3116 returning a result (though PCRE never actually uses the result).
3117
3118 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3119 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3120 returned instead of calling malloc() with an overflowing number that would
3121 most likely cause subsequent chaos.
3122
3123 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3124
3125 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3126 with this option is matched, the nmatch and pmatch options of regexec() are
3127 ignored.
3128
3129 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3130 provided in case anyone wants to the the POSIX interface with UTF-8
3131 strings.
3132
3133 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3134 C++ linking (needed for some HP-UX environments).
3135
3136 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3137 (unused parameter) and in the pcre_printint() function (omitted "default"
3138 switch label when the default is to do nothing).
3139
3140 15. Added some code to make it possible, when PCRE is compiled as a C++
3141 library, to replace subject pointers for pcre_exec() with a smart pointer
3142 class, thus making it possible to process discontinuous strings.
3143
3144 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3145 much the same function. They were added by different people who were trying
3146 to make PCRE easy to compile on non-Unix systems. It has been suggested
3147 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3148 for compiling on Windows systems. I have therefore replaced it with
3149 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3150 defaults to "extern" for C or "extern C" for C++, which works fine on
3151 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3152 SCOPE with something explicit in config.h. In addition:
3153
3154 (a) pcreposix.h still had just "extern" instead of either of these macros;
3155 I have replaced it with PCRE_DATA_SCOPE.
3156
3157 (b) Functions such as _pcre_xclass(), which are internal to the library,
3158 but external in the C sense, all had PCRE_EXPORT in their definitions.
3159 This is apparently wrong for the Windows case, so I have removed it.
3160 (It makes no difference on Unix-like systems.)
3161
3162 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3163 of recursive calls to match(). This is different to MATCH_LIMIT because
3164 that limits the total number of calls to match(), not all of which increase
3165 the depth of recursion. Limiting the recursion depth limits the amount of
3166 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3167 when PCRE is compiled, and changed at run time. A patch from Google adds
3168 this functionality to the C++ interface.
3169
3170 18. Changes to the handling of Unicode character properties:
3171
3172 (a) Updated the table to Unicode 4.1.0.
3173
3174 (b) Recognize characters that are not in the table as "Cn" (undefined).
3175
3176 (c) I revised the way the table is implemented to a much improved format
3177 which includes recognition of ranges. It now supports the ranges that
3178 are defined in UnicodeData.txt, and it also amalgamates other
3179 characters into ranges. This has reduced the number of entries in the
3180 table from around 16,000 to around 3,000, thus reducing its size
3181 considerably. I realized I did not need to use a tree structure after
3182 all - a binary chop search is just as efficient. Having reduced the
3183 number of entries, I extended their size from 6 bytes to 8 bytes to
3184 allow for more data.
3185
3186 (d) Added support for Unicode script names via properties such as \p{Han}.
3187
3188 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3189 matching that character.
3190
3191 20. When matching a repeated Unicode property with a minimum greater than zero,
3192 (for example \pL{2,}), PCRE could look past the end of the subject if it
3193 reached it while seeking the minimum number of characters. This could
3194 happen only if some of the characters were more than one byte long, because
3195 there is a check for at least the minimum number of bytes.
3196
3197 21. Refactored the implementation of \p and \P so as to be more general, to
3198 allow for more different types of property in future. This has changed the
3199 compiled form incompatibly. Anybody with saved compiled patterns that use
3200 \p or \P will have to recompile them.
3201
3202 22. Added "Any" and "L&" to the supported property types.
3203
3204 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3205 but give a compile time error if the value is greater than 0xff.
3206
3207 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3208 accidentally not being installed or uninstalled.
3209
3210 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3211 made were to insert the current release number. This seemed silly, because
3212 it made things harder for people building PCRE on systems that don't run
3213 "configure". I have turned pcre.h into a distributed file, no longer built
3214 by "configure", with the version identification directly included. There is
3215 no longer a pcre.h.in file.
3216
3217 However, this change necessitated a change to the pcre-config script as
3218 well. It is built from pcre-config.in, and one of the substitutions was the
3219 release number. I have updated configure.ac so that ./configure now finds
3220 the release number by grepping pcre.h.
3221
3222 26. Added the ability to run the tests under valgrind.
3223
3224
3225 Version 6.4 05-Sep-05
3226 ---------------------
3227
3228 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3229 "--" to be printed when multiple files were scanned, even when none of the
3230 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3231 consider it to be a bug, and have restored the previous behaviour.
3232
3233 2. A couple of code tidies to get rid of compiler warnings.
3234
3235 3. The pcretest program used to cheat by referring to symbols in the library
3236 whose names begin with _pcre_. These are internal symbols that are not
3237 really supposed to be visible externally, and in some environments it is
3238 possible to suppress them. The cheating is now confined to including
3239 certain files from the library's source, which is a bit cleaner.
3240
3241 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3242 file's purpose clearer.
3243
3244 5. Reorganized pcre_ucp_findchar().
3245
3246
3247 Version 6.3 15-Aug-05
3248 ---------------------
3249
3250 1. The file libpcre.pc.in did not have general read permission in the tarball.
3251
3252 2. There were some problems when building without C++ support:
3253
3254 (a) If C++ support was not built, "make install" and "make test" still
3255 tried to test it.
3256
3257 (b) There were problems when the value of CXX was explicitly set. Some
3258 changes have been made to try to fix these, and ...
3259
3260 (c) --disable-cpp can now be used to explicitly disable C++ support.
3261
3262 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3263 backslash in a target when C++ was disabled. This confuses some
3264 versions of "make", apparently. Using an intermediate variable solves
3265 this. (Same for CPP_LOBJ.)
3266
3267 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3268 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3269 necessary on certain architectures.
3270
3271 4. Added a setting of -export-symbols-regex to the link command to remove
3272 those symbols that are exported in the C sense, but actually are local
3273 within the library, and not documented. Their names all begin with
3274 "_pcre_". This is not a perfect job, because (a) we have to except some
3275 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3276 available (and never for static libraries). I have made a note to try to
3277 find a way round (a) in the future.
3278
3279
3280 Version 6.2 01-Aug-05
3281 ---------------------
3282
3283 1. There was no test for integer overflow of quantifier values. A construction
3284 such as {1111111111111111} would give undefined results. What is worse, if
3285 a minimum quantifier for a parenthesized subpattern overflowed and became
3286 negative, the calculation of the memory size went wrong. This could have
3287 led to memory overwriting.
3288
3289 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3290
3291 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3292 operating environments where this matters.
3293
3294 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3295 PCRE options from within the C++ wrapper.
3296
3297 5. Named capturing subpatterns were not being correctly counted when a pattern
3298 was compiled. This caused two problems: (a) If there were more than 100
3299 such subpatterns, the calculation of the memory needed for the whole
3300 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3301 back references of the form \12, where the number was greater than 9, were
3302 not recognized as back references, even though there were sufficient
3303 previous subpatterns.
3304
3305 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3306 versions of gcc, e.g. 2.95.4.
3307
3308
3309 Version 6.1 21-Jun-05
3310 ---------------------
3311
3312 1. There was one reference to the variable "posix" in pcretest.c that was not
3313 surrounded by "#if !defined NOPOSIX".
3314
3315 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3316 the cross-check on the old pcre_info() function, for the benefit of the
3317 cut-down version of PCRE that is currently imported into Exim.
3318
3319 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3320 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3321 patterns that start (?i) but I don't think that matters. The use of (?i) is
3322 just an example; this all applies to the other options as well.
3323
3324 4. Since libtool seems to echo the compile commands it is issuing, the output
3325 from "make" can be reduced a bit by putting "@" in front of each libtool
3326 compile command.
3327
3328 5. Patch from the folks at Google for configure.in to be a bit more thorough
3329 in checking for a suitable C++ installation before trying to compile the
3330 C++ stuff. This should fix a reported problem when a compiler was present,
3331 but no suitable headers.
3332
3333 6. The man pages all had just "PCRE" as their title. I have changed them to
3334 be the relevant file name. I have also arranged that these names are
3335 retained in the file doc/pcre.txt, which is a concatenation in text format
3336 of all the man pages except the little individual ones for each function.
3337
3338 7. The NON-UNIX-USE file had not been updated for the different set of source
3339 files that come with release 6. I also added a few comments about the C++
3340 wrapper.
3341
3342
3343 Version 6.0 07-Jun-05
3344 ---------------------
3345
3346 1. Some minor internal re-organization to help with my DFA experiments.
3347
3348 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3349 didn't matter for the library itself when fully configured, but did matter
3350 when compiling without UCP support, or within Exim, where the ucp files are
3351 not imported.
3352
3353 3. Refactoring of the library code to split up the various functions into
3354 different source modules. The addition of the new DFA matching code (see
3355 below) to a single monolithic source would have made it really too
3356 unwieldy, quite apart from causing all the code to be include in a
3357 statically linked application, when only some functions are used. This is
3358 relevant even without the DFA addition now that patterns can be compiled in
3359 one application and matched in another.
3360
3361 The downside of splitting up is that there have to be some external
3362 functions and data tables that are used internally in different modules of
3363 the library but which are not part of the API. These have all had their
3364 names changed to start with "_pcre_" so that they are unlikely to clash
3365 with other external names.
3366
3367 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3368 a different (DFA) algorithm. Although it is slower than the original
3369 function, it does have some advantages for certain types of matching
3370 problem.
3371
3372 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3373 including restarting after a partial match.
3374
3375 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3376 defined when compiling for Windows was sent to me. I have put it into the
3377 code, though I have no means of testing or verifying it.
3378
3379 7. Added the pcre_refcount() auxiliary function.
3380
3381 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3382 match before or at the first newline in the subject string. In pcretest,
3383 the /f option on a pattern can be used to set this.
3384
3385 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3386 would behave wrongly. This has been present in PCRE since release 4.0.
3387
3388 10. A number of changes to the pcregrep command:
3389
3390 (a) Refactored how -x works; insert ^(...)$ instead of setting
3391 PCRE_ANCHORED and checking the length, in preparation for adding
3392 something similar for -w.
3393
3394 (b) Added the -w (match as a word) option.
3395
3396 (c) Refactored the way lines are read and buffered so as to have more
3397 than one at a time available.
3398
3399 (d) Implemented a pcregrep test script.
3400
3401 (e) Added the -M (multiline match) option. This allows patterns to match
3402 over several lines of the subject. The buffering ensures that at least
3403 8K, or the rest of the document (whichever is the shorter) is available
3404 for matching (and similarly the previous 8K for lookbehind assertions).
3405
3406 (f) Changed the --help output so that it now says
3407
3408 -w, --word-regex(p)
3409
3410 instead of two lines, one with "regex" and the other with "regexp"
3411 because that confused at least one person since the short forms are the
3412 same. (This required a bit of code, as the output is generated
3413 automatically from a table. It wasn't just a text change.)
3414
3415 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3416 option but starts with a hyphen. Could be a pattern or a path name
3417 starting with a hyphen, for instance.
3418
3419 (h) "-" can be given as a file name to represent stdin.
3420
3421 (i) When file names are being printed, "(standard input)" is used for
3422 the standard input, for compatibility with GNU grep. Previously
3423 "<stdin>" was used.
3424
3425 (j) The option --label=xxx can be used to supply a name to be used for
3426 stdin when file names are being printed. There is no short form.
3427
3428 (k) Re-factored the options decoding logic because we are going to add
3429 two more options that take data. Such options can now be given in four
3430 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3431
3432 (l) Added the -A, -B, and -C options for requesting that lines of context
3433 around matches be printed.
3434
3435 (m) Added the -L option to print the names of files that do not contain
3436 any matching lines, that is, the complement of -l.
3437
3438 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3439 continue to scan other files.
3440
3441 (o) The -s option was incorrectly implemented. For compatibility with other
3442 greps, it now suppresses the error message for a non-existent or non-
3443 accessible file (but not the return code). There is a new option called
3444 -q that suppresses the output of matching lines, which was what -s was
3445 previously doing.
3446
3447 (p) Added --include and --exclude options to specify files for inclusion
3448 and exclusion when recursing.
3449
3450 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3451 Hopefully, it now does.
3452
3453 12. Missing cast in pcre_study().
3454
3455 13. Added an "uninstall" target to the makefile.
3456
3457 14. Replaced "extern" in the function prototypes in Makefile.in with
3458 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3459 world, but is set differently for Windows.
3460
3461 15. Added a second compiling function called pcre_compile2(). The only
3462 difference is that it has an extra argument, which is a pointer to an
3463 integer error code. When there is a compile-time failure, this is set
3464 non-zero, in addition to the error test pointer being set to point to an
3465 error message. The new argument may be NULL if no error number is required
3466 (but then you may as well call pcre_compile(), which is now just a
3467 wrapper). This facility is provided because some applications need a
3468 numeric error indication, but it has also enabled me to tidy up the way
3469 compile-time errors are handled in the POSIX wrapper.
3470
3471 16. Added VPATH=.libs to the makefile; this should help when building with one
3472 prefix path and installing with another. (Or so I'm told by someone who
3473 knows more about this stuff than I do.)
3474
3475 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3476 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3477 match everything, including newlines. This is not POSIX-compatible, but
3478 somebody wanted the feature. From pcretest it can be activated by using
3479 both the P and the s flags.
3480
3481 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3482
3483 19. libpcre.pc was being incorrectly installed as executable.
3484
3485 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3486 it now also looks for '\r' so that it will work unmodified on Windows.
3487
3488 21. Added Google's contributed C++ wrapper to the distribution.
3489
3490 22. Added some untidy missing memory free() calls in pcretest, to keep
3491 Electric Fence happy when testing.
3492
3493
3494
3495 Version 5.0 13-Sep-04
3496 ---------------------
3497
3498 1. Internal change: literal characters are no longer packed up into items
3499 containing multiple characters in a single byte-string. Each character
3500 is now matched using a separate opcode. However, there may be more than one
3501 byte in the character in UTF-8 mode.
3502
3503 2. The pcre_callout_block structure has two new fields: pattern_position and
3504 next_item_length. These contain the offset in the pattern to the next match
3505 item, and its length, respectively.
3506
3507 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3508 insertion of callouts before each pattern item. Added the /C option to
3509 pcretest to make use of this.
3510
3511 4. On the advice of a Windows user, the lines
3512
3513 #if defined(_WIN32) || defined(WIN32)
3514 _setmode( _fileno( stdout ), 0x8000 );
3515 #endif /* defined(_WIN32) || defined(WIN32) */
3516
3517 have been added to the source of pcretest. This apparently does useful
3518 magic in relation to line terminators.
3519
3520 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3521 for the benefit of those environments where the "b" makes a difference.
3522
3523 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3524 to know about it. I have put a hack into configure.in that adds in code
3525 to set GCC=yes if CC=icc. This seems to end up at a point in the
3526 generated configure script that is early enough to affect the setting of
3527 compiler options, which is what is needed, but I have no means of testing
3528 whether it really works. (The user who reported this had patched the
3529 generated configure script, which of course I cannot do.)
3530
3531 LATER: After change 22 below (new libtool files), the configure script
3532 seems to know about icc (and also ecc). Therefore, I have commented out
3533 this hack in configure.in.
3534
3535 7. Added support for pkg-config (2 patches were sent in).
3536
3537 8. Negated POSIX character classes that used a combination of internal tables
3538 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3539 [[:^ascii]]. Typically, they would match almost any characters. The other
3540 POSIX classes were not broken in this way.
3541
3542 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3543 to find the match, as PCRE was deluded into thinking that the match had to
3544 start at the start point or following a newline. The same bug applied to
3545 patterns with negative forward assertions or any backward assertions
3546 preceding ".*" at the start, unless the pattern required a fixed first
3547 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3548
3549 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3550 starting at the last subject character, bytes beyond the end of the subject
3551 string were read.
3552
3553 11. Renamed the variable "class" as "classbits" to make life easier for C++
3554 users. (Previously there was a macro definition, but it apparently wasn't
3555 enough.)
3556
3557 12. Added the new field "tables" to the extra data so that tables can be passed
3558 in at exec time, or the internal tables can be re-selected. This allows
3559 a compiled regex to be saved and re-used at a later time by a different
3560 program that might have everything at different addresses.
3561
3562 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3563 -R library as well as a -L library.
3564
3565 14. The debugging options of pcretest (-d on the command line or D on a
3566 pattern) showed incorrect output for anything following an extended class
3567 that contained multibyte characters and which was followed by a quantifier.
3568
3569 15. Added optional support for general category Unicode character properties
3570 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3571 support. It adds about 90K to the size of the library. The meanings of the
3572 inbuilt class escapes such as \d and \s have NOT been changed.
3573
3574 16. Updated pcredemo.c to include calls to free() to release the memory for the
3575 compiled pattern.
3576
3577 17. The generated file chartables.c was being created in the source directory
3578 instead of in the building directory. This caused the build to fail if the
3579 source directory was different from the building directory, and was
3580 read-only.
3581
3582 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3583 file. No doubt somebody will tell me if they don't make sense... Also added
3584 Dan Mooney's comments about building on OpenVMS.
3585
3586 19. Added support for partial matching via the PCRE_PARTIAL option for
3587 pcre_exec() and the \P data escape in pcretest.
3588
3589 20. Extended pcretest with 3 new pattern features:
3590
3591 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3592 write the compiled pattern to the file whose name is "rest-of-line".
3593 This is a straight binary dump of the data, with the saved pointer to
3594 the character tables forced to be NULL. The study data, if any, is
3595 written too. After writing, pcretest reads a new pattern.
3596
3597 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3598 compiled pattern from the given file. There must not be any
3599 occurrences of "<" in the file name (pretty unlikely); if there are,
3600 pcretest will instead treat the initial "<" as a pattern delimiter.
3601 After reading in the pattern, pcretest goes on to read data lines as
3602 usual.
3603
3604 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3605 and 16-bit fields in a compiled pattern, to simulate a pattern that
3606 was compiled on a host of opposite endianness.
3607
3608 21. The pcre-exec() function can now cope with patterns that were compiled on
3609 hosts of opposite endianness, with this restriction:
3610
3611 As for any compiled expression that is saved and used later, the tables
3612 pointer field cannot be preserved; the extra_data field in the arguments
3613 to pcre_exec() should be used to pass in a tables address if a value
3614 other than the default internal tables were used at compile time.
3615
3616 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3617 now diagnosed as an error. Previously, most of the time, a negative number
3618 would have been treated as zero, but if in addition "ovector" was passed as
3619 NULL, a crash could occur.
3620
3621 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3622 new versions from the libtool 1.5 distribution (the last one is a copy of
3623 a file called libtool.m4). This seems to have fixed the need to patch
3624 "configure" to support Darwin 1.3 (which I used to do). However, I still
3625 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3626 workstation).
3627
3628 24. Changed the PCRE licence to be the more standard "BSD" licence.
3629
3630
3631 Version 4.5 01-Dec-03
3632 ---------------------
3633
3634 1. There has been some re-arrangement of the code for the match() function so
3635 that it can be compiled in a version that does not call itself recursively.
3636 Instead, it keeps those local variables that need separate instances for
3637 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3638 needs to "recurse". Keeping track of where control must go is done by means
3639 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3640 hide most of the details from the main code, and operates only if
3641 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3642 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3643 operating.
3644
3645 To make it easier for callers to provide specially tailored get/free
3646 functions for this usage, two new functions, pcre_stack_malloc, and
3647 pcre_stack_free, are used. They are always called in strict stacking order,
3648 and the size of block requested is always the same.
3649
3650 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3651 PCRE has been compiled to use the stack or the heap for recursion. The
3652 -C option of pcretest uses this to show which version is compiled.
3653
3654 A new data escape \S, is added to pcretest; it causes the amounts of store
3655 obtained and freed by both kinds of malloc/free at match time to be added
3656 to the output.
3657
3658 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3659 what's available on my current Linux desktop machine.
3660
3661 3. When matching a UTF-8 string, the test for a valid string at the start has
3662 been extended. If start_offset is not zero, PCRE now checks that it points
3663 to a byte that is the start of a UTF-8 character. If not, it returns
3664 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3665 this is necessary because there may be backward assertions in the pattern.
3666 When matching the same subject several times, it may save resources to use
3667 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3668
3669 4. The code for checking the validity of UTF-8 strings has been tightened so
3670 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3671 containing "overlong sequences".
3672
3673 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3674 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3675 should have been "&", but it just so happened that all the cases this let
3676 through by mistake were picked up later in the function.
3677
3678 6. I had used a variable called "isblank" - this is a C99 function, causing
3679 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3680
3681 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3682 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3683 so the version is shown; (c) stop "make test" echoing "./RunTest".
3684
3685 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3686
3687 9. The prototype for memmove() for systems that don't have it was using
3688 size_t, but the inclusion of the header that defines size_t was later. I've
3689 moved the #includes for the C headers earlier to avoid this.
3690
3691 10. Added some adjustments to the code to make it easier to compiler on certain
3692 special systems:
3693
3694 (a) Some "const" qualifiers were missing.
3695 (b) Added the macro EXPORT before all exported functions; by default this
3696 is defined to be empty.
3697 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3698 that it reads its output file name as an argument instead of writing
3699 to the standard output and assuming this can be redirected.
3700
3701 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3702 class containing characters with values greater than 255, PCRE compilation
3703 went into a loop.
3704
3705 12. A recursive reference to a subpattern that was within another subpattern
3706 that had a minimum quantifier of zero caused PCRE to crash. For example,
3707 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3708 recursion. If the recursively-called subpattern itself had a zero repeat,
3709 that was OK.
3710
3711 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3712 buffer into which it was copied (for escape processing) was still set at
3713 1024, so long lines caused crashes.
3714
3715 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3716 "internal error: code overflow...". This applied to any character class
3717 that was followed by a possessive quantifier.
3718
3719 15. Modified the Makefile to add libpcre.la as a prerequisite for
3720 libpcreposix.la because I was told this is needed for a parallel build to
3721 work.
3722
3723 16. If a pattern that contained .* following optional items at the start was
3724 studied, the wrong optimizing data was generated, leading to matching
3725 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3726 matching string must start with a or b or c. The correct conclusion for
3727 this pattern is that a match can start with any character.
3728
3729
3730 Version 4.4 13-Aug-03
3731 ---------------------
3732
3733 1. In UTF-8 mode, a character class containing characters with values between
3734 127 and 255 was not handled correctly if the compiled pattern was studied.
3735 In fixing this, I have also improved the studying algorithm for such
3736 classes (slightly).
3737
3738 2. Three internal functions had redundant arguments passed to them. Removal
3739 might give a very teeny performance improvement.
3740
3741 3. Documentation bug: the value of the capture_top field in a callout is *one
3742 more than* the number of the hightest numbered captured substring.
3743
3744 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3745 in incorrectly linking with a previously installed version. They now link
3746 explicitly with libpcre.la.
3747
3748 5. configure.in no longer needs to recognize Cygwin specially.
3749
3750 6. A problem in pcre.in for Windows platforms is fixed.
3751
3752 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3753 pcretest, it used to include the size of the study block as part of its
3754 output. Unfortunately, the structure contains a field that has a different
3755 size on different hardware architectures. This meant that the tests that
3756 showed this size failed. As the block is currently always of a fixed size,
3757 this information isn't actually particularly useful in pcretest output, so
3758 I have just removed it.
3759
3760 8. Three pre-processor statements accidentally did not start in column 1.
3761 Sadly, there are *still* compilers around that complain, even though
3762 standard C has not required this for well over a decade. Sigh.
3763
3764 9. In pcretest, the code for checking callouts passed small integers in the
3765 callout_data field, which is a void * field. However, some picky compilers
3766 complained about the casts involved for this on 64-bit systems. Now
3767 pcretest passes the address of the small integer instead, which should get
3768 rid of the warnings.
3769
3770 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3771 both compile and run time, and gives an error if an invalid UTF-8 sequence
3772 is found. There is a option for disabling this check in cases where the
3773 string is known to be correct and/or the maximum performance is wanted.
3774
3775 11. In response to a bug report, I changed one line in Makefile.in from
3776
3777 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3778 to
3779 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3780
3781 to look similar to other lines, but I have no way of telling whether this
3782 is the right thing to do, as I do not use Windows. No doubt I'll get told
3783 if it's wrong...
3784
3785
3786 Version 4.3 21-May-03
3787 ---------------------
3788
3789 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3790 Makefile.
3791
3792 2. Some refactoring to improve the quality of the code:
3793
3794 (i) The utf8_table... variables are now declared "const".
3795
3796 (ii) The code for \cx, which used the "case flipping" table to upper case
3797 lower case letters, now just substracts 32. This is ASCII-specific,
3798 but the whole concept of \cx is ASCII-specific, so it seems
3799 reasonable.
3800
3801 (iii) PCRE was using its character types table to recognize decimal and
3802 hexadecimal digits in the pattern. This is silly, because it handles
3803 only 0-9, a-f, and A-F, but the character types table is locale-
3804 specific, which means strange things might happen. A private
3805 table is now used for this - though it costs 256 bytes, a table is
3806 much faster than multiple explicit tests. Of course, the standard
3807 character types table is still used for matching digits in subject
3808 strings against \d.
3809
3810 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3811 ending in _t are). So I've renamed it as ESC_tee.
3812
3813 3. The first argument for regexec() in the POSIX wrapper should have been
3814 defined as "const".
3815
3816 4. Changed pcretest to use malloc() for its buffers so that they can be
3817 Electric Fenced for debugging.
3818
3819 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3820 to read one or more bytes before the start of the subject string. Often this
3821 had no effect on PCRE's behaviour, but in some circumstances it could
3822 provoke a segmentation fault.
3823
3824 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3825 to try to read one or more bytes before the start of the subject string.
3826
3827 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3828 UTF-8 support could misbehave in various ways if the subject string
3829 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3830 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3831 back over UTF-8 characters.)
3832
3833
3834 Version 4.2 14-Apr-03
3835 ---------------------
3836
3837 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3838
3839 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3840 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3841 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3842 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3843 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3844 and BUILD_EXEEXT
3845 Note: automatic setting of the BUILD variables is not yet working
3846 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3847 compile-time but not at link-time
3848 [LINK]: use for linking executables only
3849 make different versions for Windows and non-Windows
3850 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3851 libraries
3852 [LINK_FOR_BUILD]: new variable
3853 [OBJEXT]: use throughout
3854 [EXEEXT]: use throughout
3855 <winshared>: new target
3856 <wininstall>: new target
3857 <dftables.o>: use native compiler
3858 <dftables>: use native linker
3859 <install>: handle Windows platform correctly
3860 <clean>: ditto
3861 <check>: ditto
3862 copy DLL to top builddir before testing
3863
3864 As part of these changes, -no-undefined was removed again. This was reported
3865 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3866 in any case.
3867
3868 3. Some tidies to get rid of compiler warnings:
3869
3870 . In the match_data structure, match_limit was an unsigned long int, whereas
3871 match_call_count was an int. I've made them both unsigned long ints.
3872
3873 . In pcretest the fact that a const uschar * doesn't automatically cast to
3874 a void * provoked a warning.
3875
3876 . Turning on some more compiler warnings threw up some "shadow" variables
3877 and a few more missing casts.
3878
3879 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3880 option, a class that contained a single character with a value between 128
3881 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3882
3883 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3884 option, a class that contained several characters, but with at least one
3885 whose value was between 128 and 255 caused PCRE to crash.
3886
3887
3888 Version 4.1 12-Mar-03
3889 ---------------------
3890
3891 1. Compiling with gcc -pedantic found a couple of places where casts were
3892 needed, and a string in dftables.c that was longer than standard compilers are
3893 required to support.
3894
3895 2. Compiling with Sun's compiler found a few more places where the code could
3896 be tidied up in order to avoid warnings.
3897
3898 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3899 first of these names is deprecated in the latest Autoconf in favour of the name
3900 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3901 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3902 analogy I have changed it to CFLAGS_FOR_BUILD.
3903
3904 4. Added -no-undefined to the linking command in the Makefile, because this is
3905 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3906 linking step for the pcreposix library.
3907
3908 5. PCRE was failing to diagnose the case of two named groups with the same
3909 name.
3910
3911 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3912 literal character that is needed in the subject for a match, and scans along to
3913 ensure that it is present before embarking on the full matching process. This
3914 saves time in cases of nested unlimited repeats that are never going to match.
3915 Problem: the scan can take a lot of time if the subject is very long (e.g.
3916 megabytes), thus penalizing straightforward matches. It is now done only if the
3917 amount of subject to be scanned is less than 1000 bytes.
3918
3919 7. A lesser problem with the same optimization is that it was recording the
3920 first character of an anchored pattern as "needed", thus provoking a search
3921 right along the subject, even when the first match of the pattern was going to
3922 fail. The "needed" character is now not set for anchored patterns, unless it
3923 follows something in the pattern that is of non-fixed length. Thus, it still
3924 fulfils its original purpose of finding quick non-matches in cases of nested
3925 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3926
3927
3928 Version 4.0 17-Feb-03
3929 ---------------------
3930
3931 1. If a comment in an extended regex that started immediately after a meta-item
3932 extended to the end of string, PCRE compiled incorrect data. This could lead to
3933 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3934
3935 2. Moved to autoconf 2.53 and libtool 1.4.2.
3936
3937 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3938 the special perltest8 script is no longer needed - all the tests can be run
3939 from a single perltest script.
3940
3941 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3942 by \s. It has now been removed in PCRE. This means it isn't recognized as
3943 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3944 class [:space:] *does* include VT, thereby creating a mess.
3945
3946 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3947 space and tab.
3948
3949 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3950 its new features into the main test script, reducing the number of scripts.
3951
3952 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3953 were backward compatible, and made the (?i) apply to the whole pattern, as if
3954 /i were given. Now it behaves more logically, and applies the option setting
3955 only to what follows. PCRE has been changed to follow suit. However, if it
3956 finds options settings right at the start of the pattern, it extracts them into
3957 the global options, as before. Thus, they show up in the info data.
3958
3959 8. Added support for the \Q...\E escape sequence. Characters in between are
3960 treated as literals. This is slightly different from Perl in that $ and @ are
3961 also handled as literals inside the quotes. In Perl, they will cause variable
3962 interpolation. Note the following examples:
3963
3964 Pattern PCRE matches Perl matches
3965
3966 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3967 \Qabc\$xyz\E abc\$xyz abc\$xyz
3968 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3969
3970 For compatibility with Perl, \Q...\E sequences are recognized inside character
3971 classes as well as outside them.
3972
3973 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3974 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3975 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3976 signed/unsigned warnings.
3977
3978 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3979 option for pcretest, so I've replaced it by a simple function that does just
3980 that job.
3981
3982 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3983 "pcregrep -".
3984
3985 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3986 Java package. This provides some syntactic sugar for simple cases of what my
3987 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3988 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3989 item, you can use this simplified notation. Note that only makes sense with
3990 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3991 greediness, whatever the setting of the PCRE_UNGREEDY option.
3992
3993 13. A change of greediness default within a pattern was not taking effect at
3994 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3995 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3996 was abstracted outside.
3997
3998 14. PCRE now supports the \G assertion. It is true when the current matching
3999 position is at the start point of the match. This differs from \A when the
4000 starting offset is non-zero. Used with the /g option of pcretest (or similar
4001 code), it works in the same way as it does for Perl's /g option. If all
4002 alternatives of a regex begin with \G, the expression is anchored to the start
4003 match position, and the "anchored" flag is set in the compiled expression.
4004
4005 15. Some bugs concerning the handling of certain option changes within patterns
4006 have been fixed. These applied to options other than (?ims). For example,
4007 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4008 been the other way round. Some of this was related to change 7 above.
4009
4010 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4011 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4012 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4013 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4014
4015 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4016 mode. Unlike ".", it always matches newline, whatever the setting of
4017 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4018 assertions. Perl allows it, but it doesn't (in general) work because it can't
4019 calculate the length of the lookbehind. At least, that's the case for Perl
4020 5.8.0 - I've been told they are going to document that it doesn't work in
4021 future.
4022
4023 18. Added an error diagnosis for escapes that PCRE does not support: these are
4024 \L, \l, \N, \P, \p, \U, \u, and \X.
4025
4026 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4027 reading past the end of the pattern in cases such as /[abcd/.
4028
4029 20. PCRE was getting more memory than necessary for patterns with classes that
4030 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4031
4032 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4033 compiling PCRE for use with Virtual Pascal.
4034
4035 22. Small fix to the Makefile to make it work properly if the build is done
4036 outside the source tree.
4037
4038 23. Added a new extension: a condition to go with recursion. If a conditional
4039 subpattern starts with (?(R) the "true" branch is used if recursion has
4040 happened, whereas the "false" branch is used only at the top level.
4041
4042 24. When there was a very long string of literal characters (over 255 bytes
4043 without UTF support, over 250 bytes with UTF support), the computation of how
4044 much memory was required could be incorrect, leading to segfaults or other
4045 strange effects.
4046
4047 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4048 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4049 there was a subsequent back reference to those brackets. This meant that, for
4050 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4051 possible to check for precisely this case. All we can do is abandon the
4052 optimization if .* occurs inside capturing brackets when there are any back
4053 references whatsoever. (See below for a better fix that came later.)
4054
4055 26. The handling of the optimization for finding the first character of a
4056 non-anchored pattern, and for finding a character that is required later in the
4057 match were failing in some cases. This didn't break the matching; it just
4058 failed to optimize when it could. The way this is done has been re-implemented.
4059
4060 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4061
4062 28. Added a new feature that provides some of the functionality that Perl
4063 provides with (?{...}). The facility is termed a "callout". The way it is done
4064 in PCRE is for the caller to provide an optional function, by setting
4065 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4066 global variable. By default it is unset, which disables all calling out. To get
4067 the function called, the regex must include (?C) at appropriate points. This
4068 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4069 This provides a means of identifying different callout points. When PCRE
4070 reaches such a point in the regex, if pcre_callout has been set, the external
4071 function is called. It is provided with data in a structure called
4072 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4073 matching continues; if it returns a non-zero value, the match at the current
4074 point fails. However, backtracking will occur if possible. [This was changed
4075 later and other features added - see item 49 below.]
4076
4077 29. pcretest is upgraded to test the callout functionality. It provides a
4078 callout function that displays information. By default, it shows the start of
4079 the match and the current position in the text. There are some new data escapes
4080 to vary what happens:
4081
4082 \C+ in addition, show current contents of captured substrings
4083 \C- do not supply a callout function
4084 \C!n return 1 when callout number n is reached
4085 \C!n!m return 1 when callout number n is reached for the mth time
4086
4087 30. If pcregrep was called with the -l option and just a single file name, it
4088 output "<stdin>" if a match was found, instead of the file name.
4089
4090 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4091 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4092 pcre_exec(). This saves a malloc/free per call. The default value of
4093 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4094 when configuring.
4095
4096 32. The default maximum size of a compiled pattern is 64K. There have been a
4097 few cases of people hitting this limit. The code now uses macros to handle the
4098 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4099 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4100 configuring. Tests 2 and 5 work only with 2-byte links because they output
4101 debugging information about compiled patterns.
4102
4103 33. Internal code re-arrangements:
4104
4105 (a) Moved the debugging function for printing out a compiled regex into
4106 its own source file (printint.c) and used #include to pull it into
4107 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4108 separate copies.
4109
4110 (b) Defined the list of op-code names for debugging as a macro in
4111 internal.h so that it is next to the definition of the opcodes.
4112
4113 (c) Defined a table of op-code lengths for simpler skipping along compiled
4114 code. This is again a macro in internal.h so that it is next to the
4115 definition of the opcodes.
4116
4117 34. Added support for recursive calls to individual subpatterns, along the
4118 lines of Robin Houston's patch (but implemented somewhat differently).
4119
4120 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4121 allow it to read and process whole directories in Win32. This code was
4122 contributed by Lionel Fourquaux; it has not been tested by me.
4123
4124 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4125 used to name a group. Names consist of alphanumerics and underscores, and must
4126 be unique. Back references use the syntax (?P=name) and recursive calls use
4127 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4128 numbers. The function pcre_fullinfo() can be used after compilation to extract
4129 a name/number map. There are three relevant calls:
4130
4131 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4132 PCRE_INFO_NAMECOUNT yields the number of entries
4133 PCRE_INFO_NAMETABLE yields a pointer to the map.
4134
4135 The map is a vector of fixed-size entries. The size of each entry depends on
4136 the length of the longest name used. The first two bytes of each entry are the
4137 group number, most significant byte first. There follows the corresponding
4138 name, zero terminated. The names are in alphabetical order.
4139
4140 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4141 case instead of 255. Making it the same both with and without UTF-8 support
4142 means that the same test output works with both.
4143
4144 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4145 calling malloc() with a zero argument.
4146
4147 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4148 optimization. I've improved things by keeping a bitmap of backreferences with
4149 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4150 fact referenced, the optimization can be applied. It is unlikely that a
4151 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4152 the match to follow \n) will appear inside brackets with a number greater than
4153 31, but if it does, any back reference > 31 suppresses the optimization.
4154
4155 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4156 of disabling numbered capturing parentheses. Any opening parenthesis that is
4157 not followed by ? behaves as if it were followed by ?: but named parentheses
4158 can still be used for capturing (and they will acquire numbers in the usual
4159 way).
4160
4161 41. Redesigned the return codes from the match() function into yes/no/error so
4162 that errors can be passed back from deep inside the nested calls. A malloc
4163 failure while inside a recursive subpattern call now causes the
4164 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4165
4166 42. It is now possible to set a limit on the number of times the match()
4167 function is called in a call to pcre_exec(). This facility makes it possible to
4168 limit the amount of recursion and backtracking, though not in a directly
4169 obvious way, because the match() function is used in a number of different
4170 circumstances. The count starts from zero for each position in the subject
4171 string (for non-anchored patterns). The default limit is, for compatibility, a
4172 large number, namely 10 000 000. You can change this in two ways:
4173
4174 (a) When configuring PCRE before making, you can use --with-match-limit=n
4175 to set a default value for the compiled library.
4176
4177 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4178 a different value is set. See 45 below.
4179
4180 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4181
4182 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4183 of things that can be changed at compile time. The first argument specifies
4184 what is wanted and the second points to where the information is to be placed.
4185 The current list of available information is:
4186
4187 PCRE_CONFIG_UTF8
4188
4189 The output is an integer that is set to one if UTF-8 support is available;
4190 otherwise it is set to zero.
4191
4192 PCRE_CONFIG_NEWLINE
4193
4194 The output is an integer that it set to the value of the code that is used for
4195 newline. It is either LF (10) or CR (13).
4196
4197 PCRE_CONFIG_LINK_SIZE
4198
4199 The output is an integer that contains the number of bytes used for internal
4200 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4201
4202 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4203
4204 The output is an integer that contains the threshold above which the POSIX
4205 interface uses malloc() for output vectors. See item 31 above.
4206
4207 PCRE_CONFIG_MATCH_LIMIT
4208
4209 The output is an unsigned integer that contains the default limit of the number
4210 of match() calls in a pcre_exec() execution. See 42 above.
4211
4212 44. pcretest has been upgraded by the addition of the -C option. This causes it
4213 to extract all the available output from the new pcre_config() function, and to
4214 output it. The program then exits immediately.
4215
4216 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4217 order to support additional features. One way would have been to define
4218 pcre_exec2() (for example) with extra arguments, but this would not have been
4219 extensible, and would also have required all calls to the original function to
4220 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4221 is used for passing in "extra" data from pcre_study().
4222
4223 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4224 contains the following fields:
4225
4226 flags a bitmap indicating which of the following fields are set
4227 study_data opaque data from pcre_study()
4228 match_limit a way of specifying a limit on match() calls for a specific
4229 call to pcre_exec()
4230 callout_data data for callouts (see 49 below)
4231
4232 The flag bits are also defined in pcre.h, and are
4233
4234 PCRE_EXTRA_STUDY_DATA
4235 PCRE_EXTRA_MATCH_LIMIT
4236 PCRE_EXTRA_CALLOUT_DATA
4237
4238 The pcre_study() function now returns one of these new pcre_extra blocks, with
4239 the actual study data pointed to by the study_data field, and the
4240 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4241 before. That is, this change is entirely upwards-compatible and requires no
4242 change to existing code.
4243
4244 If you want to pass in additional data to pcre_exec(), you can either place it
4245 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4246 block.
4247
4248 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4249 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4250 times with different match limits, until it finds the minimum value needed for
4251 pcre_exec() to complete. The value is then output. This can be instructive; for
4252 most simple matches the number is quite small, but for pathological cases it
4253 gets very large very quickly.
4254
4255 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4256 returns the size of the data block pointed to by the study_data field in a
4257 pcre_extra block, that is, the value that was passed as the argument to
4258 pcre_malloc() when PCRE was getting memory in which to place the information
4259 created by pcre_study(). The fourth argument should point to a size_t variable.
4260 pcretest has been extended so that this information is shown after a successful
4261 pcre_study() call when information about the compiled regex is being displayed.
4262
4263 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4264 because what follows is always an absolute path. (Later: it turns out that this
4265 is more than cosmetic for MinGW, because it doesn't like empty path
4266 components.)
4267
4268 49. Some changes have been made to the callout feature (see 28 above):
4269
4270 (i) A callout function now has three choices for what it returns:
4271
4272 0 => success, carry on matching
4273 > 0 => failure at this point, but backtrack if possible
4274 < 0 => serious error, return this value from pcre_exec()
4275
4276 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4277 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4278 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4279 use by callout functions. It will never be used by PCRE itself.
4280
4281 (ii) The pcre_extra structure (see 45 above) has a void * field called
4282 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4283 pcre_callout_block structure has a field of the same name. The contents of
4284 the field passed in the pcre_extra structure are passed to the callout
4285 function in the corresponding field in the callout block. This makes it
4286 easier to use the same callout-containing regex from multiple threads. For
4287 testing, the pcretest program has a new data escape
4288
4289 \C*n pass the number n (may be negative) as callout_data
4290
4291 If the callout function in pcretest receives a non-zero value as
4292 callout_data, it returns that value.
4293
4294 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4295 there were some redundant $(CFLAGS) in commands that are now specified as
4296 $(LINK), which already includes $(CFLAGS).
4297
4298 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4299 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4300 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4301 one-byte characters throughout. Note that case-insensitive matching applies
4302 only to characters whose values are less than 256. PCRE doesn't support the
4303 notion of cases for higher-valued characters.
4304
4305 (i) A character class whose characters are all within 0-255 is handled as
4306 a bit map, and the map is inverted for negative classes. Previously, a
4307 character > 255 always failed to match such a class; however it should
4308 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4309
4310 (ii) A negated character class with a single character < 255 is coded as
4311 "not this character" (OP_NOT). This wasn't working properly when the test
4312 character was multibyte, either singly or repeated.
4313
4314 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4315 mode, for example: \x{100}{2,3}.
4316
4317 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4318 singly or repeated) now correctly test multibyte characters. However,
4319 PCRE doesn't recognize any characters with values greater than 255 as
4320 digits, spaces, or word characters. Such characters always match \D, \S,
4321 and \W, and never match \d, \s, or \w.
4322
4323 (v) Classes may now contain characters and character ranges with values
4324 greater than 255. For example: [ab\x{100}-\x{400}].
4325
4326 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4327 PCRE in UTF-8 mode.
4328
4329 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4330 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4331 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4332 value.)
4333
4334 53. The single man page has become too large. I have therefore split it up into
4335 a number of separate man pages. These also give rise to individual HTML pages;
4336 these are now put in a separate directory, and there is an index.html page that
4337 lists them all. Some hyperlinking between the pages has been installed.
4338
4339 54. Added convenience functions for handling named capturing parentheses.
4340
4341 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4342 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4343 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4344 are faulted.
4345
4346 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4347 calling configure. These values are used when compiling the dftables.c program
4348 which is run to generate the source of the default character tables. They
4349 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4350 you will need to set these values.
4351
4352 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4353
4354
4355 Version 3.9 02-Jan-02
4356 ---------------------
4357
4358 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4359
4360 2. If --disable-static was given, the building process failed when trying to
4361 build pcretest and pcregrep. (For some reason it was using libtool to compile
4362 them, which is not right, as they aren't part of the library.)
4363
4364
4365 Version 3.8 18-Dec-01
4366 ---------------------
4367
4368 1. The experimental UTF-8 code was completely screwed up. It was packing the
4369 bytes in the wrong order. How dumb can you get?
4370
4371
4372 Version 3.7 29-Oct-01
4373 ---------------------
4374
4375 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4376 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4377 this didn't happen under Solaris 8, where I normally test things.
4378
4379 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4380 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4381 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4382 this fix an hour or so after the initial 3.7 release.)
4383
4384
4385 Version 3.6 23-Oct-01
4386 ---------------------
4387
4388 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4389 offsets passed as NULL with zero offset count.
4390
4391 2. The config.guess and config.sub files had not been updated when I moved to
4392 the latest autoconf.
4393
4394
4395 Version 3.5 15-Aug-01
4396 ---------------------
4397
4398 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4399 had been forgotten.
4400
4401 2. By using declared but undefined structures, we can avoid using "void"
4402 definitions in pcre.h while keeping the internal definitions of the structures
4403 private.
4404
4405 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4406 user point of view, this means that both static and shared libraries are built
4407 by default, but this can be individually controlled. More of the work of
4408 handling this static/shared cases is now inside libtool instead of PCRE's make
4409 file.
4410
4411 4. The pcretest utility is now installed along with pcregrep because it is
4412 useful for users (to test regexs) and by doing this, it automatically gets
4413 relinked by libtool. The documentation has been turned into a man page, so
4414 there are now .1, .txt, and .html versions in /doc.
4415
4416 5. Upgrades to pcregrep:
4417 (i) Added long-form option names like gnu grep.
4418 (ii) Added --help to list all options with an explanatory phrase.
4419 (iii) Added -r, --recursive to recurse into sub-directories.
4420 (iv) Added -f, --file to read patterns from a file.
4421
4422 6. pcre_exec() was referring to its "code" argument before testing that
4423 argument for NULL (and giving an error if it was NULL).
4424
4425 7. Upgraded Makefile.in to allow for compiling in a different directory from
4426 the source directory.
4427
4428 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4429 options bits, the pointer it was passed was to an int instead of to an unsigned
4430 long int. This mattered only on 64-bit systems.
4431
4432 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4433 generated) instead of pcre.in, which it its source. Also made the same change
4434 in several of the .c files.
4435
4436 10. A new release of gcc defines printf() as a macro, which broke pcretest
4437 because it had an ifdef in the middle of a string argument for printf(). Fixed
4438 by using separate calls to printf().
4439
4440 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4441 script, to force use of CR or LF instead of \n in the source. On non-Unix
4442 systems, the value can be set in config.h.
4443
4444 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4445 absolute limit. Changed the text of the error message to make this clear, and
4446 likewise updated the man page.
4447
4448 13. The limit of 99 on the number of capturing subpatterns has been removed.
4449 The new limit is 65535, which I hope will not be a "real" limit.
4450
4451
4452 Version 3.4 22-Aug-00
4453 ---------------------
4454
4455 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4456
4457 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4458
4459
4460 Version 3.3 01-Aug-00
4461 ---------------------
4462
4463 1. If an octal character was given, but the value was greater than \377, it
4464 was not getting masked to the least significant bits, as documented. This could
4465 lead to crashes in some systems.
4466
4467 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4468 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4469
4470 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4471 These just pass their arguments on to (pcre_free)(), but they are provided
4472 because some uses of PCRE bind it to non-C systems that can call its functions,
4473 but cannot call free() or pcre_free() directly.
4474
4475 4. Add "make test" as a synonym for "make check". Corrected some comments in
4476 the Makefile.
4477
4478 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4479 Makefile.
4480
4481 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4482 command called pgrep for grepping around the active processes.
4483
4484 7. Added the beginnings of support for UTF-8 character strings.
4485
4486 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4487 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4488 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4489 out for the ar command.)
4490
4491
4492 Version 3.2 12-May-00
4493 ---------------------
4494
4495 This is purely a bug fixing release.
4496
4497 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4498 of ZA. This was just one example of several cases that could provoke this bug,
4499 which was introduced by change 9 of version 2.00. The code for breaking
4500 infinite loops after an iteration that matches an empty string was't working
4501 correctly.
4502
4503 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4504 when matched against abbab (for example). After matching an empty string, it
4505 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4506 caused it to match further down the string than it should.
4507
4508 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4509 was there because it doesn't seem to be needed, and it causes trouble on some
4510 systems, as it is not a Standard C header. It has been removed.
4511
4512 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4513 were reported on the Macintosh. The changes were from
4514
4515 while ((c = *(++ptr)) != 0 && c != '\n');
4516 to
4517 while ((c = *(++ptr)) != 0 && c != '\n') ;
4518
4519 Totally extraordinary, but if that's what it takes...
4520
4521 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4522 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4523 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4524 assumes the way PCRE uses memmove() (always moving upwards).
4525
4526 6. PCRE is being used in one environment where strchr() is not available. There
4527 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4528 faster code anyway.
4529
4530
4531 Version 3.1 09-Feb-00
4532 ---------------------
4533
4534 The only change in this release is the fixing of some bugs in Makefile.in for
4535 the "install" target:
4536
4537 (1) It was failing to install pcreposix.h.
4538
4539 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4540
4541
4542 Version 3.0 01-Feb-00
4543 ---------------------
4544
4545 1. Add support for the /+ modifier to perltest (to output $` like it does in
4546 pcretest).
4547
4548 2. Add support for the /g modifier to perltest.
4549
4550 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4551 matches null strings.
4552
4553 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4554 pattern. Perl treats empty patterns specially - it reuses the most recent
4555 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4556 effect.
4557
4558 5. The POSIX interface was broken in that it was just handing over the POSIX
4559 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4560 required a bigger vector, with some working space on the end. This means that
4561 the POSIX wrapper now has to get and free some memory, and copy the results.
4562
4563 6. Added some simple autoconf support, placing the test data and the
4564 documentation in separate directories, re-organizing some of the
4565 information files, and making it build pcre-config (a GNU standard). Also added
4566 libtool support for building PCRE as a shared library, which is now the
4567 default.
4568
4569 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4570 09 are not valid octal constants. Single digits will be used for minor values
4571 less than 10.
4572
4573 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4574 existing programs that set these in the POSIX interface can use PCRE without
4575 modification.
4576
4577 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4578 return all that pcre_info() returns, plus additional data. The pcre_info()
4579 function is retained for compatibility, but is considered to be obsolete.
4580
4581 10. Added experimental recursion feature (?R) to handle one common case that
4582 Perl 5.6 will be able to do with (?p{...}).
4583
4584 11. Added support for POSIX character classes like [:alpha:], which Perl is
4585 adopting.
4586
4587
4588 Version 2.08 31-Aug-99
4589 ----------------------
4590
4591 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4592 trying to match at the startoffset position, but instead was moving forward to
4593 the next newline as if a previous match had failed.
4594
4595 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4596 and could get into a loop if a null string was matched other than at the start
4597 of the subject.
4598
4599 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4600 be distinguished at compile time, and for completeness also added PCRE_DATE.
4601
4602 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4603 in GnuWin32 environments.
4604
4605
4606 Version 2.07 29-Jul-99
4607 ----------------------
4608
4609 1. The documentation is now supplied in plain text form and HTML as well as in
4610 the form of man page sources.
4611
4612 2. C++ compilers don't like assigning (void *) values to other pointer types.
4613 In particular this affects malloc(). Although there is no problem in Standard
4614 C, I've put in casts to keep C++ compilers happy.
4615
4616 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4617 should be (const char *).
4618
4619 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4620 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4621 However, I haven't made this a standard facility. The documentation doesn't
4622 mention it, and the Makefile doesn't support it.
4623
4624 5. The Makefile now contains an "install" target, with editable destinations at
4625 the top of the file. The pcretest program is not installed.
4626
4627 6. pgrep -V now gives the PCRE version number and date.
4628
4629 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4630 causing the entire string to be ignored, instead of just the last character.
4631
4632 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4633 non-matching string, it can take a very, very long time, even for strings of
4634 quite modest length, because of the nested recursion. PCRE now does better in
4635 some of these cases. It does this by remembering the last required literal
4636 character in the pattern, and pre-searching the subject to ensure it is present
4637 before running the real match. In other words, it applies a heuristic to detect
4638 some types of certain failure quickly, and in the above example, if presented
4639 with a string that has no trailing " it gives "no match" very quickly.
4640
4641 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4642 other alternatives are tried instead.
4643
4644
4645 Version 2.06 09-Jun-99
4646 ----------------------
4647
4648 1. Change pcretest's output for amount of store used to show just the code
4649 space, because the remainder (the data block) varies in size between 32-bit and
4650 64-bit systems.
4651
4652 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4653 start matching at. This allows lookbehinds to work when searching for multiple
4654 occurrences in a string.
4655
4656 3. Added additional options to pcretest for testing multiple occurrences:
4657
4658 /+ outputs the rest of the string that follows a match
4659 /g loops for multiple occurrences, using the new startoffset argument
4660 /G loops for multiple occurrences by passing an incremented pointer
4661
4662 4. PCRE wasn't doing the "first character" optimization for patterns starting
4663 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4664 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4665 the letter 'x'. On long subject strings, this gives a significant speed-up.
4666
4667
4668 Version 2.05 21-Apr-99
4669 ----------------------
4670
4671 1. Changed the type of magic_number from int to long int so that it works
4672 properly on 16-bit systems.
4673
4674 2. Fixed a bug which caused patterns starting with .* not to work correctly
4675 when the subject string contained newline characters. PCRE was assuming
4676 anchoring for such patterns in all cases, which is not correct because .* will
4677 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4678 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4679 must be retried after every newline in the subject.
4680
4681
4682 Version 2.04 18-Feb-99
4683 ----------------------
4684
4685 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4686 computation of the store needed to hold the pattern was incorrect (too large).
4687 If such patterns were nested a few deep, this could multiply and become a real
4688 problem.
4689
4690 2. Added /M option to pcretest to show the memory requirement of a specific
4691 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4692
4693 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4694 compiled in such a way that the backtracking after subsequent failure was
4695 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4696 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4697
4698
4699 Version 2.03 02-Feb-99
4700 ----------------------
4701
4702 1. Fixed typo and small mistake in man page.
4703
4704 2. Added 4th condition (GPL supersedes if conflict) and created separate
4705 LICENCE file containing the conditions.
4706
4707 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4708 Perl, that is the internal \ allows the delimiter to be included in the
4709 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4710 the final delimiter, add \ to the end of the pattern (to test the error).
4711
4712 4. Added the convenience functions for extracting substrings after a successful
4713 match. Updated pcretest to make it able to test these functions.
4714
4715
4716 Version 2.02 14-Jan-99
4717 ----------------------
4718
4719 1. Initialized the working variables associated with each extraction so that
4720 their saving and restoring doesn't refer to uninitialized store.
4721
4722 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4723 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4724 fix the problem.
4725
4726 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4727 calls, and wasn't printing the correct value for compiling calls. Increased the
4728 default value of LOOPREPEAT, and the number of significant figures in the
4729 times.
4730
4731 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4732
4733 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4734 a building problem on Windows NT with a FAT file system.
4735
4736
4737 Version 2.01 21-Oct-98
4738 ----------------------
4739
4740 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4741 to character tables built by pcre_maketables() in the current locale. If NULL
4742 is passed, the default tables are used.
4743
4744
4745 Version 2.00 24-Sep-98
4746 ----------------------
4747
4748 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4749 it any more.
4750
4751 2. Allow quantification of (?>) groups, and make it work correctly.
4752
4753 3. The first character computation wasn't working for (?>) groups.
4754
4755 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4756 end of the subject) and add 5.005's \z, which really does match only at the
4757 very end of the subject.
4758
4759 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4760
4761 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4762 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4763 localized options. All options to pcre_study() were also removed.
4764
4765 7. Add other new features from 5.005:
4766
4767 $(?<= positive lookbehind
4768 $(?<! negative lookbehind
4769 (?imsx-imsx) added the unsetting capability
4770 such a setting is global if at outer level; local otherwise
4771 (?imsx-imsx:) non-capturing groups with option setting
4772 (?(cond)re|re) conditional pattern matching
4773
4774 A backreference to itself in a repeated group matches the previous
4775 captured string.
4776
4777 8. General tidying up of studying (both automatic and via "study")
4778 consequential on the addition of new assertions.
4779
4780 9. As in 5.005, unlimited repeated groups that could match an empty substring
4781 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4782 runtime if any iteration does actually match an empty substring.
4783
4784 10. Include the RunTest script in the distribution.
4785
4786 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4787 discrepancies, some of which were old and were also with respect to 5.004. They
4788 have now been fixed.
4789
4790
4791 Version 1.09 28-Apr-98
4792 ----------------------
4793
4794 1. A negated single character class followed by a quantifier with a minimum
4795 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4796 program crashes, or just wrong answers. This did not apply to negated classes
4797 containing more than one character, or to minima other than one.
4798
4799
4800 Version 1.08 27-Mar-98
4801 ----------------------
4802
4803 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4804
4805 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4806 latter must appear before anything that relies on it in the pattern.
4807
4808
4809 Version 1.07 16-Feb-98
4810 ----------------------
4811
4812 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4813 repeat of a potentially empty string).
4814
4815
4816 Version 1.06 23-Jan-98
4817 ----------------------
4818
4819 1. Added Markus Oberhumer's little patches for C++.
4820
4821 2. Literal strings longer than 255 characters were broken.
4822
4823
4824 Version 1.05 23-Dec-97
4825 ----------------------
4826
4827 1. Negated character classes containing more than one character were failing if
4828 PCRE_CASELESS was set at run time.
4829
4830
4831 Version 1.04 19-Dec-97
4832 ----------------------
4833
4834 1. Corrected the man page, where some "const" qualifiers had been omitted.
4835
4836 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4837 input syntax.
4838
4839 3. Fixed memory leak which occurred when a regex with back references was
4840 matched with an offsets vector that wasn't big enough. The temporary memory
4841 that is used in this case wasn't being freed if the match failed.
4842
4843 4. Tidied pcretest to ensure it frees memory that it gets.
4844
4845 5. Temporary memory was being obtained in the case where the passed offsets
4846 vector was exactly big enough.
4847
4848 6. Corrected definition of offsetof() from change 5 below.
4849
4850 7. I had screwed up change 6 below and broken the rules for the use of
4851 setjmp(). Now fixed.
4852
4853
4854 Version 1.03 18-Dec-97
4855 ----------------------
4856
4857 1. A erroneous regex with a missing opening parenthesis was correctly
4858 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4859 on some systems.
4860
4861 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4862 it was reported that one broken compiler failed on the former because "code" is
4863 also an independent variable.
4864
4865 3. The erroneous regex a[]b caused an array overrun reference.
4866
4867 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4868 fail on data ending with that character. (It was going on too far, and checking
4869 the next character, typically a binary zero.) This was specific to the
4870 optimized code for single-character negative classes.
4871
4872 5. Added a contributed patch from the TIN world which does the following:
4873
4874 + Add an undef for memmove, in case the the system defines a macro for it.
4875
4876 + Add a definition of offsetof(), in case there isn't one. (I don't know
4877 the reason behind this - offsetof() is part of the ANSI standard - but
4878 it does no harm).
4879
4880 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4881 most of the places where whitespace preceded '#'. I have given up and
4882 allowed the remaining 2 cases to be at the margin.
4883
4884 + Rename some variables in pcre to eliminate shadowing. This seems very
4885 pedantic, but does no harm, of course.
4886
4887 6. Moved the call to setjmp() into its own function, to get rid of warnings
4888 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4889
4890 7. Constructs such as \d{8,} were compiling into the equivalent of
4891 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4892 outcome, but in this particular case used more store than had been allocated,
4893 which caused the bug to be discovered because it threw up an internal error.
4894
4895 8. The debugging code in both pcre and pcretest for outputting the compiled
4896 form of a regex was going wrong in the case of back references followed by
4897 curly-bracketed repeats.
4898
4899
4900 Version 1.02 12-Dec-97
4901 ----------------------
4902
4903 1. Typos in pcre.3 and comments in the source fixed.
4904
4905 2. Applied a contributed patch to get rid of places where it used to remove
4906 'const' from variables, and fixed some signed/unsigned and uninitialized
4907 variable warnings.
4908
4909 3. Added the "runtest" target to Makefile.
4910
4911 4. Set default compiler flag to -O2 rather than just -O.
4912
4913
4914 Version 1.01 19-Nov-97
4915 ----------------------
4916
4917 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4918 like /([ab]*)*/, that is, for classes with more than one character in them.
4919
4920 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4921 as /((?>a*))*/ (a PCRE_EXTRA facility).
4922
4923
4924 Version 1.00 18-Nov-97
4925 ----------------------
4926
4927 1. Added compile-time macros to support systems such as SunOS4 which don't have
4928 memmove() or strerror() but have other things that can be used instead.
4929
4930 2. Arranged that "make clean" removes the executables.
4931
4932
4933 Version 0.99 27-Oct-97
4934 ----------------------
4935
4936 1. Fixed bug in code for optimizing classes with only one character. It was
4937 initializing a 32-byte map regardless, which could cause it to run off the end
4938 of the memory it had got.
4939
4940 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4941
4942
4943 Version 0.98 22-Oct-97
4944 ----------------------
4945
4946 1. Fixed bug in code for handling temporary memory usage when there are more
4947 back references than supplied space in the ovector. This could cause segfaults.
4948
4949
4950 Version 0.97 21-Oct-97
4951 ----------------------
4952
4953 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4954
4955 2. Optimized negated single characters not to use a bit map.
4956
4957 3. Brought error texts together as macro definitions; clarified some of them;
4958 fixed one that was wrong - it said "range out of order" when it meant "invalid
4959 escape sequence".
4960
4961 4. Changed some char * arguments to const char *.
4962
4963 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4964
4965 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4966 pcretest.
4967
4968
4969 Version 0.96 16-Oct-97
4970 ----------------------
4971
4972 1. Added a simple "pgrep" utility to the distribution.
4973
4974 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4975 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4976 where "ddd" means "one or more decimal digits".
4977
4978 3. Fixed serious bug. If a pattern had a back reference, but the call to
4979 pcre_exec() didn't supply a large enough ovector to record the related
4980 identifying subpattern, the match always failed. PCRE now remembers the number
4981 of the largest back reference, and gets some temporary memory in which to save
4982 the offsets during matching if necessary, in order to ensure that
4983 backreferences always work.
4984
4985 4. Increased the compatibility with Perl in a number of ways:
4986
4987 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4988 to request this handling. The option can be set at compile or exec time.
4989
4990 (b) $ matches before a terminating newline by default; an option
4991 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4992 mode). The option can be set at compile or exec time.
4993
4994 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4995 the same as Perl's. If the decimal number it represents is less than 10
4996 or there aren't that many previous left capturing parentheses, an octal
4997 escape is read. Inside a character class, it's always an octal escape,
4998 even if it is a single digit.
4999
5000 (d) An escaped but undefined alphabetic character is taken as a literal,
5001 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5002 escapes.
5003
5004 (e) {0} is now permitted. (The previous item is removed from the compiled
5005 pattern).
5006
5007 5. Changed all the names of code files so that the basic parts are no longer
5008 than 10 characters, and abolished the teeny "globals.c" file.
5009
5010 6. Changed the handling of character classes; they are now done with a 32-byte
5011 bit map always.
5012
5013 7. Added the -d and /D options to pcretest to make it possible to look at the
5014 internals of compilation without having to recompile pcre.
5015
5016
5017 Version 0.95 23-Sep-97
5018 ----------------------
5019
5020 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5021 \x20 at the start of a run of normal characters. These were being treated as
5022 real characters, instead of the source characters being re-checked.
5023
5024
5025 Version 0.94 18-Sep-97
5026 ----------------------
5027
5028 1. The functions are now thread-safe, with the caveat that the global variables
5029 containing pointers to malloc() and free() or alternative functions are the
5030 same for all threads.
5031
5032 2. Get pcre_study() to generate a bitmap of initial characters for non-
5033 anchored patterns when this is possible, and use it if passed to pcre_exec().
5034
5035
5036 Version 0.93 15-Sep-97
5037 ----------------------
5038
5039 1. /(b)|(:+)/ was computing an incorrect first character.
5040
5041 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5042 but not actually doing anything yet.
5043
5044 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5045 as Perl does (e.g. [-az] or [az-]).
5046
5047 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5048 all possible positions.
5049
5050 5. Split up into different modules to avoid including unneeded functions in a
5051 compiled binary. However, compile and exec are still in one module. The "study"
5052 function is split off.
5053
5054 6. The character tables are now in a separate module whose source is generated
5055 by an auxiliary program - but can then be edited by hand if required. There are
5056 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5057 toupper() in the code.
5058
5059 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5060 make them global. Abolish the function for setting them, as the caller can now
5061 set them directly.
5062
5063
5064 Version 0.92 11-Sep-97
5065 ----------------------
5066
5067 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5068 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5069
5070 2. Caseless matching was not working in character classes if the characters in
5071 the pattern were in upper case.
5072
5073 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5074
5075 4. Make PCRE_ANCHORED public and accept as a compile option.
5076
5077 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5078 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5079 pass them.
5080
5081 6. Give an error if bad option bits passed at compile or run time.
5082
5083 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5084 pcretest to cause it to pass that flag.
5085
5086 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5087 options, and the first character, if set.
5088
5089 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5090
5091
5092 Version 0.91 10-Sep-97
5093 ----------------------
5094
5095 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5096 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5097
5098 2. PCRE was looping on encountering an indefinitely repeated back reference to
5099 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5100 Perl does - treats the match as successful.
5101
5102 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5