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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1272 - (show annotations)
Thu Mar 7 11:30:01 2013 UTC (6 years, 8 months ago) by zherczeg
File size: 244427 byte(s)
Error occurred while calculating annotation data.
(*PRUNE) is now supported by the JIT compiler.
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
103 Version 8.32 30-November-2012
104 -----------------------------
105
106 1. Improved JIT compiler optimizations for first character search and single
107 character iterators.
108
109 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
110 Patch by Daniel Richard G.
111
112 3. Single character iterator optimizations in the JIT compiler.
113
114 4. Improved JIT compiler optimizations for character ranges.
115
116 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
117 Reported by Giuseppe D'Angelo.
118
119 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
120 of a line, was being set incorrectly in cases where .* appeared inside
121 atomic brackets at the start of a pattern, or where there was a subsequent
122 *PRUNE or *SKIP.
123
124 7. Improved instruction cache flush for POWER/PowerPC.
125 Patch by Daniel Richard G.
126
127 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
128 grep:
129
130 (a) There is now no limit to the number of patterns to be matched.
131
132 (b) An error is given if a pattern is too long.
133
134 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
135 are now supported.
136
137 (d) --exclude-from and --include-from (multiple use) have been added.
138
139 (e) Exclusions and inclusions now apply to all files and directories, not
140 just to those obtained from scanning a directory recursively.
141
142 (f) Multiple uses of -f and --file-list are now supported.
143
144 (g) In a Windows environment, the default for -d has been changed from
145 "read" (the GNU grep default) to "skip", because otherwise the presence
146 of a directory in the file list provokes an error.
147
148 (h) The documentation has been revised and clarified in places.
149
150 9. Improve the matching speed of capturing brackets.
151
152 10. Changed the meaning of \X so that it now matches a Unicode extended
153 grapheme cluster.
154
155 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
156 out POSIX threads when JIT support is configured.
157
158 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
159
160 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
161 structure could go wrong in environments where size_t is not the same size
162 as int.
163
164 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
165 set.
166
167 15. The EBCDIC support had decayed; later updates to the code had included
168 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
169 general tidy up of EBCDIC-related issues, and the documentation was also
170 not quite right. There is now a test that can be run on ASCII systems to
171 check some of the EBCDIC-related things (but is it not a full test).
172
173 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
174 in a small tidy to the code.
175
176 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
177
178 18. If the --only-matching (-o) option in pcregrep is specified multiple
179 times, each one causes appropriate output. For example, -o1 -o2 outputs the
180 substrings matched by the 1st and 2nd capturing parentheses. A separating
181 string can be specified by --om-separator (default empty).
182
183 19. Improving the first n character searches.
184
185 20. Turn case lists for horizontal and vertical white space into macros so that
186 they are defined only once.
187
188 21. This set of changes together give more compatible Unicode case-folding
189 behaviour for characters that have more than one other case when UCP
190 support is available.
191
192 (a) The Unicode property table now has offsets into a new table of sets of
193 three or more characters that are case-equivalent. The MultiStage2.py
194 script that generates these tables (the pcre_ucd.c file) now scans
195 CaseFolding.txt instead of UnicodeData.txt for character case
196 information.
197
198 (b) The code for adding characters or ranges of characters to a character
199 class has been abstracted into a generalized function that also handles
200 case-independence. In UTF-mode with UCP support, this uses the new data
201 to handle characters with more than one other case.
202
203 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
204 whose other case is greater than 256 are now correctly matched
205 caselessly. Previously, the high codepoint matched the low one, but not
206 vice versa.
207
208 (d) The processing of \h, \H, \v, and \ in character classes now makes use
209 of the new class addition function, using character lists defined as
210 macros alongside the case definitions of 20 above.
211
212 (e) Caseless back references now work with characters that have more than
213 one other case.
214
215 (f) General caseless matching of characters with more than one other case
216 is supported.
217
218 22. Unicode character properties were updated from Unicode 6.2.0
219
220 23. Improved CMake support under Windows. Patch by Daniel Richard G.
221
222 24. Add support for 32-bit character strings, and UTF-32
223
224 25. Major JIT compiler update (code refactoring and bugfixing).
225 Experimental Sparc 32 support is added.
226
227 26. Applied a modified version of Daniel Richard G's patch to create
228 pcre.h.generic and config.h.generic by "make" instead of in the
229 PrepareRelease script.
230
231 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
232 pcre_compile.c when checking for a zero character.
233
234 28. Introducing a native interface for JIT. Through this interface, the compiled
235 machine code can be directly executed. The purpose of this interface is to
236 provide fast pattern matching, so several sanity checks are not performed.
237 However, feature tests are still performed. The new interface provides
238 1.4x speedup compared to the old one.
239
240 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
241 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
242 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
243
244 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
245 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
246 the "old" RFC 2279). Instead, it ended up passing a negative length to
247 pcre_exec().
248
249 31. Add support for GCC's visibility feature to hide internal functions.
250
251 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
252 "unknown -C option" after outputting 0 or 1.
253
254 33. There is now support for generating a code coverage report for the test
255 suite in environments where gcc is the compiler and lcov is installed. This
256 is mainly for the benefit of the developers.
257
258 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
259 unaddressable using valgrind annotations, allowing valgrind to detect
260 invalid memory accesses. This is mainly for the benefit of the developers.
261
262 25. (*UTF) can now be used to start a pattern in any of the three libraries.
263
264 26. Give configure error if --enable-cpp but no C++ compiler found.
265
266
267 Version 8.31 06-July-2012
268 -------------------------
269
270 1. Fixing a wrong JIT test case and some compiler warnings.
271
272 2. Removed a bashism from the RunTest script.
273
274 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
275 to unsigned type, result still unsigned" that was given by an MS compiler
276 on encountering the code "-sizeof(xxx)".
277
278 4. Partial matching support is added to the JIT compiler.
279
280 5. Fixed several bugs concerned with partial matching of items that consist
281 of more than one character:
282
283 (a) /^(..)\1/ did not partially match "aba" because checking references was
284 done on an "all or nothing" basis. This also applied to repeated
285 references.
286
287 (b) \R did not give a hard partial match if \r was found at the end of the
288 subject.
289
290 (c) \X did not give a hard partial match after matching one or more
291 characters at the end of the subject.
292
293 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
294 a partial match for the string "\r".
295
296 (e) When newline was set to CRLF, the metacharacter "." did not recognize
297 a partial match for a CR character at the end of the subject string.
298
299 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
300 running pcretest, the text "(JIT)" added to the output whenever JIT is
301 actually used to run the match.
302
303 7. Individual JIT compile options can be set in pcretest by following -s+[+]
304 or /S+[+] with a digit between 1 and 7.
305
306 8. OP_NOT now supports any UTF character not just single-byte ones.
307
308 9. (*MARK) control verb is now supported by the JIT compiler.
309
310 10. The command "./RunTest list" lists the available tests without actually
311 running any of them. (Because I keep forgetting what they all are.)
312
313 11. Add PCRE_INFO_MAXLOOKBEHIND.
314
315 12. Applied a (slightly modified) user-supplied patch that improves performance
316 when the heap is used for recursion (compiled with --disable-stack-for-
317 recursion). Instead of malloc and free for each heap frame each time a
318 logical recursion happens, frames are retained on a chain and re-used where
319 possible. This sometimes gives as much as 30% improvement.
320
321 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
322 call.
323
324 14. As documented, (*COMMIT) is now confined to within a positive assertion.
325
326 15. It is now possible to link pcretest with libedit as an alternative to
327 libreadline.
328
329 16. (*COMMIT) control verb is now supported by the JIT compiler.
330
331 17. The Unicode data tables have been updated to Unicode 6.1.0.
332
333 18. Added --file-list option to pcregrep.
334
335 19. Added binary file support to pcregrep, including the -a, --binary-files,
336 -I, and --text options.
337
338 20. The madvise function is renamed for posix_madvise for QNX compatibility
339 reasons. Fixed by Giuseppe D'Angelo.
340
341 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
342
343 22. Changed the option for creating symbolic links for 16-bit man pages from
344 -s to -sf so that re-installing does not cause issues.
345
346 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
347
348 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
349 matches in certain environments (the workspace was not being correctly
350 retained). Also added to pcre_dfa_exec() a simple plausibility check on
351 some of the workspace data at the beginning of a restart.
352
353 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
354 was not doing so when it should - probably a typo introduced by SVN 528
355 (change 8.10/14).
356
357 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
358 \w+ when the character tables indicated that \x{c4} was a word character.
359 There were several related cases, all because the tests for doing a table
360 lookup were testing for characters less than 127 instead of 255.
361
362 27. If a pattern contains capturing parentheses that are not used in a match,
363 their slots in the ovector are set to -1. For those that are higher than
364 any matched groups, this happens at the end of processing. In the case when
365 there were back references that the ovector was too small to contain
366 (causing temporary malloc'd memory to be used during matching), and the
367 highest capturing number was not used, memory off the end of the ovector
368 was incorrectly being set to -1. (It was using the size of the temporary
369 memory instead of the true size.)
370
371 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
372 ovector size, it uses memory at the end of the block that it has got.
373
374 29. Check for an overlong MARK name and give an error at compile time. The
375 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
376
377 30. JIT compiler update.
378
379 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
380 Rill for the patch.
381
382 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
383
384 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
385
386 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
387 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
388 was enabled.
389
390 35. Improve JIT code generation for greedy plus quantifier.
391
392 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
393 1 to "aa" instead of to an empty string. The bug affected repeated groups
394 that could potentially match an empty string.
395
396 37. Optimizing single character iterators in JIT.
397
398 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
399 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
400 codepoints that are too big for the mode are faulted, and in a UTF mode,
401 disallowed codepoints are also faulted.
402
403 39. If PCRE was compiled with UTF support, in three places in the DFA
404 matcher there was code that should only have been obeyed in UTF mode, but
405 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
406 processing when bytes with values greater than 127 were present. In 16-bit
407 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
408 both cases the values are those that cannot be the first data item in a UTF
409 character. The three items that might have provoked this were recursions,
410 possessively repeated groups, and atomic groups.
411
412 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
413 and pcregrep, because some OS require shared objects to be explicitly
414 passed to ld, causing the link step to fail if they are not.
415
416 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
417 mode, patterns that started with \h* or \R* might be incorrectly matched.
418
419
420 Version 8.30 04-February-2012
421 -----------------------------
422
423 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
424 name is defined in ctype.h.
425
426 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
427 only in quite long subpatterns.
428
429 3. Removed the function pcre_info(), which has been obsolete and deprecated
430 since it was replaced by pcre_fullinfo() in February 2000.
431
432 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
433 match a (*MARK), and the match failed at the start of the subject, a
434 reference to memory before the start of the subject could occur. This bug
435 was introduced by fix 17 of release 8.21.
436
437 5. A reference to an unset group with zero minimum repetition was giving
438 totally wrong answers (in non-JavaScript-compatibility mode). For example,
439 /(another)?(\1?)test/ matched against "hello world test". This bug was
440 introduced in release 8.13.
441
442 6. Add support for 16-bit character strings (a large amount of work involving
443 many changes and refactorings).
444
445 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
446 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
447 from a file.
448
449 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
450 rounding is not applied in this particular case).
451
452 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
453 if they appear, or are escaped, in patterns.
454
455 10. Get rid of a number of -Wunused-but-set-variable warnings.
456
457 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
458 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
459 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
460 also returns the mark "x". This bug applied to capturing parentheses,
461 non-capturing parentheses, and atomic parentheses. It also applied to some
462 assertions.
463
464 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
465 information out of configure.ac instead of relying on pcre.h.generic, which
466 is not stored in the repository.
467
468 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
469 -lreadline.
470
471 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
472
473 15. Applied Graycode's patch to put the top-level frame on the stack rather
474 than the heap when not using the stack for recursion. This gives a
475 performance improvement in many cases when recursion is not deep.
476
477 16. Experimental code added to "pcretest -C" to output the stack frame size.
478
479
480 Version 8.21 12-Dec-2011
481 ------------------------
482
483 1. Updating the JIT compiler.
484
485 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
486 are added as well.
487
488 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
489 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
490 calling _pcre_jit_exec. Some extra comments are added.
491
492 4. (*MARK) settings inside atomic groups that do not contain any capturing
493 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
494 was introduced by change 18 for 8.20.
495
496 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
497 ECMA-262 standard.
498
499 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
500 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
501 This bug was probably introduced by change 9 of 8.13.
502
503 7. While fixing 6 above, I noticed that a number of other items were being
504 incorrectly rejected as "not fixed length". This arose partly because newer
505 opcodes had not been added to the fixed-length checking code. I have (a)
506 corrected the bug and added tests for these items, and (b) arranged for an
507 error to occur if an unknown opcode is encountered while checking for fixed
508 length instead of just assuming "not fixed length". The items that were
509 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
510 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
511 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
512
513 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
514 being incorrectly compiled and would have given unpredicatble results.
515
516 9. A possessively repeated subpattern with minimum repeat count greater than
517 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
518 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
519 the first (A) could occur when it should not.
520
521 10. Add a cast and remove a redundant test from the code.
522
523 11. JIT should use pcre_malloc/pcre_free for allocation.
524
525 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
526 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
527 is anything other than /usr, -L is still shown).
528
529 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
530
531 14. Perl does not support \N without a following name in a [] class; PCRE now
532 also gives an error.
533
534 15. If a forward reference was repeated with an upper limit of around 2000,
535 it caused the error "internal error: overran compiling workspace". The
536 maximum number of forward references (including repeats) was limited by the
537 internal workspace, and dependent on the LINK_SIZE. The code has been
538 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
539 the default depends on LINK_SIZE. There is a new upper limit (for safety)
540 of around 200,000 forward references. While doing this, I also speeded up
541 the filling in of repeated forward references.
542
543 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
544 incorrectly expecting the subject to contain another "a" after the start.
545
546 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
547 in the match, the SKIP should be ignored. This was not happening; instead
548 the SKIP was being treated as NOMATCH. For patterns such as
549 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
550 tested.
551
552 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
553 now much more compatible with Perl, in particular in cases where the result
554 is a non-match for a non-anchored pattern. For example, if
555 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
556 "m", where previously it did not return a name. A side effect of this
557 change is that for partial matches, the last encountered mark name is
558 returned, as for non matches. A number of tests that were previously not
559 Perl-compatible have been moved into the Perl-compatible test files. The
560 refactoring has had the pleasing side effect of removing one argument from
561 the match() function, thus reducing its stack requirements.
562
563 19. If the /S+ option was used in pcretest to study a pattern using JIT,
564 subsequent uses of /S (without +) incorrectly behaved like /S+.
565
566 21. Retrieve executable code size support for the JIT compiler and fixing
567 some warnings.
568
569 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
570 not work when the shorter character appeared right at the end of the
571 subject string.
572
573 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
574 systems.
575
576 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
577 output it when the /M option is used in pcretest.
578
579 25. The CheckMan script was not being included in the distribution. Also, added
580 an explicit "perl" to run Perl scripts from the PrepareRelease script
581 because this is reportedly needed in Windows.
582
583 26. If study data was being save in a file and studying had not found a set of
584 "starts with" bytes for the pattern, the data written to the file (though
585 never used) was taken from uninitialized memory and so caused valgrind to
586 complain.
587
588 27. Updated RunTest.bat as provided by Sheri Pierce.
589
590 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
591
592 29. Computation of memory usage for the table of capturing group names was
593 giving an unnecessarily large value.
594
595
596 Version 8.20 21-Oct-2011
597 ------------------------
598
599 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
600 a POSIX class. After further experiments with Perl, which convinced me that
601 Perl has bugs and confusions, a closing square bracket is no longer allowed
602 in a POSIX name. This bug also affected patterns with classes that started
603 with full stops.
604
605 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
606 captured substring, but while checking the failing first alternative,
607 substring 1 is temporarily captured. If the output vector supplied to
608 pcre_exec() was not big enough for this capture, the yield of the function
609 was still zero ("insufficient space for captured substrings"). This cannot
610 be totally fixed without adding another stack variable, which seems a lot
611 of expense for a edge case. However, I have improved the situation in cases
612 such as /(a)(b)x|abc/ matched against "abc", where the return code
613 indicates that fewer than the maximum number of slots in the ovector have
614 been set.
615
616 3. Related to (2) above: when there are more back references in a pattern than
617 slots in the output vector, pcre_exec() uses temporary memory during
618 matching, and copies in the captures as far as possible afterwards. It was
619 using the entire output vector, but this conflicts with the specification
620 that only 2/3 is used for passing back captured substrings. Now it uses
621 only the first 2/3, for compatibility. This is, of course, another edge
622 case.
623
624 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
625 main code base, and can be used by building with --enable-jit. When this is
626 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
627 runtime --no-jit option is given.
628
629 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
630 ovector, the return from the function was zero, implying that there were
631 other matches that did not fit. The correct "exactly full" value is now
632 returned.
633
634 6. If a subpattern that was called recursively or as a subroutine contained
635 (*PRUNE) or any other control that caused it to give a non-standard return,
636 invalid errors such as "Error -26 (nested recursion at the same subject
637 position)" or even infinite loops could occur.
638
639 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
640 computing the minimum length on reaching *ACCEPT, and so ended up with the
641 wrong value of 1 rather than 0. Further investigation indicates that
642 computing a minimum subject length in the presence of *ACCEPT is difficult
643 (think back references, subroutine calls), and so I have changed the code
644 so that no minimum is registered for a pattern that contains *ACCEPT.
645
646 8. If (*THEN) was present in the first (true) branch of a conditional group,
647 it was not handled as intended. [But see 16 below.]
648
649 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
650 Sheri Pierce.
651
652 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
653 the first byte in a match must be "a".
654
655 11. Change 17 for 8.13 increased the recursion depth for patterns like
656 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
657 pattern contains any instances of (*THEN). If it does not, the old
658 optimizations are restored. It would be nice to do this on a per-group
659 basis, but at the moment that is not feasible.
660
661 12. In some environments, the output of pcretest -C is CRLF terminated. This
662 broke RunTest's code that checks for the link size. A single white space
663 character after the value is now allowed for.
664
665 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
666 For "fr", it uses the Windows-specific input and output files.
667
668 14. If (*THEN) appeared in a group that was called recursively or as a
669 subroutine, it did not work as intended. [But see next item.]
670
671 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
672 pattern fragments (but not containing any | characters). If A and B are
673 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
674 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
675 D. In other words, Perl considers parentheses that do not contain any |
676 characters to be part of a surrounding alternative, whereas PCRE was
677 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
678 differently. PCRE now behaves in the same way as Perl, except in the case
679 of subroutine/recursion calls such as (?1) which have in any case always
680 been different (but PCRE had them first :-).
681
682 16. Related to 15 above: Perl does not treat the | in a conditional group as
683 creating alternatives. Such a group is treated in the same way as an
684 ordinary group without any | characters when processing (*THEN). PCRE has
685 been changed to match Perl's behaviour.
686
687 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
688 RunGrepTest script failed.
689
690 18. Change 22 for version 13 caused atomic groups to use more stack. This is
691 inevitable for groups that contain captures, but it can lead to a lot of
692 stack use in large patterns. The old behaviour has been restored for atomic
693 groups that do not contain any capturing parentheses.
694
695 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
696 suppress the check for a minimum subject length at run time. (If it was
697 given to pcre_exec() or pcre_dfa_exec() it did work.)
698
699 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
700 fail to work when decoding hex characters in data strings in EBCDIC
701 environments.
702
703 21. It appears that in at least one Mac OS environment, the isxdigit() function
704 is implemented as a macro that evaluates to its argument more than once,
705 contravening the C 90 Standard (I haven't checked a later standard). There
706 was an instance in pcretest which caused it to go wrong when processing
707 \x{...} escapes in subject strings. The has been rewritten to avoid using
708 things like p++ in the argument of isxdigit().
709
710
711 Version 8.13 16-Aug-2011
712 ------------------------
713
714 1. The Unicode data tables have been updated to Unicode 6.0.0.
715
716 2. Two minor typos in pcre_internal.h have been fixed.
717
718 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
719 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
720 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
721
722 4. There were a number of related bugs in the code for matching backrefences
723 caselessly in UTF-8 mode when codes for the characters concerned were
724 different numbers of bytes. For example, U+023A and U+2C65 are an upper
725 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
726 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
727 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
728 2-byte code at the end of the subject (it thought there wasn't enough data
729 left).
730
731 5. Comprehensive information about what went wrong is now returned by
732 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
733 as the output vector has at least 2 elements. The offset of the start of
734 the failing character and a reason code are placed in the vector.
735
736 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
737 now returned is for the first byte of the failing character, instead of the
738 last byte inspected. This is an incompatible change, but I hope it is small
739 enough not to be a problem. It makes the returned offset consistent with
740 pcre_exec() and pcre_dfa_exec().
741
742 7. pcretest now gives a text phrase as well as the error number when
743 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
744 failure, the offset and reason code are output.
745
746 8. When \R was used with a maximizing quantifier it failed to skip backwards
747 over a \r\n pair if the subsequent match failed. Instead, it just skipped
748 back over a single character (\n). This seems wrong (because it treated the
749 two characters as a single entity when going forwards), conflicts with the
750 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
751 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
752 has been changed.
753
754 9. Some internal refactoring has changed the processing so that the handling
755 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
756 time (the PCRE_DOTALL option was changed this way some time ago: version
757 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
758 which was always a bit of a fudge. It also means that there is one less
759 argument for the match() function, which reduces its stack requirements
760 slightly. This change also fixes an incompatibility with Perl: the pattern
761 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
762
763 10. More internal refactoring has drastically reduced the number of recursive
764 calls to match() for possessively repeated groups such as (abc)++ when
765 using pcre_exec().
766
767 11. While implementing 10, a number of bugs in the handling of groups were
768 discovered and fixed:
769
770 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
771 (a|)*(?1) gave a compile-time internal error.
772 ((a|)+)+ did not notice that the outer group could match an empty string.
773 (^a|^)+ was not marked as anchored.
774 (.*a|.*)+ was not marked as matching at start or after a newline.
775
776 12. Yet more internal refactoring has removed another argument from the match()
777 function. Special calls to this function are now indicated by setting a
778 value in a variable in the "match data" data block.
779
780 13. Be more explicit in pcre_study() instead of relying on "default" for
781 opcodes that mean there is no starting character; this means that when new
782 ones are added and accidentally left out of pcre_study(), testing should
783 pick them up.
784
785 14. The -s option of pcretest has been documented for ages as being an old
786 synonym of -m (show memory usage). I have changed it to mean "force study
787 for every regex", that is, assume /S for every regex. This is similar to -i
788 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
789 using it. It makes it easier to run collections of tests with and without
790 study enabled, and thereby test pcre_study() more easily. All the standard
791 tests are now run with and without -s (but some patterns can be marked as
792 "never study" - see 20 below).
793
794 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
795 restoration of the capturing data to the outer values was not happening
796 correctly.
797
798 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
799 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
800 pattern had matched an empty string, and so incorrectly returned a no
801 match.
802
803 17. There was optimizing code for the last branch of non-capturing parentheses,
804 and also for the obeyed branch of a conditional subexpression, which used
805 tail recursion to cut down on stack usage. Unfortunately, now that there is
806 the possibility of (*THEN) occurring in these branches, tail recursion is
807 no longer possible because the return has to be checked for (*THEN). These
808 two optimizations have therefore been removed. [But see 8.20/11 above.]
809
810 18. If a pattern containing \R was studied, it was assumed that \R always
811 matched two bytes, thus causing the minimum subject length to be
812 incorrectly computed because \R can also match just one byte.
813
814 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
815 was incorrectly computed.
816
817 20. If /S is present twice on a test pattern in pcretest input, it now
818 *disables* studying, thereby overriding the use of -s on the command line
819 (see 14 above). This is necessary for one or two tests to keep the output
820 identical in both cases.
821
822 21. When (*ACCEPT) was used in an assertion that matched an empty string and
823 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
824
825 22. When an atomic group that contained a capturing parenthesis was
826 successfully matched, but the branch in which it appeared failed, the
827 capturing was not being forgotten if a higher numbered group was later
828 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
829 group 1 to "a", when in fact it should be unset. This applied to multi-
830 branched capturing and non-capturing groups, repeated or not, and also to
831 positive assertions (capturing in negative assertions does not happen
832 in PCRE) and also to nested atomic groups.
833
834 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
835 subject after a captured substring, to make it easier to tell which of a
836 number of identical substrings has been captured.
837
838 24. The way atomic groups are processed by pcre_exec() has been changed so that
839 if they are repeated, backtracking one repetition now resets captured
840 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
841 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
842 "aaa". Previously, it would have been "a". As part of this code
843 refactoring, the way recursive calls are handled has also been changed.
844
845 25. If an assertion condition captured any substrings, they were not passed
846 back unless some other capturing happened later. For example, if
847 (?(?=(a))a) was matched against "a", no capturing was returned.
848
849 26. When studying a pattern that contained subroutine calls or assertions,
850 the code for finding the minimum length of a possible match was handling
851 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
852 group 1 called group 2 while simultaneously a separate group 2 called group
853 1). A stack overflow occurred in this case. I have fixed this by limiting
854 the recursion depth to 10.
855
856 27. Updated RunTest.bat in the distribution to the version supplied by Tom
857 Fortmann. This supports explicit test numbers on the command line, and has
858 argument validation and error reporting.
859
860 28. An instance of \X with an unlimited repeat could fail if at any point the
861 first character it looked at was a mark character.
862
863 29. Some minor code refactoring concerning Unicode properties and scripts
864 should reduce the stack requirement of match() slightly.
865
866 30. Added the '=' option to pcretest to check the setting of unused capturing
867 slots at the end of the pattern, which are documented as being -1, but are
868 not included in the return count.
869
870 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
871 compiled something random. Now it gives a compile-time error (as does
872 Perl).
873
874 32. A *MARK encountered during the processing of a positive assertion is now
875 recorded and passed back (compatible with Perl).
876
877 33. If --only-matching or --colour was set on a pcregrep call whose pattern
878 had alternative anchored branches, the search for a second match in a line
879 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
880 matched the line "0102" twice. The same bug affected patterns that started
881 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
882 twice.
883
884 34. Previously, PCRE did not allow quantification of assertions. However, Perl
885 does, and because of capturing effects, quantifying parenthesized
886 assertions may at times be useful. Quantifiers are now allowed for
887 parenthesized assertions.
888
889 35. A minor code tidy in pcre_compile() when checking options for \R usage.
890
891 36. \g was being checked for fancy things in a character class, when it should
892 just be a literal "g".
893
894 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
895 appearance of a nested POSIX class supersedes an apparent external class.
896 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
897 unescaped square brackets may also appear as part of class names. For
898 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
899 more like Perl. (But see 8.20/1 above.)
900
901 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
902 was because it thought it was \N{name}, which is not supported).
903
904 39. Add minix to OS list not supporting the -S option in pcretest.
905
906 40. PCRE tries to detect cases of infinite recursion at compile time, but it
907 cannot analyze patterns in sufficient detail to catch mutual recursions
908 such as ((?1))((?2)). There is now a runtime test that gives an error if a
909 subgroup is called recursively as a subpattern for a second time at the
910 same position in the subject string. In previous releases this might have
911 been caught by the recursion limit, or it might have run out of stack.
912
913 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
914 happen only once. PCRE was, however incorrectly giving a compile time error
915 "recursive call could loop indefinitely" because it cannot analyze the
916 pattern in sufficient detail. The compile time test no longer happens when
917 PCRE is compiling a conditional subpattern, but actual runaway loops are
918 now caught at runtime (see 40 above).
919
920 42. It seems that Perl allows any characters other than a closing parenthesis
921 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
922 has been changed to be the same.
923
924 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
925 as not to get warnings when autogen.sh is called. Also changed
926 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
927
928 44. To help people who use pcregrep to scan files containing exceedingly long
929 lines, the following changes have been made:
930
931 (a) The default value of the buffer size parameter has been increased from
932 8K to 20K. (The actual buffer used is three times this size.)
933
934 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
935 PCRE is built.
936
937 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
938 to be set at run time.
939
940 (d) Numerical values in pcregrep options can be followed by K or M, for
941 example --buffer-size=50K.
942
943 (e) If a line being scanned overflows pcregrep's buffer, an error is now
944 given and the return code is set to 2.
945
946 45. Add a pointer to the latest mark to the callout data block.
947
948 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
949 partial match of an empty string instead of no match. This was specific to
950 the use of ".".
951
952 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
953 complete match instead of a partial match. This bug was dependent on both
954 the PCRE_UTF8 and PCRE_DOTALL options being set.
955
956 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
957 starting byte set, because \b was not being ignored.
958
959
960 Version 8.12 15-Jan-2011
961 ------------------------
962
963 1. Fixed some typos in the markup of the man pages, and wrote a script that
964 checks for such things as part of the documentation building process.
965
966 2. On a big-endian 64-bit system, pcregrep did not correctly process the
967 --match-limit and --recursion-limit options (added for 8.11). In
968 particular, this made one of the standard tests fail. (The integer value
969 went into the wrong half of a long int.)
970
971 3. If the --colour option was given to pcregrep with -v (invert match), it
972 did strange things, either producing crazy output, or crashing. It should,
973 of course, ignore a request for colour when reporting lines that do not
974 match.
975
976 4. Another pcregrep bug caused similar problems if --colour was specified with
977 -M (multiline) and the pattern match finished with a line ending.
978
979 5. In pcregrep, when a pattern that ended with a literal newline sequence was
980 matched in multiline mode, the following line was shown as part of the
981 match. This seems wrong, so I have changed it.
982
983 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
984 the check for further matches in the same line (so they could be coloured)
985 to overrun the end of the current line. If another match was found, it was
986 incorrectly shown (and then shown again when found in the next line).
987
988 7. If pcregrep was compiled under Windows, there was a reference to the
989 function pcregrep_exit() before it was defined. I am assuming this was
990 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
991 reported by a user. I've moved the definition above the reference.
992
993
994 Version 8.11 10-Dec-2010
995 ------------------------
996
997 1. (*THEN) was not working properly if there were untried alternatives prior
998 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
999 backtracked to try for "b" instead of moving to the next alternative branch
1000 at the same level (in this case, to look for "c"). The Perl documentation
1001 is clear that when (*THEN) is backtracked onto, it goes to the "next
1002 alternative in the innermost enclosing group".
1003
1004 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1005 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1006 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1007 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1008 (*THEN).
1009
1010 3. If \s appeared in a character class, it removed the VT character from
1011 the class, even if it had been included by some previous item, for example
1012 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1013 of \s, but is part of the POSIX "space" class.)
1014
1015 4. A partial match never returns an empty string (because you can always
1016 match an empty string at the end of the subject); however the checking for
1017 an empty string was starting at the "start of match" point. This has been
1018 changed to the "earliest inspected character" point, because the returned
1019 data for a partial match starts at this character. This means that, for
1020 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1021 (previously it gave "no match").
1022
1023 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1024 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1025 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1026 has an implication that the given string is incomplete (because a partial
1027 match is preferred over a full match). For this reason, these items now
1028 give a partial match in this situation. [Aside: previously, the one case
1029 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1030 match rather than a full match, which was wrong by the old rules, but is
1031 now correct.]
1032
1033 6. There was a bug in the handling of #-introduced comments, recognized when
1034 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1035 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1036 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1037 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1038 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1039 places in pcre_compile().
1040
1041 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1042 comments when looking ahead for named forward references to subpatterns,
1043 the only newline sequence it recognized was NL. It now handles newlines
1044 according to the set newline convention.
1045
1046 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1047 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1048 cater for a lack of strerror(). These oversights have been fixed.
1049
1050 9. Added --match-limit and --recursion-limit to pcregrep.
1051
1052 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1053
1054 11. When the -o option was used, pcregrep was setting a return code of 1, even
1055 when matches were found, and --line-buffered was not being honoured.
1056
1057 12. Added an optional parentheses number to the -o and --only-matching options
1058 of pcregrep.
1059
1060 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1061 can match an empty string. The code to do it in pcretest and pcredemo
1062 needed fixing:
1063
1064 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1065 only one byte after an empty string match just before CRLF (this case
1066 just got forgotten; "any" and "anycrlf" were OK).
1067
1068 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1069 mode when an empty string match preceded an ASCII character followed by
1070 a non-ASCII character. (The code for advancing by one character rather
1071 than one byte was nonsense.)
1072
1073 (c) The pcredemo.c sample program did not have any code at all to handle
1074 the cases when CRLF is a valid newline sequence.
1075
1076 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1077 as a starting offset was within the subject string. There is now a new
1078 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1079 negative or greater than the length of the string. In order to test this,
1080 pcretest is extended to allow the setting of negative starting offsets.
1081
1082 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1083 starting offset points to the beginning of a UTF-8 character was
1084 unnecessarily clumsy. I tidied it up.
1085
1086 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1087 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1088
1089 17. Nobody had reported that the --include_dir option, which was added in
1090 release 7.7 should have been called --include-dir (hyphen, not underscore)
1091 for compatibility with GNU grep. I have changed it to --include-dir, but
1092 left --include_dir as an undocumented synonym, and the same for
1093 --exclude-dir, though that is not available in GNU grep, at least as of
1094 release 2.5.4.
1095
1096 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1097 characters from a string of bytes) have been redefined so as not to use
1098 loops, in order to improve performance in some environments. At the same
1099 time, I abstracted some of the common code into auxiliary macros to save
1100 repetition (this should not affect the compiled code).
1101
1102 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1103 compile-time error is now given if \c is not followed by an ASCII
1104 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1105 different, and any byte value is allowed.)
1106
1107 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1108 START_OPTIMIZE option, which is now allowed at compile time - but just
1109 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1110 to pcregrep and other applications that have no direct access to PCRE
1111 options. The new /Y option in pcretest sets this option when calling
1112 pcre_compile().
1113
1114 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1115 back references. Groups containing recursive back references were forced to
1116 be atomic by that change, but in the case of named groups, the amount of
1117 memory required was incorrectly computed, leading to "Failed: internal
1118 error: code overflow". This has been fixed.
1119
1120 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1121 pcretest.c, to avoid build problems in some Borland environments.
1122
1123
1124 Version 8.10 25-Jun-2010
1125 ------------------------
1126
1127 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1128 THEN.
1129
1130 2. (*ACCEPT) was not working when inside an atomic group.
1131
1132 3. Inside a character class, \B is treated as a literal by default, but
1133 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1134 causes the error). The code is unchanged, but I tidied the documentation.
1135
1136 4. Inside a character class, PCRE always treated \R and \X as literals,
1137 whereas Perl faults them if its -w option is set. I have changed PCRE so
1138 that it faults them when PCRE_EXTRA is set.
1139
1140 5. Added support for \N, which always matches any character other than
1141 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1142
1143 6. When compiling pcregrep with newer versions of gcc which may have
1144 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1145 declared with attribute warn_unused_result" were given. Just casting the
1146 result to (void) does not stop the warnings; a more elaborate fudge is
1147 needed. I've used a macro to implement this.
1148
1149 7. Minor change to pcretest.c to avoid a compiler warning.
1150
1151 8. Added four artifical Unicode properties to help with an option to make
1152 \s etc use properties (see next item). The new properties are: Xan
1153 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1154
1155 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1156 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1157 this option. Modified pcretest to add /W to test this facility. Added
1158 REG_UCP to make it available via the POSIX interface.
1159
1160 10. Added --line-buffered to pcregrep.
1161
1162 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1163 studied, and the match started with a letter with a code point greater than
1164 127 whose first byte was different to the first byte of the other case of
1165 the letter, the other case of this starting letter was not recognized
1166 (#976).
1167
1168 12. If a pattern that was studied started with a repeated Unicode property
1169 test, for example, \p{Nd}+, there was the theoretical possibility of
1170 setting up an incorrect bitmap of starting bytes, but fortunately it could
1171 not have actually happened in practice until change 8 above was made (it
1172 added property types that matched character-matching opcodes).
1173
1174 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1175 possible starting bytes for non-anchored patterns.
1176
1177 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1178 \R, and also a number of cases that involve Unicode properties, both
1179 explicit and implicit when PCRE_UCP is set.
1180
1181 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1182 input, it could crash or give wrong results if characters with values
1183 greater than 0xc0 were present in the subject string. (Detail: it assumed
1184 UTF-8 input when processing these items.)
1185
1186 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1187 size_t is 64-bit (#991).
1188
1189 17. Added a check for running out of memory when PCRE is compiled with
1190 --disable-stack-for-recursion (#990).
1191
1192 18. If the last data line in a file for pcretest does not have a newline on
1193 the end, a newline was missing in the output.
1194
1195 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1196 less than 128) in its various bitmaps. However, there is a facility for
1197 generating tables according to the current locale when PCRE is compiled. It
1198 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1199 characters, are recognized by isspace() and therefore were getting set in
1200 these tables, and indeed these tables seem to approximate to ISO 8859. This
1201 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1202 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1203 which of course cannot start UTF-8 characters. I have changed the code so
1204 that only real ASCII characters (less than 128) and the correct starting
1205 bytes for UTF-8 encodings are set for characters greater than 127 when in
1206 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1207 altogether.)
1208
1209 20. Added the /T option to pcretest so as to be able to run tests with non-
1210 standard character tables, thus making it possible to include the tests
1211 used for 19 above in the standard set of tests.
1212
1213 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1214 reference to a subpattern the other side of a comment that contains an
1215 opening parenthesis caused either an internal compiling error, or a
1216 reference to the wrong subpattern.
1217
1218
1219 Version 8.02 19-Mar-2010
1220 ------------------------
1221
1222 1. The Unicode data tables have been updated to Unicode 5.2.0.
1223
1224 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1225 configured.
1226
1227 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1228 original author of that file, following a query about its status.
1229
1230 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1231 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1232
1233 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1234 quantifier applied to a forward-referencing subroutine call, could compile
1235 incorrect code or give the error "internal error: previously-checked
1236 referenced subpattern not found".
1237
1238 6. Both MS Visual Studio and Symbian OS have problems with initializing
1239 variables to point to external functions. For these systems, therefore,
1240 pcre_malloc etc. are now initialized to local functions that call the
1241 relevant global functions.
1242
1243 7. There were two entries missing in the vectors called coptable and poptable
1244 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1245 I've fixed the data, and added a kludgy way of testing at compile time that
1246 the lengths are correct (equal to the number of opcodes).
1247
1248 8. Following on from 7, I added a similar kludge to check the length of the
1249 eint vector in pcreposix.c.
1250
1251 9. Error texts for pcre_compile() are held as one long string to avoid too
1252 much relocation at load time. To find a text, the string is searched,
1253 counting zeros. There was no check for running off the end of the string,
1254 which could happen if a new error number was added without updating the
1255 string.
1256
1257 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1258
1259 11. \K was not working if it appeared in an atomic group or in a group that
1260 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1261 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1262 the assertion is positive, but not if it is negative.
1263
1264 12. Change 11 fortuitously reduced the size of the stack frame used in the
1265 "match()" function of pcre_exec.c by one pointer. Forthcoming
1266 implementation of support for (*MARK) will need an extra pointer on the
1267 stack; I have reserved it now, so that the stack frame size does not
1268 decrease.
1269
1270 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1271 item in branch that calls a recursion is a subroutine call - as in the
1272 second branch in the above example - was incorrectly given the compile-
1273 time error "recursive call could loop indefinitely" because pcre_compile()
1274 was not correctly checking the subroutine for matching a non-empty string.
1275
1276 14. The checks for overrunning compiling workspace could trigger after an
1277 overrun had occurred. This is a "should never occur" error, but it can be
1278 triggered by pathological patterns such as hundreds of nested parentheses.
1279 The checks now trigger 100 bytes before the end of the workspace.
1280
1281 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1282
1283
1284 Version 8.01 19-Jan-2010
1285 ------------------------
1286
1287 1. If a pattern contained a conditional subpattern with only one branch (in
1288 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1289 computed the wrong minimum data length (which is of course zero for such
1290 subpatterns). This could cause incorrect "no match" results.
1291
1292 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1293 the pattern is reset in the first branch, pcre_compile() failed with
1294 "internal error: code overflow at offset...". This happened only when
1295 the reset was to the original external option setting. (An optimization
1296 abstracts leading options settings into an external setting, which was the
1297 cause of this.)
1298
1299 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1300 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1301 assertion pattern did not match (meaning that the assertion was true), it
1302 was incorrectly treated as false if the SKIP had been reached during the
1303 matching. This also applied to assertions used as conditions.
1304
1305 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1306 assertion subpattern, including such a pattern used as a condition,
1307 unpredictable results occurred, instead of the error return
1308 PCRE_ERROR_DFA_UITEM.
1309
1310 5. The C++ GlobalReplace function was not working like Perl for the special
1311 situation when an empty string is matched. It now does the fancy magic
1312 stuff that is necessary.
1313
1314 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1315 removed. (These were left over from very, very early versions of PCRE.)
1316
1317 7. Some cosmetic changes to the code to make life easier when compiling it
1318 as part of something else:
1319
1320 (a) Change DEBUG to PCRE_DEBUG.
1321
1322 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1323 called "current" as "current_branch", to prevent a collision with the
1324 Linux macro when compiled as a kernel module.
1325
1326 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1327 prevent a collision with the Linux macro when compiled as a kernel
1328 module.
1329
1330 8. In pcre_compile() there are some checks for integer overflows that used to
1331 cast potentially large values to (double). This has been changed to that
1332 when building, a check for int64_t is made, and if it is found, it is used
1333 instead, thus avoiding the use of floating point arithmetic. (There is no
1334 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1335 double.
1336
1337 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1338 2005 (difference between two addresses compared to an unsigned value).
1339
1340 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1341 custom one, because of the following reported problem in Windows:
1342
1343 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1344 under Win32.
1345 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1346 therefore missing the function definition.
1347 - The compiler thus generates a "C" signature for the test function.
1348 - The linker fails to find the "C" function.
1349 - PCRE fails to configure if asked to do so against libbz2.
1350
1351 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1352 messages were output:
1353
1354 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1355 rerunning libtoolize, to keep the correct libtool macros in-tree.
1356 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1357
1358 I have done both of these things.
1359
1360 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1361 most of the time, it *can* run out if it is given a pattern that contains a
1362 runaway infinite recursion. I updated the discussion in the pcrestack man
1363 page.
1364
1365 13. Now that we have gone to the x.xx style of version numbers, the minor
1366 version may start with zero. Using 08 or 09 is a bad idea because users
1367 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1368 interpreted as invalid octal numbers. I've updated the previous comment in
1369 configure.ac, and also added a check that gives an error if 08 or 09 are
1370 used.
1371
1372 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1373 causing partial matching to fail when the end of the subject matched \W
1374 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1375
1376 15. There were some discrepancies between the declarations in pcre_internal.h
1377 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1378 their definitions. The declarations used "const uschar *" and the
1379 definitions used USPTR. Even though USPTR is normally defined as "const
1380 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1381 reported that: "This difference in casting confuses some C++ compilers, for
1382 example, SunCC recognizes above declarations as different functions and
1383 generates broken code for hbpcre." I have changed the declarations to use
1384 USPTR.
1385
1386 16. GNU libtool is named differently on some systems. The autogen.sh script now
1387 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1388 (FreeBSD).
1389
1390 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1391 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1392 comment: "Figure out how to create a longlong from a string: strtoll and
1393 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1394 instance, but it only takes 2 args instead of 3!"
1395
1396 18. A subtle bug concerned with back references has been fixed by a change of
1397 specification, with a corresponding code fix. A pattern such as
1398 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1399 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1400 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1401 same bug. Such groups have to be quantified to be useful, or contained
1402 inside another quantified group. (If there's no repetition, the reference
1403 can never match.) The problem arises because, having left the group and
1404 moved on to the rest of the pattern, a later failure that backtracks into
1405 the group uses the captured value from the final iteration of the group
1406 rather than the correct earlier one. I have fixed this in PCRE by forcing
1407 any group that contains a reference to itself to be an atomic group; that
1408 is, there cannot be any backtracking into it once it has completed. This is
1409 similar to recursive and subroutine calls.
1410
1411
1412 Version 8.00 19-Oct-09
1413 ----------------------
1414
1415 1. The table for translating pcre_compile() error codes into POSIX error codes
1416 was out-of-date, and there was no check on the pcre_compile() error code
1417 being within the table. This could lead to an OK return being given in
1418 error.
1419
1420 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1421 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1422 in a Windows environment.
1423
1424 3. The pcregrep --count option prints the count for each file even when it is
1425 zero, as does GNU grep. However, pcregrep was also printing all files when
1426 --files-with-matches was added. Now, when both options are given, it prints
1427 counts only for those files that have at least one match. (GNU grep just
1428 prints the file name in this circumstance, but including the count seems
1429 more useful - otherwise, why use --count?) Also ensured that the
1430 combination -clh just lists non-zero counts, with no names.
1431
1432 4. The long form of the pcregrep -F option was incorrectly implemented as
1433 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1434 but it seems right to fix it, and I didn't think it was worth preserving
1435 the old behaviour.
1436
1437 5. The command line items --regex=pattern and --regexp=pattern were not
1438 recognized by pcregrep, which required --regex pattern or --regexp pattern
1439 (with a space rather than an '='). The man page documented the '=' forms,
1440 which are compatible with GNU grep; these now work.
1441
1442 6. No libpcreposix.pc file was created for pkg-config; there was just
1443 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1444
1445 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1446 when UCP support is not needed, by modifying the Python script that
1447 generates it from Unicode data files. This should not matter if the module
1448 is correctly used as a library, but I received one complaint about 50K of
1449 unwanted data. My guess is that the person linked everything into his
1450 program rather than using a library. Anyway, it does no harm.
1451
1452 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1453 was a minimum greater than 1 for a wide character in a possessive
1454 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1455 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1456 character. Chaos in the form of incorrect output or a compiling loop could
1457 result.
1458
1459 9. The restrictions on what a pattern can contain when partial matching is
1460 requested for pcre_exec() have been removed. All patterns can now be
1461 partially matched by this function. In addition, if there are at least two
1462 slots in the offset vector, the offset of the earliest inspected character
1463 for the match and the offset of the end of the subject are set in them when
1464 PCRE_ERROR_PARTIAL is returned.
1465
1466 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1467 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1468 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1469 and may be more useful for multi-segment matching.
1470
1471 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1472 used to be given if ever the end of the subject was reached; now it is
1473 given only if matching could not proceed because another character was
1474 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1475 string "Z", which now yields "no match" instead of "partial match". In the
1476 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1477 final character ended with (*FAIL).
1478
1479 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1480 if the pattern had a "must contain" character that was already found in the
1481 earlier partial match, unless partial matching was again requested. For
1482 example, with the pattern /dog.(body)?/, the "must contain" character is
1483 "g". If the first part-match was for the string "dog", restarting with
1484 "sbody" failed. This bug has been fixed.
1485
1486 13. The string returned by pcre_dfa_exec() after a partial match has been
1487 changed so that it starts at the first inspected character rather than the
1488 first character of the match. This makes a difference only if the pattern
1489 starts with a lookbehind assertion or \b or \B (\K is not supported by
1490 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1491 matching functions compatible, and I think it's the right thing to do.
1492
1493 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1494 so that the demonstration program is easily available in environments where
1495 PCRE has not been installed from source.
1496
1497 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1498 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1499 library.
1500
1501 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1502 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1503 is not the first non-POSIX option to be added. Clearly some people find
1504 these options useful.
1505
1506 17. If a caller to the POSIX matching function regexec() passes a non-zero
1507 value for nmatch with a NULL value for pmatch, the value of
1508 nmatch is forced to zero.
1509
1510 18. RunGrepTest did not have a test for the availability of the -u option of
1511 the diff command, as RunTest does. It now checks in the same way as
1512 RunTest, and also checks for the -b option.
1513
1514 19. If an odd number of negated classes containing just a single character
1515 interposed, within parentheses, between a forward reference to a named
1516 subpattern and the definition of the subpattern, compilation crashed with
1517 an internal error, complaining that it could not find the referenced
1518 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1519 [The bug was that it was starting one character too far in when skipping
1520 over the character class, thus treating the ] as data rather than
1521 terminating the class. This meant it could skip too much.]
1522
1523 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1524 /g option in pcretest when the pattern contains \K, which makes it possible
1525 to have an empty string match not at the start, even when the pattern is
1526 anchored. Updated pcretest and pcredemo to use this option.
1527
1528 21. If the maximum number of capturing subpatterns in a recursion was greater
1529 than the maximum at the outer level, the higher number was returned, but
1530 with unset values at the outer level. The correct (outer level) value is
1531 now given.
1532
1533 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1534 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1535 make it do so. The string so far is captured, making this feature
1536 compatible with Perl.
1537
1538 23. The tests have been re-organized, adding tests 11 and 12, to make it
1539 possible to check the Perl 5.10 features against Perl 5.10.
1540
1541 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1542 pattern matches a fixed length string. PCRE did not allow this; now it
1543 does. Neither allows recursion.
1544
1545 25. I finally figured out how to implement a request to provide the minimum
1546 length of subject string that was needed in order to match a given pattern.
1547 (It was back references and recursion that I had previously got hung up
1548 on.) This code has now been added to pcre_study(); it finds a lower bound
1549 to the length of subject needed. It is not necessarily the greatest lower
1550 bound, but using it to avoid searching strings that are too short does give
1551 some useful speed-ups. The value is available to calling programs via
1552 pcre_fullinfo().
1553
1554 26. While implementing 25, I discovered to my embarrassment that pcretest had
1555 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1556 study optimizations had never been tested with that matching function.
1557 Oops. What is worse, even when it was passed study data, there was a bug in
1558 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1559 were also very few tests of studied patterns with pcre_dfa_exec().
1560
1561 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1562 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1563 on the other side of the coin, they are no longer allowed to have different
1564 names, because these cannot be distinguished in PCRE, and this has caused
1565 confusion. (This is a difference from Perl.)
1566
1567 28. When duplicate subpattern names are present (necessarily with different
1568 numbers, as required by 27 above), and a test is made by name in a
1569 conditional pattern, either for a subpattern having been matched, or for
1570 recursion in such a pattern, all the associated numbered subpatterns are
1571 tested, and the overall condition is true if the condition is true for any
1572 one of them. This is the way Perl works, and is also more like the way
1573 testing by number works.
1574
1575
1576 Version 7.9 11-Apr-09
1577 ---------------------
1578
1579 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1580 (pcretest), all targets were linked against these libraries. This included
1581 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1582 libraries. This caused unwanted dependencies to be created. This problem
1583 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1584 pcretest is linked with readline.
1585
1586 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1587 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1588 moved outside it again, because FALSE and TRUE are already defined in AIX,
1589 but BOOL is not.
1590
1591 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1592 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1593
1594 4. The pcregrep documentation said spaces were inserted as well as colons (or
1595 hyphens) following file names and line numbers when outputting matching
1596 lines. This is not true; no spaces are inserted. I have also clarified the
1597 wording for the --colour (or --color) option.
1598
1599 5. In pcregrep, when --colour was used with -o, the list of matching strings
1600 was not coloured; this is different to GNU grep, so I have changed it to be
1601 the same.
1602
1603 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1604 each matching line was coloured. Now it goes on to look for further matches
1605 of any of the test patterns, which is the same behaviour as GNU grep.
1606
1607 7. A pattern that could match an empty string could cause pcregrep to loop; it
1608 doesn't make sense to accept an empty string match in pcregrep, so I have
1609 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1610 seems to be how GNU grep behaves.
1611
1612 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1613 start or after a newline", because the conditional assertion was not being
1614 correctly handled. The rule now is that both the assertion and what follows
1615 in the first alternative must satisfy the test.
1616
1617 9. If auto-callout was enabled in a pattern with a conditional group whose
1618 condition was an assertion, PCRE could crash during matching, both with
1619 pcre_exec() and pcre_dfa_exec().
1620
1621 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1622 used for matching.
1623
1624 11. Unicode property support in character classes was not working for
1625 characters (bytes) greater than 127 when not in UTF-8 mode.
1626
1627 12. Added the -M command line option to pcretest.
1628
1629 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1630
1631 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1632
1633 16. Added comments and documentation about mis-use of no_arg in the C++
1634 wrapper.
1635
1636 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1637 from Martin Jerabek that uses macro names for all relevant character and
1638 string constants.
1639
1640 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1641 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1642 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1643 these, but not everybody uses configure.
1644
1645 19. A conditional group that had only one branch was not being correctly
1646 recognized as an item that could match an empty string. This meant that an
1647 enclosing group might also not be so recognized, causing infinite looping
1648 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1649 with the subject "ab", where knowledge that the repeated group can match
1650 nothing is needed in order to break the loop.
1651
1652 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1653 exec(), but without supplying a callout function, matching went wrong.
1654
1655 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1656 leak if the size of the offset vector was greater than 30. When the vector
1657 is smaller, the saved offsets during recursion go onto a local stack
1658 vector, but for larger vectors malloc() is used. It was failing to free
1659 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1660 error, in fact).
1661
1662 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1663 heapframe that is used only when UTF-8 support is enabled. This caused no
1664 problem, but was untidy.
1665
1666 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1667 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1668 included within another project.
1669
1670 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1671 slightly modified by me:
1672
1673 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1674 not building pcregrep.
1675
1676 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1677 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1678
1679 25. Forward references, both numeric and by name, in patterns that made use of
1680 duplicate group numbers, could behave incorrectly or give incorrect errors,
1681 because when scanning forward to find the reference group, PCRE was not
1682 taking into account the duplicate group numbers. A pattern such as
1683 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1684
1685 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1686 the feature of a custom pointer more persuasive (as requested by a user).
1687
1688 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1689 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1690 pre-defined.
1691
1692 28. Added support for (*UTF8) at the start of a pattern.
1693
1694 29. Arrange for flags added by the "release type" setting in CMake to be shown
1695 in the configuration summary.
1696
1697
1698 Version 7.8 05-Sep-08
1699 ---------------------
1700
1701 1. Replaced UCP searching code with optimized version as implemented for Ad
1702 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1703 stage table and inline lookup instead of a function, giving speed ups of 2
1704 to 5 times on some simple patterns that I tested. Permission was given to
1705 distribute the MultiStage2.py script that generates the tables (it's not in
1706 the tarball, but is in the Subversion repository).
1707
1708 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1709 scripts.
1710
1711 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1712 a group with a zero qualifier. The result of the study could be incorrect,
1713 or the function might crash, depending on the pattern.
1714
1715 4. Caseless matching was not working for non-ASCII characters in back
1716 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1717 It now works when Unicode Property Support is available.
1718
1719 5. In pcretest, an escape such as \x{de} in the data was always generating
1720 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1721 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1722 truncation.
1723
1724 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1725
1726 7. Added two (int) casts to pcregrep when printing the difference of two
1727 pointers, in case they are 64-bit values.
1728
1729 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1730 test 2 if it fails.
1731
1732 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1733 and a #define of that name to empty if it is not externally set. This is to
1734 allow users of MSVC to set it if necessary.
1735
1736 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1737 the convenience functions in the pcre_get.c source file.
1738
1739 11. An option change at the start of a pattern that had top-level alternatives
1740 could cause overwriting and/or a crash. This command provoked a crash in
1741 some environments:
1742
1743 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1744
1745 This potential security problem was recorded as CVE-2008-2371.
1746
1747 12. For a pattern where the match had to start at the beginning or immediately
1748 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1749 pcre_dfa_exec() could read past the end of the passed subject if there was
1750 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1751 pcretest so that it places the subject at the end of its malloc-ed buffer.
1752
1753 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1754 exec() might read past the end of the data buffer in UTF-8 mode.
1755
1756 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1757 the data contained the byte 0x85 as part of a UTF-8 character within its
1758 first line. This applied both to normal and DFA matching.
1759
1760 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1761 /^[^d]*?$/8 failed to match "abc".
1762
1763 16. Added a missing copyright notice to pcrecpp_internal.h.
1764
1765 17. Make it more clear in the documentation that values returned from
1766 pcre_exec() in ovector are byte offsets, not character counts.
1767
1768 18. Tidied a few places to stop certain compilers from issuing warnings.
1769
1770 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1771 supplied by Stefan Weber. I made a further small update for 7.8 because
1772 there is a change of source arrangements: the pcre_searchfuncs.c module is
1773 replaced by pcre_ucd.c.
1774
1775
1776 Version 7.7 07-May-08
1777 ---------------------
1778
1779 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1780 a string to a long long, pretend we don't even have a long long." This is
1781 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1782
1783 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1784 pre-7.6 versions, which defined a global no_arg variable instead of putting
1785 it in the RE class. (See also #8 below.)
1786
1787 3. Remove a line of dead code, identified by coverity and reported by Nuno
1788 Lopes.
1789
1790 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1791
1792 (1) The include/exclude patterns were being applied to the whole pathnames
1793 of files, instead of just to the final components.
1794
1795 (2) If there was more than one level of directory, the subdirectories were
1796 skipped unless they satisfied the include/exclude conditions. This is
1797 inconsistent with GNU grep (and could even be seen as contrary to the
1798 pcregrep specification - which I improved to make it absolutely clear).
1799 The action now is always to scan all levels of directory, and just
1800 apply the include/exclude patterns to regular files.
1801
1802 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1803 --exclude_dir in the tests to avoid scanning .svn directories.
1804
1805 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1806 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1807 doesn't support NULs in patterns.
1808
1809 7. Added some missing "const"s to declarations of static tables in
1810 pcre_compile.c and pcre_dfa_exec.c.
1811
1812 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1813 caused by fix #2 above. (Subsequently also a second patch to fix the
1814 first patch. And a third patch - this was a messy problem.)
1815
1816 9. Applied Craig's patch to remove the use of push_back().
1817
1818 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1819 matching function regexec().
1820
1821 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1822 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1823 references. PCRE supports relative numbers with this syntax (I don't think
1824 Oniguruma does).
1825
1826 12. Previously, a group with a zero repeat such as (...){0} was completely
1827 omitted from the compiled regex. However, this means that if the group
1828 was called as a subroutine from elsewhere in the pattern, things went wrong
1829 (an internal error was given). Such groups are now left in the compiled
1830 pattern, with a new opcode that causes them to be skipped at execution
1831 time.
1832
1833 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1834 to the way PCRE behaves:
1835
1836 (a) A lone ] character is dis-allowed (Perl treats it as data).
1837
1838 (b) A back reference to an unmatched subpattern matches an empty string
1839 (Perl fails the current match path).
1840
1841 (c) A data ] in a character class must be notated as \] because if the
1842 first data character in a class is ], it defines an empty class. (In
1843 Perl it is not possible to have an empty class.) The empty class []
1844 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1845 The negative empty class [^] matches any one character, independently
1846 of the DOTALL setting.
1847
1848 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1849 non-existent subpattern following a character class starting with ']' and
1850 containing () gave an internal compiling error instead of "reference to
1851 non-existent subpattern". Fortunately, when the pattern did exist, the
1852 compiled code was correct. (When scanning forwards to check for the
1853 existencd of the subpattern, it was treating the data ']' as terminating
1854 the class, so got the count wrong. When actually compiling, the reference
1855 was subsequently set up correctly.)
1856
1857 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1858 it was being rejected as not supported by pcre_dfa_exec(), even though
1859 other assertions are supported. I have made pcre_dfa_exec() support
1860 (*FAIL).
1861
1862 16. The implementation of 13c above involved the invention of a new opcode,
1863 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1864 cannot be changed at match time, I realized I could make a small
1865 improvement to matching performance by compiling OP_ALLANY instead of
1866 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1867 on the OP_ANY path.
1868
1869 17. Compiling pcretest on Windows with readline support failed without the
1870 following two fixes: (1) Make the unistd.h include conditional on
1871 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1872
1873 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1874 ncurses library to be included for pcretest when ReadLine support is
1875 requested, but also to allow for it to be overridden. This patch came from
1876 Daniel Bergström.
1877
1878 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1879 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1880 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1881 spotting this.
1882
1883
1884 Version 7.6 28-Jan-08
1885 ---------------------
1886
1887 1. A character class containing a very large number of characters with
1888 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1889 overflow.
1890
1891 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1892 HAVE_LONG_LONG is not defined.
1893
1894 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1895 bring them up to date and include new features. This patch includes:
1896
1897 - Fixed PH's badly added libz and libbz2 support.
1898 - Fixed a problem with static linking.
1899 - Added pcredemo. [But later removed - see 7 below.]
1900 - Fixed dftables problem and added an option.
1901 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1902 HAVE_LONG_LONG.
1903 - Added readline support for pcretest.
1904 - Added an listing of the option settings after cmake has run.
1905
1906 4. A user submitted a patch to Makefile that makes it easy to create
1907 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1908 Makefile.am that cause it to include this special target, without
1909 affecting anything else. Note that the same mingw target plus all
1910 the other distribution libraries and programs are now supported
1911 when configuring with CMake (see 6 below) instead of with
1912 Configure/Make.
1913
1914 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1915 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1916 exported in the Windows port". It has not yet been confirmed that the patch
1917 solves the problem, but it does no harm.
1918
1919 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1920 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1921 with CMake, and also correct the comment about stack recursion.
1922
1923 7. Remove the automatic building of pcredemo from the ./configure system and
1924 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1925 of a program that users should build themselves after PCRE is installed, so
1926 building it automatically is not really right. What is more, it gave
1927 trouble in some build environments.
1928
1929 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1930
1931
1932 Version 7.5 10-Jan-08
1933 ---------------------
1934
1935 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1936 values in parens when parsing an RE using the C++ wrapper."
1937
1938 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1939 Characters greater than 255 were excluded from the class instead of being
1940 included.
1941
1942 3. The same bug as (2) above applied to negated POSIX classes such as
1943 [:^space:].
1944
1945 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1946 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1947 I have changed it.
1948
1949 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1950 first named subpattern) and a construct such as (?&a) would reference the
1951 first named subpattern whose name started with "a" (in other words, the
1952 length check was missing). Both these problems are fixed. "Subpattern name
1953 expected" is now given for (?&) (a zero-length name), and this patch also
1954 makes it give the same error for \k'' (previously it complained that that
1955 was a reference to a non-existent subpattern).
1956
1957 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1958 this is right because (?- can be followed by option settings as well as by
1959 digits. I have, however, made the messages clearer.
1960
1961 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1962 than the number used in the conditional) now cause a compile-time error.
1963 This is actually not compatible with Perl, which accepts such patterns, but
1964 treats the conditional as always being FALSE (as PCRE used to), but it
1965 seems to me that giving a diagnostic is better.
1966
1967 8. Change "alphameric" to the more common word "alphanumeric" in comments
1968 and messages.
1969
1970 9. Fix two occurrences of "backslash" in comments that should have been
1971 "backspace".
1972
1973 10. Remove two redundant lines of code that can never be obeyed (their function
1974 was moved elsewhere).
1975
1976 11. The program that makes PCRE's Unicode character property table had a bug
1977 which caused it to generate incorrect table entries for sequences of
1978 characters that have the same character type, but are in different scripts.
1979 It amalgamated them into a single range, with the script of the first of
1980 them. In other words, some characters were in the wrong script. There were
1981 thirteen such cases, affecting characters in the following ranges:
1982
1983 U+002b0 - U+002c1
1984 U+0060c - U+0060d
1985 U+0061e - U+00612
1986 U+0064b - U+0065e
1987 U+0074d - U+0076d
1988 U+01800 - U+01805
1989 U+01d00 - U+01d77
1990 U+01d9b - U+01dbf
1991 U+0200b - U+0200f
1992 U+030fc - U+030fe
1993 U+03260 - U+0327f
1994 U+0fb46 - U+0fbb1
1995 U+10450 - U+1049d
1996
1997 12. The -o option (show only the matching part of a line) for pcregrep was not
1998 compatible with GNU grep in that, if there was more than one match in a
1999 line, it showed only the first of them. It now behaves in the same way as
2000 GNU grep.
2001
2002 13. If the -o and -v options were combined for pcregrep, it printed a blank
2003 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2004 does the same. The return code can be used to tell if there were any
2005 non-matching lines.
2006
2007 14. Added --file-offsets and --line-offsets to pcregrep.
2008
2009 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2010 infinitely looping recursion. The bug was that positive lookaheads were not
2011 being skipped when checking for a possible empty match (negative lookaheads
2012 and both kinds of lookbehind were skipped).
2013
2014 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2015 inclusion of <windows.h> to before rather than after the definition of
2016 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2017
2018 17. Specifying a possessive quantifier with a specific limit for a Unicode
2019 character property caused pcre_compile() to compile bad code, which led at
2020 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2021 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2022 caused the error; without that there was no problem.
2023
2024 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2025
2026 19. Added --enable-pcretest-libreadline.
2027
2028 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2029 RE::GlobalReplace(). As a result, the number of replacements returned was
2030 double what it should be. I removed one of the increments, but Craig sent a
2031 later patch that removed the other one (the right fix) and added unit tests
2032 that check the return values (which was not done before).
2033
2034 21. Several CMake things:
2035
2036 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2037 the names libpcre and libpcreposix, not just pcre and pcreposix.
2038
2039 (2) The above change means that pcretest and pcregrep are now correctly
2040 linked with the newly-built libraries, not previously installed ones.
2041
2042 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2043
2044 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2045 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2046 UTF-8 newline character). The key issue is that the pattern starts .*;
2047 this means that the match must be either at the beginning, or after a
2048 newline. The bug was in the code for advancing after a failed match and
2049 checking that the new position followed a newline. It was not taking
2050 account of UTF-8 characters correctly.
2051
2052 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2053 character classes. PCRE was not treating the sequence [:...:] as a
2054 character class unless the ... were all letters. Perl, however, seems to
2055 allow any characters between [: and :], though of course it rejects as
2056 unknown any "names" that contain non-letters, because all the known class
2057 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2058 for example, whereas PCRE did not - it did not recognize a POSIX character
2059 class. This seemed a bit dangerous, so the code has been changed to be
2060 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2061 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2062 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2063 Perl does, and where it didn't before.
2064
2065 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2066 Windows environments %n is disabled by default.
2067
2068
2069 Version 7.4 21-Sep-07
2070 ---------------------
2071
2072 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2073 means that a class such as [\s] counted as "explicit reference to CR or
2074 LF". That isn't really right - the whole point of the change was to try to
2075 help when there was an actual mention of one of the two characters. So now
2076 the change happens only if \r or \n (or a literal CR or LF) character is
2077 encountered.
2078
2079 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2080 of both had grown to the point where there were only 3 bits left.
2081 Fortunately, there was spare space in the data structure, and so I have
2082 moved the internal flags into a new 16-bit field to free up more option
2083 bits.
2084
2085 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2086 but did not set the internal JCHANGED flag - either of these is enough to
2087 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2088 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2089 start sets both bits.
2090
2091 4. Added options (at build time, compile time, exec time) to change \R from
2092 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2093
2094 5. doc/pcresyntax.html was missing from the distribution.
2095
2096 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2097 compatibility, even though it is no longer used.
2098
2099 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2100 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2101 windows.h file is present (where different names are used). [This was
2102 reversed later after testing - see 16 below.]
2103
2104 8. Changed all #include <config.h> to #include "config.h". There were also
2105 some further <pcre.h> cases that I changed to "pcre.h".
2106
2107 9. When pcregrep was used with the --colour option, it missed the line ending
2108 sequence off the lines that it output.
2109
2110 10. It was pointed out to me that arrays of string pointers cause lots of
2111 relocations when a shared library is dynamically loaded. A technique of
2112 using a single long string with a table of offsets can drastically reduce
2113 these. I have refactored PCRE in four places to do this. The result is
2114 dramatic:
2115
2116 Originally: 290
2117 After changing UCP table: 187
2118 After changing error message table: 43
2119 After changing table of "verbs" 36
2120 After changing table of Posix names 22
2121
2122 Thanks to the folks working on Gregex for glib for this insight.
2123
2124 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2125 unicode-properties was also set.
2126
2127 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2128
2129 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2130 checked only for CRLF.
2131
2132 14. Added casts to pcretest.c to avoid compiler warnings.
2133
2134 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2135
2136 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2137 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2138 entirely. This removes changes made in 7 above.
2139
2140 17. The CMake files have been updated, and there is now more information about
2141 building with CMake in the NON-UNIX-USE document.
2142
2143
2144 Version 7.3 28-Aug-07
2145 ---------------------
2146
2147 1. In the rejigging of the build system that eventually resulted in 7.1, the
2148 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2149 brackets there is not right, since it causes compilers to look for an
2150 installed pcre.h, not the version that is in the source that is being
2151 compiled (which of course may be different). I have changed it back to:
2152
2153 #include "pcre.h"
2154
2155 I have a vague recollection that the change was concerned with compiling in
2156 different directories, but in the new build system, that is taken care of
2157 by the VPATH setting the Makefile.
2158
2159 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2160 when the subject happened to end in the byte 0x85 (e.g. if the last
2161 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2162 characters but of course it shouldn't be taken as a newline when it is part
2163 of another character. The bug was that, for an unlimited repeat of . in
2164 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2165 characters when looking for a newline.
2166
2167 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2168
2169 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2170 in debug output.
2171
2172 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2173 long printing in the pcrecpp unittest when running under MinGW.
2174
2175 6. ESC_K was left out of the EBCDIC table.
2176
2177 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2178 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2179 limit also applies to "virtual nesting" when a pattern is recursive, and in
2180 this case 1000 isn't so big. I have been able to remove this limit at the
2181 expense of backing off one optimization in certain circumstances. Normally,
2182 when pcre_exec() would call its internal match() function recursively and
2183 immediately return the result unconditionally, it uses a "tail recursion"
2184 feature to save stack. However, when a subpattern that can match an empty
2185 string has an unlimited repetition quantifier, it no longer makes this
2186 optimization. That gives it a stack frame in which to save the data for
2187 checking that an empty string has been matched. Previously this was taken
2188 from the 1000-entry workspace that had been reserved. So now there is no
2189 explicit limit, but more stack is used.
2190
2191 8. Applied Daniel's patches to solve problems with the import/export magic
2192 syntax that is required for Windows, and which was going wrong for the
2193 pcreposix and pcrecpp parts of the library. These were overlooked when this
2194 problem was solved for the main library.
2195
2196 9. There were some crude static tests to avoid integer overflow when computing
2197 the size of patterns that contain repeated groups with explicit upper
2198 limits. As the maximum quantifier is 65535, the maximum group length was
2199 set at 30,000 so that the product of these two numbers did not overflow a
2200 32-bit integer. However, it turns out that people want to use groups that
2201 are longer than 30,000 bytes (though not repeat them that many times).
2202 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2203 made it possible to implement the integer overflow checks in a much more
2204 dynamic way, which I have now done. The artificial limitation on group
2205 length has been removed - we now have only the limit on the total length of
2206 the compiled pattern, which depends on the LINK_SIZE setting.
2207
2208 10. Fixed a bug in the documentation for get/copy named substring when
2209 duplicate names are permitted. If none of the named substrings are set, the
2210 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2211 empty string.
2212
2213 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2214 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2215 because the ] is interpreted as the first data character and the
2216 terminating ] is not found. PCRE has been made compatible with Perl in this
2217 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2218 cause memory overwriting.
2219
2220 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2221 string has been matched (to stop an infinite loop). It was not recognizing
2222 a conditional subpattern that could match an empty string if that
2223 subpattern was within another subpattern. For example, it looped when
2224 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2225 condition was not nested. This bug has been fixed.
2226
2227 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2228 past the start of the subject in the presence of bytes with the top bit
2229 set, for example "\x8aBCD".
2230
2231 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2232 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2233
2234 14. Optimized (?!) to (*FAIL).
2235
2236 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2237 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2238 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2239 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2240 does: it's just the validity check that is more restrictive.
2241
2242 16. Inserted checks for integer overflows during escape sequence (backslash)
2243 processing, and also fixed erroneous offset values for syntax errors during
2244 backslash processing.
2245
2246 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2247 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2248
2249 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2250 caused an overrun.
2251
2252 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2253 something other than just ASCII characters) inside a group that had an
2254 unlimited repeat caused a loop at compile time (while checking to see
2255 whether the group could match an empty string).
2256
2257 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2258 [\P{Any}] did so. (Error in the code for printing property names.)
2259
2260 21. An orphan \E inside a character class could cause a crash.
2261
2262 22. A repeated capturing bracket such as (A)? could cause a wild memory
2263 reference during compilation.
2264
2265 23. There are several functions in pcre_compile() that scan along a compiled
2266 expression for various reasons (e.g. to see if it's fixed length for look
2267 behind). There were bugs in these functions when a repeated \p or \P was
2268 present in the pattern. These operators have additional parameters compared
2269 with \d, etc, and these were not being taken into account when moving along
2270 the compiled data. Specifically:
2271
2272 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2273 length.
2274
2275 (b) An item such as \pL+ within a repeated group could cause crashes or
2276 loops.
2277
2278 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2279 "reference to non-existent subpattern" error.
2280
2281 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2282
2283 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2284 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2285
2286 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2287
2288 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2289 character were causing crashes (broken optimization).
2290
2291 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2292 \p or \P) caused a compile-time loop.
2293
2294 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2295 break. For example, the unstudied pattern [\r\n]A does not match the string
2296 "\r\nA" because change 7.0/46 below moves the current point on by two
2297 characters after failing to match at the start. However, the pattern \nA
2298 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2299 the same is true. There doesn't seem any very clean way out of this, but
2300 what I have chosen to do makes the common cases work: PCRE now takes note
2301 of whether there can be an explicit match for \r or \n anywhere in the
2302 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2303 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2304 pattern has explicit CR or LF references.
2305
2306 29. Added (*CR) etc for changing newline setting at start of pattern.
2307
2308
2309 Version 7.2 19-Jun-07
2310 ---------------------
2311
2312 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2313 which is apparently normally available under Windows.
2314
2315 2. Re-jig the pcregrep tests with different newline settings in an attempt
2316 to make them independent of the local environment's newline setting.
2317
2318 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2319
2320 4. Some of the "internals" tests were previously cut out when the link size
2321 was not 2, because the output contained actual offsets. The recent new
2322 "Z" feature of pcretest means that these can be cut out, making the tests
2323 usable with all link sizes.
2324
2325 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2326 stack recursion. This gives a massive performance boost under BSD, but just
2327 a small improvement under Linux. However, it saves one field in the frame
2328 in all cases.
2329
2330 6. Added more features from the forthcoming Perl 5.10:
2331
2332 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2333 recursion call. It refers to the nth most recently opened parentheses.
2334
2335 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2336 to be opened parentheses.
2337
2338 (c) Conditions that refer to capturing parentheses can be specified
2339 relatively, for example, (?(-2)... or (?(+3)...
2340
2341 (d) \K resets the start of the current match so that everything before
2342 is not part of it.
2343
2344 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2345
2346 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2347 reference syntax.
2348
2349 (g) (?| introduces a group in which the numbering of parentheses in each
2350 alternative starts with the same number.
2351
2352 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2353
2354 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2355 PCRE_INFO_JCHANGED.
2356
2357 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2358 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2359 for detecting groups that can match an empty string.
2360
2361 9. A pattern with a very large number of alternatives (more than several
2362 hundred) was running out of internal workspace during the pre-compile
2363 phase, where pcre_compile() figures out how much memory will be needed. A
2364 bit of new cunning has reduced the workspace needed for groups with
2365 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2366 workspace instead of running out of the 4096 that are available.
2367
2368 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2369
2370 11. Applied patch from Google to remove an optimization that didn't quite work.
2371 The report of the bug said:
2372
2373 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2374 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2375 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2376
2377 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2378 it matched the wrong number of bytes.
2379
2380
2381 Version 7.1 24-Apr-07
2382 ---------------------
2383
2384 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2385 that is more "standard", making use of automake and other Autotools. There
2386 is some re-arrangement of the files and adjustment of comments consequent
2387 on this.
2388
2389 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2390 for recursive directory scanning broke on some systems because the files
2391 are not scanned in any specific order and on different systems the order
2392 was different. A call to "sort" has been inserted into RunGrepTest for the
2393 approprate test as a short-term fix. In the longer term there may be an
2394 alternative.
2395
2396 3. I had an email from Eric Raymond about problems translating some of PCRE's
2397 man pages to HTML (despite the fact that I distribute HTML pages, some
2398 people do their own conversions for various reasons). The problems
2399 concerned the use of low-level troff macros .br and .in. I have therefore
2400 removed all such uses from the man pages (some were redundant, some could
2401 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2402 HTML has been updated to handle .nf/.fi and to complain if it encounters
2403 .br or .in.
2404
2405 4. Updated comments in configure.ac that get placed in config.h.in and also
2406 arranged for config.h to be included in the distribution, with the name
2407 config.h.generic, for the benefit of those who have to compile without
2408 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2409
2410 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2411 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2412 makevp.bat for the current PCRE, using the additional files
2413 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2414
2415 6. A Windows user reported a minor discrepancy with test 2, which turned out
2416 to be caused by a trailing space on an input line that had got lost in his
2417 copy. The trailing space was an accident, so I've just removed it.
2418
2419 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2420 that is needed.
2421
2422 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2423 as "const" (a) because they are and (b) because it helps the PHP
2424 maintainers who have recently made a script to detect big data structures
2425 in the php code that should be moved to the .rodata section. I remembered
2426 to update Builducptable as well, so it won't revert if ucptable.h is ever
2427 re-created.
2428
2429 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2430 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2431 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2432 support is not required. This saves 1.5-2K of code, which is important in
2433 some applications.
2434
2435 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2436 so as not to refer to the tables, even though these functions will never be
2437 called when UTF-8 support is disabled. Otherwise there are problems with a
2438 shared library.
2439
2440 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2441
2442 (a) It was defining its arguments as char * instead of void *.
2443
2444 (b) It was assuming that all moves were upwards in memory; this was true
2445 a long time ago when I wrote it, but is no longer the case.
2446
2447 The emulated memove() is provided for those environments that have neither
2448 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2449 is clearly not the case, as these two bugs were recently reported.
2450
2451 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2452 and Detrail to create the HTML documentation, the .txt form of the man
2453 pages, and it removes trailing spaces from listed files. It also creates
2454 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2455 case, it wraps all the #defines with #ifndefs. This script should be run
2456 before "make dist".
2457
2458 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2459 with Unicode property support.
2460
2461 (a) For a maximizing quantifier, if the two different cases of the
2462 character were of different lengths in their UTF-8 codings (there are
2463 some cases like this - I found 11), and the matching function had to
2464 back up over a mixture of the two cases, it incorrectly assumed they
2465 were both the same length.
2466
2467 (b) When PCRE was configured to use the heap rather than the stack for
2468 recursion during matching, it was not correctly preserving the data for
2469 the other case of a UTF-8 character when checking ahead for a match
2470 while processing a minimizing repeat. If the check also involved
2471 matching a wide character, but failed, corruption could cause an
2472 erroneous result when trying to check for a repeat of the original
2473 character.
2474
2475 13. Some tidying changes to the testing mechanism:
2476
2477 (a) The RunTest script now detects the internal link size and whether there
2478 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2479 values substituted by "configure". (The RunGrepTest script already did
2480 this for UTF-8.) The configure.ac script no longer substitutes the
2481 relevant variables.
2482
2483 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2484 with length and offset values. This means that the output is different
2485 for different internal link sizes. Test 2 is skipped for link sizes
2486 other than 2 because of this, bypassing the problem. Unfortunately,
2487 there was also a test in test 3 (the locale tests) that used /B and
2488 failed for link sizes other than 2. Rather than cut the whole test out,
2489 I have added a new /Z option to pcretest that replaces the length and
2490 offset values with spaces. This is now used to make test 3 independent
2491 of link size. (Test 2 will be tidied up later.)
2492
2493 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2494 segmentation fault instead of returning the appropriate error message.
2495
2496 15. In multiline mode when the newline sequence was set to "any", the pattern
2497 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2498 This doesn't seem right; it now treats the CRLF combination as the line
2499 ending, and so does not match in that case. It's only a pattern such as ^$
2500 that would hit this one: something like ^ABC$ would have failed after \r
2501 and then tried again after \r\n.
2502
2503 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2504 in an attempt to make files that differ only in their line terminators
2505 compare equal. This works on Linux.
2506
2507 17. Under certain error circumstances pcregrep might try to free random memory
2508 as it exited. This is now fixed, thanks to valgrind.
2509
2510 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2511 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2512 was because its rules for how to advance for /g after matching an empty
2513 string at the end of a line did not allow for this case. They now check for
2514 it specially.
2515
2516 20. pcretest is supposed to handle patterns and data of any length, by
2517 extending its buffers when necessary. It was getting this wrong when the
2518 buffer for a data line had to be extended.
2519
2520 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2521 CRLF as a newline sequence.
2522
2523 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2524 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2525 I have nevertheless tidied it up.
2526
2527 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2528
2529 24. Added a man page for pcre-config.
2530
2531
2532 Version 7.0 19-Dec-06
2533 ---------------------
2534
2535 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2536 moving to gcc 4.1.1.
2537
2538 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2539 sys/time.h, which is documented as needed for this function. It doesn't
2540 seem to matter on Linux, but it showed up on some releases of OS X.
2541
2542 3. It seems that there are systems where bytes whose values are greater than
2543 127 match isprint() in the "C" locale. The "C" locale should be the
2544 default when a C program starts up. In most systems, only ASCII printing
2545 characters match isprint(). This difference caused the output from pcretest
2546 to vary, making some of the tests fail. I have changed pcretest so that:
2547
2548 (a) When it is outputting text in the compiled version of a pattern, bytes
2549 other than 32-126 are always shown as hex escapes.
2550
2551 (b) When it is outputting text that is a matched part of a subject string,
2552 it does the same, unless a different locale has been set for the match
2553 (using the /L modifier). In this case, it uses isprint() to decide.
2554
2555 4. Fixed a major bug that caused incorrect computation of the amount of memory
2556 required for a compiled pattern when options that changed within the
2557 pattern affected the logic of the preliminary scan that determines the
2558 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2559 that the computed length was too small. The symptoms of this bug were
2560 either the PCRE error "internal error: code overflow" from pcre_compile(),
2561 or a glibc crash with a message such as "pcretest: free(): invalid next
2562 size (fast)". Examples of patterns that provoked this bug (shown in
2563 pcretest format) are:
2564
2565 /(?-x: )/x
2566 /(?x)(?-x: \s*#\s*)/
2567 /((?i)[\x{c0}])/8
2568 /(?i:[\x{c0}])/8
2569
2570 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2571 is now done differently.
2572
2573 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2574 wrapper classes; (b) implement a new function in the C++ scanner that is
2575 more efficient than the old way of doing things because it avoids levels of
2576 recursion in the regex matching; (c) add a paragraph to the documentation
2577 for the FullMatch() function.
2578
2579 6. The escape sequence \n was being treated as whatever was defined as
2580 "newline". Not only was this contrary to the documentation, which states
2581 that \n is character 10 (hex 0A), but it also went horribly wrong when
2582 "newline" was defined as CRLF. This has been fixed.
2583
2584 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2585 was being set to -1 for the "end of line" case (supposedly a value that no
2586 character can have). Though this value is never used (the check for end of
2587 line is "zero bytes in current character"), it caused compiler complaints.
2588 I've changed it to 0xffffffff.
2589
2590 8. In pcre_version.c, the version string was being built by a sequence of
2591 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2592 string (as it is for production releases) called a macro with an empty
2593 argument. The C standard says the result of this is undefined. The gcc
2594 compiler treats it as an empty string (which was what was wanted) but it is
2595 reported that Visual C gives an error. The source has been hacked around to
2596 avoid this problem.
2597
2598 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2599 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2600 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2601 of them did).
2602
2603 10. Originally, pcretest opened its input and output without "b"; then I was
2604 told that "b" was needed in some environments, so it was added for release
2605 5.0 to both the input and output. (It makes no difference on Unix-like
2606 systems.) Later I was told that it is wrong for the input on Windows. I've
2607 now abstracted the modes into two macros, to make it easier to fiddle with
2608 them, and removed "b" from the input mode under Windows.
2609
2610 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2611
2612 12. Added -help and --help to pcretest as an official way of being reminded
2613 of the options.
2614
2615 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2616 and pcrecpp.cc because they annoy compilers at high warning levels.
2617
2618 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2619
2620 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2621 scripts are not C programs :-) and which was not noticed because it works
2622 on Linux.
2623
2624 16. pcretest is supposed to handle any length of pattern and data line (as one
2625 line or as a continued sequence of lines) by extending its input buffer if
2626 necessary. This feature was broken for very long pattern lines, leading to
2627 a string of junk being passed to pcre_compile() if the pattern was longer
2628 than about 50K.
2629
2630 17. I have done a major re-factoring of the way pcre_compile() computes the
2631 amount of memory needed for a compiled pattern. Previously, there was code
2632 that made a preliminary scan of the pattern in order to do this. That was
2633 OK when PCRE was new, but as the facilities have expanded, it has become
2634 harder and harder to keep it in step with the real compile phase, and there
2635 have been a number of bugs (see for example, 4 above). I have now found a
2636 cunning way of running the real compile function in a "fake" mode that
2637 enables it to compute how much memory it would need, while actually only
2638 ever using a few hundred bytes of working memory and without too many
2639 tests of the mode. This should make future maintenance and development
2640 easier. A side effect of this work is that the limit of 200 on the nesting
2641 depth of parentheses has been removed (though this was never a serious
2642 limitation, I suspect). However, there is a downside: pcre_compile() now
2643 runs more slowly than before (30% or more, depending on the pattern). I
2644 hope this isn't a big issue. There is no effect on runtime performance.
2645
2646 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2647 newline (only possible for the last line of a file) and it was a
2648 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2649
2650 19. Added additional timing features to pcretest. (1) The -tm option now times
2651 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2652 separate command line item, by a number that specifies the number of
2653 repeats to use when timing. The default is 50000; this gives better
2654 precision, but takes uncomfortably long for very large patterns.
2655
2656 20. Extended pcre_study() to be more clever in cases where a branch of a
2657 subpattern has no definite first character. For example, (a*|b*)[cd] would
2658 previously give no result from pcre_study(). Now it recognizes that the
2659 first character must be a, b, c, or d.
2660
2661 21. There was an incorrect error "recursive call could loop indefinitely" if
2662 a subpattern (or the entire pattern) that was being tested for matching an
2663 empty string contained only one non-empty item after a nested subpattern.
2664 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2665 incorrectly, because the \d was being skipped in the check.
2666
2667 22. The pcretest program now has a new pattern option /B and a command line
2668 option -b, which is equivalent to adding /B to every pattern. This causes
2669 it to show the compiled bytecode, without the additional information that
2670 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2671 is the same as /B/I).
2672
2673 23. A new optimization is now able automatically to treat some sequences such
2674 as a*b as a*+b. More specifically, if something simple (such as a character
2675 or a simple class like \d) has an unlimited quantifier, and is followed by
2676 something that cannot possibly match the quantified thing, the quantifier
2677 is automatically "possessified".
2678
2679 24. A recursive reference to a subpattern whose number was greater than 39
2680 went wrong under certain circumstances in UTF-8 mode. This bug could also
2681 have affected the operation of pcre_study().
2682
2683 25. Realized that a little bit of performance could be had by replacing
2684 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2685
2686 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2687
2688 27. Possessive quantifiers such as a++ were previously implemented by turning
2689 them into atomic groups such as ($>a+). Now they have their own opcodes,
2690 which improves performance. This includes the automatically created ones
2691 from 23 above.
2692
2693 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2694 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2695 the first matched character to be a colon. This applied both to named and
2696 numbered groups.
2697
2698 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2699
2700 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2701 building PCRE on Windows easier, so I have included it in the distribution.
2702
2703 31. There is now a check in pcretest against a ridiculously large number being
2704 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2705 loop, the loop is abandoned.
2706
2707 32. Forward references to subpatterns in conditions such as (?(2)...) where
2708 subpattern 2 is defined later cause pcre_compile() to search forwards in
2709 the pattern for the relevant set of parentheses. This search went wrong
2710 when there were unescaped parentheses in a character class, parentheses
2711 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2712
2713 33. "Subroutine" calls and backreferences were previously restricted to
2714 referencing subpatterns earlier in the regex. This restriction has now
2715 been removed.
2716
2717 34. Added a number of extra features that are going to be in Perl 5.10. On the
2718 whole, these are just syntactic alternatives for features that PCRE had
2719 previously implemented using the Python syntax or my own invention. The
2720 other formats are all retained for compatibility.
2721
2722 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2723 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2724 also .NET compatible.
2725
2726 (b) A recursion or subroutine call to a named group can now be defined as
2727 (?&name) as well as (?P>name).
2728
2729 (c) A backreference to a named group can now be defined as \k<name> or
2730 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2731 5.10, are also .NET compatible.
2732
2733 (d) A conditional reference to a named group can now use the syntax
2734 (?(<name>) or (?('name') as well as (?(name).
2735
2736 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2737 groups (named and numbered) that are never evaluated inline, but can be
2738 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2739 is always false. There may be only one alternative in such a group.
2740
2741 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2742 as the simple (?(R). The condition is true only if the most recent
2743 recursion is that of the given number or name. It does not search out
2744 through the entire recursion stack.
2745
2746 (g) The escape \gN or \g{N} has been added, where N is a positive or
2747 negative number, specifying an absolute or relative reference.
2748
2749 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2750 some "unreachable code" warnings.
2751
2752 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2753 things, this adds five new scripts.
2754
2755 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2756 There were also incompatibilities regarding the handling of \Q..\E inside
2757 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2758 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2759
2760 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2761 matches an empty string, and forcibly breaks the loop. There were bugs in
2762 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2763 against aaaa the result was just "a" rather than "aaaa", for example. Two
2764 separate and independent bugs (that affected different cases) have been
2765 fixed.
2766
2767 39. Refactored the code to abolish the use of different opcodes for small
2768 capturing bracket numbers. This is a tidy that I avoided doing when I
2769 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2770 The new approach is not only tidier, it makes it possible to reduce the
2771 memory needed to fix the previous bug (38).
2772
2773 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2774 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2775 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2776 mode.
2777
2778 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2779 report.
2780
2781 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2782 copy construction and assignment in the C++ wrapper.
2783
2784 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2785 couple of bugs that could have given wrong results in the "--newline=crlf"
2786 case.
2787
2788 44. Added a number of casts and did some reorganization of signed/unsigned int
2789 variables following suggestions from Dair Grant. Also renamed the variable
2790 "this" as "item" because it is a C++ keyword.
2791
2792 45. Arranged for dftables to add
2793
2794 #include "pcre_internal.h"
2795
2796 to pcre_chartables.c because without it, gcc 4.x may remove the array
2797 definition from the final binary if PCRE is built into a static library and
2798 dead code stripping is activated.
2799
2800 46. For an unanchored pattern, if a match attempt fails at the start of a
2801 newline sequence, and the newline setting is CRLF or ANY, and the next two
2802 characters are CRLF, advance by two characters instead of one.
2803
2804
2805 Version 6.7 04-Jul-06
2806 ---------------------
2807
2808 1. In order to handle tests when input lines are enormously long, pcretest has
2809 been re-factored so that it automatically extends its buffers when
2810 necessary. The code is crude, but this _is_ just a test program. The
2811 default size has been increased from 32K to 50K.
2812
2813 2. The code in pcre_study() was using the value of the re argument before
2814 testing it for NULL. (Of course, in any sensible call of the function, it
2815 won't be NULL.)
2816
2817 3. The memmove() emulation function in pcre_internal.h, which is used on
2818 systems that lack both memmove() and bcopy() - that is, hardly ever -
2819 was missing a "static" storage class specifier.
2820
2821 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2822 containing an extended class (one that cannot be represented by a bitmap
2823 because it contains high-valued characters or Unicode property items, e.g.
2824 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2825 pattern, but PCRE should not loop if you do not (it no longer does).
2826 [Detail: two cases were found: (a) a repeated subpattern containing an
2827 extended class; (b) a recursive reference to a subpattern that followed a
2828 previous extended class. It wasn't skipping over the extended class
2829 correctly when UTF-8 mode was not set.]
2830
2831 5. A negated single-character class was not being recognized as fixed-length
2832 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2833 compile error "lookbehind assertion is not fixed length".
2834
2835 6. The RunPerlTest auxiliary script was showing an unexpected difference
2836 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2837 write a Perl script that can interpret lines of an input file either as
2838 byte characters or as UTF-8, which is what "perltest" was being required to
2839 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2840 can't do is switch easily at run time between having the "use utf8;" pragma
2841 or not. In the end, I fudged it by using the RunPerlTest script to insert
2842 "use utf8;" explicitly for the UTF-8 tests.
2843
2844 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2845 the end of the subject string, contrary to the documentation and to what
2846 Perl does. This was true of both matching functions. Now it matches only at
2847 the start of the subject and immediately after *internal* newlines.
2848
2849 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2850 a pointer to an int instead of a pointer to an unsigned long int. This
2851 caused problems on 64-bit systems.
2852
2853 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2854 instance of the 'standard' template library not being so standard".
2855
2856 10. There was no check on the number of named subpatterns nor the maximum
2857 length of a subpattern name. The product of these values is used to compute
2858 the size of the memory block for a compiled pattern. By supplying a very
2859 long subpattern name and a large number of named subpatterns, the size
2860 computation could be caused to overflow. This is now prevented by limiting
2861 the length of names to 32 characters, and the number of named subpatterns
2862 to 10,000.
2863
2864 11. Subpatterns that are repeated with specific counts have to be replicated in
2865 the compiled pattern. The size of memory for this was computed from the
2866 length of the subpattern and the repeat count. The latter is limited to
2867 65535, but there was no limit on the former, meaning that integer overflow
2868 could in principle occur. The compiled length of a repeated subpattern is
2869 now limited to 30,000 bytes in order to prevent this.
2870
2871 12. Added the optional facility to have named substrings with the same name.
2872
2873 13. Added the ability to use a named substring as a condition, using the
2874 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2875 are numbers (not recommended). Forward references are permitted.
2876
2877 14. Added forward references in named backreferences (if you see what I mean).
2878
2879 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2880 pattern could run off the end of the subject. For example, the pattern
2881 "(?s)(.{1,5})"8 did this with the subject "ab".
2882
2883 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2884 PCRE_CASELESS was set when matching characters that were quantified with ?
2885 or *.
2886
2887 17. A character class other than a single negated character that had a minimum
2888 but no maximum quantifier - for example [ab]{6,} - was not handled
2889 correctly by pce_dfa_exec(). It would match only one character.
2890
2891 18. A valid (though odd) pattern that looked like a POSIX character
2892 class but used an invalid character after [ (for example [[,abc,]]) caused
2893 pcre_compile() to give the error "Failed: internal error: code overflow" or
2894 in some cases to crash with a glibc free() error. This could even happen if
2895 the pattern terminated after [[ but there just happened to be a sequence of
2896 letters, a binary zero, and a closing ] in the memory that followed.
2897
2898 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2899 over the years. Originally (before any Unicode support), just the bottom 8
2900 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2901 output from "man perlunicode" includes this:
2902
2903 The regular expression compiler produces polymorphic opcodes. That
2904 is, the pattern adapts to the data and automatically switches to
2905 the Unicode character scheme when presented with Unicode data--or
2906 instead uses a traditional byte scheme when presented with byte
2907 data.
2908
2909 Sadly, a wide octal escape does not cause a switch, and in a string with
2910 no other multibyte characters, these octal escapes are treated as before.
2911 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2912 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2913 Unicode string.
2914
2915 I have not perpetrated such confusion in PCRE. Up till now, it took just
2916 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2917 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2918 translate to the appropriate multibyte character.
2919
2920 29. Applied some refactoring to reduce the number of warnings from Microsoft
2921 and Borland compilers. This has included removing the fudge introduced
2922 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2923 a warning about an unused variable.
2924
2925 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2926 characters since release 4.0, because Perl (from release 5.004) does not.
2927 [Or at least, is documented not to: some releases seem to be in conflict
2928 with the documentation.] However, when a pattern was studied with
2929 pcre_study() and all its branches started with \s, PCRE still included VT
2930 as a possible starting character. Of course, this did no harm; it just
2931 caused an unnecessary match attempt.
2932
2933 22. Removed a now-redundant internal flag bit that recorded the fact that case
2934 dependency changed within the pattern. This was once needed for "required
2935 byte" processing, but is no longer used. This recovers a now-scarce options
2936 bit. Also moved the least significant internal flag bit to the most-
2937 significant bit of the word, which was not previously used (hangover from
2938 the days when it was an int rather than a uint) to free up another bit for
2939 the future.
2940
2941 23. Added support for CRLF line endings as well as CR and LF. As well as the
2942 default being selectable at build time, it can now be changed at runtime
2943 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2944 specify that it is scanning data with non-default line endings.
2945
2946 24. Changed the definition of CXXLINK to make it agree with the definition of
2947 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2948
2949 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2950 recursions. This makes a big different to stack usage for some patterns.
2951
2952 26. If a subpattern containing a named recursion or subroutine reference such
2953 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2954 the space required for the compiled pattern went wrong and gave too small a
2955 value. Depending on the environment, this could lead to "Failed: internal
2956 error: code overflow at offset 49" or "glibc detected double free or
2957 corruption" errors.
2958
2959 27. Applied patches from Google (a) to support the new newline modes and (b) to
2960 advance over multibyte UTF-8 characters in GlobalReplace.
2961
2962 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2963 difference for some implementation of PCRE in some Windows version.
2964
2965 29. Added some extra testing facilities to pcretest:
2966
2967 \q<number> in a data line sets the "match limit" value
2968 \Q<number> in a data line sets the "match recursion limt" value
2969 -S <number> sets the stack size, where <number> is in megabytes
2970
2971 The -S option isn't available for Windows.
2972
2973
2974 Version 6.6 06-Feb-06
2975 ---------------------
2976
2977 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2978 in pcreposix.h. I have copied the definition from pcre.h.
2979
2980 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2981 because pcre.h is no longer a built file.
2982
2983 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2984 not normally included in the compiled code.
2985
2986
2987 Version 6.5 01-Feb-06
2988 ---------------------
2989
2990 1. When using the partial match feature with pcre_dfa_exec(), it was not
2991 anchoring the second and subsequent partial matches at the new starting
2992 point. This could lead to incorrect results. For example, with the pattern
2993 /1234/, partially matching against "123" and then "a4" gave a match.
2994
2995 2. Changes to pcregrep:
2996
2997 (a) All non-match returns from pcre_exec() were being treated as failures
2998 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2999 error message is output. Some extra information is given for the
3000 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3001 probably the only errors that are likely to be caused by users (by
3002 specifying a regex that has nested indefinite repeats, for instance).
3003 If there are more than 20 of these errors, pcregrep is abandoned.
3004
3005 (b) A binary zero was treated as data while matching, but terminated the
3006 output line if it was written out. This has been fixed: binary zeroes
3007 are now no different to any other data bytes.
3008
3009 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3010 used to set a locale for matching. The --locale=xxxx long option has
3011 been added (no short equivalent) to specify a locale explicitly on the
3012 pcregrep command, overriding the environment variables.
3013
3014 (d) When -B was used with -n, some line numbers in the output were one less
3015 than they should have been.
3016
3017 (e) Added the -o (--only-matching) option.
3018
3019 (f) If -A or -C was used with -c (count only), some lines of context were
3020 accidentally printed for the final match.
3021
3022 (g) Added the -H (--with-filename) option.
3023
3024 (h) The combination of options -rh failed to suppress file names for files
3025 that were found from directory arguments.
3026
3027 (i) Added the -D (--devices) and -d (--directories) options.
3028
3029 (j) Added the -F (--fixed-strings) option.
3030
3031 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3032
3033 (l) Added the --colo(u)r option.
3034
3035 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3036 is not present by default.
3037
3038 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3039 items such as (?R) or (?1), when the recursion could match a number of
3040 alternatives. If it matched one of the alternatives, but subsequently,
3041 outside the recursion, there was a failure, the code tried to back up into
3042 the recursion. However, because of the way PCRE is implemented, this is not
3043 possible, and the result was an incorrect result from the match.
3044
3045 In order to prevent this happening, the specification of recursion has
3046 been changed so that all such subpatterns are automatically treated as
3047 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3048
3049 4. I had overlooked the fact that, in some locales, there are characters for
3050 which isalpha() is true but neither isupper() nor islower() are true. In
3051 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3052 and ordfeminine) are like this. This affected the treatment of \w and \W
3053 when they appeared in character classes, but not when they appeared outside
3054 a character class. The bit map for "word" characters is now created
3055 separately from the results of isalnum() instead of just taking it from the
3056 upper, lower, and digit maps. (Plus the underscore character, of course.)
3057
3058 5. The above bug also affected the handling of POSIX character classes such as
3059 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3060 permanent tables. Instead, the bit maps for such a class were previously
3061 created as the appropriate unions of the upper, lower, and digit bitmaps.
3062 Now they are created by subtraction from the [[:word:]] class, which has
3063 its own bitmap.
3064
3065 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3066 It is created by subtracting the vertical space characters (\x09, \x0a,
3067 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3068 subtraction was done in the overall bitmap for a character class, meaning
3069 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3070 be recognized. This bug has been fixed.
3071
3072 7. Patches from the folks at Google:
3073
3074 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3075 real life, but is still worth protecting against".
3076
3077 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3078 regular expressions".
3079
3080 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3081 have it.
3082
3083 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3084 "configure" and the latter not, in order to fix a problem somebody had
3085 with compiling the Arg class on HP-UX.
3086
3087 (e) Improve the error-handling of the C++ wrapper a little bit.
3088
3089 (f) New tests for checking recursion limiting.
3090
3091 8. The pcre_memmove() function, which is used only if the environment does not
3092 have a standard memmove() function (and is therefore rarely compiled),
3093 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3094 returning a result (though PCRE never actually uses the result).
3095
3096 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3097 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3098 returned instead of calling malloc() with an overflowing number that would
3099 most likely cause subsequent chaos.
3100
3101 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3102
3103 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3104 with this option is matched, the nmatch and pmatch options of regexec() are
3105 ignored.
3106
3107 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3108 provided in case anyone wants to the the POSIX interface with UTF-8
3109 strings.
3110
3111 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3112 C++ linking (needed for some HP-UX environments).
3113
3114 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3115 (unused parameter) and in the pcre_printint() function (omitted "default"
3116 switch label when the default is to do nothing).
3117
3118 15. Added some code to make it possible, when PCRE is compiled as a C++
3119 library, to replace subject pointers for pcre_exec() with a smart pointer
3120 class, thus making it possible to process discontinuous strings.
3121
3122 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3123 much the same function. They were added by different people who were trying
3124 to make PCRE easy to compile on non-Unix systems. It has been suggested
3125 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3126 for compiling on Windows systems. I have therefore replaced it with
3127 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3128 defaults to "extern" for C or "extern C" for C++, which works fine on
3129 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3130 SCOPE with something explicit in config.h. In addition:
3131
3132 (a) pcreposix.h still had just "extern" instead of either of these macros;
3133 I have replaced it with PCRE_DATA_SCOPE.
3134
3135 (b) Functions such as _pcre_xclass(), which are internal to the library,
3136 but external in the C sense, all had PCRE_EXPORT in their definitions.
3137 This is apparently wrong for the Windows case, so I have removed it.
3138 (It makes no difference on Unix-like systems.)
3139
3140 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3141 of recursive calls to match(). This is different to MATCH_LIMIT because
3142 that limits the total number of calls to match(), not all of which increase
3143 the depth of recursion. Limiting the recursion depth limits the amount of
3144 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3145 when PCRE is compiled, and changed at run time. A patch from Google adds
3146 this functionality to the C++ interface.
3147
3148 18. Changes to the handling of Unicode character properties:
3149
3150 (a) Updated the table to Unicode 4.1.0.
3151
3152 (b) Recognize characters that are not in the table as "Cn" (undefined).
3153
3154 (c) I revised the way the table is implemented to a much improved format
3155 which includes recognition of ranges. It now supports the ranges that
3156 are defined in UnicodeData.txt, and it also amalgamates other
3157 characters into ranges. This has reduced the number of entries in the
3158 table from around 16,000 to around 3,000, thus reducing its size
3159 considerably. I realized I did not need to use a tree structure after
3160 all - a binary chop search is just as efficient. Having reduced the
3161 number of entries, I extended their size from 6 bytes to 8 bytes to
3162 allow for more data.
3163
3164 (d) Added support for Unicode script names via properties such as \p{Han}.
3165
3166 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3167 matching that character.
3168
3169 20. When matching a repeated Unicode property with a minimum greater than zero,
3170 (for example \pL{2,}), PCRE could look past the end of the subject if it
3171 reached it while seeking the minimum number of characters. This could
3172 happen only if some of the characters were more than one byte long, because
3173 there is a check for at least the minimum number of bytes.
3174
3175 21. Refactored the implementation of \p and \P so as to be more general, to
3176 allow for more different types of property in future. This has changed the
3177 compiled form incompatibly. Anybody with saved compiled patterns that use
3178 \p or \P will have to recompile them.
3179
3180 22. Added "Any" and "L&" to the supported property types.
3181
3182 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3183 but give a compile time error if the value is greater than 0xff.
3184
3185 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3186 accidentally not being installed or uninstalled.
3187
3188 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3189 made were to insert the current release number. This seemed silly, because
3190 it made things harder for people building PCRE on systems that don't run
3191 "configure". I have turned pcre.h into a distributed file, no longer built
3192 by "configure", with the version identification directly included. There is
3193 no longer a pcre.h.in file.
3194
3195 However, this change necessitated a change to the pcre-config script as
3196 well. It is built from pcre-config.in, and one of the substitutions was the
3197 release number. I have updated configure.ac so that ./configure now finds
3198 the release number by grepping pcre.h.
3199
3200 26. Added the ability to run the tests under valgrind.
3201
3202
3203 Version 6.4 05-Sep-05
3204 ---------------------
3205
3206 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3207 "--" to be printed when multiple files were scanned, even when none of the
3208 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3209 consider it to be a bug, and have restored the previous behaviour.
3210
3211 2. A couple of code tidies to get rid of compiler warnings.
3212
3213 3. The pcretest program used to cheat by referring to symbols in the library
3214 whose names begin with _pcre_. These are internal symbols that are not
3215 really supposed to be visible externally, and in some environments it is
3216 possible to suppress them. The cheating is now confined to including
3217 certain files from the library's source, which is a bit cleaner.
3218
3219 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3220 file's purpose clearer.
3221
3222 5. Reorganized pcre_ucp_findchar().
3223
3224
3225 Version 6.3 15-Aug-05
3226 ---------------------
3227
3228 1. The file libpcre.pc.in did not have general read permission in the tarball.
3229
3230 2. There were some problems when building without C++ support:
3231
3232 (a) If C++ support was not built, "make install" and "make test" still
3233 tried to test it.
3234
3235 (b) There were problems when the value of CXX was explicitly set. Some
3236 changes have been made to try to fix these, and ...
3237
3238 (c) --disable-cpp can now be used to explicitly disable C++ support.
3239
3240 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3241 backslash in a target when C++ was disabled. This confuses some
3242 versions of "make", apparently. Using an intermediate variable solves
3243 this. (Same for CPP_LOBJ.)
3244
3245 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3246 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3247 necessary on certain architectures.
3248
3249 4. Added a setting of -export-symbols-regex to the link command to remove
3250 those symbols that are exported in the C sense, but actually are local
3251 within the library, and not documented. Their names all begin with
3252 "_pcre_". This is not a perfect job, because (a) we have to except some
3253 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3254 available (and never for static libraries). I have made a note to try to
3255 find a way round (a) in the future.
3256
3257
3258 Version 6.2 01-Aug-05
3259 ---------------------
3260
3261 1. There was no test for integer overflow of quantifier values. A construction
3262 such as {1111111111111111} would give undefined results. What is worse, if
3263 a minimum quantifier for a parenthesized subpattern overflowed and became
3264 negative, the calculation of the memory size went wrong. This could have
3265 led to memory overwriting.
3266
3267 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3268
3269 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3270 operating environments where this matters.
3271
3272 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3273 PCRE options from within the C++ wrapper.
3274
3275 5. Named capturing subpatterns were not being correctly counted when a pattern
3276 was compiled. This caused two problems: (a) If there were more than 100
3277 such subpatterns, the calculation of the memory needed for the whole
3278 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3279 back references of the form \12, where the number was greater than 9, were
3280 not recognized as back references, even though there were sufficient
3281 previous subpatterns.
3282
3283 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3284 versions of gcc, e.g. 2.95.4.
3285
3286
3287 Version 6.1 21-Jun-05
3288 ---------------------
3289
3290 1. There was one reference to the variable "posix" in pcretest.c that was not
3291 surrounded by "#if !defined NOPOSIX".
3292
3293 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3294 the cross-check on the old pcre_info() function, for the benefit of the
3295 cut-down version of PCRE that is currently imported into Exim.
3296
3297 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3298 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3299 patterns that start (?i) but I don't think that matters. The use of (?i) is
3300 just an example; this all applies to the other options as well.
3301
3302 4. Since libtool seems to echo the compile commands it is issuing, the output
3303 from "make" can be reduced a bit by putting "@" in front of each libtool
3304 compile command.
3305
3306 5. Patch from the folks at Google for configure.in to be a bit more thorough
3307 in checking for a suitable C++ installation before trying to compile the
3308 C++ stuff. This should fix a reported problem when a compiler was present,
3309 but no suitable headers.
3310
3311 6. The man pages all had just "PCRE" as their title. I have changed them to
3312 be the relevant file name. I have also arranged that these names are
3313 retained in the file doc/pcre.txt, which is a concatenation in text format
3314 of all the man pages except the little individual ones for each function.
3315
3316 7. The NON-UNIX-USE file had not been updated for the different set of source
3317 files that come with release 6. I also added a few comments about the C++
3318 wrapper.
3319
3320
3321 Version 6.0 07-Jun-05
3322 ---------------------
3323
3324 1. Some minor internal re-organization to help with my DFA experiments.
3325
3326 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3327 didn't matter for the library itself when fully configured, but did matter
3328 when compiling without UCP support, or within Exim, where the ucp files are
3329 not imported.
3330
3331 3. Refactoring of the library code to split up the various functions into
3332 different source modules. The addition of the new DFA matching code (see
3333 below) to a single monolithic source would have made it really too
3334 unwieldy, quite apart from causing all the code to be include in a
3335 statically linked application, when only some functions are used. This is
3336 relevant even without the DFA addition now that patterns can be compiled in
3337 one application and matched in another.
3338
3339 The downside of splitting up is that there have to be some external
3340 functions and data tables that are used internally in different modules of
3341 the library but which are not part of the API. These have all had their
3342 names changed to start with "_pcre_" so that they are unlikely to clash
3343 with other external names.
3344
3345 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3346 a different (DFA) algorithm. Although it is slower than the original
3347 function, it does have some advantages for certain types of matching
3348 problem.
3349
3350 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3351 including restarting after a partial match.
3352
3353 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3354 defined when compiling for Windows was sent to me. I have put it into the
3355 code, though I have no means of testing or verifying it.
3356
3357 7. Added the pcre_refcount() auxiliary function.
3358
3359 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3360 match before or at the first newline in the subject string. In pcretest,
3361 the /f option on a pattern can be used to set this.
3362
3363 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3364 would behave wrongly. This has been present in PCRE since release 4.0.
3365
3366 10. A number of changes to the pcregrep command:
3367
3368 (a) Refactored how -x works; insert ^(...)$ instead of setting
3369 PCRE_ANCHORED and checking the length, in preparation for adding
3370 something similar for -w.
3371
3372 (b) Added the -w (match as a word) option.
3373
3374 (c) Refactored the way lines are read and buffered so as to have more
3375 than one at a time available.
3376
3377 (d) Implemented a pcregrep test script.
3378
3379 (e) Added the -M (multiline match) option. This allows patterns to match
3380 over several lines of the subject. The buffering ensures that at least
3381 8K, or the rest of the document (whichever is the shorter) is available
3382 for matching (and similarly the previous 8K for lookbehind assertions).
3383
3384 (f) Changed the --help output so that it now says
3385
3386 -w, --word-regex(p)
3387
3388 instead of two lines, one with "regex" and the other with "regexp"
3389 because that confused at least one person since the short forms are the
3390 same. (This required a bit of code, as the output is generated
3391 automatically from a table. It wasn't just a text change.)
3392
3393 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3394 option but starts with a hyphen. Could be a pattern or a path name
3395 starting with a hyphen, for instance.
3396
3397 (h) "-" can be given as a file name to represent stdin.
3398
3399 (i) When file names are being printed, "(standard input)" is used for
3400 the standard input, for compatibility with GNU grep. Previously
3401 "<stdin>" was used.
3402
3403 (j) The option --label=xxx can be used to supply a name to be used for
3404 stdin when file names are being printed. There is no short form.
3405
3406 (k) Re-factored the options decoding logic because we are going to add
3407 two more options that take data. Such options can now be given in four
3408 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3409
3410 (l) Added the -A, -B, and -C options for requesting that lines of context
3411 around matches be printed.
3412
3413 (m) Added the -L option to print the names of files that do not contain
3414 any matching lines, that is, the complement of -l.
3415
3416 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3417 continue to scan other files.
3418
3419 (o) The -s option was incorrectly implemented. For compatibility with other
3420 greps, it now suppresses the error message for a non-existent or non-
3421 accessible file (but not the return code). There is a new option called
3422 -q that suppresses the output of matching lines, which was what -s was
3423 previously doing.
3424
3425 (p) Added --include and --exclude options to specify files for inclusion
3426 and exclusion when recursing.
3427
3428 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3429 Hopefully, it now does.
3430
3431 12. Missing cast in pcre_study().
3432
3433 13. Added an "uninstall" target to the makefile.
3434
3435 14. Replaced "extern" in the function prototypes in Makefile.in with
3436 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3437 world, but is set differently for Windows.
3438
3439 15. Added a second compiling function called pcre_compile2(). The only
3440 difference is that it has an extra argument, which is a pointer to an
3441 integer error code. When there is a compile-time failure, this is set
3442 non-zero, in addition to the error test pointer being set to point to an
3443 error message. The new argument may be NULL if no error number is required
3444 (but then you may as well call pcre_compile(), which is now just a
3445 wrapper). This facility is provided because some applications need a
3446 numeric error indication, but it has also enabled me to tidy up the way
3447 compile-time errors are handled in the POSIX wrapper.
3448
3449 16. Added VPATH=.libs to the makefile; this should help when building with one
3450 prefix path and installing with another. (Or so I'm told by someone who
3451 knows more about this stuff than I do.)
3452
3453 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3454 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3455 match everything, including newlines. This is not POSIX-compatible, but
3456 somebody wanted the feature. From pcretest it can be activated by using
3457 both the P and the s flags.
3458
3459 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3460
3461 19. libpcre.pc was being incorrectly installed as executable.
3462
3463 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3464 it now also looks for '\r' so that it will work unmodified on Windows.
3465
3466 21. Added Google's contributed C++ wrapper to the distribution.
3467
3468 22. Added some untidy missing memory free() calls in pcretest, to keep
3469 Electric Fence happy when testing.
3470
3471
3472
3473 Version 5.0 13-Sep-04
3474 ---------------------
3475
3476 1. Internal change: literal characters are no longer packed up into items
3477 containing multiple characters in a single byte-string. Each character
3478 is now matched using a separate opcode. However, there may be more than one
3479 byte in the character in UTF-8 mode.
3480
3481 2. The pcre_callout_block structure has two new fields: pattern_position and
3482 next_item_length. These contain the offset in the pattern to the next match
3483 item, and its length, respectively.
3484
3485 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3486 insertion of callouts before each pattern item. Added the /C option to
3487 pcretest to make use of this.
3488
3489 4. On the advice of a Windows user, the lines
3490
3491 #if defined(_WIN32) || defined(WIN32)
3492 _setmode( _fileno( stdout ), 0x8000 );
3493 #endif /* defined(_WIN32) || defined(WIN32) */
3494
3495 have been added to the source of pcretest. This apparently does useful
3496 magic in relation to line terminators.
3497
3498 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3499 for the benefit of those environments where the "b" makes a difference.
3500
3501 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3502 to know about it. I have put a hack into configure.in that adds in code
3503 to set GCC=yes if CC=icc. This seems to end up at a point in the
3504 generated configure script that is early enough to affect the setting of
3505 compiler options, which is what is needed, but I have no means of testing
3506 whether it really works. (The user who reported this had patched the
3507 generated configure script, which of course I cannot do.)
3508
3509 LATER: After change 22 below (new libtool files), the configure script
3510 seems to know about icc (and also ecc). Therefore, I have commented out
3511 this hack in configure.in.
3512
3513 7. Added support for pkg-config (2 patches were sent in).
3514
3515 8. Negated POSIX character classes that used a combination of internal tables
3516 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3517 [[:^ascii]]. Typically, they would match almost any characters. The other
3518 POSIX classes were not broken in this way.
3519
3520 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3521 to find the match, as PCRE was deluded into thinking that the match had to
3522 start at the start point or following a newline. The same bug applied to
3523 patterns with negative forward assertions or any backward assertions
3524 preceding ".*" at the start, unless the pattern required a fixed first
3525 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3526
3527 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3528 starting at the last subject character, bytes beyond the end of the subject
3529 string were read.
3530
3531 11. Renamed the variable "class" as "classbits" to make life easier for C++
3532 users. (Previously there was a macro definition, but it apparently wasn't
3533 enough.)
3534
3535 12. Added the new field "tables" to the extra data so that tables can be passed
3536 in at exec time, or the internal tables can be re-selected. This allows
3537 a compiled regex to be saved and re-used at a later time by a different
3538 program that might have everything at different addresses.
3539
3540 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3541 -R library as well as a -L library.
3542
3543 14. The debugging options of pcretest (-d on the command line or D on a
3544 pattern) showed incorrect output for anything following an extended class
3545 that contained multibyte characters and which was followed by a quantifier.
3546
3547 15. Added optional support for general category Unicode character properties
3548 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3549 support. It adds about 90K to the size of the library. The meanings of the
3550 inbuilt class escapes such as \d and \s have NOT been changed.
3551
3552 16. Updated pcredemo.c to include calls to free() to release the memory for the
3553 compiled pattern.
3554
3555 17. The generated file chartables.c was being created in the source directory
3556 instead of in the building directory. This caused the build to fail if the
3557 source directory was different from the building directory, and was
3558 read-only.
3559
3560 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3561 file. No doubt somebody will tell me if they don't make sense... Also added
3562 Dan Mooney's comments about building on OpenVMS.
3563
3564 19. Added support for partial matching via the PCRE_PARTIAL option for
3565 pcre_exec() and the \P data escape in pcretest.
3566
3567 20. Extended pcretest with 3 new pattern features:
3568
3569 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3570 write the compiled pattern to the file whose name is "rest-of-line".
3571 This is a straight binary dump of the data, with the saved pointer to
3572 the character tables forced to be NULL. The study data, if any, is
3573 written too. After writing, pcretest reads a new pattern.
3574
3575 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3576 compiled pattern from the given file. There must not be any
3577 occurrences of "<" in the file name (pretty unlikely); if there are,
3578 pcretest will instead treat the initial "<" as a pattern delimiter.
3579 After reading in the pattern, pcretest goes on to read data lines as
3580 usual.
3581
3582 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3583 and 16-bit fields in a compiled pattern, to simulate a pattern that
3584 was compiled on a host of opposite endianness.
3585
3586 21. The pcre-exec() function can now cope with patterns that were compiled on
3587 hosts of opposite endianness, with this restriction:
3588
3589 As for any compiled expression that is saved and used later, the tables
3590 pointer field cannot be preserved; the extra_data field in the arguments
3591 to pcre_exec() should be used to pass in a tables address if a value
3592 other than the default internal tables were used at compile time.
3593
3594 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3595 now diagnosed as an error. Previously, most of the time, a negative number
3596 would have been treated as zero, but if in addition "ovector" was passed as
3597 NULL, a crash could occur.
3598
3599 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3600 new versions from the libtool 1.5 distribution (the last one is a copy of
3601 a file called libtool.m4). This seems to have fixed the need to patch
3602 "configure" to support Darwin 1.3 (which I used to do). However, I still
3603 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3604 workstation).
3605
3606 24. Changed the PCRE licence to be the more standard "BSD" licence.
3607
3608
3609 Version 4.5 01-Dec-03
3610 ---------------------
3611
3612 1. There has been some re-arrangement of the code for the match() function so
3613 that it can be compiled in a version that does not call itself recursively.
3614 Instead, it keeps those local variables that need separate instances for
3615 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3616 needs to "recurse". Keeping track of where control must go is done by means
3617 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3618 hide most of the details from the main code, and operates only if
3619 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3620 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3621 operating.
3622
3623 To make it easier for callers to provide specially tailored get/free
3624 functions for this usage, two new functions, pcre_stack_malloc, and
3625 pcre_stack_free, are used. They are always called in strict stacking order,
3626 and the size of block requested is always the same.
3627
3628 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3629 PCRE has been compiled to use the stack or the heap for recursion. The
3630 -C option of pcretest uses this to show which version is compiled.
3631
3632 A new data escape \S, is added to pcretest; it causes the amounts of store
3633 obtained and freed by both kinds of malloc/free at match time to be added
3634 to the output.
3635
3636 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3637 what's available on my current Linux desktop machine.
3638
3639 3. When matching a UTF-8 string, the test for a valid string at the start has
3640 been extended. If start_offset is not zero, PCRE now checks that it points
3641 to a byte that is the start of a UTF-8 character. If not, it returns
3642 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3643 this is necessary because there may be backward assertions in the pattern.
3644 When matching the same subject several times, it may save resources to use
3645 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3646
3647 4. The code for checking the validity of UTF-8 strings has been tightened so
3648 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3649 containing "overlong sequences".
3650
3651 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3652 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3653 should have been "&", but it just so happened that all the cases this let
3654 through by mistake were picked up later in the function.
3655
3656 6. I had used a variable called "isblank" - this is a C99 function, causing
3657 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3658
3659 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3660 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3661 so the version is shown; (c) stop "make test" echoing "./RunTest".
3662
3663 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3664
3665 9. The prototype for memmove() for systems that don't have it was using
3666 size_t, but the inclusion of the header that defines size_t was later. I've
3667 moved the #includes for the C headers earlier to avoid this.
3668
3669 10. Added some adjustments to the code to make it easier to compiler on certain
3670 special systems:
3671
3672 (a) Some "const" qualifiers were missing.
3673 (b) Added the macro EXPORT before all exported functions; by default this
3674 is defined to be empty.
3675 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3676 that it reads its output file name as an argument instead of writing
3677 to the standard output and assuming this can be redirected.
3678
3679 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3680 class containing characters with values greater than 255, PCRE compilation
3681 went into a loop.
3682
3683 12. A recursive reference to a subpattern that was within another subpattern
3684 that had a minimum quantifier of zero caused PCRE to crash. For example,
3685 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3686 recursion. If the recursively-called subpattern itself had a zero repeat,
3687 that was OK.
3688
3689 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3690 buffer into which it was copied (for escape processing) was still set at
3691 1024, so long lines caused crashes.
3692
3693 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3694 "internal error: code overflow...". This applied to any character class
3695 that was followed by a possessive quantifier.
3696
3697 15. Modified the Makefile to add libpcre.la as a prerequisite for
3698 libpcreposix.la because I was told this is needed for a parallel build to
3699 work.
3700
3701 16. If a pattern that contained .* following optional items at the start was
3702 studied, the wrong optimizing data was generated, leading to matching
3703 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3704 matching string must start with a or b or c. The correct conclusion for
3705 this pattern is that a match can start with any character.
3706
3707
3708 Version 4.4 13-Aug-03
3709 ---------------------
3710
3711 1. In UTF-8 mode, a character class containing characters with values between
3712 127 and 255 was not handled correctly if the compiled pattern was studied.
3713 In fixing this, I have also improved the studying algorithm for such
3714 classes (slightly).
3715
3716 2. Three internal functions had redundant arguments passed to them. Removal
3717 might give a very teeny performance improvement.
3718
3719 3. Documentation bug: the value of the capture_top field in a callout is *one
3720 more than* the number of the hightest numbered captured substring.
3721
3722 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3723 in incorrectly linking with a previously installed version. They now link
3724 explicitly with libpcre.la.
3725
3726 5. configure.in no longer needs to recognize Cygwin specially.
3727
3728 6. A problem in pcre.in for Windows platforms is fixed.
3729
3730 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3731 pcretest, it used to include the size of the study block as part of its
3732 output. Unfortunately, the structure contains a field that has a different
3733 size on different hardware architectures. This meant that the tests that
3734 showed this size failed. As the block is currently always of a fixed size,
3735 this information isn't actually particularly useful in pcretest output, so
3736 I have just removed it.
3737
3738 8. Three pre-processor statements accidentally did not start in column 1.
3739 Sadly, there are *still* compilers around that complain, even though
3740 standard C has not required this for well over a decade. Sigh.
3741
3742 9. In pcretest, the code for checking callouts passed small integers in the
3743 callout_data field, which is a void * field. However, some picky compilers
3744 complained about the casts involved for this on 64-bit systems. Now
3745 pcretest passes the address of the small integer instead, which should get
3746 rid of the warnings.
3747
3748 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3749 both compile and run time, and gives an error if an invalid UTF-8 sequence
3750 is found. There is a option for disabling this check in cases where the
3751 string is known to be correct and/or the maximum performance is wanted.
3752
3753 11. In response to a bug report, I changed one line in Makefile.in from
3754
3755 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3756 to
3757 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3758
3759 to look similar to other lines, but I have no way of telling whether this
3760 is the right thing to do, as I do not use Windows. No doubt I'll get told
3761 if it's wrong...
3762
3763
3764 Version 4.3 21-May-03
3765 ---------------------
3766
3767 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3768 Makefile.
3769
3770 2. Some refactoring to improve the quality of the code:
3771
3772 (i) The utf8_table... variables are now declared "const".
3773
3774 (ii) The code for \cx, which used the "case flipping" table to upper case
3775 lower case letters, now just substracts 32. This is ASCII-specific,
3776 but the whole concept of \cx is ASCII-specific, so it seems
3777 reasonable.
3778
3779 (iii) PCRE was using its character types table to recognize decimal and
3780 hexadecimal digits in the pattern. This is silly, because it handles
3781 only 0-9, a-f, and A-F, but the character types table is locale-
3782 specific, which means strange things might happen. A private
3783 table is now used for this - though it costs 256 bytes, a table is
3784 much faster than multiple explicit tests. Of course, the standard
3785 character types table is still used for matching digits in subject
3786 strings against \d.
3787
3788 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3789 ending in _t are). So I've renamed it as ESC_tee.
3790
3791 3. The first argument for regexec() in the POSIX wrapper should have been
3792 defined as "const".
3793
3794 4. Changed pcretest to use malloc() for its buffers so that they can be
3795 Electric Fenced for debugging.
3796
3797 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3798 to read one or more bytes before the start of the subject string. Often this
3799 had no effect on PCRE's behaviour, but in some circumstances it could
3800 provoke a segmentation fault.
3801
3802 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3803 to try to read one or more bytes before the start of the subject string.
3804
3805 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3806 UTF-8 support could misbehave in various ways if the subject string
3807 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3808 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3809 back over UTF-8 characters.)
3810
3811
3812 Version 4.2 14-Apr-03
3813 ---------------------
3814
3815 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3816
3817 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3818 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3819 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3820 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3821 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3822 and BUILD_EXEEXT
3823 Note: automatic setting of the BUILD variables is not yet working
3824 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3825 compile-time but not at link-time
3826 [LINK]: use for linking executables only
3827 make different versions for Windows and non-Windows
3828 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3829 libraries
3830 [LINK_FOR_BUILD]: new variable
3831 [OBJEXT]: use throughout
3832 [EXEEXT]: use throughout
3833 <winshared>: new target
3834 <wininstall>: new target
3835 <dftables.o>: use native compiler
3836 <dftables>: use native linker
3837 <install>: handle Windows platform correctly
3838 <clean>: ditto
3839 <check>: ditto
3840 copy DLL to top builddir before testing
3841
3842 As part of these changes, -no-undefined was removed again. This was reported
3843 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3844 in any case.
3845
3846 3. Some tidies to get rid of compiler warnings:
3847
3848 . In the match_data structure, match_limit was an unsigned long int, whereas
3849 match_call_count was an int. I've made them both unsigned long ints.
3850
3851 . In pcretest the fact that a const uschar * doesn't automatically cast to
3852 a void * provoked a warning.
3853
3854 . Turning on some more compiler warnings threw up some "shadow" variables
3855 and a few more missing casts.
3856
3857 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3858 option, a class that contained a single character with a value between 128
3859 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3860
3861 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3862 option, a class that contained several characters, but with at least one
3863 whose value was between 128 and 255 caused PCRE to crash.
3864
3865
3866 Version 4.1 12-Mar-03
3867 ---------------------
3868
3869 1. Compiling with gcc -pedantic found a couple of places where casts were
3870 needed, and a string in dftables.c that was longer than standard compilers are
3871 required to support.
3872
3873 2. Compiling with Sun's compiler found a few more places where the code could
3874 be tidied up in order to avoid warnings.
3875
3876 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3877 first of these names is deprecated in the latest Autoconf in favour of the name
3878 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3879 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3880 analogy I have changed it to CFLAGS_FOR_BUILD.
3881
3882 4. Added -no-undefined to the linking command in the Makefile, because this is
3883 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3884 linking step for the pcreposix library.
3885
3886 5. PCRE was failing to diagnose the case of two named groups with the same
3887 name.
3888
3889 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3890 literal character that is needed in the subject for a match, and scans along to
3891 ensure that it is present before embarking on the full matching process. This
3892 saves time in cases of nested unlimited repeats that are never going to match.
3893 Problem: the scan can take a lot of time if the subject is very long (e.g.
3894 megabytes), thus penalizing straightforward matches. It is now done only if the
3895 amount of subject to be scanned is less than 1000 bytes.
3896
3897 7. A lesser problem with the same optimization is that it was recording the
3898 first character of an anchored pattern as "needed", thus provoking a search
3899 right along the subject, even when the first match of the pattern was going to
3900 fail. The "needed" character is now not set for anchored patterns, unless it
3901 follows something in the pattern that is of non-fixed length. Thus, it still
3902 fulfils its original purpose of finding quick non-matches in cases of nested
3903 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3904
3905
3906 Version 4.0 17-Feb-03
3907 ---------------------
3908
3909 1. If a comment in an extended regex that started immediately after a meta-item
3910 extended to the end of string, PCRE compiled incorrect data. This could lead to
3911 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3912
3913 2. Moved to autoconf 2.53 and libtool 1.4.2.
3914
3915 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3916 the special perltest8 script is no longer needed - all the tests can be run
3917 from a single perltest script.
3918
3919 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3920 by \s. It has now been removed in PCRE. This means it isn't recognized as
3921 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3922 class [:space:] *does* include VT, thereby creating a mess.
3923
3924 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3925 space and tab.
3926
3927 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3928 its new features into the main test script, reducing the number of scripts.
3929
3930 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3931 were backward compatible, and made the (?i) apply to the whole pattern, as if
3932 /i were given. Now it behaves more logically, and applies the option setting
3933 only to what follows. PCRE has been changed to follow suit. However, if it
3934 finds options settings right at the start of the pattern, it extracts them into
3935 the global options, as before. Thus, they show up in the info data.
3936
3937 8. Added support for the \Q...\E escape sequence. Characters in between are
3938 treated as literals. This is slightly different from Perl in that $ and @ are
3939 also handled as literals inside the quotes. In Perl, they will cause variable
3940 interpolation. Note the following examples:
3941
3942 Pattern PCRE matches Perl matches
3943
3944 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3945 \Qabc\$xyz\E abc\$xyz abc\$xyz
3946 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3947
3948 For compatibility with Perl, \Q...\E sequences are recognized inside character
3949 classes as well as outside them.
3950
3951 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3952 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3953 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3954 signed/unsigned warnings.
3955
3956 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3957 option for pcretest, so I've replaced it by a simple function that does just
3958 that job.
3959
3960 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3961 "pcregrep -".
3962
3963 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3964 Java package. This provides some syntactic sugar for simple cases of what my
3965 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3966 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3967 item, you can use this simplified notation. Note that only makes sense with
3968 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3969 greediness, whatever the setting of the PCRE_UNGREEDY option.
3970
3971 13. A change of greediness default within a pattern was not taking effect at
3972 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3973 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3974 was abstracted outside.
3975
3976 14. PCRE now supports the \G assertion. It is true when the current matching
3977 position is at the start point of the match. This differs from \A when the
3978 starting offset is non-zero. Used with the /g option of pcretest (or similar
3979 code), it works in the same way as it does for Perl's /g option. If all
3980 alternatives of a regex begin with \G, the expression is anchored to the start
3981 match position, and the "anchored" flag is set in the compiled expression.
3982
3983 15. Some bugs concerning the handling of certain option changes within patterns
3984 have been fixed. These applied to options other than (?ims). For example,
3985 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3986 been the other way round. Some of this was related to change 7 above.
3987
3988 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3989 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3990 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3991 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3992
3993 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3994 mode. Unlike ".", it always matches newline, whatever the setting of
3995 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3996 assertions. Perl allows it, but it doesn't (in general) work because it can't
3997 calculate the length of the lookbehind. At least, that's the case for Perl
3998 5.8.0 - I've been told they are going to document that it doesn't work in
3999 future.
4000
4001 18. Added an error diagnosis for escapes that PCRE does not support: these are
4002 \L, \l, \N, \P, \p, \U, \u, and \X.
4003
4004 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4005 reading past the end of the pattern in cases such as /[abcd/.
4006
4007 20. PCRE was getting more memory than necessary for patterns with classes that
4008 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4009
4010 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4011 compiling PCRE for use with Virtual Pascal.
4012
4013 22. Small fix to the Makefile to make it work properly if the build is done
4014 outside the source tree.
4015
4016 23. Added a new extension: a condition to go with recursion. If a conditional
4017 subpattern starts with (?(R) the "true" branch is used if recursion has
4018 happened, whereas the "false" branch is used only at the top level.
4019
4020 24. When there was a very long string of literal characters (over 255 bytes
4021 without UTF support, over 250 bytes with UTF support), the computation of how
4022 much memory was required could be incorrect, leading to segfaults or other
4023 strange effects.
4024
4025 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4026 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4027 there was a subsequent back reference to those brackets. This meant that, for
4028 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4029 possible to check for precisely this case. All we can do is abandon the
4030 optimization if .* occurs inside capturing brackets when there are any back
4031 references whatsoever. (See below for a better fix that came later.)
4032
4033 26. The handling of the optimization for finding the first character of a
4034 non-anchored pattern, and for finding a character that is required later in the
4035 match were failing in some cases. This didn't break the matching; it just
4036 failed to optimize when it could. The way this is done has been re-implemented.
4037
4038 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4039
4040 28. Added a new feature that provides some of the functionality that Perl
4041 provides with (?{...}). The facility is termed a "callout". The way it is done
4042 in PCRE is for the caller to provide an optional function, by setting
4043 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4044 global variable. By default it is unset, which disables all calling out. To get
4045 the function called, the regex must include (?C) at appropriate points. This
4046 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4047 This provides a means of identifying different callout points. When PCRE
4048 reaches such a point in the regex, if pcre_callout has been set, the external
4049 function is called. It is provided with data in a structure called
4050 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4051 matching continues; if it returns a non-zero value, the match at the current
4052 point fails. However, backtracking will occur if possible. [This was changed
4053 later and other features added - see item 49 below.]
4054
4055 29. pcretest is upgraded to test the callout functionality. It provides a
4056 callout function that displays information. By default, it shows the start of
4057 the match and the current position in the text. There are some new data escapes
4058 to vary what happens:
4059
4060 \C+ in addition, show current contents of captured substrings
4061 \C- do not supply a callout function
4062 \C!n return 1 when callout number n is reached
4063 \C!n!m return 1 when callout number n is reached for the mth time
4064
4065 30. If pcregrep was called with the -l option and just a single file name, it
4066 output "<stdin>" if a match was found, instead of the file name.
4067
4068 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4069 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4070 pcre_exec(). This saves a malloc/free per call. The default value of
4071 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4072 when configuring.
4073
4074 32. The default maximum size of a compiled pattern is 64K. There have been a
4075 few cases of people hitting this limit. The code now uses macros to handle the
4076 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4077 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4078 configuring. Tests 2 and 5 work only with 2-byte links because they output
4079 debugging information about compiled patterns.
4080
4081 33. Internal code re-arrangements:
4082
4083 (a) Moved the debugging function for printing out a compiled regex into
4084 its own source file (printint.c) and used #include to pull it into
4085 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4086 separate copies.
4087
4088 (b) Defined the list of op-code names for debugging as a macro in
4089 internal.h so that it is next to the definition of the opcodes.
4090
4091 (c) Defined a table of op-code lengths for simpler skipping along compiled
4092 code. This is again a macro in internal.h so that it is next to the
4093 definition of the opcodes.
4094
4095 34. Added support for recursive calls to individual subpatterns, along the
4096 lines of Robin Houston's patch (but implemented somewhat differently).
4097
4098 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4099 allow it to read and process whole directories in Win32. This code was
4100 contributed by Lionel Fourquaux; it has not been tested by me.
4101
4102 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4103 used to name a group. Names consist of alphanumerics and underscores, and must
4104 be unique. Back references use the syntax (?P=name) and recursive calls use
4105 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4106 numbers. The function pcre_fullinfo() can be used after compilation to extract
4107 a name/number map. There are three relevant calls:
4108
4109 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4110 PCRE_INFO_NAMECOUNT yields the number of entries
4111 PCRE_INFO_NAMETABLE yields a pointer to the map.
4112
4113 The map is a vector of fixed-size entries. The size of each entry depends on
4114 the length of the longest name used. The first two bytes of each entry are the
4115 group number, most significant byte first. There follows the corresponding
4116 name, zero terminated. The names are in alphabetical order.
4117
4118 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4119 case instead of 255. Making it the same both with and without UTF-8 support
4120 means that the same test output works with both.
4121
4122 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4123 calling malloc() with a zero argument.
4124
4125 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4126 optimization. I've improved things by keeping a bitmap of backreferences with
4127 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4128 fact referenced, the optimization can be applied. It is unlikely that a
4129 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4130 the match to follow \n) will appear inside brackets with a number greater than
4131 31, but if it does, any back reference > 31 suppresses the optimization.
4132
4133 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4134 of disabling numbered capturing parentheses. Any opening parenthesis that is
4135 not followed by ? behaves as if it were followed by ?: but named parentheses
4136 can still be used for capturing (and they will acquire numbers in the usual
4137 way).
4138
4139 41. Redesigned the return codes from the match() function into yes/no/error so
4140 that errors can be passed back from deep inside the nested calls. A malloc
4141 failure while inside a recursive subpattern call now causes the
4142 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4143
4144 42. It is now possible to set a limit on the number of times the match()
4145 function is called in a call to pcre_exec(). This facility makes it possible to
4146 limit the amount of recursion and backtracking, though not in a directly
4147 obvious way, because the match() function is used in a number of different
4148 circumstances. The count starts from zero for each position in the subject
4149 string (for non-anchored patterns). The default limit is, for compatibility, a
4150 large number, namely 10 000 000. You can change this in two ways:
4151
4152 (a) When configuring PCRE before making, you can use --with-match-limit=n
4153 to set a default value for the compiled library.
4154
4155 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4156 a different value is set. See 45 below.
4157
4158 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4159
4160 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4161 of things that can be changed at compile time. The first argument specifies
4162 what is wanted and the second points to where the information is to be placed.
4163 The current list of available information is:
4164
4165 PCRE_CONFIG_UTF8
4166
4167 The output is an integer that is set to one if UTF-8 support is available;
4168 otherwise it is set to zero.
4169
4170 PCRE_CONFIG_NEWLINE
4171
4172 The output is an integer that it set to the value of the code that is used for
4173 newline. It is either LF (10) or CR (13).
4174
4175 PCRE_CONFIG_LINK_SIZE
4176
4177 The output is an integer that contains the number of bytes used for internal
4178 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4179
4180 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4181
4182 The output is an integer that contains the threshold above which the POSIX
4183 interface uses malloc() for output vectors. See item 31 above.
4184
4185 PCRE_CONFIG_MATCH_LIMIT
4186
4187 The output is an unsigned integer that contains the default limit of the number
4188 of match() calls in a pcre_exec() execution. See 42 above.
4189
4190 44. pcretest has been upgraded by the addition of the -C option. This causes it
4191 to extract all the available output from the new pcre_config() function, and to
4192 output it. The program then exits immediately.
4193
4194 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4195 order to support additional features. One way would have been to define
4196 pcre_exec2() (for example) with extra arguments, but this would not have been
4197 extensible, and would also have required all calls to the original function to
4198 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4199 is used for passing in "extra" data from pcre_study().
4200
4201 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4202 contains the following fields:
4203
4204 flags a bitmap indicating which of the following fields are set
4205 study_data opaque data from pcre_study()
4206 match_limit a way of specifying a limit on match() calls for a specific
4207 call to pcre_exec()
4208 callout_data data for callouts (see 49 below)
4209
4210 The flag bits are also defined in pcre.h, and are
4211
4212 PCRE_EXTRA_STUDY_DATA
4213 PCRE_EXTRA_MATCH_LIMIT
4214 PCRE_EXTRA_CALLOUT_DATA
4215
4216 The pcre_study() function now returns one of these new pcre_extra blocks, with
4217 the actual study data pointed to by the study_data field, and the
4218 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4219 before. That is, this change is entirely upwards-compatible and requires no
4220 change to existing code.
4221
4222 If you want to pass in additional data to pcre_exec(), you can either place it
4223 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4224 block.
4225
4226 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4227 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4228 times with different match limits, until it finds the minimum value needed for
4229 pcre_exec() to complete. The value is then output. This can be instructive; for
4230 most simple matches the number is quite small, but for pathological cases it
4231 gets very large very quickly.
4232
4233 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4234 returns the size of the data block pointed to by the study_data field in a
4235 pcre_extra block, that is, the value that was passed as the argument to
4236 pcre_malloc() when PCRE was getting memory in which to place the information
4237 created by pcre_study(). The fourth argument should point to a size_t variable.
4238 pcretest has been extended so that this information is shown after a successful
4239 pcre_study() call when information about the compiled regex is being displayed.
4240
4241 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4242 because what follows is always an absolute path. (Later: it turns out that this
4243 is more than cosmetic for MinGW, because it doesn't like empty path
4244 components.)
4245
4246 49. Some changes have been made to the callout feature (see 28 above):
4247
4248 (i) A callout function now has three choices for what it returns:
4249
4250 0 => success, carry on matching
4251 > 0 => failure at this point, but backtrack if possible
4252 < 0 => serious error, return this value from pcre_exec()
4253
4254 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4255 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4256 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4257 use by callout functions. It will never be used by PCRE itself.
4258
4259 (ii) The pcre_extra structure (see 45 above) has a void * field called
4260 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4261 pcre_callout_block structure has a field of the same name. The contents of
4262 the field passed in the pcre_extra structure are passed to the callout
4263 function in the corresponding field in the callout block. This makes it
4264 easier to use the same callout-containing regex from multiple threads. For
4265 testing, the pcretest program has a new data escape
4266
4267 \C*n pass the number n (may be negative) as callout_data
4268
4269 If the callout function in pcretest receives a non-zero value as
4270 callout_data, it returns that value.
4271
4272 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4273 there were some redundant $(CFLAGS) in commands that are now specified as
4274 $(LINK), which already includes $(CFLAGS).
4275
4276 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4277 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4278 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4279 one-byte characters throughout. Note that case-insensitive matching applies
4280 only to characters whose values are less than 256. PCRE doesn't support the
4281 notion of cases for higher-valued characters.
4282
4283 (i) A character class whose characters are all within 0-255 is handled as
4284 a bit map, and the map is inverted for negative classes. Previously, a
4285 character > 255 always failed to match such a class; however it should
4286 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4287
4288 (ii) A negated character class with a single character < 255 is coded as
4289 "not this character" (OP_NOT). This wasn't working properly when the test
4290 character was multibyte, either singly or repeated.
4291
4292 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4293 mode, for example: \x{100}{2,3}.
4294
4295 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4296 singly or repeated) now correctly test multibyte characters. However,
4297 PCRE doesn't recognize any characters with values greater than 255 as
4298 digits, spaces, or word characters. Such characters always match \D, \S,
4299 and \W, and never match \d, \s, or \w.
4300
4301 (v) Classes may now contain characters and character ranges with values
4302 greater than 255. For example: [ab\x{100}-\x{400}].
4303
4304 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4305 PCRE in UTF-8 mode.
4306
4307 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4308 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4309 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4310 value.)
4311
4312 53. The single man page has become too large. I have therefore split it up into
4313 a number of separate man pages. These also give rise to individual HTML pages;
4314 these are now put in a separate directory, and there is an index.html page that
4315 lists them all. Some hyperlinking between the pages has been installed.
4316
4317 54. Added convenience functions for handling named capturing parentheses.
4318
4319 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4320 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4321 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4322 are faulted.
4323
4324 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4325 calling configure. These values are used when compiling the dftables.c program
4326 which is run to generate the source of the default character tables. They
4327 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4328 you will need to set these values.
4329
4330 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4331
4332
4333 Version 3.9 02-Jan-02
4334 ---------------------
4335
4336 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4337
4338 2. If --disable-static was given, the building process failed when trying to
4339 build pcretest and pcregrep. (For some reason it was using libtool to compile
4340 them, which is not right, as they aren't part of the library.)
4341
4342
4343 Version 3.8 18-Dec-01
4344 ---------------------
4345
4346 1. The experimental UTF-8 code was completely screwed up. It was packing the
4347 bytes in the wrong order. How dumb can you get?
4348
4349
4350 Version 3.7 29-Oct-01
4351 ---------------------
4352
4353 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4354 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4355 this didn't happen under Solaris 8, where I normally test things.
4356
4357 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4358 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4359 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4360 this fix an hour or so after the initial 3.7 release.)
4361
4362
4363 Version 3.6 23-Oct-01
4364 ---------------------
4365
4366 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4367 offsets passed as NULL with zero offset count.
4368
4369 2. The config.guess and config.sub files had not been updated when I moved to
4370 the latest autoconf.
4371
4372
4373 Version 3.5 15-Aug-01
4374 ---------------------
4375
4376 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4377 had been forgotten.
4378
4379 2. By using declared but undefined structures, we can avoid using "void"
4380 definitions in pcre.h while keeping the internal definitions of the structures
4381 private.
4382
4383 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4384 user point of view, this means that both static and shared libraries are built
4385 by default, but this can be individually controlled. More of the work of
4386 handling this static/shared cases is now inside libtool instead of PCRE's make
4387 file.
4388
4389 4. The pcretest utility is now installed along with pcregrep because it is
4390 useful for users (to test regexs) and by doing this, it automatically gets
4391 relinked by libtool. The documentation has been turned into a man page, so
4392 there are now .1, .txt, and .html versions in /doc.
4393
4394 5. Upgrades to pcregrep:
4395 (i) Added long-form option names like gnu grep.
4396 (ii) Added --help to list all options with an explanatory phrase.
4397 (iii) Added -r, --recursive to recurse into sub-directories.
4398 (iv) Added -f, --file to read patterns from a file.
4399
4400 6. pcre_exec() was referring to its "code" argument before testing that
4401 argument for NULL (and giving an error if it was NULL).
4402
4403 7. Upgraded Makefile.in to allow for compiling in a different directory from
4404 the source directory.
4405
4406 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4407 options bits, the pointer it was passed was to an int instead of to an unsigned
4408 long int. This mattered only on 64-bit systems.
4409
4410 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4411 generated) instead of pcre.in, which it its source. Also made the same change
4412 in several of the .c files.
4413
4414 10. A new release of gcc defines printf() as a macro, which broke pcretest
4415 because it had an ifdef in the middle of a string argument for printf(). Fixed
4416 by using separate calls to printf().
4417
4418 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4419 script, to force use of CR or LF instead of \n in the source. On non-Unix
4420 systems, the value can be set in config.h.
4421
4422 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4423 absolute limit. Changed the text of the error message to make this clear, and
4424 likewise updated the man page.
4425
4426 13. The limit of 99 on the number of capturing subpatterns has been removed.
4427 The new limit is 65535, which I hope will not be a "real" limit.
4428
4429
4430 Version 3.4 22-Aug-00
4431 ---------------------
4432
4433 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4434
4435 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4436
4437
4438 Version 3.3 01-Aug-00
4439 ---------------------
4440
4441 1. If an octal character was given, but the value was greater than \377, it
4442 was not getting masked to the least significant bits, as documented. This could
4443 lead to crashes in some systems.
4444
4445 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4446 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4447
4448 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4449 These just pass their arguments on to (pcre_free)(), but they are provided
4450 because some uses of PCRE bind it to non-C systems that can call its functions,
4451 but cannot call free() or pcre_free() directly.
4452
4453 4. Add "make test" as a synonym for "make check". Corrected some comments in
4454 the Makefile.
4455
4456 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4457 Makefile.
4458
4459 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4460 command called pgrep for grepping around the active processes.
4461
4462 7. Added the beginnings of support for UTF-8 character strings.
4463
4464 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4465 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4466 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4467 out for the ar command.)
4468
4469
4470 Version 3.2 12-May-00
4471 ---------------------
4472
4473 This is purely a bug fixing release.
4474
4475 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4476 of ZA. This was just one example of several cases that could provoke this bug,
4477 which was introduced by change 9 of version 2.00. The code for breaking
4478 infinite loops after an iteration that matches an empty string was't working
4479 correctly.
4480
4481 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4482 when matched against abbab (for example). After matching an empty string, it
4483 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4484 caused it to match further down the string than it should.
4485
4486 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4487 was there because it doesn't seem to be needed, and it causes trouble on some
4488 systems, as it is not a Standard C header. It has been removed.
4489
4490 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4491 were reported on the Macintosh. The changes were from
4492
4493 while ((c = *(++ptr)) != 0 && c != '\n');
4494 to
4495 while ((c = *(++ptr)) != 0 && c != '\n') ;
4496
4497 Totally extraordinary, but if that's what it takes...
4498
4499 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4500 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4501 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4502 assumes the way PCRE uses memmove() (always moving upwards).
4503
4504 6. PCRE is being used in one environment where strchr() is not available. There
4505 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4506 faster code anyway.
4507
4508
4509 Version 3.1 09-Feb-00
4510 ---------------------
4511
4512 The only change in this release is the fixing of some bugs in Makefile.in for
4513 the "install" target:
4514
4515 (1) It was failing to install pcreposix.h.
4516
4517 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4518
4519
4520 Version 3.0 01-Feb-00
4521 ---------------------
4522
4523 1. Add support for the /+ modifier to perltest (to output $` like it does in
4524 pcretest).
4525
4526 2. Add support for the /g modifier to perltest.
4527
4528 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4529 matches null strings.
4530
4531 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4532 pattern. Perl treats empty patterns specially - it reuses the most recent
4533 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4534 effect.
4535
4536 5. The POSIX interface was broken in that it was just handing over the POSIX
4537 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4538 required a bigger vector, with some working space on the end. This means that
4539 the POSIX wrapper now has to get and free some memory, and copy the results.
4540
4541 6. Added some simple autoconf support, placing the test data and the
4542 documentation in separate directories, re-organizing some of the
4543 information files, and making it build pcre-config (a GNU standard). Also added
4544 libtool support for building PCRE as a shared library, which is now the
4545 default.
4546
4547 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4548 09 are not valid octal constants. Single digits will be used for minor values
4549 less than 10.
4550
4551 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4552 existing programs that set these in the POSIX interface can use PCRE without
4553 modification.
4554
4555 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4556 return all that pcre_info() returns, plus additional data. The pcre_info()
4557 function is retained for compatibility, but is considered to be obsolete.
4558
4559 10. Added experimental recursion feature (?R) to handle one common case that
4560 Perl 5.6 will be able to do with (?p{...}).
4561
4562 11. Added support for POSIX character classes like [:alpha:], which Perl is
4563 adopting.
4564
4565
4566 Version 2.08 31-Aug-99
4567 ----------------------
4568
4569 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4570 trying to match at the startoffset position, but instead was moving forward to
4571 the next newline as if a previous match had failed.
4572
4573 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4574 and could get into a loop if a null string was matched other than at the start
4575 of the subject.
4576
4577 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4578 be distinguished at compile time, and for completeness also added PCRE_DATE.
4579
4580 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4581 in GnuWin32 environments.
4582
4583
4584 Version 2.07 29-Jul-99
4585 ----------------------
4586
4587 1. The documentation is now supplied in plain text form and HTML as well as in
4588 the form of man page sources.
4589
4590 2. C++ compilers don't like assigning (void *) values to other pointer types.
4591 In particular this affects malloc(). Although there is no problem in Standard
4592 C, I've put in casts to keep C++ compilers happy.
4593
4594 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4595 should be (const char *).
4596
4597 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4598 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4599 However, I haven't made this a standard facility. The documentation doesn't
4600 mention it, and the Makefile doesn't support it.
4601
4602 5. The Makefile now contains an "install" target, with editable destinations at
4603 the top of the file. The pcretest program is not installed.
4604
4605 6. pgrep -V now gives the PCRE version number and date.
4606
4607 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4608 causing the entire string to be ignored, instead of just the last character.
4609
4610 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4611 non-matching string, it can take a very, very long time, even for strings of
4612 quite modest length, because of the nested recursion. PCRE now does better in
4613 some of these cases. It does this by remembering the last required literal
4614 character in the pattern, and pre-searching the subject to ensure it is present
4615 before running the real match. In other words, it applies a heuristic to detect
4616 some types of certain failure quickly, and in the above example, if presented
4617 with a string that has no trailing " it gives "no match" very quickly.
4618
4619 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4620 other alternatives are tried instead.
4621
4622
4623 Version 2.06 09-Jun-99
4624 ----------------------
4625
4626 1. Change pcretest's output for amount of store used to show just the code
4627 space, because the remainder (the data block) varies in size between 32-bit and
4628 64-bit systems.
4629
4630 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4631 start matching at. This allows lookbehinds to work when searching for multiple
4632 occurrences in a string.
4633
4634 3. Added additional options to pcretest for testing multiple occurrences:
4635
4636 /+ outputs the rest of the string that follows a match
4637 /g loops for multiple occurrences, using the new startoffset argument
4638 /G loops for multiple occurrences by passing an incremented pointer
4639
4640 4. PCRE wasn't doing the "first character" optimization for patterns starting
4641 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4642 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4643 the letter 'x'. On long subject strings, this gives a significant speed-up.
4644
4645
4646 Version 2.05 21-Apr-99
4647 ----------------------
4648
4649 1. Changed the type of magic_number from int to long int so that it works
4650 properly on 16-bit systems.
4651
4652 2. Fixed a bug which caused patterns starting with .* not to work correctly
4653 when the subject string contained newline characters. PCRE was assuming
4654 anchoring for such patterns in all cases, which is not correct because .* will
4655 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4656 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4657 must be retried after every newline in the subject.
4658
4659
4660 Version 2.04 18-Feb-99
4661 ----------------------
4662
4663 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4664 computation of the store needed to hold the pattern was incorrect (too large).
4665 If such patterns were nested a few deep, this could multiply and become a real
4666 problem.
4667
4668 2. Added /M option to pcretest to show the memory requirement of a specific
4669 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4670
4671 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4672 compiled in such a way that the backtracking after subsequent failure was
4673 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4674 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4675
4676
4677 Version 2.03 02-Feb-99
4678 ----------------------
4679
4680 1. Fixed typo and small mistake in man page.
4681
4682 2. Added 4th condition (GPL supersedes if conflict) and created separate
4683 LICENCE file containing the conditions.
4684
4685 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4686 Perl, that is the internal \ allows the delimiter to be included in the
4687 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4688 the final delimiter, add \ to the end of the pattern (to test the error).
4689
4690 4. Added the convenience functions for extracting substrings after a successful
4691 match. Updated pcretest to make it able to test these functions.
4692
4693
4694 Version 2.02 14-Jan-99
4695 ----------------------
4696
4697 1. Initialized the working variables associated with each extraction so that
4698 their saving and restoring doesn't refer to uninitialized store.
4699
4700 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4701 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4702 fix the problem.
4703
4704 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4705 calls, and wasn't printing the correct value for compiling calls. Increased the
4706 default value of LOOPREPEAT, and the number of significant figures in the
4707 times.
4708
4709 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4710
4711 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4712 a building problem on Windows NT with a FAT file system.
4713
4714
4715 Version 2.01 21-Oct-98
4716 ----------------------
4717
4718 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4719 to character tables built by pcre_maketables() in the current locale. If NULL
4720 is passed, the default tables are used.
4721
4722
4723 Version 2.00 24-Sep-98
4724 ----------------------
4725
4726 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4727 it any more.
4728
4729 2. Allow quantification of (?>) groups, and make it work correctly.
4730
4731 3. The first character computation wasn't working for (?>) groups.
4732
4733 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4734 end of the subject) and add 5.005's \z, which really does match only at the
4735 very end of the subject.
4736
4737 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4738
4739 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4740 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4741 localized options. All options to pcre_study() were also removed.
4742
4743 7. Add other new features from 5.005:
4744
4745 $(?<= positive lookbehind
4746 $(?<! negative lookbehind
4747 (?imsx-imsx) added the unsetting capability
4748 such a setting is global if at outer level; local otherwise
4749 (?imsx-imsx:) non-capturing groups with option setting
4750 (?(cond)re|re) conditional pattern matching
4751
4752 A backreference to itself in a repeated group matches the previous
4753 captured string.
4754
4755 8. General tidying up of studying (both automatic and via "study")
4756 consequential on the addition of new assertions.
4757
4758 9. As in 5.005, unlimited repeated groups that could match an empty substring
4759 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4760 runtime if any iteration does actually match an empty substring.
4761
4762 10. Include the RunTest script in the distribution.
4763
4764 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4765 discrepancies, some of which were old and were also with respect to 5.004. They
4766 have now been fixed.
4767
4768
4769 Version 1.09 28-Apr-98
4770 ----------------------
4771
4772 1. A negated single character class followed by a quantifier with a minimum
4773 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4774 program crashes, or just wrong answers. This did not apply to negated classes
4775 containing more than one character, or to minima other than one.
4776
4777
4778 Version 1.08 27-Mar-98
4779 ----------------------
4780
4781 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4782
4783 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4784 latter must appear before anything that relies on it in the pattern.
4785
4786
4787 Version 1.07 16-Feb-98
4788 ----------------------
4789
4790 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4791 repeat of a potentially empty string).
4792
4793
4794 Version 1.06 23-Jan-98
4795 ----------------------
4796
4797 1. Added Markus Oberhumer's little patches for C++.
4798
4799 2. Literal strings longer than 255 characters were broken.
4800
4801
4802 Version 1.05 23-Dec-97
4803 ----------------------
4804
4805 1. Negated character classes containing more than one character were failing if
4806 PCRE_CASELESS was set at run time.
4807
4808
4809 Version 1.04 19-Dec-97
4810 ----------------------
4811
4812 1. Corrected the man page, where some "const" qualifiers had been omitted.
4813
4814 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4815 input syntax.
4816
4817 3. Fixed memory leak which occurred when a regex with back references was
4818 matched with an offsets vector that wasn't big enough. The temporary memory
4819 that is used in this case wasn't being freed if the match failed.
4820
4821 4. Tidied pcretest to ensure it frees memory that it gets.
4822
4823 5. Temporary memory was being obtained in the case where the passed offsets
4824 vector was exactly big enough.
4825
4826 6. Corrected definition of offsetof() from change 5 below.
4827
4828 7. I had screwed up change 6 below and broken the rules for the use of
4829 setjmp(). Now fixed.
4830
4831
4832 Version 1.03 18-Dec-97
4833 ----------------------
4834
4835 1. A erroneous regex with a missing opening parenthesis was correctly
4836 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4837 on some systems.
4838
4839 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4840 it was reported that one broken compiler failed on the former because "code" is
4841 also an independent variable.
4842
4843 3. The erroneous regex a[]b caused an array overrun reference.
4844
4845 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4846 fail on data ending with that character. (It was going on too far, and checking
4847 the next character, typically a binary zero.) This was specific to the
4848 optimized code for single-character negative classes.
4849
4850 5. Added a contributed patch from the TIN world which does the following:
4851
4852 + Add an undef for memmove, in case the the system defines a macro for it.
4853
4854 + Add a definition of offsetof(), in case there isn't one. (I don't know
4855 the reason behind this - offsetof() is part of the ANSI standard - but
4856 it does no harm).
4857
4858 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4859 most of the places where whitespace preceded '#'. I have given up and
4860 allowed the remaining 2 cases to be at the margin.
4861
4862 + Rename some variables in pcre to eliminate shadowing. This seems very
4863 pedantic, but does no harm, of course.
4864
4865 6. Moved the call to setjmp() into its own function, to get rid of warnings
4866 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4867
4868 7. Constructs such as \d{8,} were compiling into the equivalent of
4869 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4870 outcome, but in this particular case used more store than had been allocated,
4871 which caused the bug to be discovered because it threw up an internal error.
4872
4873 8. The debugging code in both pcre and pcretest for outputting the compiled
4874 form of a regex was going wrong in the case of back references followed by
4875 curly-bracketed repeats.
4876
4877
4878 Version 1.02 12-Dec-97
4879 ----------------------
4880
4881 1. Typos in pcre.3 and comments in the source fixed.
4882
4883 2. Applied a contributed patch to get rid of places where it used to remove
4884 'const' from variables, and fixed some signed/unsigned and uninitialized
4885 variable warnings.
4886
4887 3. Added the "runtest" target to Makefile.
4888
4889 4. Set default compiler flag to -O2 rather than just -O.
4890
4891
4892 Version 1.01 19-Nov-97
4893 ----------------------
4894
4895 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4896 like /([ab]*)*/, that is, for classes with more than one character in them.
4897
4898 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4899 as /((?>a*))*/ (a PCRE_EXTRA facility).
4900
4901
4902 Version 1.00 18-Nov-97
4903 ----------------------
4904
4905 1. Added compile-time macros to support systems such as SunOS4 which don't have
4906 memmove() or strerror() but have other things that can be used instead.
4907
4908 2. Arranged that "make clean" removes the executables.
4909
4910
4911 Version 0.99 27-Oct-97
4912 ----------------------
4913
4914 1. Fixed bug in code for optimizing classes with only one character. It was
4915 initializing a 32-byte map regardless, which could cause it to run off the end
4916 of the memory it had got.
4917
4918 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4919
4920
4921 Version 0.98 22-Oct-97
4922 ----------------------
4923
4924 1. Fixed bug in code for handling temporary memory usage when there are more
4925 back references than supplied space in the ovector. This could cause segfaults.
4926
4927
4928 Version 0.97 21-Oct-97
4929 ----------------------
4930
4931 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4932
4933 2. Optimized negated single characters not to use a bit map.
4934
4935 3. Brought error texts together as macro definitions; clarified some of them;
4936 fixed one that was wrong - it said "range out of order" when it meant "invalid
4937 escape sequence".
4938
4939 4. Changed some char * arguments to const char *.
4940
4941 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4942
4943 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4944 pcretest.
4945
4946
4947 Version 0.96 16-Oct-97
4948 ----------------------
4949
4950 1. Added a simple "pgrep" utility to the distribution.
4951
4952 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4953 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4954 where "ddd" means "one or more decimal digits".
4955
4956 3. Fixed serious bug. If a pattern had a back reference, but the call to
4957 pcre_exec() didn't supply a large enough ovector to record the related
4958 identifying subpattern, the match always failed. PCRE now remembers the number
4959 of the largest back reference, and gets some temporary memory in which to save
4960 the offsets during matching if necessary, in order to ensure that
4961 backreferences always work.
4962
4963 4. Increased the compatibility with Perl in a number of ways:
4964
4965 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4966 to request this handling. The option can be set at compile or exec time.
4967
4968 (b) $ matches before a terminating newline by default; an option
4969 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4970 mode). The option can be set at compile or exec time.
4971
4972 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4973 the same as Perl's. If the decimal number it represents is less than 10
4974 or there aren't that many previous left capturing parentheses, an octal
4975 escape is read. Inside a character class, it's always an octal escape,
4976 even if it is a single digit.
4977
4978 (d) An escaped but undefined alphabetic character is taken as a literal,
4979 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4980 escapes.
4981
4982 (e) {0} is now permitted. (The previous item is removed from the compiled
4983 pattern).
4984
4985 5. Changed all the names of code files so that the basic parts are no longer
4986 than 10 characters, and abolished the teeny "globals.c" file.
4987
4988 6. Changed the handling of character classes; they are now done with a 32-byte
4989 bit map always.
4990
4991 7. Added the -d and /D options to pcretest to make it possible to look at the
4992 internals of compilation without having to recompile pcre.
4993
4994
4995 Version 0.95 23-Sep-97
4996 ----------------------
4997
4998 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4999 \x20 at the start of a run of normal characters. These were being treated as
5000 real characters, instead of the source characters being re-checked.
5001
5002
5003 Version 0.94 18-Sep-97
5004 ----------------------
5005
5006 1. The functions are now thread-safe, with the caveat that the global variables
5007 containing pointers to malloc() and free() or alternative functions are the
5008 same for all threads.
5009
5010 2. Get pcre_study() to generate a bitmap of initial characters for non-
5011 anchored patterns when this is possible, and use it if passed to pcre_exec().
5012
5013
5014 Version 0.93 15-Sep-97
5015 ----------------------
5016
5017 1. /(b)|(:+)/ was computing an incorrect first character.
5018
5019 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5020 but not actually doing anything yet.
5021
5022 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5023 as Perl does (e.g. [-az] or [az-]).
5024
5025 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5026 all possible positions.
5027
5028 5. Split up into different modules to avoid including unneeded functions in a
5029 compiled binary. However, compile and exec are still in one module. The "study"
5030 function is split off.
5031
5032 6. The character tables are now in a separate module whose source is generated
5033 by an auxiliary program - but can then be edited by hand if required. There are
5034 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5035 toupper() in the code.
5036
5037 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5038 make them global. Abolish the function for setting them, as the caller can now
5039 set them directly.
5040
5041
5042 Version 0.92 11-Sep-97
5043 ----------------------
5044
5045 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5046 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5047
5048 2. Caseless matching was not working in character classes if the characters in
5049 the pattern were in upper case.
5050
5051 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5052
5053 4. Make PCRE_ANCHORED public and accept as a compile option.
5054
5055 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5056 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5057 pass them.
5058
5059 6. Give an error if bad option bits passed at compile or run time.
5060
5061 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5062 pcretest to cause it to pass that flag.
5063
5064 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5065 options, and the first character, if set.
5066
5067 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5068
5069
5070 Version 0.91 10-Sep-97
5071 ----------------------
5072
5073 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5074 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5075
5076 2. PCRE was looping on encountering an indefinitely repeated back reference to
5077 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5078 Perl does - treats the match as successful.
5079
5080 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5