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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1320 - (show annotations)
Wed May 1 16:39:35 2013 UTC (6 years, 5 months ago) by ph10
File size: 246336 byte(s)
Source tidies (trails spaces, html updates) for 8.33-RC1.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 28-April-2013
5 --------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80 be expressed in certain programming languages using Universal Character
81 Names.
82
83 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84 which points out that "non characters" are not "characters that may not
85 appear in Unicode strings" but rather "characters that are reserved for
86 internal use and have only local meaning".
87
88 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89 there was a conditional group that depended on an assertion, if the
90 assertion was false, the callout that immediately followed the alternation
91 in the condition was skipped when pcre_exec() was used for matching.
92
93 23. Allow an explicit callout to be inserted before an assertion that is the
94 condition for a conditional group, for compatibility with automatic
95 callouts, which always insert a callout at this point.
96
97 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99
100 25. (*PRUNE) is now supported by the JIT compiler.
101
102 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103
104 27. Fix the case where there are two or more SKIPs with arguments that may be
105 ignored.
106
107 28. (*SKIP) is now supported by the JIT compiler.
108
109 29. (*THEN) is now supported by the JIT compiler.
110
111 30. Update RunTest with additional test selector options.
112
113 31. The way PCRE handles backtracking verbs has been changed in two ways.
114
115 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
116 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
117 backtracking. In some cases this makes it more Perl-compatible, but Perl's
118 rather obscure rules do not always do the same thing.
119
120 (2) Previously, backtracking verbs were confined within assertions. This is
121 no longer the case for positive assertions, except for (*ACCEPT). Again,
122 this sometimes improves Perl compatibility, and sometimes does not.
123
124 32. A number of tests that were in test 2 because Perl did things differently
125 have been moved to test 1, because either Perl or PCRE has changed, and
126 these tests are now compatible.
127
128 32. Control verbs are handled in the same way in JIT and interpreter.
129
130 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
131 contained a forward subroutine reference caused a compile error.
132
133 34. Auto-detect and optimize limited repetitions in JIT.
134
135 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
136 blocking (*UTF) etc.
137
138 36. In the interpreter, maximizing pattern repetitions for characters and
139 character types now use tail recursion, which reduces stack usage.
140
141 37. The value of the max lookbehind was not correctly preserved if a compiled
142 and saved regex was reloaded on a host of different endianness.
143
144 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
145 of the compiled pattern block, expand the flags field from 16 to 32 bits
146 because it was almost full.
147
148
149 Version 8.32 30-November-2012
150 -----------------------------
151
152 1. Improved JIT compiler optimizations for first character search and single
153 character iterators.
154
155 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
156 Patch by Daniel Richard G.
157
158 3. Single character iterator optimizations in the JIT compiler.
159
160 4. Improved JIT compiler optimizations for character ranges.
161
162 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
163 Reported by Giuseppe D'Angelo.
164
165 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
166 of a line, was being set incorrectly in cases where .* appeared inside
167 atomic brackets at the start of a pattern, or where there was a subsequent
168 *PRUNE or *SKIP.
169
170 7. Improved instruction cache flush for POWER/PowerPC.
171 Patch by Daniel Richard G.
172
173 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
174 grep:
175
176 (a) There is now no limit to the number of patterns to be matched.
177
178 (b) An error is given if a pattern is too long.
179
180 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
181 are now supported.
182
183 (d) --exclude-from and --include-from (multiple use) have been added.
184
185 (e) Exclusions and inclusions now apply to all files and directories, not
186 just to those obtained from scanning a directory recursively.
187
188 (f) Multiple uses of -f and --file-list are now supported.
189
190 (g) In a Windows environment, the default for -d has been changed from
191 "read" (the GNU grep default) to "skip", because otherwise the presence
192 of a directory in the file list provokes an error.
193
194 (h) The documentation has been revised and clarified in places.
195
196 9. Improve the matching speed of capturing brackets.
197
198 10. Changed the meaning of \X so that it now matches a Unicode extended
199 grapheme cluster.
200
201 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
202 out POSIX threads when JIT support is configured.
203
204 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
205
206 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
207 structure could go wrong in environments where size_t is not the same size
208 as int.
209
210 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
211 set.
212
213 15. The EBCDIC support had decayed; later updates to the code had included
214 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
215 general tidy up of EBCDIC-related issues, and the documentation was also
216 not quite right. There is now a test that can be run on ASCII systems to
217 check some of the EBCDIC-related things (but is it not a full test).
218
219 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
220 in a small tidy to the code.
221
222 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
223
224 18. If the --only-matching (-o) option in pcregrep is specified multiple
225 times, each one causes appropriate output. For example, -o1 -o2 outputs the
226 substrings matched by the 1st and 2nd capturing parentheses. A separating
227 string can be specified by --om-separator (default empty).
228
229 19. Improving the first n character searches.
230
231 20. Turn case lists for horizontal and vertical white space into macros so that
232 they are defined only once.
233
234 21. This set of changes together give more compatible Unicode case-folding
235 behaviour for characters that have more than one other case when UCP
236 support is available.
237
238 (a) The Unicode property table now has offsets into a new table of sets of
239 three or more characters that are case-equivalent. The MultiStage2.py
240 script that generates these tables (the pcre_ucd.c file) now scans
241 CaseFolding.txt instead of UnicodeData.txt for character case
242 information.
243
244 (b) The code for adding characters or ranges of characters to a character
245 class has been abstracted into a generalized function that also handles
246 case-independence. In UTF-mode with UCP support, this uses the new data
247 to handle characters with more than one other case.
248
249 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
250 whose other case is greater than 256 are now correctly matched
251 caselessly. Previously, the high codepoint matched the low one, but not
252 vice versa.
253
254 (d) The processing of \h, \H, \v, and \ in character classes now makes use
255 of the new class addition function, using character lists defined as
256 macros alongside the case definitions of 20 above.
257
258 (e) Caseless back references now work with characters that have more than
259 one other case.
260
261 (f) General caseless matching of characters with more than one other case
262 is supported.
263
264 22. Unicode character properties were updated from Unicode 6.2.0
265
266 23. Improved CMake support under Windows. Patch by Daniel Richard G.
267
268 24. Add support for 32-bit character strings, and UTF-32
269
270 25. Major JIT compiler update (code refactoring and bugfixing).
271 Experimental Sparc 32 support is added.
272
273 26. Applied a modified version of Daniel Richard G's patch to create
274 pcre.h.generic and config.h.generic by "make" instead of in the
275 PrepareRelease script.
276
277 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
278 pcre_compile.c when checking for a zero character.
279
280 28. Introducing a native interface for JIT. Through this interface, the compiled
281 machine code can be directly executed. The purpose of this interface is to
282 provide fast pattern matching, so several sanity checks are not performed.
283 However, feature tests are still performed. The new interface provides
284 1.4x speedup compared to the old one.
285
286 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
287 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
288 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
289
290 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
291 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
292 the "old" RFC 2279). Instead, it ended up passing a negative length to
293 pcre_exec().
294
295 31. Add support for GCC's visibility feature to hide internal functions.
296
297 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
298 "unknown -C option" after outputting 0 or 1.
299
300 33. There is now support for generating a code coverage report for the test
301 suite in environments where gcc is the compiler and lcov is installed. This
302 is mainly for the benefit of the developers.
303
304 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
305 unaddressable using valgrind annotations, allowing valgrind to detect
306 invalid memory accesses. This is mainly for the benefit of the developers.
307
308 25. (*UTF) can now be used to start a pattern in any of the three libraries.
309
310 26. Give configure error if --enable-cpp but no C++ compiler found.
311
312
313 Version 8.31 06-July-2012
314 -------------------------
315
316 1. Fixing a wrong JIT test case and some compiler warnings.
317
318 2. Removed a bashism from the RunTest script.
319
320 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
321 to unsigned type, result still unsigned" that was given by an MS compiler
322 on encountering the code "-sizeof(xxx)".
323
324 4. Partial matching support is added to the JIT compiler.
325
326 5. Fixed several bugs concerned with partial matching of items that consist
327 of more than one character:
328
329 (a) /^(..)\1/ did not partially match "aba" because checking references was
330 done on an "all or nothing" basis. This also applied to repeated
331 references.
332
333 (b) \R did not give a hard partial match if \r was found at the end of the
334 subject.
335
336 (c) \X did not give a hard partial match after matching one or more
337 characters at the end of the subject.
338
339 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
340 a partial match for the string "\r".
341
342 (e) When newline was set to CRLF, the metacharacter "." did not recognize
343 a partial match for a CR character at the end of the subject string.
344
345 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
346 running pcretest, the text "(JIT)" added to the output whenever JIT is
347 actually used to run the match.
348
349 7. Individual JIT compile options can be set in pcretest by following -s+[+]
350 or /S+[+] with a digit between 1 and 7.
351
352 8. OP_NOT now supports any UTF character not just single-byte ones.
353
354 9. (*MARK) control verb is now supported by the JIT compiler.
355
356 10. The command "./RunTest list" lists the available tests without actually
357 running any of them. (Because I keep forgetting what they all are.)
358
359 11. Add PCRE_INFO_MAXLOOKBEHIND.
360
361 12. Applied a (slightly modified) user-supplied patch that improves performance
362 when the heap is used for recursion (compiled with --disable-stack-for-
363 recursion). Instead of malloc and free for each heap frame each time a
364 logical recursion happens, frames are retained on a chain and re-used where
365 possible. This sometimes gives as much as 30% improvement.
366
367 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
368 call.
369
370 14. As documented, (*COMMIT) is now confined to within a positive assertion.
371
372 15. It is now possible to link pcretest with libedit as an alternative to
373 libreadline.
374
375 16. (*COMMIT) control verb is now supported by the JIT compiler.
376
377 17. The Unicode data tables have been updated to Unicode 6.1.0.
378
379 18. Added --file-list option to pcregrep.
380
381 19. Added binary file support to pcregrep, including the -a, --binary-files,
382 -I, and --text options.
383
384 20. The madvise function is renamed for posix_madvise for QNX compatibility
385 reasons. Fixed by Giuseppe D'Angelo.
386
387 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
388
389 22. Changed the option for creating symbolic links for 16-bit man pages from
390 -s to -sf so that re-installing does not cause issues.
391
392 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
393
394 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
395 matches in certain environments (the workspace was not being correctly
396 retained). Also added to pcre_dfa_exec() a simple plausibility check on
397 some of the workspace data at the beginning of a restart.
398
399 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
400 was not doing so when it should - probably a typo introduced by SVN 528
401 (change 8.10/14).
402
403 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
404 \w+ when the character tables indicated that \x{c4} was a word character.
405 There were several related cases, all because the tests for doing a table
406 lookup were testing for characters less than 127 instead of 255.
407
408 27. If a pattern contains capturing parentheses that are not used in a match,
409 their slots in the ovector are set to -1. For those that are higher than
410 any matched groups, this happens at the end of processing. In the case when
411 there were back references that the ovector was too small to contain
412 (causing temporary malloc'd memory to be used during matching), and the
413 highest capturing number was not used, memory off the end of the ovector
414 was incorrectly being set to -1. (It was using the size of the temporary
415 memory instead of the true size.)
416
417 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
418 ovector size, it uses memory at the end of the block that it has got.
419
420 29. Check for an overlong MARK name and give an error at compile time. The
421 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
422
423 30. JIT compiler update.
424
425 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
426 Rill for the patch.
427
428 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
429
430 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
431
432 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
433 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
434 was enabled.
435
436 35. Improve JIT code generation for greedy plus quantifier.
437
438 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
439 1 to "aa" instead of to an empty string. The bug affected repeated groups
440 that could potentially match an empty string.
441
442 37. Optimizing single character iterators in JIT.
443
444 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
445 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
446 codepoints that are too big for the mode are faulted, and in a UTF mode,
447 disallowed codepoints are also faulted.
448
449 39. If PCRE was compiled with UTF support, in three places in the DFA
450 matcher there was code that should only have been obeyed in UTF mode, but
451 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
452 processing when bytes with values greater than 127 were present. In 16-bit
453 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
454 both cases the values are those that cannot be the first data item in a UTF
455 character. The three items that might have provoked this were recursions,
456 possessively repeated groups, and atomic groups.
457
458 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
459 and pcregrep, because some OS require shared objects to be explicitly
460 passed to ld, causing the link step to fail if they are not.
461
462 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
463 mode, patterns that started with \h* or \R* might be incorrectly matched.
464
465
466 Version 8.30 04-February-2012
467 -----------------------------
468
469 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
470 name is defined in ctype.h.
471
472 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
473 only in quite long subpatterns.
474
475 3. Removed the function pcre_info(), which has been obsolete and deprecated
476 since it was replaced by pcre_fullinfo() in February 2000.
477
478 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
479 match a (*MARK), and the match failed at the start of the subject, a
480 reference to memory before the start of the subject could occur. This bug
481 was introduced by fix 17 of release 8.21.
482
483 5. A reference to an unset group with zero minimum repetition was giving
484 totally wrong answers (in non-JavaScript-compatibility mode). For example,
485 /(another)?(\1?)test/ matched against "hello world test". This bug was
486 introduced in release 8.13.
487
488 6. Add support for 16-bit character strings (a large amount of work involving
489 many changes and refactorings).
490
491 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
492 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
493 from a file.
494
495 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
496 rounding is not applied in this particular case).
497
498 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
499 if they appear, or are escaped, in patterns.
500
501 10. Get rid of a number of -Wunused-but-set-variable warnings.
502
503 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
504 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
505 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
506 also returns the mark "x". This bug applied to capturing parentheses,
507 non-capturing parentheses, and atomic parentheses. It also applied to some
508 assertions.
509
510 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
511 information out of configure.ac instead of relying on pcre.h.generic, which
512 is not stored in the repository.
513
514 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
515 -lreadline.
516
517 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
518
519 15. Applied Graycode's patch to put the top-level frame on the stack rather
520 than the heap when not using the stack for recursion. This gives a
521 performance improvement in many cases when recursion is not deep.
522
523 16. Experimental code added to "pcretest -C" to output the stack frame size.
524
525
526 Version 8.21 12-Dec-2011
527 ------------------------
528
529 1. Updating the JIT compiler.
530
531 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
532 are added as well.
533
534 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
535 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
536 calling _pcre_jit_exec. Some extra comments are added.
537
538 4. (*MARK) settings inside atomic groups that do not contain any capturing
539 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
540 was introduced by change 18 for 8.20.
541
542 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
543 ECMA-262 standard.
544
545 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
546 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
547 This bug was probably introduced by change 9 of 8.13.
548
549 7. While fixing 6 above, I noticed that a number of other items were being
550 incorrectly rejected as "not fixed length". This arose partly because newer
551 opcodes had not been added to the fixed-length checking code. I have (a)
552 corrected the bug and added tests for these items, and (b) arranged for an
553 error to occur if an unknown opcode is encountered while checking for fixed
554 length instead of just assuming "not fixed length". The items that were
555 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
556 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
557 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
558
559 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
560 being incorrectly compiled and would have given unpredicatble results.
561
562 9. A possessively repeated subpattern with minimum repeat count greater than
563 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
564 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
565 the first (A) could occur when it should not.
566
567 10. Add a cast and remove a redundant test from the code.
568
569 11. JIT should use pcre_malloc/pcre_free for allocation.
570
571 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
572 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
573 is anything other than /usr, -L is still shown).
574
575 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
576
577 14. Perl does not support \N without a following name in a [] class; PCRE now
578 also gives an error.
579
580 15. If a forward reference was repeated with an upper limit of around 2000,
581 it caused the error "internal error: overran compiling workspace". The
582 maximum number of forward references (including repeats) was limited by the
583 internal workspace, and dependent on the LINK_SIZE. The code has been
584 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
585 the default depends on LINK_SIZE. There is a new upper limit (for safety)
586 of around 200,000 forward references. While doing this, I also speeded up
587 the filling in of repeated forward references.
588
589 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
590 incorrectly expecting the subject to contain another "a" after the start.
591
592 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
593 in the match, the SKIP should be ignored. This was not happening; instead
594 the SKIP was being treated as NOMATCH. For patterns such as
595 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
596 tested.
597
598 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
599 now much more compatible with Perl, in particular in cases where the result
600 is a non-match for a non-anchored pattern. For example, if
601 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
602 "m", where previously it did not return a name. A side effect of this
603 change is that for partial matches, the last encountered mark name is
604 returned, as for non matches. A number of tests that were previously not
605 Perl-compatible have been moved into the Perl-compatible test files. The
606 refactoring has had the pleasing side effect of removing one argument from
607 the match() function, thus reducing its stack requirements.
608
609 19. If the /S+ option was used in pcretest to study a pattern using JIT,
610 subsequent uses of /S (without +) incorrectly behaved like /S+.
611
612 21. Retrieve executable code size support for the JIT compiler and fixing
613 some warnings.
614
615 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
616 not work when the shorter character appeared right at the end of the
617 subject string.
618
619 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
620 systems.
621
622 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
623 output it when the /M option is used in pcretest.
624
625 25. The CheckMan script was not being included in the distribution. Also, added
626 an explicit "perl" to run Perl scripts from the PrepareRelease script
627 because this is reportedly needed in Windows.
628
629 26. If study data was being save in a file and studying had not found a set of
630 "starts with" bytes for the pattern, the data written to the file (though
631 never used) was taken from uninitialized memory and so caused valgrind to
632 complain.
633
634 27. Updated RunTest.bat as provided by Sheri Pierce.
635
636 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
637
638 29. Computation of memory usage for the table of capturing group names was
639 giving an unnecessarily large value.
640
641
642 Version 8.20 21-Oct-2011
643 ------------------------
644
645 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
646 a POSIX class. After further experiments with Perl, which convinced me that
647 Perl has bugs and confusions, a closing square bracket is no longer allowed
648 in a POSIX name. This bug also affected patterns with classes that started
649 with full stops.
650
651 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
652 captured substring, but while checking the failing first alternative,
653 substring 1 is temporarily captured. If the output vector supplied to
654 pcre_exec() was not big enough for this capture, the yield of the function
655 was still zero ("insufficient space for captured substrings"). This cannot
656 be totally fixed without adding another stack variable, which seems a lot
657 of expense for a edge case. However, I have improved the situation in cases
658 such as /(a)(b)x|abc/ matched against "abc", where the return code
659 indicates that fewer than the maximum number of slots in the ovector have
660 been set.
661
662 3. Related to (2) above: when there are more back references in a pattern than
663 slots in the output vector, pcre_exec() uses temporary memory during
664 matching, and copies in the captures as far as possible afterwards. It was
665 using the entire output vector, but this conflicts with the specification
666 that only 2/3 is used for passing back captured substrings. Now it uses
667 only the first 2/3, for compatibility. This is, of course, another edge
668 case.
669
670 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
671 main code base, and can be used by building with --enable-jit. When this is
672 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
673 runtime --no-jit option is given.
674
675 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
676 ovector, the return from the function was zero, implying that there were
677 other matches that did not fit. The correct "exactly full" value is now
678 returned.
679
680 6. If a subpattern that was called recursively or as a subroutine contained
681 (*PRUNE) or any other control that caused it to give a non-standard return,
682 invalid errors such as "Error -26 (nested recursion at the same subject
683 position)" or even infinite loops could occur.
684
685 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
686 computing the minimum length on reaching *ACCEPT, and so ended up with the
687 wrong value of 1 rather than 0. Further investigation indicates that
688 computing a minimum subject length in the presence of *ACCEPT is difficult
689 (think back references, subroutine calls), and so I have changed the code
690 so that no minimum is registered for a pattern that contains *ACCEPT.
691
692 8. If (*THEN) was present in the first (true) branch of a conditional group,
693 it was not handled as intended. [But see 16 below.]
694
695 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
696 Sheri Pierce.
697
698 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
699 the first byte in a match must be "a".
700
701 11. Change 17 for 8.13 increased the recursion depth for patterns like
702 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
703 pattern contains any instances of (*THEN). If it does not, the old
704 optimizations are restored. It would be nice to do this on a per-group
705 basis, but at the moment that is not feasible.
706
707 12. In some environments, the output of pcretest -C is CRLF terminated. This
708 broke RunTest's code that checks for the link size. A single white space
709 character after the value is now allowed for.
710
711 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
712 For "fr", it uses the Windows-specific input and output files.
713
714 14. If (*THEN) appeared in a group that was called recursively or as a
715 subroutine, it did not work as intended. [But see next item.]
716
717 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
718 pattern fragments (but not containing any | characters). If A and B are
719 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
720 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
721 D. In other words, Perl considers parentheses that do not contain any |
722 characters to be part of a surrounding alternative, whereas PCRE was
723 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
724 differently. PCRE now behaves in the same way as Perl, except in the case
725 of subroutine/recursion calls such as (?1) which have in any case always
726 been different (but PCRE had them first :-).
727
728 16. Related to 15 above: Perl does not treat the | in a conditional group as
729 creating alternatives. Such a group is treated in the same way as an
730 ordinary group without any | characters when processing (*THEN). PCRE has
731 been changed to match Perl's behaviour.
732
733 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
734 RunGrepTest script failed.
735
736 18. Change 22 for version 13 caused atomic groups to use more stack. This is
737 inevitable for groups that contain captures, but it can lead to a lot of
738 stack use in large patterns. The old behaviour has been restored for atomic
739 groups that do not contain any capturing parentheses.
740
741 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
742 suppress the check for a minimum subject length at run time. (If it was
743 given to pcre_exec() or pcre_dfa_exec() it did work.)
744
745 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
746 fail to work when decoding hex characters in data strings in EBCDIC
747 environments.
748
749 21. It appears that in at least one Mac OS environment, the isxdigit() function
750 is implemented as a macro that evaluates to its argument more than once,
751 contravening the C 90 Standard (I haven't checked a later standard). There
752 was an instance in pcretest which caused it to go wrong when processing
753 \x{...} escapes in subject strings. The has been rewritten to avoid using
754 things like p++ in the argument of isxdigit().
755
756
757 Version 8.13 16-Aug-2011
758 ------------------------
759
760 1. The Unicode data tables have been updated to Unicode 6.0.0.
761
762 2. Two minor typos in pcre_internal.h have been fixed.
763
764 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
765 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
766 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
767
768 4. There were a number of related bugs in the code for matching backrefences
769 caselessly in UTF-8 mode when codes for the characters concerned were
770 different numbers of bytes. For example, U+023A and U+2C65 are an upper
771 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
772 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
773 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
774 2-byte code at the end of the subject (it thought there wasn't enough data
775 left).
776
777 5. Comprehensive information about what went wrong is now returned by
778 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
779 as the output vector has at least 2 elements. The offset of the start of
780 the failing character and a reason code are placed in the vector.
781
782 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
783 now returned is for the first byte of the failing character, instead of the
784 last byte inspected. This is an incompatible change, but I hope it is small
785 enough not to be a problem. It makes the returned offset consistent with
786 pcre_exec() and pcre_dfa_exec().
787
788 7. pcretest now gives a text phrase as well as the error number when
789 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
790 failure, the offset and reason code are output.
791
792 8. When \R was used with a maximizing quantifier it failed to skip backwards
793 over a \r\n pair if the subsequent match failed. Instead, it just skipped
794 back over a single character (\n). This seems wrong (because it treated the
795 two characters as a single entity when going forwards), conflicts with the
796 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
797 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
798 has been changed.
799
800 9. Some internal refactoring has changed the processing so that the handling
801 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
802 time (the PCRE_DOTALL option was changed this way some time ago: version
803 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
804 which was always a bit of a fudge. It also means that there is one less
805 argument for the match() function, which reduces its stack requirements
806 slightly. This change also fixes an incompatibility with Perl: the pattern
807 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
808
809 10. More internal refactoring has drastically reduced the number of recursive
810 calls to match() for possessively repeated groups such as (abc)++ when
811 using pcre_exec().
812
813 11. While implementing 10, a number of bugs in the handling of groups were
814 discovered and fixed:
815
816 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
817 (a|)*(?1) gave a compile-time internal error.
818 ((a|)+)+ did not notice that the outer group could match an empty string.
819 (^a|^)+ was not marked as anchored.
820 (.*a|.*)+ was not marked as matching at start or after a newline.
821
822 12. Yet more internal refactoring has removed another argument from the match()
823 function. Special calls to this function are now indicated by setting a
824 value in a variable in the "match data" data block.
825
826 13. Be more explicit in pcre_study() instead of relying on "default" for
827 opcodes that mean there is no starting character; this means that when new
828 ones are added and accidentally left out of pcre_study(), testing should
829 pick them up.
830
831 14. The -s option of pcretest has been documented for ages as being an old
832 synonym of -m (show memory usage). I have changed it to mean "force study
833 for every regex", that is, assume /S for every regex. This is similar to -i
834 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
835 using it. It makes it easier to run collections of tests with and without
836 study enabled, and thereby test pcre_study() more easily. All the standard
837 tests are now run with and without -s (but some patterns can be marked as
838 "never study" - see 20 below).
839
840 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
841 restoration of the capturing data to the outer values was not happening
842 correctly.
843
844 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
845 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
846 pattern had matched an empty string, and so incorrectly returned a no
847 match.
848
849 17. There was optimizing code for the last branch of non-capturing parentheses,
850 and also for the obeyed branch of a conditional subexpression, which used
851 tail recursion to cut down on stack usage. Unfortunately, now that there is
852 the possibility of (*THEN) occurring in these branches, tail recursion is
853 no longer possible because the return has to be checked for (*THEN). These
854 two optimizations have therefore been removed. [But see 8.20/11 above.]
855
856 18. If a pattern containing \R was studied, it was assumed that \R always
857 matched two bytes, thus causing the minimum subject length to be
858 incorrectly computed because \R can also match just one byte.
859
860 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
861 was incorrectly computed.
862
863 20. If /S is present twice on a test pattern in pcretest input, it now
864 *disables* studying, thereby overriding the use of -s on the command line
865 (see 14 above). This is necessary for one or two tests to keep the output
866 identical in both cases.
867
868 21. When (*ACCEPT) was used in an assertion that matched an empty string and
869 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
870
871 22. When an atomic group that contained a capturing parenthesis was
872 successfully matched, but the branch in which it appeared failed, the
873 capturing was not being forgotten if a higher numbered group was later
874 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
875 group 1 to "a", when in fact it should be unset. This applied to multi-
876 branched capturing and non-capturing groups, repeated or not, and also to
877 positive assertions (capturing in negative assertions does not happen
878 in PCRE) and also to nested atomic groups.
879
880 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
881 subject after a captured substring, to make it easier to tell which of a
882 number of identical substrings has been captured.
883
884 24. The way atomic groups are processed by pcre_exec() has been changed so that
885 if they are repeated, backtracking one repetition now resets captured
886 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
887 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
888 "aaa". Previously, it would have been "a". As part of this code
889 refactoring, the way recursive calls are handled has also been changed.
890
891 25. If an assertion condition captured any substrings, they were not passed
892 back unless some other capturing happened later. For example, if
893 (?(?=(a))a) was matched against "a", no capturing was returned.
894
895 26. When studying a pattern that contained subroutine calls or assertions,
896 the code for finding the minimum length of a possible match was handling
897 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
898 group 1 called group 2 while simultaneously a separate group 2 called group
899 1). A stack overflow occurred in this case. I have fixed this by limiting
900 the recursion depth to 10.
901
902 27. Updated RunTest.bat in the distribution to the version supplied by Tom
903 Fortmann. This supports explicit test numbers on the command line, and has
904 argument validation and error reporting.
905
906 28. An instance of \X with an unlimited repeat could fail if at any point the
907 first character it looked at was a mark character.
908
909 29. Some minor code refactoring concerning Unicode properties and scripts
910 should reduce the stack requirement of match() slightly.
911
912 30. Added the '=' option to pcretest to check the setting of unused capturing
913 slots at the end of the pattern, which are documented as being -1, but are
914 not included in the return count.
915
916 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
917 compiled something random. Now it gives a compile-time error (as does
918 Perl).
919
920 32. A *MARK encountered during the processing of a positive assertion is now
921 recorded and passed back (compatible with Perl).
922
923 33. If --only-matching or --colour was set on a pcregrep call whose pattern
924 had alternative anchored branches, the search for a second match in a line
925 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
926 matched the line "0102" twice. The same bug affected patterns that started
927 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
928 twice.
929
930 34. Previously, PCRE did not allow quantification of assertions. However, Perl
931 does, and because of capturing effects, quantifying parenthesized
932 assertions may at times be useful. Quantifiers are now allowed for
933 parenthesized assertions.
934
935 35. A minor code tidy in pcre_compile() when checking options for \R usage.
936
937 36. \g was being checked for fancy things in a character class, when it should
938 just be a literal "g".
939
940 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
941 appearance of a nested POSIX class supersedes an apparent external class.
942 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
943 unescaped square brackets may also appear as part of class names. For
944 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
945 more like Perl. (But see 8.20/1 above.)
946
947 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
948 was because it thought it was \N{name}, which is not supported).
949
950 39. Add minix to OS list not supporting the -S option in pcretest.
951
952 40. PCRE tries to detect cases of infinite recursion at compile time, but it
953 cannot analyze patterns in sufficient detail to catch mutual recursions
954 such as ((?1))((?2)). There is now a runtime test that gives an error if a
955 subgroup is called recursively as a subpattern for a second time at the
956 same position in the subject string. In previous releases this might have
957 been caught by the recursion limit, or it might have run out of stack.
958
959 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
960 happen only once. PCRE was, however incorrectly giving a compile time error
961 "recursive call could loop indefinitely" because it cannot analyze the
962 pattern in sufficient detail. The compile time test no longer happens when
963 PCRE is compiling a conditional subpattern, but actual runaway loops are
964 now caught at runtime (see 40 above).
965
966 42. It seems that Perl allows any characters other than a closing parenthesis
967 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
968 has been changed to be the same.
969
970 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
971 as not to get warnings when autogen.sh is called. Also changed
972 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
973
974 44. To help people who use pcregrep to scan files containing exceedingly long
975 lines, the following changes have been made:
976
977 (a) The default value of the buffer size parameter has been increased from
978 8K to 20K. (The actual buffer used is three times this size.)
979
980 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
981 PCRE is built.
982
983 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
984 to be set at run time.
985
986 (d) Numerical values in pcregrep options can be followed by K or M, for
987 example --buffer-size=50K.
988
989 (e) If a line being scanned overflows pcregrep's buffer, an error is now
990 given and the return code is set to 2.
991
992 45. Add a pointer to the latest mark to the callout data block.
993
994 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
995 partial match of an empty string instead of no match. This was specific to
996 the use of ".".
997
998 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
999 complete match instead of a partial match. This bug was dependent on both
1000 the PCRE_UTF8 and PCRE_DOTALL options being set.
1001
1002 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1003 starting byte set, because \b was not being ignored.
1004
1005
1006 Version 8.12 15-Jan-2011
1007 ------------------------
1008
1009 1. Fixed some typos in the markup of the man pages, and wrote a script that
1010 checks for such things as part of the documentation building process.
1011
1012 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1013 --match-limit and --recursion-limit options (added for 8.11). In
1014 particular, this made one of the standard tests fail. (The integer value
1015 went into the wrong half of a long int.)
1016
1017 3. If the --colour option was given to pcregrep with -v (invert match), it
1018 did strange things, either producing crazy output, or crashing. It should,
1019 of course, ignore a request for colour when reporting lines that do not
1020 match.
1021
1022 4. Another pcregrep bug caused similar problems if --colour was specified with
1023 -M (multiline) and the pattern match finished with a line ending.
1024
1025 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1026 matched in multiline mode, the following line was shown as part of the
1027 match. This seems wrong, so I have changed it.
1028
1029 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1030 the check for further matches in the same line (so they could be coloured)
1031 to overrun the end of the current line. If another match was found, it was
1032 incorrectly shown (and then shown again when found in the next line).
1033
1034 7. If pcregrep was compiled under Windows, there was a reference to the
1035 function pcregrep_exit() before it was defined. I am assuming this was
1036 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1037 reported by a user. I've moved the definition above the reference.
1038
1039
1040 Version 8.11 10-Dec-2010
1041 ------------------------
1042
1043 1. (*THEN) was not working properly if there were untried alternatives prior
1044 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1045 backtracked to try for "b" instead of moving to the next alternative branch
1046 at the same level (in this case, to look for "c"). The Perl documentation
1047 is clear that when (*THEN) is backtracked onto, it goes to the "next
1048 alternative in the innermost enclosing group".
1049
1050 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1051 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1052 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1053 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1054 (*THEN).
1055
1056 3. If \s appeared in a character class, it removed the VT character from
1057 the class, even if it had been included by some previous item, for example
1058 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1059 of \s, but is part of the POSIX "space" class.)
1060
1061 4. A partial match never returns an empty string (because you can always
1062 match an empty string at the end of the subject); however the checking for
1063 an empty string was starting at the "start of match" point. This has been
1064 changed to the "earliest inspected character" point, because the returned
1065 data for a partial match starts at this character. This means that, for
1066 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1067 (previously it gave "no match").
1068
1069 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1070 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1071 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1072 has an implication that the given string is incomplete (because a partial
1073 match is preferred over a full match). For this reason, these items now
1074 give a partial match in this situation. [Aside: previously, the one case
1075 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1076 match rather than a full match, which was wrong by the old rules, but is
1077 now correct.]
1078
1079 6. There was a bug in the handling of #-introduced comments, recognized when
1080 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1081 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1082 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1083 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1084 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1085 places in pcre_compile().
1086
1087 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1088 comments when looking ahead for named forward references to subpatterns,
1089 the only newline sequence it recognized was NL. It now handles newlines
1090 according to the set newline convention.
1091
1092 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1093 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1094 cater for a lack of strerror(). These oversights have been fixed.
1095
1096 9. Added --match-limit and --recursion-limit to pcregrep.
1097
1098 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1099
1100 11. When the -o option was used, pcregrep was setting a return code of 1, even
1101 when matches were found, and --line-buffered was not being honoured.
1102
1103 12. Added an optional parentheses number to the -o and --only-matching options
1104 of pcregrep.
1105
1106 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1107 can match an empty string. The code to do it in pcretest and pcredemo
1108 needed fixing:
1109
1110 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1111 only one byte after an empty string match just before CRLF (this case
1112 just got forgotten; "any" and "anycrlf" were OK).
1113
1114 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1115 mode when an empty string match preceded an ASCII character followed by
1116 a non-ASCII character. (The code for advancing by one character rather
1117 than one byte was nonsense.)
1118
1119 (c) The pcredemo.c sample program did not have any code at all to handle
1120 the cases when CRLF is a valid newline sequence.
1121
1122 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1123 as a starting offset was within the subject string. There is now a new
1124 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1125 negative or greater than the length of the string. In order to test this,
1126 pcretest is extended to allow the setting of negative starting offsets.
1127
1128 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1129 starting offset points to the beginning of a UTF-8 character was
1130 unnecessarily clumsy. I tidied it up.
1131
1132 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1133 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1134
1135 17. Nobody had reported that the --include_dir option, which was added in
1136 release 7.7 should have been called --include-dir (hyphen, not underscore)
1137 for compatibility with GNU grep. I have changed it to --include-dir, but
1138 left --include_dir as an undocumented synonym, and the same for
1139 --exclude-dir, though that is not available in GNU grep, at least as of
1140 release 2.5.4.
1141
1142 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1143 characters from a string of bytes) have been redefined so as not to use
1144 loops, in order to improve performance in some environments. At the same
1145 time, I abstracted some of the common code into auxiliary macros to save
1146 repetition (this should not affect the compiled code).
1147
1148 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1149 compile-time error is now given if \c is not followed by an ASCII
1150 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1151 different, and any byte value is allowed.)
1152
1153 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1154 START_OPTIMIZE option, which is now allowed at compile time - but just
1155 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1156 to pcregrep and other applications that have no direct access to PCRE
1157 options. The new /Y option in pcretest sets this option when calling
1158 pcre_compile().
1159
1160 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1161 back references. Groups containing recursive back references were forced to
1162 be atomic by that change, but in the case of named groups, the amount of
1163 memory required was incorrectly computed, leading to "Failed: internal
1164 error: code overflow". This has been fixed.
1165
1166 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1167 pcretest.c, to avoid build problems in some Borland environments.
1168
1169
1170 Version 8.10 25-Jun-2010
1171 ------------------------
1172
1173 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1174 THEN.
1175
1176 2. (*ACCEPT) was not working when inside an atomic group.
1177
1178 3. Inside a character class, \B is treated as a literal by default, but
1179 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1180 causes the error). The code is unchanged, but I tidied the documentation.
1181
1182 4. Inside a character class, PCRE always treated \R and \X as literals,
1183 whereas Perl faults them if its -w option is set. I have changed PCRE so
1184 that it faults them when PCRE_EXTRA is set.
1185
1186 5. Added support for \N, which always matches any character other than
1187 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1188
1189 6. When compiling pcregrep with newer versions of gcc which may have
1190 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1191 declared with attribute warn_unused_result" were given. Just casting the
1192 result to (void) does not stop the warnings; a more elaborate fudge is
1193 needed. I've used a macro to implement this.
1194
1195 7. Minor change to pcretest.c to avoid a compiler warning.
1196
1197 8. Added four artifical Unicode properties to help with an option to make
1198 \s etc use properties (see next item). The new properties are: Xan
1199 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1200
1201 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1202 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1203 this option. Modified pcretest to add /W to test this facility. Added
1204 REG_UCP to make it available via the POSIX interface.
1205
1206 10. Added --line-buffered to pcregrep.
1207
1208 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1209 studied, and the match started with a letter with a code point greater than
1210 127 whose first byte was different to the first byte of the other case of
1211 the letter, the other case of this starting letter was not recognized
1212 (#976).
1213
1214 12. If a pattern that was studied started with a repeated Unicode property
1215 test, for example, \p{Nd}+, there was the theoretical possibility of
1216 setting up an incorrect bitmap of starting bytes, but fortunately it could
1217 not have actually happened in practice until change 8 above was made (it
1218 added property types that matched character-matching opcodes).
1219
1220 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1221 possible starting bytes for non-anchored patterns.
1222
1223 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1224 \R, and also a number of cases that involve Unicode properties, both
1225 explicit and implicit when PCRE_UCP is set.
1226
1227 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1228 input, it could crash or give wrong results if characters with values
1229 greater than 0xc0 were present in the subject string. (Detail: it assumed
1230 UTF-8 input when processing these items.)
1231
1232 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1233 size_t is 64-bit (#991).
1234
1235 17. Added a check for running out of memory when PCRE is compiled with
1236 --disable-stack-for-recursion (#990).
1237
1238 18. If the last data line in a file for pcretest does not have a newline on
1239 the end, a newline was missing in the output.
1240
1241 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1242 less than 128) in its various bitmaps. However, there is a facility for
1243 generating tables according to the current locale when PCRE is compiled. It
1244 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1245 characters, are recognized by isspace() and therefore were getting set in
1246 these tables, and indeed these tables seem to approximate to ISO 8859. This
1247 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1248 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1249 which of course cannot start UTF-8 characters. I have changed the code so
1250 that only real ASCII characters (less than 128) and the correct starting
1251 bytes for UTF-8 encodings are set for characters greater than 127 when in
1252 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1253 altogether.)
1254
1255 20. Added the /T option to pcretest so as to be able to run tests with non-
1256 standard character tables, thus making it possible to include the tests
1257 used for 19 above in the standard set of tests.
1258
1259 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1260 reference to a subpattern the other side of a comment that contains an
1261 opening parenthesis caused either an internal compiling error, or a
1262 reference to the wrong subpattern.
1263
1264
1265 Version 8.02 19-Mar-2010
1266 ------------------------
1267
1268 1. The Unicode data tables have been updated to Unicode 5.2.0.
1269
1270 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1271 configured.
1272
1273 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1274 original author of that file, following a query about its status.
1275
1276 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1277 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1278
1279 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1280 quantifier applied to a forward-referencing subroutine call, could compile
1281 incorrect code or give the error "internal error: previously-checked
1282 referenced subpattern not found".
1283
1284 6. Both MS Visual Studio and Symbian OS have problems with initializing
1285 variables to point to external functions. For these systems, therefore,
1286 pcre_malloc etc. are now initialized to local functions that call the
1287 relevant global functions.
1288
1289 7. There were two entries missing in the vectors called coptable and poptable
1290 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1291 I've fixed the data, and added a kludgy way of testing at compile time that
1292 the lengths are correct (equal to the number of opcodes).
1293
1294 8. Following on from 7, I added a similar kludge to check the length of the
1295 eint vector in pcreposix.c.
1296
1297 9. Error texts for pcre_compile() are held as one long string to avoid too
1298 much relocation at load time. To find a text, the string is searched,
1299 counting zeros. There was no check for running off the end of the string,
1300 which could happen if a new error number was added without updating the
1301 string.
1302
1303 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1304
1305 11. \K was not working if it appeared in an atomic group or in a group that
1306 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1307 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1308 the assertion is positive, but not if it is negative.
1309
1310 12. Change 11 fortuitously reduced the size of the stack frame used in the
1311 "match()" function of pcre_exec.c by one pointer. Forthcoming
1312 implementation of support for (*MARK) will need an extra pointer on the
1313 stack; I have reserved it now, so that the stack frame size does not
1314 decrease.
1315
1316 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1317 item in branch that calls a recursion is a subroutine call - as in the
1318 second branch in the above example - was incorrectly given the compile-
1319 time error "recursive call could loop indefinitely" because pcre_compile()
1320 was not correctly checking the subroutine for matching a non-empty string.
1321
1322 14. The checks for overrunning compiling workspace could trigger after an
1323 overrun had occurred. This is a "should never occur" error, but it can be
1324 triggered by pathological patterns such as hundreds of nested parentheses.
1325 The checks now trigger 100 bytes before the end of the workspace.
1326
1327 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1328
1329
1330 Version 8.01 19-Jan-2010
1331 ------------------------
1332
1333 1. If a pattern contained a conditional subpattern with only one branch (in
1334 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1335 computed the wrong minimum data length (which is of course zero for such
1336 subpatterns). This could cause incorrect "no match" results.
1337
1338 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1339 the pattern is reset in the first branch, pcre_compile() failed with
1340 "internal error: code overflow at offset...". This happened only when
1341 the reset was to the original external option setting. (An optimization
1342 abstracts leading options settings into an external setting, which was the
1343 cause of this.)
1344
1345 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1346 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1347 assertion pattern did not match (meaning that the assertion was true), it
1348 was incorrectly treated as false if the SKIP had been reached during the
1349 matching. This also applied to assertions used as conditions.
1350
1351 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1352 assertion subpattern, including such a pattern used as a condition,
1353 unpredictable results occurred, instead of the error return
1354 PCRE_ERROR_DFA_UITEM.
1355
1356 5. The C++ GlobalReplace function was not working like Perl for the special
1357 situation when an empty string is matched. It now does the fancy magic
1358 stuff that is necessary.
1359
1360 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1361 removed. (These were left over from very, very early versions of PCRE.)
1362
1363 7. Some cosmetic changes to the code to make life easier when compiling it
1364 as part of something else:
1365
1366 (a) Change DEBUG to PCRE_DEBUG.
1367
1368 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1369 called "current" as "current_branch", to prevent a collision with the
1370 Linux macro when compiled as a kernel module.
1371
1372 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1373 prevent a collision with the Linux macro when compiled as a kernel
1374 module.
1375
1376 8. In pcre_compile() there are some checks for integer overflows that used to
1377 cast potentially large values to (double). This has been changed to that
1378 when building, a check for int64_t is made, and if it is found, it is used
1379 instead, thus avoiding the use of floating point arithmetic. (There is no
1380 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1381 double.
1382
1383 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1384 2005 (difference between two addresses compared to an unsigned value).
1385
1386 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1387 custom one, because of the following reported problem in Windows:
1388
1389 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1390 under Win32.
1391 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1392 therefore missing the function definition.
1393 - The compiler thus generates a "C" signature for the test function.
1394 - The linker fails to find the "C" function.
1395 - PCRE fails to configure if asked to do so against libbz2.
1396
1397 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1398 messages were output:
1399
1400 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1401 rerunning libtoolize, to keep the correct libtool macros in-tree.
1402 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1403
1404 I have done both of these things.
1405
1406 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1407 most of the time, it *can* run out if it is given a pattern that contains a
1408 runaway infinite recursion. I updated the discussion in the pcrestack man
1409 page.
1410
1411 13. Now that we have gone to the x.xx style of version numbers, the minor
1412 version may start with zero. Using 08 or 09 is a bad idea because users
1413 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1414 interpreted as invalid octal numbers. I've updated the previous comment in
1415 configure.ac, and also added a check that gives an error if 08 or 09 are
1416 used.
1417
1418 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1419 causing partial matching to fail when the end of the subject matched \W
1420 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1421
1422 15. There were some discrepancies between the declarations in pcre_internal.h
1423 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1424 their definitions. The declarations used "const uschar *" and the
1425 definitions used USPTR. Even though USPTR is normally defined as "const
1426 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1427 reported that: "This difference in casting confuses some C++ compilers, for
1428 example, SunCC recognizes above declarations as different functions and
1429 generates broken code for hbpcre." I have changed the declarations to use
1430 USPTR.
1431
1432 16. GNU libtool is named differently on some systems. The autogen.sh script now
1433 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1434 (FreeBSD).
1435
1436 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1437 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1438 comment: "Figure out how to create a longlong from a string: strtoll and
1439 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1440 instance, but it only takes 2 args instead of 3!"
1441
1442 18. A subtle bug concerned with back references has been fixed by a change of
1443 specification, with a corresponding code fix. A pattern such as
1444 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1445 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1446 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1447 same bug. Such groups have to be quantified to be useful, or contained
1448 inside another quantified group. (If there's no repetition, the reference
1449 can never match.) The problem arises because, having left the group and
1450 moved on to the rest of the pattern, a later failure that backtracks into
1451 the group uses the captured value from the final iteration of the group
1452 rather than the correct earlier one. I have fixed this in PCRE by forcing
1453 any group that contains a reference to itself to be an atomic group; that
1454 is, there cannot be any backtracking into it once it has completed. This is
1455 similar to recursive and subroutine calls.
1456
1457
1458 Version 8.00 19-Oct-09
1459 ----------------------
1460
1461 1. The table for translating pcre_compile() error codes into POSIX error codes
1462 was out-of-date, and there was no check on the pcre_compile() error code
1463 being within the table. This could lead to an OK return being given in
1464 error.
1465
1466 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1467 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1468 in a Windows environment.
1469
1470 3. The pcregrep --count option prints the count for each file even when it is
1471 zero, as does GNU grep. However, pcregrep was also printing all files when
1472 --files-with-matches was added. Now, when both options are given, it prints
1473 counts only for those files that have at least one match. (GNU grep just
1474 prints the file name in this circumstance, but including the count seems
1475 more useful - otherwise, why use --count?) Also ensured that the
1476 combination -clh just lists non-zero counts, with no names.
1477
1478 4. The long form of the pcregrep -F option was incorrectly implemented as
1479 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1480 but it seems right to fix it, and I didn't think it was worth preserving
1481 the old behaviour.
1482
1483 5. The command line items --regex=pattern and --regexp=pattern were not
1484 recognized by pcregrep, which required --regex pattern or --regexp pattern
1485 (with a space rather than an '='). The man page documented the '=' forms,
1486 which are compatible with GNU grep; these now work.
1487
1488 6. No libpcreposix.pc file was created for pkg-config; there was just
1489 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1490
1491 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1492 when UCP support is not needed, by modifying the Python script that
1493 generates it from Unicode data files. This should not matter if the module
1494 is correctly used as a library, but I received one complaint about 50K of
1495 unwanted data. My guess is that the person linked everything into his
1496 program rather than using a library. Anyway, it does no harm.
1497
1498 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1499 was a minimum greater than 1 for a wide character in a possessive
1500 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1501 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1502 character. Chaos in the form of incorrect output or a compiling loop could
1503 result.
1504
1505 9. The restrictions on what a pattern can contain when partial matching is
1506 requested for pcre_exec() have been removed. All patterns can now be
1507 partially matched by this function. In addition, if there are at least two
1508 slots in the offset vector, the offset of the earliest inspected character
1509 for the match and the offset of the end of the subject are set in them when
1510 PCRE_ERROR_PARTIAL is returned.
1511
1512 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1513 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1514 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1515 and may be more useful for multi-segment matching.
1516
1517 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1518 used to be given if ever the end of the subject was reached; now it is
1519 given only if matching could not proceed because another character was
1520 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1521 string "Z", which now yields "no match" instead of "partial match". In the
1522 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1523 final character ended with (*FAIL).
1524
1525 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1526 if the pattern had a "must contain" character that was already found in the
1527 earlier partial match, unless partial matching was again requested. For
1528 example, with the pattern /dog.(body)?/, the "must contain" character is
1529 "g". If the first part-match was for the string "dog", restarting with
1530 "sbody" failed. This bug has been fixed.
1531
1532 13. The string returned by pcre_dfa_exec() after a partial match has been
1533 changed so that it starts at the first inspected character rather than the
1534 first character of the match. This makes a difference only if the pattern
1535 starts with a lookbehind assertion or \b or \B (\K is not supported by
1536 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1537 matching functions compatible, and I think it's the right thing to do.
1538
1539 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1540 so that the demonstration program is easily available in environments where
1541 PCRE has not been installed from source.
1542
1543 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1544 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1545 library.
1546
1547 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1548 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1549 is not the first non-POSIX option to be added. Clearly some people find
1550 these options useful.
1551
1552 17. If a caller to the POSIX matching function regexec() passes a non-zero
1553 value for nmatch with a NULL value for pmatch, the value of
1554 nmatch is forced to zero.
1555
1556 18. RunGrepTest did not have a test for the availability of the -u option of
1557 the diff command, as RunTest does. It now checks in the same way as
1558 RunTest, and also checks for the -b option.
1559
1560 19. If an odd number of negated classes containing just a single character
1561 interposed, within parentheses, between a forward reference to a named
1562 subpattern and the definition of the subpattern, compilation crashed with
1563 an internal error, complaining that it could not find the referenced
1564 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1565 [The bug was that it was starting one character too far in when skipping
1566 over the character class, thus treating the ] as data rather than
1567 terminating the class. This meant it could skip too much.]
1568
1569 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1570 /g option in pcretest when the pattern contains \K, which makes it possible
1571 to have an empty string match not at the start, even when the pattern is
1572 anchored. Updated pcretest and pcredemo to use this option.
1573
1574 21. If the maximum number of capturing subpatterns in a recursion was greater
1575 than the maximum at the outer level, the higher number was returned, but
1576 with unset values at the outer level. The correct (outer level) value is
1577 now given.
1578
1579 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1580 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1581 make it do so. The string so far is captured, making this feature
1582 compatible with Perl.
1583
1584 23. The tests have been re-organized, adding tests 11 and 12, to make it
1585 possible to check the Perl 5.10 features against Perl 5.10.
1586
1587 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1588 pattern matches a fixed length string. PCRE did not allow this; now it
1589 does. Neither allows recursion.
1590
1591 25. I finally figured out how to implement a request to provide the minimum
1592 length of subject string that was needed in order to match a given pattern.
1593 (It was back references and recursion that I had previously got hung up
1594 on.) This code has now been added to pcre_study(); it finds a lower bound
1595 to the length of subject needed. It is not necessarily the greatest lower
1596 bound, but using it to avoid searching strings that are too short does give
1597 some useful speed-ups. The value is available to calling programs via
1598 pcre_fullinfo().
1599
1600 26. While implementing 25, I discovered to my embarrassment that pcretest had
1601 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1602 study optimizations had never been tested with that matching function.
1603 Oops. What is worse, even when it was passed study data, there was a bug in
1604 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1605 were also very few tests of studied patterns with pcre_dfa_exec().
1606
1607 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1608 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1609 on the other side of the coin, they are no longer allowed to have different
1610 names, because these cannot be distinguished in PCRE, and this has caused
1611 confusion. (This is a difference from Perl.)
1612
1613 28. When duplicate subpattern names are present (necessarily with different
1614 numbers, as required by 27 above), and a test is made by name in a
1615 conditional pattern, either for a subpattern having been matched, or for
1616 recursion in such a pattern, all the associated numbered subpatterns are
1617 tested, and the overall condition is true if the condition is true for any
1618 one of them. This is the way Perl works, and is also more like the way
1619 testing by number works.
1620
1621
1622 Version 7.9 11-Apr-09
1623 ---------------------
1624
1625 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1626 (pcretest), all targets were linked against these libraries. This included
1627 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1628 libraries. This caused unwanted dependencies to be created. This problem
1629 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1630 pcretest is linked with readline.
1631
1632 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1633 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1634 moved outside it again, because FALSE and TRUE are already defined in AIX,
1635 but BOOL is not.
1636
1637 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1638 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1639
1640 4. The pcregrep documentation said spaces were inserted as well as colons (or
1641 hyphens) following file names and line numbers when outputting matching
1642 lines. This is not true; no spaces are inserted. I have also clarified the
1643 wording for the --colour (or --color) option.
1644
1645 5. In pcregrep, when --colour was used with -o, the list of matching strings
1646 was not coloured; this is different to GNU grep, so I have changed it to be
1647 the same.
1648
1649 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1650 each matching line was coloured. Now it goes on to look for further matches
1651 of any of the test patterns, which is the same behaviour as GNU grep.
1652
1653 7. A pattern that could match an empty string could cause pcregrep to loop; it
1654 doesn't make sense to accept an empty string match in pcregrep, so I have
1655 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1656 seems to be how GNU grep behaves.
1657
1658 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1659 start or after a newline", because the conditional assertion was not being
1660 correctly handled. The rule now is that both the assertion and what follows
1661 in the first alternative must satisfy the test.
1662
1663 9. If auto-callout was enabled in a pattern with a conditional group whose
1664 condition was an assertion, PCRE could crash during matching, both with
1665 pcre_exec() and pcre_dfa_exec().
1666
1667 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1668 used for matching.
1669
1670 11. Unicode property support in character classes was not working for
1671 characters (bytes) greater than 127 when not in UTF-8 mode.
1672
1673 12. Added the -M command line option to pcretest.
1674
1675 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1676
1677 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1678
1679 16. Added comments and documentation about mis-use of no_arg in the C++
1680 wrapper.
1681
1682 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1683 from Martin Jerabek that uses macro names for all relevant character and
1684 string constants.
1685
1686 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1687 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1688 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1689 these, but not everybody uses configure.
1690
1691 19. A conditional group that had only one branch was not being correctly
1692 recognized as an item that could match an empty string. This meant that an
1693 enclosing group might also not be so recognized, causing infinite looping
1694 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1695 with the subject "ab", where knowledge that the repeated group can match
1696 nothing is needed in order to break the loop.
1697
1698 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1699 exec(), but without supplying a callout function, matching went wrong.
1700
1701 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1702 leak if the size of the offset vector was greater than 30. When the vector
1703 is smaller, the saved offsets during recursion go onto a local stack
1704 vector, but for larger vectors malloc() is used. It was failing to free
1705 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1706 error, in fact).
1707
1708 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1709 heapframe that is used only when UTF-8 support is enabled. This caused no
1710 problem, but was untidy.
1711
1712 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1713 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1714 included within another project.
1715
1716 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1717 slightly modified by me:
1718
1719 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1720 not building pcregrep.
1721
1722 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1723 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1724
1725 25. Forward references, both numeric and by name, in patterns that made use of
1726 duplicate group numbers, could behave incorrectly or give incorrect errors,
1727 because when scanning forward to find the reference group, PCRE was not
1728 taking into account the duplicate group numbers. A pattern such as
1729 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1730
1731 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1732 the feature of a custom pointer more persuasive (as requested by a user).
1733
1734 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1735 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1736 pre-defined.
1737
1738 28. Added support for (*UTF8) at the start of a pattern.
1739
1740 29. Arrange for flags added by the "release type" setting in CMake to be shown
1741 in the configuration summary.
1742
1743
1744 Version 7.8 05-Sep-08
1745 ---------------------
1746
1747 1. Replaced UCP searching code with optimized version as implemented for Ad
1748 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1749 stage table and inline lookup instead of a function, giving speed ups of 2
1750 to 5 times on some simple patterns that I tested. Permission was given to
1751 distribute the MultiStage2.py script that generates the tables (it's not in
1752 the tarball, but is in the Subversion repository).
1753
1754 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1755 scripts.
1756
1757 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1758 a group with a zero qualifier. The result of the study could be incorrect,
1759 or the function might crash, depending on the pattern.
1760
1761 4. Caseless matching was not working for non-ASCII characters in back
1762 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1763 It now works when Unicode Property Support is available.
1764
1765 5. In pcretest, an escape such as \x{de} in the data was always generating
1766 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1767 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1768 truncation.
1769
1770 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1771
1772 7. Added two (int) casts to pcregrep when printing the difference of two
1773 pointers, in case they are 64-bit values.
1774
1775 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1776 test 2 if it fails.
1777
1778 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1779 and a #define of that name to empty if it is not externally set. This is to
1780 allow users of MSVC to set it if necessary.
1781
1782 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1783 the convenience functions in the pcre_get.c source file.
1784
1785 11. An option change at the start of a pattern that had top-level alternatives
1786 could cause overwriting and/or a crash. This command provoked a crash in
1787 some environments:
1788
1789 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1790
1791 This potential security problem was recorded as CVE-2008-2371.
1792
1793 12. For a pattern where the match had to start at the beginning or immediately
1794 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1795 pcre_dfa_exec() could read past the end of the passed subject if there was
1796 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1797 pcretest so that it places the subject at the end of its malloc-ed buffer.
1798
1799 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1800 exec() might read past the end of the data buffer in UTF-8 mode.
1801
1802 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1803 the data contained the byte 0x85 as part of a UTF-8 character within its
1804 first line. This applied both to normal and DFA matching.
1805
1806 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1807 /^[^d]*?$/8 failed to match "abc".
1808
1809 16. Added a missing copyright notice to pcrecpp_internal.h.
1810
1811 17. Make it more clear in the documentation that values returned from
1812 pcre_exec() in ovector are byte offsets, not character counts.
1813
1814 18. Tidied a few places to stop certain compilers from issuing warnings.
1815
1816 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1817 supplied by Stefan Weber. I made a further small update for 7.8 because
1818 there is a change of source arrangements: the pcre_searchfuncs.c module is
1819 replaced by pcre_ucd.c.
1820
1821
1822 Version 7.7 07-May-08
1823 ---------------------
1824
1825 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1826 a string to a long long, pretend we don't even have a long long." This is
1827 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1828
1829 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1830 pre-7.6 versions, which defined a global no_arg variable instead of putting
1831 it in the RE class. (See also #8 below.)
1832
1833 3. Remove a line of dead code, identified by coverity and reported by Nuno
1834 Lopes.
1835
1836 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1837
1838 (1) The include/exclude patterns were being applied to the whole pathnames
1839 of files, instead of just to the final components.
1840
1841 (2) If there was more than one level of directory, the subdirectories were
1842 skipped unless they satisfied the include/exclude conditions. This is
1843 inconsistent with GNU grep (and could even be seen as contrary to the
1844 pcregrep specification - which I improved to make it absolutely clear).
1845 The action now is always to scan all levels of directory, and just
1846 apply the include/exclude patterns to regular files.
1847
1848 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1849 --exclude_dir in the tests to avoid scanning .svn directories.
1850
1851 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1852 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1853 doesn't support NULs in patterns.
1854
1855 7. Added some missing "const"s to declarations of static tables in
1856 pcre_compile.c and pcre_dfa_exec.c.
1857
1858 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1859 caused by fix #2 above. (Subsequently also a second patch to fix the
1860 first patch. And a third patch - this was a messy problem.)
1861
1862 9. Applied Craig's patch to remove the use of push_back().
1863
1864 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1865 matching function regexec().
1866
1867 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1868 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1869 references. PCRE supports relative numbers with this syntax (I don't think
1870 Oniguruma does).
1871
1872 12. Previously, a group with a zero repeat such as (...){0} was completely
1873 omitted from the compiled regex. However, this means that if the group
1874 was called as a subroutine from elsewhere in the pattern, things went wrong
1875 (an internal error was given). Such groups are now left in the compiled
1876 pattern, with a new opcode that causes them to be skipped at execution
1877 time.
1878
1879 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1880 to the way PCRE behaves:
1881
1882 (a) A lone ] character is dis-allowed (Perl treats it as data).
1883
1884 (b) A back reference to an unmatched subpattern matches an empty string
1885 (Perl fails the current match path).
1886
1887 (c) A data ] in a character class must be notated as \] because if the
1888 first data character in a class is ], it defines an empty class. (In
1889 Perl it is not possible to have an empty class.) The empty class []
1890 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1891 The negative empty class [^] matches any one character, independently
1892 of the DOTALL setting.
1893
1894 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1895 non-existent subpattern following a character class starting with ']' and
1896 containing () gave an internal compiling error instead of "reference to
1897 non-existent subpattern". Fortunately, when the pattern did exist, the
1898 compiled code was correct. (When scanning forwards to check for the
1899 existencd of the subpattern, it was treating the data ']' as terminating
1900 the class, so got the count wrong. When actually compiling, the reference
1901 was subsequently set up correctly.)
1902
1903 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1904 it was being rejected as not supported by pcre_dfa_exec(), even though
1905 other assertions are supported. I have made pcre_dfa_exec() support
1906 (*FAIL).
1907
1908 16. The implementation of 13c above involved the invention of a new opcode,
1909 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1910 cannot be changed at match time, I realized I could make a small
1911 improvement to matching performance by compiling OP_ALLANY instead of
1912 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1913 on the OP_ANY path.
1914
1915 17. Compiling pcretest on Windows with readline support failed without the
1916 following two fixes: (1) Make the unistd.h include conditional on
1917 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1918
1919 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1920 ncurses library to be included for pcretest when ReadLine support is
1921 requested, but also to allow for it to be overridden. This patch came from
1922 Daniel Bergström.
1923
1924 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1925 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1926 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1927 spotting this.
1928
1929
1930 Version 7.6 28-Jan-08
1931 ---------------------
1932
1933 1. A character class containing a very large number of characters with
1934 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1935 overflow.
1936
1937 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1938 HAVE_LONG_LONG is not defined.
1939
1940 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1941 bring them up to date and include new features. This patch includes:
1942
1943 - Fixed PH's badly added libz and libbz2 support.
1944 - Fixed a problem with static linking.
1945 - Added pcredemo. [But later removed - see 7 below.]
1946 - Fixed dftables problem and added an option.
1947 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1948 HAVE_LONG_LONG.
1949 - Added readline support for pcretest.
1950 - Added an listing of the option settings after cmake has run.
1951
1952 4. A user submitted a patch to Makefile that makes it easy to create
1953 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1954 Makefile.am that cause it to include this special target, without
1955 affecting anything else. Note that the same mingw target plus all
1956 the other distribution libraries and programs are now supported
1957 when configuring with CMake (see 6 below) instead of with
1958 Configure/Make.
1959
1960 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1961 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1962 exported in the Windows port". It has not yet been confirmed that the patch
1963 solves the problem, but it does no harm.
1964
1965 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1966 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1967 with CMake, and also correct the comment about stack recursion.
1968
1969 7. Remove the automatic building of pcredemo from the ./configure system and
1970 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1971 of a program that users should build themselves after PCRE is installed, so
1972 building it automatically is not really right. What is more, it gave
1973 trouble in some build environments.
1974
1975 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1976
1977
1978 Version 7.5 10-Jan-08
1979 ---------------------
1980
1981 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1982 values in parens when parsing an RE using the C++ wrapper."
1983
1984 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1985 Characters greater than 255 were excluded from the class instead of being
1986 included.
1987
1988 3. The same bug as (2) above applied to negated POSIX classes such as
1989 [:^space:].
1990
1991 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1992 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1993 I have changed it.
1994
1995 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1996 first named subpattern) and a construct such as (?&a) would reference the
1997 first named subpattern whose name started with "a" (in other words, the
1998 length check was missing). Both these problems are fixed. "Subpattern name
1999 expected" is now given for (?&) (a zero-length name), and this patch also
2000 makes it give the same error for \k'' (previously it complained that that
2001 was a reference to a non-existent subpattern).
2002
2003 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2004 this is right because (?- can be followed by option settings as well as by
2005 digits. I have, however, made the messages clearer.
2006
2007 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2008 than the number used in the conditional) now cause a compile-time error.
2009 This is actually not compatible with Perl, which accepts such patterns, but
2010 treats the conditional as always being FALSE (as PCRE used to), but it
2011 seems to me that giving a diagnostic is better.
2012
2013 8. Change "alphameric" to the more common word "alphanumeric" in comments
2014 and messages.
2015
2016 9. Fix two occurrences of "backslash" in comments that should have been
2017 "backspace".
2018
2019 10. Remove two redundant lines of code that can never be obeyed (their function
2020 was moved elsewhere).
2021
2022 11. The program that makes PCRE's Unicode character property table had a bug
2023 which caused it to generate incorrect table entries for sequences of
2024 characters that have the same character type, but are in different scripts.
2025 It amalgamated them into a single range, with the script of the first of
2026 them. In other words, some characters were in the wrong script. There were
2027 thirteen such cases, affecting characters in the following ranges:
2028
2029 U+002b0 - U+002c1
2030 U+0060c - U+0060d
2031 U+0061e - U+00612
2032 U+0064b - U+0065e
2033 U+0074d - U+0076d
2034 U+01800 - U+01805
2035 U+01d00 - U+01d77
2036 U+01d9b - U+01dbf
2037 U+0200b - U+0200f
2038 U+030fc - U+030fe
2039 U+03260 - U+0327f
2040 U+0fb46 - U+0fbb1
2041 U+10450 - U+1049d
2042
2043 12. The -o option (show only the matching part of a line) for pcregrep was not
2044 compatible with GNU grep in that, if there was more than one match in a
2045 line, it showed only the first of them. It now behaves in the same way as
2046 GNU grep.
2047
2048 13. If the -o and -v options were combined for pcregrep, it printed a blank
2049 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2050 does the same. The return code can be used to tell if there were any
2051 non-matching lines.
2052
2053 14. Added --file-offsets and --line-offsets to pcregrep.
2054
2055 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2056 infinitely looping recursion. The bug was that positive lookaheads were not
2057 being skipped when checking for a possible empty match (negative lookaheads
2058 and both kinds of lookbehind were skipped).
2059
2060 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2061 inclusion of <windows.h> to before rather than after the definition of
2062 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2063
2064 17. Specifying a possessive quantifier with a specific limit for a Unicode
2065 character property caused pcre_compile() to compile bad code, which led at
2066 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2067 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2068 caused the error; without that there was no problem.
2069
2070 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2071
2072 19. Added --enable-pcretest-libreadline.
2073
2074 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2075 RE::GlobalReplace(). As a result, the number of replacements returned was
2076 double what it should be. I removed one of the increments, but Craig sent a
2077 later patch that removed the other one (the right fix) and added unit tests
2078 that check the return values (which was not done before).
2079
2080 21. Several CMake things:
2081
2082 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2083 the names libpcre and libpcreposix, not just pcre and pcreposix.
2084
2085 (2) The above change means that pcretest and pcregrep are now correctly
2086 linked with the newly-built libraries, not previously installed ones.
2087
2088 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2089
2090 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2091 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2092 UTF-8 newline character). The key issue is that the pattern starts .*;
2093 this means that the match must be either at the beginning, or after a
2094 newline. The bug was in the code for advancing after a failed match and
2095 checking that the new position followed a newline. It was not taking
2096 account of UTF-8 characters correctly.
2097
2098 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2099 character classes. PCRE was not treating the sequence [:...:] as a
2100 character class unless the ... were all letters. Perl, however, seems to
2101 allow any characters between [: and :], though of course it rejects as
2102 unknown any "names" that contain non-letters, because all the known class
2103 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2104 for example, whereas PCRE did not - it did not recognize a POSIX character
2105 class. This seemed a bit dangerous, so the code has been changed to be
2106 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2107 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2108 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2109 Perl does, and where it didn't before.
2110
2111 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2112 Windows environments %n is disabled by default.
2113
2114
2115 Version 7.4 21-Sep-07
2116 ---------------------
2117
2118 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2119 means that a class such as [\s] counted as "explicit reference to CR or
2120 LF". That isn't really right - the whole point of the change was to try to
2121 help when there was an actual mention of one of the two characters. So now
2122 the change happens only if \r or \n (or a literal CR or LF) character is
2123 encountered.
2124
2125 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2126 of both had grown to the point where there were only 3 bits left.
2127 Fortunately, there was spare space in the data structure, and so I have
2128 moved the internal flags into a new 16-bit field to free up more option
2129 bits.
2130
2131 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2132 but did not set the internal JCHANGED flag - either of these is enough to
2133 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2134 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2135 start sets both bits.
2136
2137 4. Added options (at build time, compile time, exec time) to change \R from
2138 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2139
2140 5. doc/pcresyntax.html was missing from the distribution.
2141
2142 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2143 compatibility, even though it is no longer used.
2144
2145 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2146 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2147 windows.h file is present (where different names are used). [This was
2148 reversed later after testing - see 16 below.]
2149
2150 8. Changed all #include <config.h> to #include "config.h". There were also
2151 some further <pcre.h> cases that I changed to "pcre.h".
2152
2153 9. When pcregrep was used with the --colour option, it missed the line ending
2154 sequence off the lines that it output.
2155
2156 10. It was pointed out to me that arrays of string pointers cause lots of
2157 relocations when a shared library is dynamically loaded. A technique of
2158 using a single long string with a table of offsets can drastically reduce
2159 these. I have refactored PCRE in four places to do this. The result is
2160 dramatic:
2161
2162 Originally: 290
2163 After changing UCP table: 187
2164 After changing error message table: 43
2165 After changing table of "verbs" 36
2166 After changing table of Posix names 22
2167
2168 Thanks to the folks working on Gregex for glib for this insight.
2169
2170 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2171 unicode-properties was also set.
2172
2173 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2174
2175 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2176 checked only for CRLF.
2177
2178 14. Added casts to pcretest.c to avoid compiler warnings.
2179
2180 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2181
2182 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2183 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2184 entirely. This removes changes made in 7 above.
2185
2186 17. The CMake files have been updated, and there is now more information about
2187 building with CMake in the NON-UNIX-USE document.
2188
2189
2190 Version 7.3 28-Aug-07
2191 ---------------------
2192
2193 1. In the rejigging of the build system that eventually resulted in 7.1, the
2194 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2195 brackets there is not right, since it causes compilers to look for an
2196 installed pcre.h, not the version that is in the source that is being
2197 compiled (which of course may be different). I have changed it back to:
2198
2199 #include "pcre.h"
2200
2201 I have a vague recollection that the change was concerned with compiling in
2202 different directories, but in the new build system, that is taken care of
2203 by the VPATH setting the Makefile.
2204
2205 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2206 when the subject happened to end in the byte 0x85 (e.g. if the last
2207 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2208 characters but of course it shouldn't be taken as a newline when it is part
2209 of another character. The bug was that, for an unlimited repeat of . in
2210 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2211 characters when looking for a newline.
2212
2213 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2214
2215 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2216 in debug output.
2217
2218 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2219 long printing in the pcrecpp unittest when running under MinGW.
2220
2221 6. ESC_K was left out of the EBCDIC table.
2222
2223 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2224 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2225 limit also applies to "virtual nesting" when a pattern is recursive, and in
2226 this case 1000 isn't so big. I have been able to remove this limit at the
2227 expense of backing off one optimization in certain circumstances. Normally,
2228 when pcre_exec() would call its internal match() function recursively and
2229 immediately return the result unconditionally, it uses a "tail recursion"
2230 feature to save stack. However, when a subpattern that can match an empty
2231 string has an unlimited repetition quantifier, it no longer makes this
2232 optimization. That gives it a stack frame in which to save the data for
2233 checking that an empty string has been matched. Previously this was taken
2234 from the 1000-entry workspace that had been reserved. So now there is no
2235 explicit limit, but more stack is used.
2236
2237 8. Applied Daniel's patches to solve problems with the import/export magic
2238 syntax that is required for Windows, and which was going wrong for the
2239 pcreposix and pcrecpp parts of the library. These were overlooked when this
2240 problem was solved for the main library.
2241
2242 9. There were some crude static tests to avoid integer overflow when computing
2243 the size of patterns that contain repeated groups with explicit upper
2244 limits. As the maximum quantifier is 65535, the maximum group length was
2245 set at 30,000 so that the product of these two numbers did not overflow a
2246 32-bit integer. However, it turns out that people want to use groups that
2247 are longer than 30,000 bytes (though not repeat them that many times).
2248 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2249 made it possible to implement the integer overflow checks in a much more
2250 dynamic way, which I have now done. The artificial limitation on group
2251 length has been removed - we now have only the limit on the total length of
2252 the compiled pattern, which depends on the LINK_SIZE setting.
2253
2254 10. Fixed a bug in the documentation for get/copy named substring when
2255 duplicate names are permitted. If none of the named substrings are set, the
2256 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2257 empty string.
2258
2259 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2260 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2261 because the ] is interpreted as the first data character and the
2262 terminating ] is not found. PCRE has been made compatible with Perl in this
2263 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2264 cause memory overwriting.
2265
2266 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2267 string has been matched (to stop an infinite loop). It was not recognizing
2268 a conditional subpattern that could match an empty string if that
2269 subpattern was within another subpattern. For example, it looped when
2270 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2271 condition was not nested. This bug has been fixed.
2272
2273 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2274 past the start of the subject in the presence of bytes with the top bit
2275 set, for example "\x8aBCD".
2276
2277 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2278 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2279
2280 14. Optimized (?!) to (*FAIL).
2281
2282 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2283 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2284 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2285 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2286 does: it's just the validity check that is more restrictive.
2287
2288 16. Inserted checks for integer overflows during escape sequence (backslash)
2289 processing, and also fixed erroneous offset values for syntax errors during
2290 backslash processing.
2291
2292 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2293 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2294
2295 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2296 caused an overrun.
2297
2298 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2299 something other than just ASCII characters) inside a group that had an
2300 unlimited repeat caused a loop at compile time (while checking to see
2301 whether the group could match an empty string).
2302
2303 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2304 [\P{Any}] did so. (Error in the code for printing property names.)
2305
2306 21. An orphan \E inside a character class could cause a crash.
2307
2308 22. A repeated capturing bracket such as (A)? could cause a wild memory
2309 reference during compilation.
2310
2311 23. There are several functions in pcre_compile() that scan along a compiled
2312 expression for various reasons (e.g. to see if it's fixed length for look
2313 behind). There were bugs in these functions when a repeated \p or \P was
2314 present in the pattern. These operators have additional parameters compared
2315 with \d, etc, and these were not being taken into account when moving along
2316 the compiled data. Specifically:
2317
2318 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2319 length.
2320
2321 (b) An item such as \pL+ within a repeated group could cause crashes or
2322 loops.
2323
2324 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2325 "reference to non-existent subpattern" error.
2326
2327 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2328
2329 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2330 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2331
2332 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2333
2334 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2335 character were causing crashes (broken optimization).
2336
2337 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2338 \p or \P) caused a compile-time loop.
2339
2340 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2341 break. For example, the unstudied pattern [\r\n]A does not match the string
2342 "\r\nA" because change 7.0/46 below moves the current point on by two
2343 characters after failing to match at the start. However, the pattern \nA
2344 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2345 the same is true. There doesn't seem any very clean way out of this, but
2346 what I have chosen to do makes the common cases work: PCRE now takes note
2347 of whether there can be an explicit match for \r or \n anywhere in the
2348 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2349 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2350 pattern has explicit CR or LF references.
2351
2352 29. Added (*CR) etc for changing newline setting at start of pattern.
2353
2354
2355 Version 7.2 19-Jun-07
2356 ---------------------
2357
2358 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2359 which is apparently normally available under Windows.
2360
2361 2. Re-jig the pcregrep tests with different newline settings in an attempt
2362 to make them independent of the local environment's newline setting.
2363
2364 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2365
2366 4. Some of the "internals" tests were previously cut out when the link size
2367 was not 2, because the output contained actual offsets. The recent new
2368 "Z" feature of pcretest means that these can be cut out, making the tests
2369 usable with all link sizes.
2370
2371 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2372 stack recursion. This gives a massive performance boost under BSD, but just
2373 a small improvement under Linux. However, it saves one field in the frame
2374 in all cases.
2375
2376 6. Added more features from the forthcoming Perl 5.10:
2377
2378 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2379 recursion call. It refers to the nth most recently opened parentheses.
2380
2381 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2382 to be opened parentheses.
2383
2384 (c) Conditions that refer to capturing parentheses can be specified
2385 relatively, for example, (?(-2)... or (?(+3)...
2386
2387 (d) \K resets the start of the current match so that everything before
2388 is not part of it.
2389
2390 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2391
2392 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2393 reference syntax.
2394
2395 (g) (?| introduces a group in which the numbering of parentheses in each
2396 alternative starts with the same number.
2397
2398 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2399
2400 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2401 PCRE_INFO_JCHANGED.
2402
2403 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2404 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2405 for detecting groups that can match an empty string.
2406
2407 9. A pattern with a very large number of alternatives (more than several
2408 hundred) was running out of internal workspace during the pre-compile
2409 phase, where pcre_compile() figures out how much memory will be needed. A
2410 bit of new cunning has reduced the workspace needed for groups with
2411 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2412 workspace instead of running out of the 4096 that are available.
2413
2414 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2415
2416 11. Applied patch from Google to remove an optimization that didn't quite work.
2417 The report of the bug said:
2418
2419 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2420 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2421 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2422
2423 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2424 it matched the wrong number of bytes.
2425
2426
2427 Version 7.1 24-Apr-07
2428 ---------------------
2429
2430 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2431 that is more "standard", making use of automake and other Autotools. There
2432 is some re-arrangement of the files and adjustment of comments consequent
2433 on this.
2434
2435 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2436 for recursive directory scanning broke on some systems because the files
2437 are not scanned in any specific order and on different systems the order
2438 was different. A call to "sort" has been inserted into RunGrepTest for the
2439 approprate test as a short-term fix. In the longer term there may be an
2440 alternative.
2441
2442 3. I had an email from Eric Raymond about problems translating some of PCRE's
2443 man pages to HTML (despite the fact that I distribute HTML pages, some
2444 people do their own conversions for various reasons). The problems
2445 concerned the use of low-level troff macros .br and .in. I have therefore
2446 removed all such uses from the man pages (some were redundant, some could
2447 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2448 HTML has been updated to handle .nf/.fi and to complain if it encounters
2449 .br or .in.
2450
2451 4. Updated comments in configure.ac that get placed in config.h.in and also
2452 arranged for config.h to be included in the distribution, with the name
2453 config.h.generic, for the benefit of those who have to compile without
2454 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2455
2456 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2457 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2458 makevp.bat for the current PCRE, using the additional files
2459 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2460
2461 6. A Windows user reported a minor discrepancy with test 2, which turned out
2462 to be caused by a trailing space on an input line that had got lost in his
2463 copy. The trailing space was an accident, so I've just removed it.
2464
2465 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2466 that is needed.
2467
2468 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2469 as "const" (a) because they are and (b) because it helps the PHP
2470 maintainers who have recently made a script to detect big data structures
2471 in the php code that should be moved to the .rodata section. I remembered
2472 to update Builducptable as well, so it won't revert if ucptable.h is ever
2473 re-created.
2474
2475 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2476 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2477 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2478 support is not required. This saves 1.5-2K of code, which is important in
2479 some applications.
2480
2481 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2482 so as not to refer to the tables, even though these functions will never be
2483 called when UTF-8 support is disabled. Otherwise there are problems with a
2484 shared library.
2485
2486 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2487
2488 (a) It was defining its arguments as char * instead of void *.
2489
2490 (b) It was assuming that all moves were upwards in memory; this was true
2491 a long time ago when I wrote it, but is no longer the case.
2492
2493 The emulated memove() is provided for those environments that have neither
2494 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2495 is clearly not the case, as these two bugs were recently reported.
2496
2497 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2498 and Detrail to create the HTML documentation, the .txt form of the man
2499 pages, and it removes trailing spaces from listed files. It also creates
2500 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2501 case, it wraps all the #defines with #ifndefs. This script should be run
2502 before "make dist".
2503
2504 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2505 with Unicode property support.
2506
2507 (a) For a maximizing quantifier, if the two different cases of the
2508 character were of different lengths in their UTF-8 codings (there are
2509 some cases like this - I found 11), and the matching function had to
2510 back up over a mixture of the two cases, it incorrectly assumed they
2511 were both the same length.
2512
2513 (b) When PCRE was configured to use the heap rather than the stack for
2514 recursion during matching, it was not correctly preserving the data for
2515 the other case of a UTF-8 character when checking ahead for a match
2516 while processing a minimizing repeat. If the check also involved
2517 matching a wide character, but failed, corruption could cause an
2518 erroneous result when trying to check for a repeat of the original
2519 character.
2520
2521 13. Some tidying changes to the testing mechanism:
2522
2523 (a) The RunTest script now detects the internal link size and whether there
2524 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2525 values substituted by "configure". (The RunGrepTest script already did
2526 this for UTF-8.) The configure.ac script no longer substitutes the
2527 relevant variables.
2528
2529 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2530 with length and offset values. This means that the output is different
2531 for different internal link sizes. Test 2 is skipped for link sizes
2532 other than 2 because of this, bypassing the problem. Unfortunately,
2533 there was also a test in test 3 (the locale tests) that used /B and
2534 failed for link sizes other than 2. Rather than cut the whole test out,
2535 I have added a new /Z option to pcretest that replaces the length and
2536 offset values with spaces. This is now used to make test 3 independent
2537 of link size. (Test 2 will be tidied up later.)
2538
2539 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2540 segmentation fault instead of returning the appropriate error message.
2541
2542 15. In multiline mode when the newline sequence was set to "any", the pattern
2543 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2544 This doesn't seem right; it now treats the CRLF combination as the line
2545 ending, and so does not match in that case. It's only a pattern such as ^$
2546 that would hit this one: something like ^ABC$ would have failed after \r
2547 and then tried again after \r\n.
2548
2549 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2550 in an attempt to make files that differ only in their line terminators
2551 compare equal. This works on Linux.
2552
2553 17. Under certain error circumstances pcregrep might try to free random memory
2554 as it exited. This is now fixed, thanks to valgrind.
2555
2556 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2557 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2558 was because its rules for how to advance for /g after matching an empty
2559 string at the end of a line did not allow for this case. They now check for
2560 it specially.
2561
2562 20. pcretest is supposed to handle patterns and data of any length, by
2563 extending its buffers when necessary. It was getting this wrong when the
2564 buffer for a data line had to be extended.
2565
2566 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2567 CRLF as a newline sequence.
2568
2569 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2570 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2571 I have nevertheless tidied it up.
2572
2573 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2574
2575 24. Added a man page for pcre-config.
2576
2577
2578 Version 7.0 19-Dec-06
2579 ---------------------
2580
2581 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2582 moving to gcc 4.1.1.
2583
2584 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2585 sys/time.h, which is documented as needed for this function. It doesn't
2586 seem to matter on Linux, but it showed up on some releases of OS X.
2587
2588 3. It seems that there are systems where bytes whose values are greater than
2589 127 match isprint() in the "C" locale. The "C" locale should be the
2590 default when a C program starts up. In most systems, only ASCII printing
2591 characters match isprint(). This difference caused the output from pcretest
2592 to vary, making some of the tests fail. I have changed pcretest so that:
2593
2594 (a) When it is outputting text in the compiled version of a pattern, bytes
2595 other than 32-126 are always shown as hex escapes.
2596
2597 (b) When it is outputting text that is a matched part of a subject string,
2598 it does the same, unless a different locale has been set for the match
2599 (using the /L modifier). In this case, it uses isprint() to decide.
2600
2601 4. Fixed a major bug that caused incorrect computation of the amount of memory
2602 required for a compiled pattern when options that changed within the
2603 pattern affected the logic of the preliminary scan that determines the
2604 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2605 that the computed length was too small. The symptoms of this bug were
2606 either the PCRE error "internal error: code overflow" from pcre_compile(),
2607 or a glibc crash with a message such as "pcretest: free(): invalid next
2608 size (fast)". Examples of patterns that provoked this bug (shown in
2609 pcretest format) are:
2610
2611 /(?-x: )/x
2612 /(?x)(?-x: \s*#\s*)/
2613 /((?i)[\x{c0}])/8
2614 /(?i:[\x{c0}])/8
2615
2616 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2617 is now done differently.
2618
2619 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2620 wrapper classes; (b) implement a new function in the C++ scanner that is
2621 more efficient than the old way of doing things because it avoids levels of
2622 recursion in the regex matching; (c) add a paragraph to the documentation
2623 for the FullMatch() function.
2624
2625 6. The escape sequence \n was being treated as whatever was defined as
2626 "newline". Not only was this contrary to the documentation, which states
2627 that \n is character 10 (hex 0A), but it also went horribly wrong when
2628 "newline" was defined as CRLF. This has been fixed.
2629
2630 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2631 was being set to -1 for the "end of line" case (supposedly a value that no
2632 character can have). Though this value is never used (the check for end of
2633 line is "zero bytes in current character"), it caused compiler complaints.
2634 I've changed it to 0xffffffff.
2635
2636 8. In pcre_version.c, the version string was being built by a sequence of
2637 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2638 string (as it is for production releases) called a macro with an empty
2639 argument. The C standard says the result of this is undefined. The gcc
2640 compiler treats it as an empty string (which was what was wanted) but it is
2641 reported that Visual C gives an error. The source has been hacked around to
2642 avoid this problem.
2643
2644 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2645 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2646 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2647 of them did).
2648
2649 10. Originally, pcretest opened its input and output without "b"; then I was
2650 told that "b" was needed in some environments, so it was added for release
2651 5.0 to both the input and output. (It makes no difference on Unix-like
2652 systems.) Later I was told that it is wrong for the input on Windows. I've
2653 now abstracted the modes into two macros, to make it easier to fiddle with
2654 them, and removed "b" from the input mode under Windows.
2655
2656 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2657
2658 12. Added -help and --help to pcretest as an official way of being reminded
2659 of the options.
2660
2661 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2662 and pcrecpp.cc because they annoy compilers at high warning levels.
2663
2664 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2665
2666 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2667 scripts are not C programs :-) and which was not noticed because it works
2668 on Linux.
2669
2670 16. pcretest is supposed to handle any length of pattern and data line (as one
2671 line or as a continued sequence of lines) by extending its input buffer if
2672 necessary. This feature was broken for very long pattern lines, leading to
2673 a string of junk being passed to pcre_compile() if the pattern was longer
2674 than about 50K.
2675
2676 17. I have done a major re-factoring of the way pcre_compile() computes the
2677 amount of memory needed for a compiled pattern. Previously, there was code
2678 that made a preliminary scan of the pattern in order to do this. That was
2679 OK when PCRE was new, but as the facilities have expanded, it has become
2680 harder and harder to keep it in step with the real compile phase, and there
2681 have been a number of bugs (see for example, 4 above). I have now found a
2682 cunning way of running the real compile function in a "fake" mode that
2683 enables it to compute how much memory it would need, while actually only
2684 ever using a few hundred bytes of working memory and without too many
2685 tests of the mode. This should make future maintenance and development
2686 easier. A side effect of this work is that the limit of 200 on the nesting
2687 depth of parentheses has been removed (though this was never a serious
2688 limitation, I suspect). However, there is a downside: pcre_compile() now
2689 runs more slowly than before (30% or more, depending on the pattern). I
2690 hope this isn't a big issue. There is no effect on runtime performance.
2691
2692 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2693 newline (only possible for the last line of a file) and it was a
2694 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2695
2696 19. Added additional timing features to pcretest. (1) The -tm option now times
2697 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2698 separate command line item, by a number that specifies the number of
2699 repeats to use when timing. The default is 50000; this gives better
2700 precision, but takes uncomfortably long for very large patterns.
2701
2702 20. Extended pcre_study() to be more clever in cases where a branch of a
2703 subpattern has no definite first character. For example, (a*|b*)[cd] would
2704 previously give no result from pcre_study(). Now it recognizes that the
2705 first character must be a, b, c, or d.
2706
2707 21. There was an incorrect error "recursive call could loop indefinitely" if
2708 a subpattern (or the entire pattern) that was being tested for matching an
2709 empty string contained only one non-empty item after a nested subpattern.
2710 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2711 incorrectly, because the \d was being skipped in the check.
2712
2713 22. The pcretest program now has a new pattern option /B and a command line
2714 option -b, which is equivalent to adding /B to every pattern. This causes
2715 it to show the compiled bytecode, without the additional information that
2716 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2717 is the same as /B/I).
2718
2719 23. A new optimization is now able automatically to treat some sequences such
2720 as a*b as a*+b. More specifically, if something simple (such as a character
2721 or a simple class like \d) has an unlimited quantifier, and is followed by
2722 something that cannot possibly match the quantified thing, the quantifier
2723 is automatically "possessified".
2724
2725 24. A recursive reference to a subpattern whose number was greater than 39
2726 went wrong under certain circumstances in UTF-8 mode. This bug could also
2727 have affected the operation of pcre_study().
2728
2729 25. Realized that a little bit of performance could be had by replacing
2730 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2731
2732 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2733
2734 27. Possessive quantifiers such as a++ were previously implemented by turning
2735 them into atomic groups such as ($>a+). Now they have their own opcodes,
2736 which improves performance. This includes the automatically created ones
2737 from 23 above.
2738
2739 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2740 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2741 the first matched character to be a colon. This applied both to named and
2742 numbered groups.
2743
2744 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2745
2746 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2747 building PCRE on Windows easier, so I have included it in the distribution.
2748
2749 31. There is now a check in pcretest against a ridiculously large number being
2750 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2751 loop, the loop is abandoned.
2752
2753 32. Forward references to subpatterns in conditions such as (?(2)...) where
2754 subpattern 2 is defined later cause pcre_compile() to search forwards in
2755 the pattern for the relevant set of parentheses. This search went wrong
2756 when there were unescaped parentheses in a character class, parentheses
2757 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2758
2759 33. "Subroutine" calls and backreferences were previously restricted to
2760 referencing subpatterns earlier in the regex. This restriction has now
2761 been removed.
2762
2763 34. Added a number of extra features that are going to be in Perl 5.10. On the
2764 whole, these are just syntactic alternatives for features that PCRE had
2765 previously implemented using the Python syntax or my own invention. The
2766 other formats are all retained for compatibility.
2767
2768 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2769 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2770 also .NET compatible.
2771
2772 (b) A recursion or subroutine call to a named group can now be defined as
2773 (?&name) as well as (?P>name).
2774
2775 (c) A backreference to a named group can now be defined as \k<name> or
2776 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2777 5.10, are also .NET compatible.
2778
2779 (d) A conditional reference to a named group can now use the syntax
2780 (?(<name>) or (?('name') as well as (?(name).
2781
2782 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2783 groups (named and numbered) that are never evaluated inline, but can be
2784 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2785 is always false. There may be only one alternative in such a group.
2786
2787 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2788 as the simple (?(R). The condition is true only if the most recent
2789 recursion is that of the given number or name. It does not search out
2790 through the entire recursion stack.
2791
2792 (g) The escape \gN or \g{N} has been added, where N is a positive or
2793 negative number, specifying an absolute or relative reference.
2794
2795 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2796 some "unreachable code" warnings.
2797
2798 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2799 things, this adds five new scripts.
2800
2801 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2802 There were also incompatibilities regarding the handling of \Q..\E inside
2803 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2804 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2805
2806 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2807 matches an empty string, and forcibly breaks the loop. There were bugs in
2808 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2809 against aaaa the result was just "a" rather than "aaaa", for example. Two
2810 separate and independent bugs (that affected different cases) have been
2811 fixed.
2812
2813 39. Refactored the code to abolish the use of different opcodes for small
2814 capturing bracket numbers. This is a tidy that I avoided doing when I
2815 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2816 The new approach is not only tidier, it makes it possible to reduce the
2817 memory needed to fix the previous bug (38).
2818
2819 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2820 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2821 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2822 mode.
2823
2824 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2825 report.
2826
2827 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2828 copy construction and assignment in the C++ wrapper.
2829
2830 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2831 couple of bugs that could have given wrong results in the "--newline=crlf"
2832 case.
2833
2834 44. Added a number of casts and did some reorganization of signed/unsigned int
2835 variables following suggestions from Dair Grant. Also renamed the variable
2836 "this" as "item" because it is a C++ keyword.
2837
2838 45. Arranged for dftables to add
2839
2840 #include "pcre_internal.h"
2841
2842 to pcre_chartables.c because without it, gcc 4.x may remove the array
2843 definition from the final binary if PCRE is built into a static library and
2844 dead code stripping is activated.
2845
2846 46. For an unanchored pattern, if a match attempt fails at the start of a
2847 newline sequence, and the newline setting is CRLF or ANY, and the next two
2848 characters are CRLF, advance by two characters instead of one.
2849
2850
2851 Version 6.7 04-Jul-06
2852 ---------------------
2853
2854 1. In order to handle tests when input lines are enormously long, pcretest has
2855 been re-factored so that it automatically extends its buffers when
2856 necessary. The code is crude, but this _is_ just a test program. The
2857 default size has been increased from 32K to 50K.
2858
2859 2. The code in pcre_study() was using the value of the re argument before
2860 testing it for NULL. (Of course, in any sensible call of the function, it
2861 won't be NULL.)
2862
2863 3. The memmove() emulation function in pcre_internal.h, which is used on
2864 systems that lack both memmove() and bcopy() - that is, hardly ever -
2865 was missing a "static" storage class specifier.
2866
2867 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2868 containing an extended class (one that cannot be represented by a bitmap
2869 because it contains high-valued characters or Unicode property items, e.g.
2870 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2871 pattern, but PCRE should not loop if you do not (it no longer does).
2872 [Detail: two cases were found: (a) a repeated subpattern containing an
2873 extended class; (b) a recursive reference to a subpattern that followed a
2874 previous extended class. It wasn't skipping over the extended class
2875 correctly when UTF-8 mode was not set.]
2876
2877 5. A negated single-character class was not being recognized as fixed-length
2878 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2879 compile error "lookbehind assertion is not fixed length".
2880
2881 6. The RunPerlTest auxiliary script was showing an unexpected difference
2882 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2883 write a Perl script that can interpret lines of an input file either as
2884 byte characters or as UTF-8, which is what "perltest" was being required to
2885 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2886 can't do is switch easily at run time between having the "use utf8;" pragma
2887 or not. In the end, I fudged it by using the RunPerlTest script to insert
2888 "use utf8;" explicitly for the UTF-8 tests.
2889
2890 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2891 the end of the subject string, contrary to the documentation and to what
2892 Perl does. This was true of both matching functions. Now it matches only at
2893 the start of the subject and immediately after *internal* newlines.
2894
2895 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2896 a pointer to an int instead of a pointer to an unsigned long int. This
2897 caused problems on 64-bit systems.
2898
2899 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2900 instance of the 'standard' template library not being so standard".
2901
2902 10. There was no check on the number of named subpatterns nor the maximum
2903 length of a subpattern name. The product of these values is used to compute
2904 the size of the memory block for a compiled pattern. By supplying a very
2905 long subpattern name and a large number of named subpatterns, the size
2906 computation could be caused to overflow. This is now prevented by limiting
2907 the length of names to 32 characters, and the number of named subpatterns
2908 to 10,000.
2909
2910 11. Subpatterns that are repeated with specific counts have to be replicated in
2911 the compiled pattern. The size of memory for this was computed from the
2912 length of the subpattern and the repeat count. The latter is limited to
2913 65535, but there was no limit on the former, meaning that integer overflow
2914 could in principle occur. The compiled length of a repeated subpattern is
2915 now limited to 30,000 bytes in order to prevent this.
2916
2917 12. Added the optional facility to have named substrings with the same name.
2918
2919 13. Added the ability to use a named substring as a condition, using the
2920 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2921 are numbers (not recommended). Forward references are permitted.
2922
2923 14. Added forward references in named backreferences (if you see what I mean).
2924
2925 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2926 pattern could run off the end of the subject. For example, the pattern
2927 "(?s)(.{1,5})"8 did this with the subject "ab".
2928
2929 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2930 PCRE_CASELESS was set when matching characters that were quantified with ?
2931 or *.
2932
2933 17. A character class other than a single negated character that had a minimum
2934 but no maximum quantifier - for example [ab]{6,} - was not handled
2935 correctly by pce_dfa_exec(). It would match only one character.
2936
2937 18. A valid (though odd) pattern that looked like a POSIX character
2938 class but used an invalid character after [ (for example [[,abc,]]) caused
2939 pcre_compile() to give the error "Failed: internal error: code overflow" or
2940 in some cases to crash with a glibc free() error. This could even happen if
2941 the pattern terminated after [[ but there just happened to be a sequence of
2942 letters, a binary zero, and a closing ] in the memory that followed.
2943
2944 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2945 over the years. Originally (before any Unicode support), just the bottom 8
2946 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2947 output from "man perlunicode" includes this:
2948
2949 The regular expression compiler produces polymorphic opcodes. That
2950 is, the pattern adapts to the data and automatically switches to
2951 the Unicode character scheme when presented with Unicode data--or
2952 instead uses a traditional byte scheme when presented with byte
2953 data.
2954
2955 Sadly, a wide octal escape does not cause a switch, and in a string with
2956 no other multibyte characters, these octal escapes are treated as before.
2957 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2958 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2959 Unicode string.
2960
2961 I have not perpetrated such confusion in PCRE. Up till now, it took just
2962 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2963 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2964 translate to the appropriate multibyte character.
2965
2966 29. Applied some refactoring to reduce the number of warnings from Microsoft
2967 and Borland compilers. This has included removing the fudge introduced
2968 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2969 a warning about an unused variable.
2970
2971 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2972 characters since release 4.0, because Perl (from release 5.004) does not.
2973 [Or at least, is documented not to: some releases seem to be in conflict
2974 with the documentation.] However, when a pattern was studied with
2975 pcre_study() and all its branches started with \s, PCRE still included VT
2976 as a possible starting character. Of course, this did no harm; it just
2977 caused an unnecessary match attempt.
2978
2979 22. Removed a now-redundant internal flag bit that recorded the fact that case
2980 dependency changed within the pattern. This was once needed for "required
2981 byte" processing, but is no longer used. This recovers a now-scarce options
2982 bit. Also moved the least significant internal flag bit to the most-
2983 significant bit of the word, which was not previously used (hangover from
2984 the days when it was an int rather than a uint) to free up another bit for
2985 the future.
2986
2987 23. Added support for CRLF line endings as well as CR and LF. As well as the
2988 default being selectable at build time, it can now be changed at runtime
2989 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2990 specify that it is scanning data with non-default line endings.
2991
2992 24. Changed the definition of CXXLINK to make it agree with the definition of
2993 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2994
2995 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2996 recursions. This makes a big different to stack usage for some patterns.
2997
2998 26. If a subpattern containing a named recursion or subroutine reference such
2999 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3000 the space required for the compiled pattern went wrong and gave too small a
3001 value. Depending on the environment, this could lead to "Failed: internal
3002 error: code overflow at offset 49" or "glibc detected double free or
3003 corruption" errors.
3004
3005 27. Applied patches from Google (a) to support the new newline modes and (b) to
3006 advance over multibyte UTF-8 characters in GlobalReplace.
3007
3008 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3009 difference for some implementation of PCRE in some Windows version.
3010
3011 29. Added some extra testing facilities to pcretest:
3012
3013 \q<number> in a data line sets the "match limit" value
3014 \Q<number> in a data line sets the "match recursion limt" value
3015 -S <number> sets the stack size, where <number> is in megabytes
3016
3017 The -S option isn't available for Windows.
3018
3019
3020 Version 6.6 06-Feb-06
3021 ---------------------
3022
3023 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3024 in pcreposix.h. I have copied the definition from pcre.h.
3025
3026 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3027 because pcre.h is no longer a built file.
3028
3029 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3030 not normally included in the compiled code.
3031
3032
3033 Version 6.5 01-Feb-06
3034 ---------------------
3035
3036 1. When using the partial match feature with pcre_dfa_exec(), it was not
3037 anchoring the second and subsequent partial matches at the new starting
3038 point. This could lead to incorrect results. For example, with the pattern
3039 /1234/, partially matching against "123" and then "a4" gave a match.
3040
3041 2. Changes to pcregrep:
3042
3043 (a) All non-match returns from pcre_exec() were being treated as failures
3044 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3045 error message is output. Some extra information is given for the
3046 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3047 probably the only errors that are likely to be caused by users (by
3048 specifying a regex that has nested indefinite repeats, for instance).
3049 If there are more than 20 of these errors, pcregrep is abandoned.
3050
3051 (b) A binary zero was treated as data while matching, but terminated the
3052 output line if it was written out. This has been fixed: binary zeroes
3053 are now no different to any other data bytes.
3054
3055 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3056 used to set a locale for matching. The --locale=xxxx long option has
3057 been added (no short equivalent) to specify a locale explicitly on the
3058 pcregrep command, overriding the environment variables.
3059
3060 (d) When -B was used with -n, some line numbers in the output were one less
3061 than they should have been.
3062
3063 (e) Added the -o (--only-matching) option.
3064
3065 (f) If -A or -C was used with -c (count only), some lines of context were
3066 accidentally printed for the final match.
3067
3068 (g) Added the -H (--with-filename) option.
3069
3070 (h) The combination of options -rh failed to suppress file names for files
3071 that were found from directory arguments.
3072
3073 (i) Added the -D (--devices) and -d (--directories) options.
3074
3075 (j) Added the -F (--fixed-strings) option.
3076
3077 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3078
3079 (l) Added the --colo(u)r option.
3080
3081 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3082 is not present by default.
3083
3084 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3085 items such as (?R) or (?1), when the recursion could match a number of
3086 alternatives. If it matched one of the alternatives, but subsequently,
3087 outside the recursion, there was a failure, the code tried to back up into
3088 the recursion. However, because of the way PCRE is implemented, this is not
3089 possible, and the result was an incorrect result from the match.
3090
3091 In order to prevent this happening, the specification of recursion has
3092 been changed so that all such subpatterns are automatically treated as
3093 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3094
3095 4. I had overlooked the fact that, in some locales, there are characters for
3096 which isalpha() is true but neither isupper() nor islower() are true. In
3097 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3098 and ordfeminine) are like this. This affected the treatment of \w and \W
3099 when they appeared in character classes, but not when they appeared outside
3100 a character class. The bit map for "word" characters is now created
3101 separately from the results of isalnum() instead of just taking it from the
3102 upper, lower, and digit maps. (Plus the underscore character, of course.)
3103
3104 5. The above bug also affected the handling of POSIX character classes such as
3105 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3106 permanent tables. Instead, the bit maps for such a class were previously
3107 created as the appropriate unions of the upper, lower, and digit bitmaps.
3108 Now they are created by subtraction from the [[:word:]] class, which has
3109 its own bitmap.
3110
3111 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3112 It is created by subtracting the vertical space characters (\x09, \x0a,
3113 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3114 subtraction was done in the overall bitmap for a character class, meaning
3115 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3116 be recognized. This bug has been fixed.
3117
3118 7. Patches from the folks at Google:
3119
3120 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3121 real life, but is still worth protecting against".
3122
3123 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3124 regular expressions".
3125
3126 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3127 have it.
3128
3129 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3130 "configure" and the latter not, in order to fix a problem somebody had
3131 with compiling the Arg class on HP-UX.
3132
3133 (e) Improve the error-handling of the C++ wrapper a little bit.
3134
3135 (f) New tests for checking recursion limiting.
3136
3137 8. The pcre_memmove() function, which is used only if the environment does not
3138 have a standard memmove() function (and is therefore rarely compiled),
3139 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3140 returning a result (though PCRE never actually uses the result).
3141
3142 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3143 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3144 returned instead of calling malloc() with an overflowing number that would
3145 most likely cause subsequent chaos.
3146
3147 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3148
3149 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3150 with this option is matched, the nmatch and pmatch options of regexec() are
3151 ignored.
3152
3153 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3154 provided in case anyone wants to the the POSIX interface with UTF-8
3155 strings.
3156
3157 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3158 C++ linking (needed for some HP-UX environments).
3159
3160 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3161 (unused parameter) and in the pcre_printint() function (omitted "default"
3162 switch label when the default is to do nothing).
3163
3164 15. Added some code to make it possible, when PCRE is compiled as a C++
3165 library, to replace subject pointers for pcre_exec() with a smart pointer
3166 class, thus making it possible to process discontinuous strings.
3167
3168 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3169 much the same function. They were added by different people who were trying
3170 to make PCRE easy to compile on non-Unix systems. It has been suggested
3171 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3172 for compiling on Windows systems. I have therefore replaced it with
3173 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3174 defaults to "extern" for C or "extern C" for C++, which works fine on
3175 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3176 SCOPE with something explicit in config.h. In addition:
3177
3178 (a) pcreposix.h still had just "extern" instead of either of these macros;
3179 I have replaced it with PCRE_DATA_SCOPE.
3180
3181 (b) Functions such as _pcre_xclass(), which are internal to the library,
3182 but external in the C sense, all had PCRE_EXPORT in their definitions.
3183 This is apparently wrong for the Windows case, so I have removed it.
3184 (It makes no difference on Unix-like systems.)
3185
3186 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3187 of recursive calls to match(). This is different to MATCH_LIMIT because
3188 that limits the total number of calls to match(), not all of which increase
3189 the depth of recursion. Limiting the recursion depth limits the amount of
3190 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3191 when PCRE is compiled, and changed at run time. A patch from Google adds
3192 this functionality to the C++ interface.
3193
3194 18. Changes to the handling of Unicode character properties:
3195
3196 (a) Updated the table to Unicode 4.1.0.
3197
3198 (b) Recognize characters that are not in the table as "Cn" (undefined).
3199
3200 (c) I revised the way the table is implemented to a much improved format
3201 which includes recognition of ranges. It now supports the ranges that
3202 are defined in UnicodeData.txt, and it also amalgamates other
3203 characters into ranges. This has reduced the number of entries in the
3204 table from around 16,000 to around 3,000, thus reducing its size
3205 considerably. I realized I did not need to use a tree structure after
3206 all - a binary chop search is just as efficient. Having reduced the
3207 number of entries, I extended their size from 6 bytes to 8 bytes to
3208 allow for more data.
3209
3210 (d) Added support for Unicode script names via properties such as \p{Han}.
3211
3212 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3213 matching that character.
3214
3215 20. When matching a repeated Unicode property with a minimum greater than zero,
3216 (for example \pL{2,}), PCRE could look past the end of the subject if it
3217 reached it while seeking the minimum number of characters. This could
3218 happen only if some of the characters were more than one byte long, because
3219 there is a check for at least the minimum number of bytes.
3220
3221 21. Refactored the implementation of \p and \P so as to be more general, to
3222 allow for more different types of property in future. This has changed the
3223 compiled form incompatibly. Anybody with saved compiled patterns that use
3224 \p or \P will have to recompile them.
3225
3226 22. Added "Any" and "L&" to the supported property types.
3227
3228 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3229 but give a compile time error if the value is greater than 0xff.
3230
3231 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3232 accidentally not being installed or uninstalled.
3233
3234 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3235 made were to insert the current release number. This seemed silly, because
3236 it made things harder for people building PCRE on systems that don't run
3237 "configure". I have turned pcre.h into a distributed file, no longer built
3238 by "configure", with the version identification directly included. There is
3239 no longer a pcre.h.in file.
3240
3241 However, this change necessitated a change to the pcre-config script as
3242 well. It is built from pcre-config.in, and one of the substitutions was the
3243 release number. I have updated configure.ac so that ./configure now finds
3244 the release number by grepping pcre.h.
3245
3246 26. Added the ability to run the tests under valgrind.
3247
3248
3249 Version 6.4 05-Sep-05
3250 ---------------------
3251
3252 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3253 "--" to be printed when multiple files were scanned, even when none of the
3254 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3255 consider it to be a bug, and have restored the previous behaviour.
3256
3257 2. A couple of code tidies to get rid of compiler warnings.
3258
3259 3. The pcretest program used to cheat by referring to symbols in the library
3260 whose names begin with _pcre_. These are internal symbols that are not
3261 really supposed to be visible externally, and in some environments it is
3262 possible to suppress them. The cheating is now confined to including
3263 certain files from the library's source, which is a bit cleaner.
3264
3265 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3266 file's purpose clearer.
3267
3268 5. Reorganized pcre_ucp_findchar().
3269
3270
3271 Version 6.3 15-Aug-05
3272 ---------------------
3273
3274 1. The file libpcre.pc.in did not have general read permission in the tarball.
3275
3276 2. There were some problems when building without C++ support:
3277
3278 (a) If C++ support was not built, "make install" and "make test" still
3279 tried to test it.
3280
3281 (b) There were problems when the value of CXX was explicitly set. Some
3282 changes have been made to try to fix these, and ...
3283
3284 (c) --disable-cpp can now be used to explicitly disable C++ support.
3285
3286 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3287 backslash in a target when C++ was disabled. This confuses some
3288 versions of "make", apparently. Using an intermediate variable solves
3289 this. (Same for CPP_LOBJ.)
3290
3291 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3292 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3293 necessary on certain architectures.
3294
3295 4. Added a setting of -export-symbols-regex to the link command to remove
3296 those symbols that are exported in the C sense, but actually are local
3297 within the library, and not documented. Their names all begin with
3298 "_pcre_". This is not a perfect job, because (a) we have to except some
3299 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3300 available (and never for static libraries). I have made a note to try to
3301 find a way round (a) in the future.
3302
3303
3304 Version 6.2 01-Aug-05
3305 ---------------------
3306
3307 1. There was no test for integer overflow of quantifier values. A construction
3308 such as {1111111111111111} would give undefined results. What is worse, if
3309 a minimum quantifier for a parenthesized subpattern overflowed and became
3310 negative, the calculation of the memory size went wrong. This could have
3311 led to memory overwriting.
3312
3313 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3314
3315 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3316 operating environments where this matters.
3317
3318 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3319 PCRE options from within the C++ wrapper.
3320
3321 5. Named capturing subpatterns were not being correctly counted when a pattern
3322 was compiled. This caused two problems: (a) If there were more than 100
3323 such subpatterns, the calculation of the memory needed for the whole
3324 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3325 back references of the form \12, where the number was greater than 9, were
3326 not recognized as back references, even though there were sufficient
3327 previous subpatterns.
3328
3329 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3330 versions of gcc, e.g. 2.95.4.
3331
3332
3333 Version 6.1 21-Jun-05
3334 ---------------------
3335
3336 1. There was one reference to the variable "posix" in pcretest.c that was not
3337 surrounded by "#if !defined NOPOSIX".
3338
3339 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3340 the cross-check on the old pcre_info() function, for the benefit of the
3341 cut-down version of PCRE that is currently imported into Exim.
3342
3343 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3344 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3345 patterns that start (?i) but I don't think that matters. The use of (?i) is
3346 just an example; this all applies to the other options as well.
3347
3348 4. Since libtool seems to echo the compile commands it is issuing, the output
3349 from "make" can be reduced a bit by putting "@" in front of each libtool
3350 compile command.
3351
3352 5. Patch from the folks at Google for configure.in to be a bit more thorough
3353 in checking for a suitable C++ installation before trying to compile the
3354 C++ stuff. This should fix a reported problem when a compiler was present,
3355 but no suitable headers.
3356
3357 6. The man pages all had just "PCRE" as their title. I have changed them to
3358 be the relevant file name. I have also arranged that these names are
3359 retained in the file doc/pcre.txt, which is a concatenation in text format
3360 of all the man pages except the little individual ones for each function.
3361
3362 7. The NON-UNIX-USE file had not been updated for the different set of source
3363 files that come with release 6. I also added a few comments about the C++
3364 wrapper.
3365
3366
3367 Version 6.0 07-Jun-05
3368 ---------------------
3369
3370 1. Some minor internal re-organization to help with my DFA experiments.
3371
3372 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3373 didn't matter for the library itself when fully configured, but did matter
3374 when compiling without UCP support, or within Exim, where the ucp files are
3375 not imported.
3376
3377 3. Refactoring of the library code to split up the various functions into
3378 different source modules. The addition of the new DFA matching code (see
3379 below) to a single monolithic source would have made it really too
3380 unwieldy, quite apart from causing all the code to be include in a
3381 statically linked application, when only some functions are used. This is
3382 relevant even without the DFA addition now that patterns can be compiled in
3383 one application and matched in another.
3384
3385 The downside of splitting up is that there have to be some external
3386 functions and data tables that are used internally in different modules of
3387 the library but which are not part of the API. These have all had their
3388 names changed to start with "_pcre_" so that they are unlikely to clash
3389 with other external names.
3390
3391 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3392 a different (DFA) algorithm. Although it is slower than the original
3393 function, it does have some advantages for certain types of matching
3394 problem.
3395
3396 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3397 including restarting after a partial match.
3398
3399 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3400 defined when compiling for Windows was sent to me. I have put it into the
3401 code, though I have no means of testing or verifying it.
3402
3403 7. Added the pcre_refcount() auxiliary function.
3404
3405 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3406 match before or at the first newline in the subject string. In pcretest,
3407 the /f option on a pattern can be used to set this.
3408
3409 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3410 would behave wrongly. This has been present in PCRE since release 4.0.
3411
3412 10. A number of changes to the pcregrep command:
3413
3414 (a) Refactored how -x works; insert ^(...)$ instead of setting
3415 PCRE_ANCHORED and checking the length, in preparation for adding
3416 something similar for -w.
3417
3418 (b) Added the -w (match as a word) option.
3419
3420 (c) Refactored the way lines are read and buffered so as to have more
3421 than one at a time available.
3422
3423 (d) Implemented a pcregrep test script.
3424
3425 (e) Added the -M (multiline match) option. This allows patterns to match
3426 over several lines of the subject. The buffering ensures that at least
3427 8K, or the rest of the document (whichever is the shorter) is available
3428 for matching (and similarly the previous 8K for lookbehind assertions).
3429
3430 (f) Changed the --help output so that it now says
3431
3432 -w, --word-regex(p)
3433
3434 instead of two lines, one with "regex" and the other with "regexp"
3435 because that confused at least one person since the short forms are the
3436 same. (This required a bit of code, as the output is generated
3437 automatically from a table. It wasn't just a text change.)
3438
3439 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3440 option but starts with a hyphen. Could be a pattern or a path name
3441 starting with a hyphen, for instance.
3442
3443 (h) "-" can be given as a file name to represent stdin.
3444
3445 (i) When file names are being printed, "(standard input)" is used for
3446 the standard input, for compatibility with GNU grep. Previously
3447 "<stdin>" was used.
3448
3449 (j) The option --label=xxx can be used to supply a name to be used for
3450 stdin when file names are being printed. There is no short form.
3451
3452 (k) Re-factored the options decoding logic because we are going to add
3453 two more options that take data. Such options can now be given in four
3454 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3455
3456 (l) Added the -A, -B, and -C options for requesting that lines of context
3457 around matches be printed.
3458
3459 (m) Added the -L option to print the names of files that do not contain
3460 any matching lines, that is, the complement of -l.
3461
3462 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3463 continue to scan other files.
3464
3465 (o) The -s option was incorrectly implemented. For compatibility with other
3466 greps, it now suppresses the error message for a non-existent or non-
3467 accessible file (but not the return code). There is a new option called
3468 -q that suppresses the output of matching lines, which was what -s was
3469 previously doing.
3470
3471 (p) Added --include and --exclude options to specify files for inclusion
3472 and exclusion when recursing.
3473
3474 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3475 Hopefully, it now does.
3476
3477 12. Missing cast in pcre_study().
3478
3479 13. Added an "uninstall" target to the makefile.
3480
3481 14. Replaced "extern" in the function prototypes in Makefile.in with
3482 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3483 world, but is set differently for Windows.
3484
3485 15. Added a second compiling function called pcre_compile2(). The only
3486 difference is that it has an extra argument, which is a pointer to an
3487 integer error code. When there is a compile-time failure, this is set
3488 non-zero, in addition to the error test pointer being set to point to an
3489 error message. The new argument may be NULL if no error number is required
3490 (but then you may as well call pcre_compile(), which is now just a
3491 wrapper). This facility is provided because some applications need a
3492 numeric error indication, but it has also enabled me to tidy up the way
3493 compile-time errors are handled in the POSIX wrapper.
3494
3495 16. Added VPATH=.libs to the makefile; this should help when building with one
3496 prefix path and installing with another. (Or so I'm told by someone who
3497 knows more about this stuff than I do.)
3498
3499 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3500 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3501 match everything, including newlines. This is not POSIX-compatible, but
3502 somebody wanted the feature. From pcretest it can be activated by using
3503 both the P and the s flags.
3504
3505 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3506
3507 19. libpcre.pc was being incorrectly installed as executable.
3508
3509 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3510 it now also looks for '\r' so that it will work unmodified on Windows.
3511
3512 21. Added Google's contributed C++ wrapper to the distribution.
3513
3514 22. Added some untidy missing memory free() calls in pcretest, to keep
3515 Electric Fence happy when testing.
3516
3517
3518
3519 Version 5.0 13-Sep-04
3520 ---------------------
3521
3522 1. Internal change: literal characters are no longer packed up into items
3523 containing multiple characters in a single byte-string. Each character
3524 is now matched using a separate opcode. However, there may be more than one
3525 byte in the character in UTF-8 mode.
3526
3527 2. The pcre_callout_block structure has two new fields: pattern_position and
3528 next_item_length. These contain the offset in the pattern to the next match
3529 item, and its length, respectively.
3530
3531 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3532 insertion of callouts before each pattern item. Added the /C option to
3533 pcretest to make use of this.
3534
3535 4. On the advice of a Windows user, the lines
3536
3537 #if defined(_WIN32) || defined(WIN32)
3538 _setmode( _fileno( stdout ), 0x8000 );
3539 #endif /* defined(_WIN32) || defined(WIN32) */
3540
3541 have been added to the source of pcretest. This apparently does useful
3542 magic in relation to line terminators.
3543
3544 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3545 for the benefit of those environments where the "b" makes a difference.
3546
3547 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3548 to know about it. I have put a hack into configure.in that adds in code
3549 to set GCC=yes if CC=icc. This seems to end up at a point in the
3550 generated configure script that is early enough to affect the setting of
3551 compiler options, which is what is needed, but I have no means of testing
3552 whether it really works. (The user who reported this had patched the
3553 generated configure script, which of course I cannot do.)
3554
3555 LATER: After change 22 below (new libtool files), the configure script
3556 seems to know about icc (and also ecc). Therefore, I have commented out
3557 this hack in configure.in.
3558
3559 7. Added support for pkg-config (2 patches were sent in).
3560
3561 8. Negated POSIX character classes that used a combination of internal tables
3562 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3563 [[:^ascii]]. Typically, they would match almost any characters. The other
3564 POSIX classes were not broken in this way.
3565
3566 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3567 to find the match, as PCRE was deluded into thinking that the match had to
3568 start at the start point or following a newline. The same bug applied to
3569 patterns with negative forward assertions or any backward assertions
3570 preceding ".*" at the start, unless the pattern required a fixed first
3571 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3572
3573 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3574 starting at the last subject character, bytes beyond the end of the subject
3575 string were read.
3576
3577 11. Renamed the variable "class" as "classbits" to make life easier for C++
3578 users. (Previously there was a macro definition, but it apparently wasn't
3579 enough.)
3580
3581 12. Added the new field "tables" to the extra data so that tables can be passed
3582 in at exec time, or the internal tables can be re-selected. This allows
3583 a compiled regex to be saved and re-used at a later time by a different
3584 program that might have everything at different addresses.
3585
3586 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3587 -R library as well as a -L library.
3588
3589 14. The debugging options of pcretest (-d on the command line or D on a
3590 pattern) showed incorrect output for anything following an extended class
3591 that contained multibyte characters and which was followed by a quantifier.
3592
3593 15. Added optional support for general category Unicode character properties
3594 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3595 support. It adds about 90K to the size of the library. The meanings of the
3596 inbuilt class escapes such as \d and \s have NOT been changed.
3597
3598 16. Updated pcredemo.c to include calls to free() to release the memory for the
3599 compiled pattern.
3600
3601 17. The generated file chartables.c was being created in the source directory
3602 instead of in the building directory. This caused the build to fail if the
3603 source directory was different from the building directory, and was
3604 read-only.
3605
3606 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3607 file. No doubt somebody will tell me if they don't make sense... Also added
3608 Dan Mooney's comments about building on OpenVMS.
3609
3610 19. Added support for partial matching via the PCRE_PARTIAL option for
3611 pcre_exec() and the \P data escape in pcretest.
3612
3613 20. Extended pcretest with 3 new pattern features:
3614
3615 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3616 write the compiled pattern to the file whose name is "rest-of-line".
3617 This is a straight binary dump of the data, with the saved pointer to
3618 the character tables forced to be NULL. The study data, if any, is
3619 written too. After writing, pcretest reads a new pattern.
3620
3621 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3622 compiled pattern from the given file. There must not be any
3623 occurrences of "<" in the file name (pretty unlikely); if there are,
3624 pcretest will instead treat the initial "<" as a pattern delimiter.
3625 After reading in the pattern, pcretest goes on to read data lines as
3626 usual.
3627
3628 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3629 and 16-bit fields in a compiled pattern, to simulate a pattern that
3630 was compiled on a host of opposite endianness.
3631
3632 21. The pcre-exec() function can now cope with patterns that were compiled on
3633 hosts of opposite endianness, with this restriction:
3634
3635 As for any compiled expression that is saved and used later, the tables
3636 pointer field cannot be preserved; the extra_data field in the arguments
3637 to pcre_exec() should be used to pass in a tables address if a value
3638 other than the default internal tables were used at compile time.
3639
3640 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3641 now diagnosed as an error. Previously, most of the time, a negative number
3642 would have been treated as zero, but if in addition "ovector" was passed as
3643 NULL, a crash could occur.
3644
3645 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3646 new versions from the libtool 1.5 distribution (the last one is a copy of
3647 a file called libtool.m4). This seems to have fixed the need to patch
3648 "configure" to support Darwin 1.3 (which I used to do). However, I still
3649 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3650 workstation).
3651
3652 24. Changed the PCRE licence to be the more standard "BSD" licence.
3653
3654
3655 Version 4.5 01-Dec-03
3656 ---------------------
3657
3658 1. There has been some re-arrangement of the code for the match() function so
3659 that it can be compiled in a version that does not call itself recursively.
3660 Instead, it keeps those local variables that need separate instances for
3661 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3662 needs to "recurse". Keeping track of where control must go is done by means
3663 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3664 hide most of the details from the main code, and operates only if
3665 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3666 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3667 operating.
3668
3669 To make it easier for callers to provide specially tailored get/free
3670 functions for this usage, two new functions, pcre_stack_malloc, and
3671 pcre_stack_free, are used. They are always called in strict stacking order,
3672 and the size of block requested is always the same.
3673
3674 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3675 PCRE has been compiled to use the stack or the heap for recursion. The
3676 -C option of pcretest uses this to show which version is compiled.
3677
3678 A new data escape \S, is added to pcretest; it causes the amounts of store
3679 obtained and freed by both kinds of malloc/free at match time to be added
3680 to the output.
3681
3682 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3683 what's available on my current Linux desktop machine.
3684
3685 3. When matching a UTF-8 string, the test for a valid string at the start has
3686 been extended. If start_offset is not zero, PCRE now checks that it points
3687 to a byte that is the start of a UTF-8 character. If not, it returns
3688 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3689 this is necessary because there may be backward assertions in the pattern.
3690 When matching the same subject several times, it may save resources to use
3691 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3692
3693 4. The code for checking the validity of UTF-8 strings has been tightened so
3694 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3695 containing "overlong sequences".
3696
3697 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3698 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3699 should have been "&", but it just so happened that all the cases this let
3700 through by mistake were picked up later in the function.
3701
3702 6. I had used a variable called "isblank" - this is a C99 function, causing
3703 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3704
3705 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3706 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3707 so the version is shown; (c) stop "make test" echoing "./RunTest".
3708
3709 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3710
3711 9. The prototype for memmove() for systems that don't have it was using
3712 size_t, but the inclusion of the header that defines size_t was later. I've
3713 moved the #includes for the C headers earlier to avoid this.
3714
3715 10. Added some adjustments to the code to make it easier to compiler on certain
3716 special systems:
3717
3718 (a) Some "const" qualifiers were missing.
3719 (b) Added the macro EXPORT before all exported functions; by default this
3720 is defined to be empty.
3721 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3722 that it reads its output file name as an argument instead of writing
3723 to the standard output and assuming this can be redirected.
3724
3725 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3726 class containing characters with values greater than 255, PCRE compilation
3727 went into a loop.
3728
3729 12. A recursive reference to a subpattern that was within another subpattern
3730 that had a minimum quantifier of zero caused PCRE to crash. For example,
3731 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3732 recursion. If the recursively-called subpattern itself had a zero repeat,
3733 that was OK.
3734
3735 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3736 buffer into which it was copied (for escape processing) was still set at
3737 1024, so long lines caused crashes.
3738
3739 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3740 "internal error: code overflow...". This applied to any character class
3741 that was followed by a possessive quantifier.
3742
3743 15. Modified the Makefile to add libpcre.la as a prerequisite for
3744 libpcreposix.la because I was told this is needed for a parallel build to
3745 work.
3746
3747 16. If a pattern that contained .* following optional items at the start was
3748 studied, the wrong optimizing data was generated, leading to matching
3749 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3750 matching string must start with a or b or c. The correct conclusion for
3751 this pattern is that a match can start with any character.
3752
3753
3754 Version 4.4 13-Aug-03
3755 ---------------------
3756
3757 1. In UTF-8 mode, a character class containing characters with values between
3758 127 and 255 was not handled correctly if the compiled pattern was studied.
3759 In fixing this, I have also improved the studying algorithm for such
3760 classes (slightly).
3761
3762 2. Three internal functions had redundant arguments passed to them. Removal
3763 might give a very teeny performance improvement.
3764
3765 3. Documentation bug: the value of the capture_top field in a callout is *one
3766 more than* the number of the hightest numbered captured substring.
3767
3768 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3769 in incorrectly linking with a previously installed version. They now link
3770 explicitly with libpcre.la.
3771
3772 5. configure.in no longer needs to recognize Cygwin specially.
3773
3774 6. A problem in pcre.in for Windows platforms is fixed.
3775
3776 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3777 pcretest, it used to include the size of the study block as part of its
3778 output. Unfortunately, the structure contains a field that has a different
3779 size on different hardware architectures. This meant that the tests that
3780 showed this size failed. As the block is currently always of a fixed size,
3781 this information isn't actually particularly useful in pcretest output, so
3782 I have just removed it.
3783
3784 8. Three pre-processor statements accidentally did not start in column 1.
3785 Sadly, there are *still* compilers around that complain, even though
3786 standard C has not required this for well over a decade. Sigh.
3787
3788 9. In pcretest, the code for checking callouts passed small integers in the
3789 callout_data field, which is a void * field. However, some picky compilers
3790 complained about the casts involved for this on 64-bit systems. Now
3791 pcretest passes the address of the small integer instead, which should get
3792 rid of the warnings.
3793
3794 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3795 both compile and run time, and gives an error if an invalid UTF-8 sequence
3796 is found. There is a option for disabling this check in cases where the
3797 string is known to be correct and/or the maximum performance is wanted.
3798
3799 11. In response to a bug report, I changed one line in Makefile.in from
3800
3801 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3802 to
3803 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3804
3805 to look similar to other lines, but I have no way of telling whether this
3806 is the right thing to do, as I do not use Windows. No doubt I'll get told
3807 if it's wrong...
3808
3809
3810 Version 4.3 21-May-03
3811 ---------------------
3812
3813 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3814 Makefile.
3815
3816 2. Some refactoring to improve the quality of the code:
3817
3818 (i) The utf8_table... variables are now declared "const".
3819
3820 (ii) The code for \cx, which used the "case flipping" table to upper case
3821 lower case letters, now just substracts 32. This is ASCII-specific,
3822 but the whole concept of \cx is ASCII-specific, so it seems
3823 reasonable.
3824
3825 (iii) PCRE was using its character types table to recognize decimal and
3826 hexadecimal digits in the pattern. This is silly, because it handles
3827 only 0-9, a-f, and A-F, but the character types table is locale-
3828 specific, which means strange things might happen. A private
3829 table is now used for this - though it costs 256 bytes, a table is
3830 much faster than multiple explicit tests. Of course, the standard
3831 character types table is still used for matching digits in subject
3832 strings against \d.
3833
3834 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3835 ending in _t are). So I've renamed it as ESC_tee.
3836
3837 3. The first argument for regexec() in the POSIX wrapper should have been
3838 defined as "const".
3839
3840 4. Changed pcretest to use malloc() for its buffers so that they can be
3841 Electric Fenced for debugging.
3842
3843 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3844 to read one or more bytes before the start of the subject string. Often this
3845 had no effect on PCRE's behaviour, but in some circumstances it could
3846 provoke a segmentation fault.
3847
3848 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3849 to try to read one or more bytes before the start of the subject string.
3850
3851 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3852 UTF-8 support could misbehave in various ways if the subject string
3853 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3854 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3855 back over UTF-8 characters.)
3856
3857
3858 Version 4.2 14-Apr-03
3859 ---------------------
3860
3861 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3862
3863 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3864 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3865 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3866 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3867 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3868 and BUILD_EXEEXT
3869 Note: automatic setting of the BUILD variables is not yet working
3870 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3871 compile-time but not at link-time
3872 [LINK]: use for linking executables only
3873 make different versions for Windows and non-Windows
3874 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3875 libraries
3876 [LINK_FOR_BUILD]: new variable
3877 [OBJEXT]: use throughout
3878 [EXEEXT]: use throughout
3879 <winshared>: new target
3880 <wininstall>: new target
3881 <dftables.o>: use native compiler
3882 <dftables>: use native linker
3883 <install>: handle Windows platform correctly
3884 <clean>: ditto
3885 <check>: ditto
3886 copy DLL to top builddir before testing
3887
3888 As part of these changes, -no-undefined was removed again. This was reported
3889 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3890 in any case.
3891
3892 3. Some tidies to get rid of compiler warnings:
3893
3894 . In the match_data structure, match_limit was an unsigned long int, whereas
3895 match_call_count was an int. I've made them both unsigned long ints.
3896
3897 . In pcretest the fact that a const uschar * doesn't automatically cast to
3898 a void * provoked a warning.
3899
3900 . Turning on some more compiler warnings threw up some "shadow" variables
3901 and a few more missing casts.
3902
3903 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3904 option, a class that contained a single character with a value between 128
3905 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3906
3907 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3908 option, a class that contained several characters, but with at least one
3909 whose value was between 128 and 255 caused PCRE to crash.
3910
3911
3912 Version 4.1 12-Mar-03
3913 ---------------------
3914
3915 1. Compiling with gcc -pedantic found a couple of places where casts were
3916 needed, and a string in dftables.c that was longer than standard compilers are
3917 required to support.
3918
3919 2. Compiling with Sun's compiler found a few more places where the code could
3920 be tidied up in order to avoid warnings.
3921
3922 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3923 first of these names is deprecated in the latest Autoconf in favour of the name
3924 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3925 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3926 analogy I have changed it to CFLAGS_FOR_BUILD.
3927
3928 4. Added -no-undefined to the linking command in the Makefile, because this is
3929 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3930 linking step for the pcreposix library.
3931
3932 5. PCRE was failing to diagnose the case of two named groups with the same
3933 name.
3934
3935 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3936 literal character that is needed in the subject for a match, and scans along to
3937 ensure that it is present before embarking on the full matching process. This
3938 saves time in cases of nested unlimited repeats that are never going to match.
3939 Problem: the scan can take a lot of time if the subject is very long (e.g.
3940 megabytes), thus penalizing straightforward matches. It is now done only if the
3941 amount of subject to be scanned is less than 1000 bytes.
3942
3943 7. A lesser problem with the same optimization is that it was recording the
3944 first character of an anchored pattern as "needed", thus provoking a search
3945 right along the subject, even when the first match of the pattern was going to
3946 fail. The "needed" character is now not set for anchored patterns, unless it
3947 follows something in the pattern that is of non-fixed length. Thus, it still
3948 fulfils its original purpose of finding quick non-matches in cases of nested
3949 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3950
3951
3952 Version 4.0 17-Feb-03
3953 ---------------------
3954
3955 1. If a comment in an extended regex that started immediately after a meta-item
3956 extended to the end of string, PCRE compiled incorrect data. This could lead to
3957 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3958
3959 2. Moved to autoconf 2.53 and libtool 1.4.2.
3960
3961 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3962 the special perltest8 script is no longer needed - all the tests can be run
3963 from a single perltest script.
3964
3965 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3966 by \s. It has now been removed in PCRE. This means it isn't recognized as
3967 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3968 class [:space:] *does* include VT, thereby creating a mess.
3969
3970 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3971 space and tab.
3972
3973 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3974 its new features into the main test script, reducing the number of scripts.
3975
3976 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3977 were backward compatible, and made the (?i) apply to the whole pattern, as if
3978 /i were given. Now it behaves more logically, and applies the option setting
3979 only to what follows. PCRE has been changed to follow suit. However, if it
3980 finds options settings right at the start of the pattern, it extracts them into
3981 the global options, as before. Thus, they show up in the info data.
3982
3983 8. Added support for the \Q...\E escape sequence. Characters in between are
3984 treated as literals. This is slightly different from Perl in that $ and @ are
3985 also handled as literals inside the quotes. In Perl, they will cause variable
3986 interpolation. Note the following examples:
3987
3988 Pattern PCRE matches Perl matches
3989
3990 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3991 \Qabc\$xyz\E abc\$xyz abc\$xyz
3992 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3993
3994 For compatibility with Perl, \Q...\E sequences are recognized inside character
3995 classes as well as outside them.
3996
3997 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3998 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3999 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4000 signed/unsigned warnings.
4001
4002 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4003 option for pcretest, so I've replaced it by a simple function that does just
4004 that job.
4005
4006 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4007 "pcregrep -".
4008
4009 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4010 Java package. This provides some syntactic sugar for simple cases of what my
4011 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4012 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4013 item, you can use this simplified notation. Note that only makes sense with
4014 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4015 greediness, whatever the setting of the PCRE_UNGREEDY option.
4016
4017 13. A change of greediness default within a pattern was not taking effect at
4018 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4019 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4020 was abstracted outside.
4021
4022 14. PCRE now supports the \G assertion. It is true when the current matching
4023 position is at the start point of the match. This differs from \A when the
4024 starting offset is non-zero. Used with the /g option of pcretest (or similar
4025 code), it works in the same way as it does for Perl's /g option. If all
4026 alternatives of a regex begin with \G, the expression is anchored to the start
4027 match position, and the "anchored" flag is set in the compiled expression.
4028
4029 15. Some bugs concerning the handling of certain option changes within patterns
4030 have been fixed. These applied to options other than (?ims). For example,
4031 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4032 been the other way round. Some of this was related to change 7 above.
4033
4034 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4035 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4036 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4037 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4038
4039 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4040 mode. Unlike ".", it always matches newline, whatever the setting of
4041 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4042 assertions. Perl allows it, but it doesn't (in general) work because it can't
4043 calculate the length of the lookbehind. At least, that's the case for Perl
4044 5.8.0 - I've been told they are going to document that it doesn't work in
4045 future.
4046
4047 18. Added an error diagnosis for escapes that PCRE does not support: these are
4048 \L, \l, \N, \P, \p, \U, \u, and \X.
4049
4050 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4051 reading past the end of the pattern in cases such as /[abcd/.
4052
4053 20. PCRE was getting more memory than necessary for patterns with classes that
4054 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4055
4056 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4057 compiling PCRE for use with Virtual Pascal.
4058
4059 22. Small fix to the Makefile to make it work properly if the build is done
4060 outside the source tree.
4061
4062 23. Added a new extension: a condition to go with recursion. If a conditional
4063 subpattern starts with (?(R) the "true" branch is used if recursion has
4064 happened, whereas the "false" branch is used only at the top level.
4065
4066 24. When there was a very long string of literal characters (over 255 bytes
4067 without UTF support, over 250 bytes with UTF support), the computation of how
4068 much memory was required could be incorrect, leading to segfaults or other
4069 strange effects.
4070
4071 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4072 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4073 there was a subsequent back reference to those brackets. This meant that, for
4074 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4075 possible to check for precisely this case. All we can do is abandon the
4076 optimization if .* occurs inside capturing brackets when there are any back
4077 references whatsoever. (See below for a better fix that came later.)
4078
4079 26. The handling of the optimization for finding the first character of a
4080 non-anchored pattern, and for finding a character that is required later in the
4081 match were failing in some cases. This didn't break the matching; it just
4082 failed to optimize when it could. The way this is done has been re-implemented.
4083
4084 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4085
4086 28. Added a new feature that provides some of the functionality that Perl
4087 provides with (?{...}). The facility is termed a "callout". The way it is done
4088 in PCRE is for the caller to provide an optional function, by setting
4089 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4090 global variable. By default it is unset, which disables all calling out. To get
4091 the function called, the regex must include (?C) at appropriate points. This
4092 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4093 This provides a means of identifying different callout points. When PCRE
4094 reaches such a point in the regex, if pcre_callout has been set, the external
4095 function is called. It is provided with data in a structure called
4096 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4097 matching continues; if it returns a non-zero value, the match at the current
4098 point fails. However, backtracking will occur if possible. [This was changed
4099 later and other features added - see item 49 below.]
4100
4101 29. pcretest is upgraded to test the callout functionality. It provides a
4102 callout function that displays information. By default, it shows the start of
4103 the match and the current position in the text. There are some new data escapes
4104 to vary what happens:
4105
4106 \C+ in addition, show current contents of captured substrings
4107 \C- do not supply a callout function
4108 \C!n return 1 when callout number n is reached
4109 \C!n!m return 1 when callout number n is reached for the mth time
4110
4111 30. If pcregrep was called with the -l option and just a single file name, it
4112 output "<stdin>" if a match was found, instead of the file name.
4113
4114 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4115 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4116 pcre_exec(). This saves a malloc/free per call. The default value of
4117 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4118 when configuring.
4119
4120 32. The default maximum size of a compiled pattern is 64K. There have been a
4121 few cases of people hitting this limit. The code now uses macros to handle the
4122 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4123 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4124 configuring. Tests 2 and 5 work only with 2-byte links because they output
4125 debugging information about compiled patterns.
4126
4127 33. Internal code re-arrangements:
4128
4129 (a) Moved the debugging function for printing out a compiled regex into
4130 its own source file (printint.c) and used #include to pull it into
4131 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4132 separate copies.
4133
4134 (b) Defined the list of op-code names for debugging as a macro in
4135 internal.h so that it is next to the definition of the opcodes.
4136
4137 (c) Defined a table of op-code lengths for simpler skipping along compiled
4138 code. This is again a macro in internal.h so that it is next to the
4139 definition of the opcodes.
4140
4141 34. Added support for recursive calls to individual subpatterns, along the
4142 lines of Robin Houston's patch (but implemented somewhat differently).
4143
4144 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4145 allow it to read and process whole directories in Win32. This code was
4146 contributed by Lionel Fourquaux; it has not been tested by me.
4147
4148 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4149 used to name a group. Names consist of alphanumerics and underscores, and must
4150 be unique. Back references use the syntax (?P=name) and recursive calls use
4151 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4152 numbers. The function pcre_fullinfo() can be used after compilation to extract
4153 a name/number map. There are three relevant calls:
4154
4155 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4156 PCRE_INFO_NAMECOUNT yields the number of entries
4157 PCRE_INFO_NAMETABLE yields a pointer to the map.
4158
4159 The map is a vector of fixed-size entries. The size of each entry depends on
4160 the length of the longest name used. The first two bytes of each entry are the
4161 group number, most significant byte first. There follows the corresponding
4162 name, zero terminated. The names are in alphabetical order.
4163
4164 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4165 case instead of 255. Making it the same both with and without UTF-8 support
4166 means that the same test output works with both.
4167
4168 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4169 calling malloc() with a zero argument.
4170
4171 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4172 optimization. I've improved things by keeping a bitmap of backreferences with
4173 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4174 fact referenced, the optimization can be applied. It is unlikely that a
4175 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4176 the match to follow \n) will appear inside brackets with a number greater than
4177 31, but if it does, any back reference > 31 suppresses the optimization.
4178
4179 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4180 of disabling numbered capturing parentheses. Any opening parenthesis that is
4181 not followed by ? behaves as if it were followed by ?: but named parentheses
4182 can still be used for capturing (and they will acquire numbers in the usual
4183 way).
4184
4185 41. Redesigned the return codes from the match() function into yes/no/error so
4186 that errors can be passed back from deep inside the nested calls. A malloc
4187 failure while inside a recursive subpattern call now causes the
4188 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4189
4190 42. It is now possible to set a limit on the number of times the match()
4191 function is called in a call to pcre_exec(). This facility makes it possible to
4192 limit the amount of recursion and backtracking, though not in a directly
4193 obvious way, because the match() function is used in a number of different
4194 circumstances. The count starts from zero for each position in the subject
4195 string (for non-anchored patterns). The default limit is, for compatibility, a
4196 large number, namely 10 000 000. You can change this in two ways:
4197
4198 (a) When configuring PCRE before making, you can use --with-match-limit=n
4199 to set a default value for the compiled library.
4200
4201 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4202 a different value is set. See 45 below.
4203
4204 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4205
4206 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4207 of things that can be changed at compile time. The first argument specifies
4208 what is wanted and the second points to where the information is to be placed.
4209 The current list of available information is:
4210
4211 PCRE_CONFIG_UTF8
4212
4213 The output is an integer that is set to one if UTF-8 support is available;
4214 otherwise it is set to zero.
4215
4216 PCRE_CONFIG_NEWLINE
4217
4218 The output is an integer that it set to the value of the code that is used for
4219 newline. It is either LF (10) or CR (13).
4220
4221 PCRE_CONFIG_LINK_SIZE
4222
4223 The output is an integer that contains the number of bytes used for internal
4224 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4225
4226 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4227
4228 The output is an integer that contains the threshold above which the POSIX
4229 interface uses malloc() for output vectors. See item 31 above.
4230
4231 PCRE_CONFIG_MATCH_LIMIT
4232
4233 The output is an unsigned integer that contains the default limit of the number
4234 of match() calls in a pcre_exec() execution. See 42 above.
4235
4236 44. pcretest has been upgraded by the addition of the -C option. This causes it
4237 to extract all the available output from the new pcre_config() function, and to
4238 output it. The program then exits immediately.
4239
4240 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4241 order to support additional features. One way would have been to define
4242 pcre_exec2() (for example) with extra arguments, but this would not have been
4243 extensible, and would also have required all calls to the original function to
4244 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4245 is used for passing in "extra" data from pcre_study().
4246
4247 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4248 contains the following fields:
4249
4250 flags a bitmap indicating which of the following fields are set
4251 study_data opaque data from pcre_study()
4252 match_limit a way of specifying a limit on match() calls for a specific
4253 call to pcre_exec()
4254 callout_data data for callouts (see 49 below)
4255
4256 The flag bits are also defined in pcre.h, and are
4257
4258 PCRE_EXTRA_STUDY_DATA
4259 PCRE_EXTRA_MATCH_LIMIT
4260 PCRE_EXTRA_CALLOUT_DATA
4261
4262 The pcre_study() function now returns one of these new pcre_extra blocks, with
4263 the actual study data pointed to by the study_data field, and the
4264 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4265 before. That is, this change is entirely upwards-compatible and requires no
4266 change to existing code.
4267
4268 If you want to pass in additional data to pcre_exec(), you can either place it
4269 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4270 block.
4271
4272 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4273 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4274 times with different match limits, until it finds the minimum value needed for
4275 pcre_exec() to complete. The value is then output. This can be instructive; for
4276 most simple matches the number is quite small, but for pathological cases it
4277 gets very large very quickly.
4278
4279 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4280 returns the size of the data block pointed to by the study_data field in a
4281 pcre_extra block, that is, the value that was passed as the argument to
4282 pcre_malloc() when PCRE was getting memory in which to place the information
4283 created by pcre_study(). The fourth argument should point to a size_t variable.
4284 pcretest has been extended so that this information is shown after a successful
4285 pcre_study() call when information about the compiled regex is being displayed.
4286
4287 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4288 because what follows is always an absolute path. (Later: it turns out that this
4289 is more than cosmetic for MinGW, because it doesn't like empty path
4290 components.)
4291
4292 49. Some changes have been made to the callout feature (see 28 above):
4293
4294 (i) A callout function now has three choices for what it returns:
4295
4296 0 => success, carry on matching
4297 > 0 => failure at this point, but backtrack if possible
4298 < 0 => serious error, return this value from pcre_exec()
4299
4300 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4301 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4302 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4303 use by callout functions. It will never be used by PCRE itself.
4304
4305 (ii) The pcre_extra structure (see 45 above) has a void * field called
4306 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4307 pcre_callout_block structure has a field of the same name. The contents of
4308 the field passed in the pcre_extra structure are passed to the callout
4309 function in the corresponding field in the callout block. This makes it
4310 easier to use the same callout-containing regex from multiple threads. For
4311 testing, the pcretest program has a new data escape
4312
4313 \C*n pass the number n (may be negative) as callout_data
4314
4315 If the callout function in pcretest receives a non-zero value as
4316 callout_data, it returns that value.
4317
4318 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4319 there were some redundant $(CFLAGS) in commands that are now specified as
4320 $(LINK), which already includes $(CFLAGS).
4321
4322 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4323 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4324 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4325 one-byte characters throughout. Note that case-insensitive matching applies
4326 only to characters whose values are less than 256. PCRE doesn't support the
4327 notion of cases for higher-valued characters.
4328
4329 (i) A character class whose characters are all within 0-255 is handled as
4330 a bit map, and the map is inverted for negative classes. Previously, a
4331 character > 255 always failed to match such a class; however it should
4332 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4333
4334 (ii) A negated character class with a single character < 255 is coded as
4335 "not this character" (OP_NOT). This wasn't working properly when the test
4336 character was multibyte, either singly or repeated.
4337
4338 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4339 mode, for example: \x{100}{2,3}.
4340
4341 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4342 singly or repeated) now correctly test multibyte characters. However,
4343 PCRE doesn't recognize any characters with values greater than 255 as
4344 digits, spaces, or word characters. Such characters always match \D, \S,
4345 and \W, and never match \d, \s, or \w.
4346
4347 (v) Classes may now contain characters and character ranges with values
4348 greater than 255. For example: [ab\x{100}-\x{400}].
4349
4350 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4351 PCRE in UTF-8 mode.
4352
4353 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4354 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4355 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4356 value.)
4357
4358 53. The single man page has become too large. I have therefore split it up into
4359 a number of separate man pages. These also give rise to individual HTML pages;
4360 these are now put in a separate directory, and there is an index.html page that
4361 lists them all. Some hyperlinking between the pages has been installed.
4362
4363 54. Added convenience functions for handling named capturing parentheses.
4364
4365 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4366 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4367 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4368 are faulted.
4369
4370 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4371 calling configure. These values are used when compiling the dftables.c program
4372 which is run to generate the source of the default character tables. They
4373 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4374 you will need to set these values.
4375
4376 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4377
4378
4379 Version 3.9 02-Jan-02
4380 ---------------------
4381
4382 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4383
4384 2. If --disable-static was given, the building process failed when trying to
4385 build pcretest and pcregrep. (For some reason it was using libtool to compile
4386 them, which is not right, as they aren't part of the library.)
4387
4388
4389 Version 3.8 18-Dec-01
4390 ---------------------
4391
4392 1. The experimental UTF-8 code was completely screwed up. It was packing the
4393 bytes in the wrong order. How dumb can you get?
4394
4395
4396 Version 3.7 29-Oct-01
4397 ---------------------
4398
4399 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4400 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4401 this didn't happen under Solaris 8, where I normally test things.
4402
4403 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4404 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4405 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4406 this fix an hour or so after the initial 3.7 release.)
4407
4408
4409 Version 3.6 23-Oct-01
4410 ---------------------
4411
4412 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4413 offsets passed as NULL with zero offset count.
4414
4415 2. The config.guess and config.sub files had not been updated when I moved to
4416 the latest autoconf.
4417
4418
4419 Version 3.5 15-Aug-01
4420 ---------------------
4421
4422 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4423 had been forgotten.
4424
4425 2. By using declared but undefined structures, we can avoid using "void"
4426 definitions in pcre.h while keeping the internal definitions of the structures
4427 private.
4428
4429 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4430 user point of view, this means that both static and shared libraries are built
4431 by default, but this can be individually controlled. More of the work of
4432 handling this static/shared cases is now inside libtool instead of PCRE's make
4433 file.
4434
4435 4. The pcretest utility is now installed along with pcregrep because it is
4436 useful for users (to test regexs) and by doing this, it automatically gets
4437 relinked by libtool. The documentation has been turned into a man page, so
4438 there are now .1, .txt, and .html versions in /doc.
4439
4440 5. Upgrades to pcregrep:
4441 (i) Added long-form option names like gnu grep.
4442 (ii) Added --help to list all options with an explanatory phrase.
4443 (iii) Added -r, --recursive to recurse into sub-directories.
4444 (iv) Added -f, --file to read patterns from a file.
4445
4446 6. pcre_exec() was referring to its "code" argument before testing that
4447 argument for NULL (and giving an error if it was NULL).
4448
4449 7. Upgraded Makefile.in to allow for compiling in a different directory from
4450 the source directory.
4451
4452 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4453 options bits, the pointer it was passed was to an int instead of to an unsigned
4454 long int. This mattered only on 64-bit systems.
4455
4456 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4457 generated) instead of pcre.in, which it its source. Also made the same change
4458 in several of the .c files.
4459
4460 10. A new release of gcc defines printf() as a macro, which broke pcretest
4461 because it had an ifdef in the middle of a string argument for printf(). Fixed
4462 by using separate calls to printf().
4463
4464 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4465 script, to force use of CR or LF instead of \n in the source. On non-Unix
4466 systems, the value can be set in config.h.
4467
4468 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4469 absolute limit. Changed the text of the error message to make this clear, and
4470 likewise updated the man page.
4471
4472 13. The limit of 99 on the number of capturing subpatterns has been removed.
4473 The new limit is 65535, which I hope will not be a "real" limit.
4474
4475
4476 Version 3.4 22-Aug-00
4477 ---------------------
4478
4479 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4480
4481 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4482
4483
4484 Version 3.3 01-Aug-00
4485 ---------------------
4486
4487 1. If an octal character was given, but the value was greater than \377, it
4488 was not getting masked to the least significant bits, as documented. This could
4489 lead to crashes in some systems.
4490
4491 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4492 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4493
4494 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4495 These just pass their arguments on to (pcre_free)(), but they are provided
4496 because some uses of PCRE bind it to non-C systems that can call its functions,
4497 but cannot call free() or pcre_free() directly.
4498
4499 4. Add "make test" as a synonym for "make check". Corrected some comments in
4500 the Makefile.
4501
4502 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4503 Makefile.
4504
4505 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4506 command called pgrep for grepping around the active processes.
4507
4508 7. Added the beginnings of support for UTF-8 character strings.
4509
4510 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4511 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4512 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4513 out for the ar command.)
4514
4515
4516 Version 3.2 12-May-00
4517 ---------------------
4518
4519 This is purely a bug fixing release.
4520
4521 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4522 of ZA. This was just one example of several cases that could provoke this bug,
4523 which was introduced by change 9 of version 2.00. The code for breaking
4524 infinite loops after an iteration that matches an empty string was't working
4525 correctly.
4526
4527 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4528 when matched against abbab (for example). After matching an empty string, it
4529 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4530 caused it to match further down the string than it should.
4531
4532 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4533 was there because it doesn't seem to be needed, and it causes trouble on some
4534 systems, as it is not a Standard C header. It has been removed.
4535
4536 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4537 were reported on the Macintosh. The changes were from
4538
4539 while ((c = *(++ptr)) != 0 && c != '\n');
4540 to
4541 while ((c = *(++ptr)) != 0 && c != '\n') ;
4542
4543 Totally extraordinary, but if that's what it takes...
4544
4545 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4546 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4547 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4548 assumes the way PCRE uses memmove() (always moving upwards).
4549
4550 6. PCRE is being used in one environment where strchr() is not available. There
4551 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4552 faster code anyway.
4553
4554
4555 Version 3.1 09-Feb-00
4556 ---------------------
4557
4558 The only change in this release is the fixing of some bugs in Makefile.in for
4559 the "install" target:
4560
4561 (1) It was failing to install pcreposix.h.
4562
4563 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4564
4565
4566 Version 3.0 01-Feb-00
4567 ---------------------
4568
4569 1. Add support for the /+ modifier to perltest (to output $` like it does in
4570 pcretest).
4571
4572 2. Add support for the /g modifier to perltest.
4573
4574 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4575 matches null strings.
4576
4577 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4578 pattern. Perl treats empty patterns specially - it reuses the most recent
4579 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4580 effect.
4581
4582 5. The POSIX interface was broken in that it was just handing over the POSIX
4583 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4584 required a bigger vector, with some working space on the end. This means that
4585 the POSIX wrapper now has to get and free some memory, and copy the results.
4586
4587 6. Added some simple autoconf support, placing the test data and the
4588 documentation in separate directories, re-organizing some of the
4589 information files, and making it build pcre-config (a GNU standard). Also added
4590 libtool support for building PCRE as a shared library, which is now the
4591 default.
4592
4593 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4594 09 are not valid octal constants. Single digits will be used for minor values
4595 less than 10.
4596
4597 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4598 existing programs that set these in the POSIX interface can use PCRE without
4599 modification.
4600
4601 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4602 return all that pcre_info() returns, plus additional data. The pcre_info()
4603 function is retained for compatibility, but is considered to be obsolete.
4604
4605 10. Added experimental recursion feature (?R) to handle one common case that
4606 Perl 5.6 will be able to do with (?p{...}).
4607
4608 11. Added support for POSIX character classes like [:alpha:], which Perl is
4609 adopting.
4610
4611
4612 Version 2.08 31-Aug-99
4613 ----------------------
4614
4615 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4616 trying to match at the startoffset position, but instead was moving forward to
4617 the next newline as if a previous match had failed.
4618
4619 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4620 and could get into a loop if a null string was matched other than at the start
4621 of the subject.
4622
4623 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4624 be distinguished at compile time, and for completeness also added PCRE_DATE.
4625
4626 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4627 in GnuWin32 environments.
4628
4629
4630 Version 2.07 29-Jul-99
4631 ----------------------
4632
4633 1. The documentation is now supplied in plain text form and HTML as well as in
4634 the form of man page sources.
4635
4636 2. C++ compilers don't like assigning (void *) values to other pointer types.
4637 In particular this affects malloc(). Although there is no problem in Standard
4638 C, I've put in casts to keep C++ compilers happy.
4639
4640 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4641 should be (const char *).
4642
4643 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4644 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4645 However, I haven't made this a standard facility. The documentation doesn't
4646 mention it, and the Makefile doesn't support it.
4647
4648 5. The Makefile now contains an "install" target, with editable destinations at
4649 the top of the file. The pcretest program is not installed.
4650
4651 6. pgrep -V now gives the PCRE version number and date.
4652
4653 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4654 causing the entire string to be ignored, instead of just the last character.
4655
4656 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4657 non-matching string, it can take a very, very long time, even for strings of
4658 quite modest length, because of the nested recursion. PCRE now does better in
4659 some of these cases. It does this by remembering the last required literal
4660 character in the pattern, and pre-searching the subject to ensure it is present
4661 before running the real match. In other words, it applies a heuristic to detect
4662 some types of certain failure quickly, and in the above example, if presented
4663 with a string that has no trailing " it gives "no match" very quickly.
4664
4665 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4666 other alternatives are tried instead.
4667
4668
4669 Version 2.06 09-Jun-99
4670 ----------------------
4671
4672 1. Change pcretest's output for amount of store used to show just the code
4673 space, because the remainder (the data block) varies in size between 32-bit and
4674 64-bit systems.
4675
4676 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4677 start matching at. This allows lookbehinds to work when searching for multiple
4678 occurrences in a string.
4679
4680 3. Added additional options to pcretest for testing multiple occurrences:
4681
4682 /+ outputs the rest of the string that follows a match
4683 /g loops for multiple occurrences, using the new startoffset argument
4684 /G loops for multiple occurrences by passing an incremented pointer
4685
4686 4. PCRE wasn't doing the "first character" optimization for patterns starting
4687 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4688 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4689 the letter 'x'. On long subject strings, this gives a significant speed-up.
4690
4691
4692 Version 2.05 21-Apr-99
4693 ----------------------
4694
4695 1. Changed the type of magic_number from int to long int so that it works
4696 properly on 16-bit systems.
4697
4698 2. Fixed a bug which caused patterns starting with .* not to work correctly
4699 when the subject string contained newline characters. PCRE was assuming
4700 anchoring for such patterns in all cases, which is not correct because .* will
4701 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4702 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4703 must be retried after every newline in the subject.
4704
4705
4706 Version 2.04 18-Feb-99
4707 ----------------------
4708
4709 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4710 computation of the store needed to hold the pattern was incorrect (too large).
4711 If such patterns were nested a few deep, this could multiply and become a real
4712 problem.
4713
4714 2. Added /M option to pcretest to show the memory requirement of a specific
4715 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4716
4717 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4718 compiled in such a way that the backtracking after subsequent failure was
4719 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4720 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4721
4722
4723 Version 2.03 02-Feb-99
4724 ----------------------
4725
4726 1. Fixed typo and small mistake in man page.
4727
4728 2. Added 4th condition (GPL supersedes if conflict) and created separate
4729 LICENCE file containing the conditions.
4730
4731 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4732 Perl, that is the internal \ allows the delimiter to be included in the
4733 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4734 the final delimiter, add \ to the end of the pattern (to test the error).
4735
4736 4. Added the convenience functions for extracting substrings after a successful
4737 match. Updated pcretest to make it able to test these functions.
4738
4739
4740 Version 2.02 14-Jan-99
4741 ----------------------
4742
4743 1. Initialized the working variables associated with each extraction so that
4744 their saving and restoring doesn't refer to uninitialized store.
4745
4746 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4747 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4748 fix the problem.
4749
4750 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4751 calls, and wasn't printing the correct value for compiling calls. Increased the
4752 default value of LOOPREPEAT, and the number of significant figures in the
4753 times.
4754
4755 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4756
4757 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4758 a building problem on Windows NT with a FAT file system.
4759
4760
4761 Version 2.01 21-Oct-98
4762 ----------------------
4763
4764 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4765 to character tables built by pcre_maketables() in the current locale. If NULL
4766 is passed, the default tables are used.
4767
4768
4769 Version 2.00 24-Sep-98
4770 ----------------------
4771
4772 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4773 it any more.
4774
4775 2. Allow quantification of (?>) groups, and make it work correctly.
4776
4777 3. The first character computation wasn't working for (?>) groups.
4778
4779 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4780 end of the subject) and add 5.005's \z, which really does match only at the
4781 very end of the subject.
4782
4783 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4784
4785 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4786 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4787 localized options. All options to pcre_study() were also removed.
4788
4789 7. Add other new features from 5.005:
4790
4791 $(?<= positive lookbehind
4792 $(?<! negative lookbehind
4793 (?imsx-imsx) added the unsetting capability
4794 such a setting is global if at outer level; local otherwise
4795 (?imsx-imsx:) non-capturing groups with option setting
4796 (?(cond)re|re) conditional pattern matching
4797
4798 A backreference to itself in a repeated group matches the previous
4799 captured string.
4800
4801 8. General tidying up of studying (both automatic and via "study")
4802 consequential on the addition of new assertions.
4803
4804 9. As in 5.005, unlimited repeated groups that could match an empty substring
4805 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4806 runtime if any iteration does actually match an empty substring.
4807
4808 10. Include the RunTest script in the distribution.
4809
4810 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4811 discrepancies, some of which were old and were also with respect to 5.004. They
4812 have now been fixed.
4813
4814
4815 Version 1.09 28-Apr-98
4816 ----------------------
4817
4818 1. A negated single character class followed by a quantifier with a minimum
4819 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4820 program crashes, or just wrong answers. This did not apply to negated classes
4821 containing more than one character, or to minima other than one.
4822
4823
4824 Version 1.08 27-Mar-98
4825 ----------------------
4826
4827 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4828
4829 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4830 latter must appear before anything that relies on it in the pattern.
4831
4832
4833 Version 1.07 16-Feb-98
4834 ----------------------
4835
4836 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4837 repeat of a potentially empty string).
4838
4839
4840 Version 1.06 23-Jan-98
4841 ----------------------
4842
4843 1. Added Markus Oberhumer's little patches for C++.
4844
4845 2. Literal strings longer than 255 characters were broken.
4846
4847
4848 Version 1.05 23-Dec-97
4849 ----------------------
4850
4851 1. Negated character classes containing more than one character were failing if
4852 PCRE_CASELESS was set at run time.
4853
4854
4855 Version 1.04 19-Dec-97
4856 ----------------------
4857
4858 1. Corrected the man page, where some "const" qualifiers had been omitted.
4859
4860 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4861 input syntax.
4862
4863 3. Fixed memory leak which occurred when a regex with back references was
4864 matched with an offsets vector that wasn't big enough. The temporary memory
4865 that is used in this case wasn't being freed if the match failed.
4866
4867 4. Tidied pcretest to ensure it frees memory that it gets.
4868
4869 5. Temporary memory was being obtained in the case where the passed offsets
4870 vector was exactly big enough.
4871
4872 6. Corrected definition of offsetof() from change 5 below.
4873
4874 7. I had screwed up change 6 below and broken the rules for the use of
4875 setjmp(). Now fixed.
4876
4877
4878 Version 1.03 18-Dec-97
4879 ----------------------
4880
4881 1. A erroneous regex with a missing opening parenthesis was correctly
4882 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4883 on some systems.
4884
4885 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4886 it was reported that one broken compiler failed on the former because "code" is
4887 also an independent variable.
4888
4889 3. The erroneous regex a[]b caused an array overrun reference.
4890
4891 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4892 fail on data ending with that character. (It was going on too far, and checking
4893 the next character, typically a binary zero.) This was specific to the
4894 optimized code for single-character negative classes.
4895
4896 5. Added a contributed patch from the TIN world which does the following:
4897
4898 + Add an undef for memmove, in case the the system defines a macro for it.
4899
4900 + Add a definition of offsetof(), in case there isn't one. (I don't know
4901 the reason behind this - offsetof() is part of the ANSI standard - but
4902 it does no harm).
4903
4904 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4905 most of the places where whitespace preceded '#'. I have given up and
4906 allowed the remaining 2 cases to be at the margin.
4907
4908 + Rename some variables in pcre to eliminate shadowing. This seems very
4909 pedantic, but does no harm, of course.
4910
4911 6. Moved the call to setjmp() into its own function, to get rid of warnings
4912 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4913
4914 7. Constructs such as \d{8,} were compiling into the equivalent of
4915 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4916 outcome, but in this particular case used more store than had been allocated,
4917 which caused the bug to be discovered because it threw up an internal error.
4918
4919 8. The debugging code in both pcre and pcretest for outputting the compiled
4920 form of a regex was going wrong in the case of back references followed by
4921 curly-bracketed repeats.
4922
4923
4924 Version 1.02 12-Dec-97
4925 ----------------------
4926
4927 1. Typos in pcre.3 and comments in the source fixed.
4928
4929 2. Applied a contributed patch to get rid of places where it used to remove
4930 'const' from variables, and fixed some signed/unsigned and uninitialized
4931 variable warnings.
4932
4933 3. Added the "runtest" target to Makefile.
4934
4935 4. Set default compiler flag to -O2 rather than just -O.
4936
4937
4938 Version 1.01 19-Nov-97
4939 ----------------------
4940
4941 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4942 like /([ab]*)*/, that is, for classes with more than one character in them.
4943
4944 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4945 as /((?>a*))*/ (a PCRE_EXTRA facility).
4946
4947
4948 Version 1.00 18-Nov-97
4949 ----------------------
4950
4951 1. Added compile-time macros to support systems such as SunOS4 which don't have
4952 memmove() or strerror() but have other things that can be used instead.
4953
4954 2. Arranged that "make clean" removes the executables.
4955
4956
4957 Version 0.99 27-Oct-97
4958 ----------------------
4959
4960 1. Fixed bug in code for optimizing classes with only one character. It was
4961 initializing a 32-byte map regardless, which could cause it to run off the end
4962 of the memory it had got.
4963
4964 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4965
4966
4967 Version 0.98 22-Oct-97
4968 ----------------------
4969
4970 1. Fixed bug in code for handling temporary memory usage when there are more
4971 back references than supplied space in the ovector. This could cause segfaults.
4972
4973
4974 Version 0.97 21-Oct-97
4975 ----------------------
4976
4977 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4978
4979 2. Optimized negated single characters not to use a bit map.
4980
4981 3. Brought error texts together as macro definitions; clarified some of them;
4982 fixed one that was wrong - it said "range out of order" when it meant "invalid
4983 escape sequence".
4984
4985 4. Changed some char * arguments to const char *.
4986
4987 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4988
4989 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4990 pcretest.
4991
4992
4993 Version 0.96 16-Oct-97
4994 ----------------------
4995
4996 1. Added a simple "pgrep" utility to the distribution.
4997
4998 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4999 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5000 where "ddd" means "one or more decimal digits".
5001
5002 3. Fixed serious bug. If a pattern had a back reference, but the call to
5003 pcre_exec() didn't supply a large enough ovector to record the related
5004 identifying subpattern, the match always failed. PCRE now remembers the number
5005 of the largest back reference, and gets some temporary memory in which to save
5006 the offsets during matching if necessary, in order to ensure that
5007 backreferences always work.
5008
5009 4. Increased the compatibility with Perl in a number of ways:
5010
5011 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5012 to request this handling. The option can be set at compile or exec time.
5013
5014 (b) $ matches before a terminating newline by default; an option
5015 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5016 mode). The option can be set at compile or exec time.
5017
5018 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5019 the same as Perl's. If the decimal number it represents is less than 10
5020 or there aren't that many previous left capturing parentheses, an octal
5021 escape is read. Inside a character class, it's always an octal escape,
5022 even if it is a single digit.
5023
5024 (d) An escaped but undefined alphabetic character is taken as a literal,
5025 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5026 escapes.
5027
5028 (e) {0} is now permitted. (The previous item is removed from the compiled
5029 pattern).
5030
5031 5. Changed all the names of code files so that the basic parts are no longer
5032 than 10 characters, and abolished the teeny "globals.c" file.
5033
5034 6. Changed the handling of character classes; they are now done with a 32-byte
5035 bit map always.
5036
5037 7. Added the -d and /D options to pcretest to make it possible to look at the
5038 internals of compilation without having to recompile pcre.
5039
5040
5041 Version 0.95 23-Sep-97
5042 ----------------------
5043
5044 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5045 \x20 at the start of a run of normal characters. These were being treated as
5046 real characters, instead of the source characters being re-checked.
5047
5048
5049 Version 0.94 18-Sep-97
5050 ----------------------
5051
5052 1. The functions are now thread-safe, with the caveat that the global variables
5053 containing pointers to malloc() and free() or alternative functions are the
5054 same for all threads.
5055
5056 2. Get pcre_study() to generate a bitmap of initial characters for non-
5057 anchored patterns when this is possible, and use it if passed to pcre_exec().
5058
5059
5060 Version 0.93 15-Sep-97
5061 ----------------------
5062
5063 1. /(b)|(:+)/ was computing an incorrect first character.
5064
5065 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5066 but not actually doing anything yet.
5067
5068 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5069 as Perl does (e.g. [-az] or [az-]).
5070
5071 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5072 all possible positions.
5073
5074 5. Split up into different modules to avoid including unneeded functions in a
5075 compiled binary. However, compile and exec are still in one module. The "study"
5076 function is split off.
5077
5078 6. The character tables are now in a separate module whose source is generated
5079 by an auxiliary program - but can then be edited by hand if required. There are
5080 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5081 toupper() in the code.
5082
5083 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5084 make them global. Abolish the function for setting them, as the caller can now
5085 set them directly.
5086
5087
5088 Version 0.92 11-Sep-97
5089 ----------------------
5090
5091 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5092 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5093
5094 2. Caseless matching was not working in character classes if the characters in
5095 the pattern were in upper case.
5096
5097 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5098
5099 4. Make PCRE_ANCHORED public and accept as a compile option.
5100
5101 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5102 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5103 pass them.
5104
5105 6. Give an error if bad option bits passed at compile or run time.
5106
5107 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5108 pcretest to cause it to pass that flag.
5109
5110 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5111 options, and the first character, if set.
5112
5113 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5114
5115
5116 Version 0.91 10-Sep-97
5117 ----------------------
5118
5119 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5120 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5121
5122 2. PCRE was looping on encountering an indefinitely repeated back reference to
5123 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5124 Perl does - treats the match as successful.
5125
5126 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5