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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1254 - (show annotations)
Fri Feb 22 12:46:26 2013 UTC (6 years, 9 months ago) by ph10
File size: 243088 byte(s)
Error occurred while calculating annotation data.
Added some #ifdef __VMS code to pcretest.c.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76
77 Version 8.32 30-November-2012
78 -----------------------------
79
80 1. Improved JIT compiler optimizations for first character search and single
81 character iterators.
82
83 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
84 Patch by Daniel Richard G.
85
86 3. Single character iterator optimizations in the JIT compiler.
87
88 4. Improved JIT compiler optimizations for character ranges.
89
90 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
91 Reported by Giuseppe D'Angelo.
92
93 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
94 of a line, was being set incorrectly in cases where .* appeared inside
95 atomic brackets at the start of a pattern, or where there was a subsequent
96 *PRUNE or *SKIP.
97
98 7. Improved instruction cache flush for POWER/PowerPC.
99 Patch by Daniel Richard G.
100
101 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
102 grep:
103
104 (a) There is now no limit to the number of patterns to be matched.
105
106 (b) An error is given if a pattern is too long.
107
108 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
109 are now supported.
110
111 (d) --exclude-from and --include-from (multiple use) have been added.
112
113 (e) Exclusions and inclusions now apply to all files and directories, not
114 just to those obtained from scanning a directory recursively.
115
116 (f) Multiple uses of -f and --file-list are now supported.
117
118 (g) In a Windows environment, the default for -d has been changed from
119 "read" (the GNU grep default) to "skip", because otherwise the presence
120 of a directory in the file list provokes an error.
121
122 (h) The documentation has been revised and clarified in places.
123
124 9. Improve the matching speed of capturing brackets.
125
126 10. Changed the meaning of \X so that it now matches a Unicode extended
127 grapheme cluster.
128
129 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
130 out POSIX threads when JIT support is configured.
131
132 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
133
134 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
135 structure could go wrong in environments where size_t is not the same size
136 as int.
137
138 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
139 set.
140
141 15. The EBCDIC support had decayed; later updates to the code had included
142 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
143 general tidy up of EBCDIC-related issues, and the documentation was also
144 not quite right. There is now a test that can be run on ASCII systems to
145 check some of the EBCDIC-related things (but is it not a full test).
146
147 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
148 in a small tidy to the code.
149
150 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
151
152 18. If the --only-matching (-o) option in pcregrep is specified multiple
153 times, each one causes appropriate output. For example, -o1 -o2 outputs the
154 substrings matched by the 1st and 2nd capturing parentheses. A separating
155 string can be specified by --om-separator (default empty).
156
157 19. Improving the first n character searches.
158
159 20. Turn case lists for horizontal and vertical white space into macros so that
160 they are defined only once.
161
162 21. This set of changes together give more compatible Unicode case-folding
163 behaviour for characters that have more than one other case when UCP
164 support is available.
165
166 (a) The Unicode property table now has offsets into a new table of sets of
167 three or more characters that are case-equivalent. The MultiStage2.py
168 script that generates these tables (the pcre_ucd.c file) now scans
169 CaseFolding.txt instead of UnicodeData.txt for character case
170 information.
171
172 (b) The code for adding characters or ranges of characters to a character
173 class has been abstracted into a generalized function that also handles
174 case-independence. In UTF-mode with UCP support, this uses the new data
175 to handle characters with more than one other case.
176
177 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
178 whose other case is greater than 256 are now correctly matched
179 caselessly. Previously, the high codepoint matched the low one, but not
180 vice versa.
181
182 (d) The processing of \h, \H, \v, and \ in character classes now makes use
183 of the new class addition function, using character lists defined as
184 macros alongside the case definitions of 20 above.
185
186 (e) Caseless back references now work with characters that have more than
187 one other case.
188
189 (f) General caseless matching of characters with more than one other case
190 is supported.
191
192 22. Unicode character properties were updated from Unicode 6.2.0
193
194 23. Improved CMake support under Windows. Patch by Daniel Richard G.
195
196 24. Add support for 32-bit character strings, and UTF-32
197
198 25. Major JIT compiler update (code refactoring and bugfixing).
199 Experimental Sparc 32 support is added.
200
201 26. Applied a modified version of Daniel Richard G's patch to create
202 pcre.h.generic and config.h.generic by "make" instead of in the
203 PrepareRelease script.
204
205 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
206 pcre_compile.c when checking for a zero character.
207
208 28. Introducing a native interface for JIT. Through this interface, the compiled
209 machine code can be directly executed. The purpose of this interface is to
210 provide fast pattern matching, so several sanity checks are not performed.
211 However, feature tests are still performed. The new interface provides
212 1.4x speedup compared to the old one.
213
214 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
215 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
216 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
217
218 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
219 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
220 the "old" RFC 2279). Instead, it ended up passing a negative length to
221 pcre_exec().
222
223 31. Add support for GCC's visibility feature to hide internal functions.
224
225 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
226 "unknown -C option" after outputting 0 or 1.
227
228 33. There is now support for generating a code coverage report for the test
229 suite in environments where gcc is the compiler and lcov is installed. This
230 is mainly for the benefit of the developers.
231
232 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
233 unaddressable using valgrind annotations, allowing valgrind to detect
234 invalid memory accesses. This is mainly for the benefit of the developers.
235
236 25. (*UTF) can now be used to start a pattern in any of the three libraries.
237
238 26. Give configure error if --enable-cpp but no C++ compiler found.
239
240
241 Version 8.31 06-July-2012
242 -------------------------
243
244 1. Fixing a wrong JIT test case and some compiler warnings.
245
246 2. Removed a bashism from the RunTest script.
247
248 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
249 to unsigned type, result still unsigned" that was given by an MS compiler
250 on encountering the code "-sizeof(xxx)".
251
252 4. Partial matching support is added to the JIT compiler.
253
254 5. Fixed several bugs concerned with partial matching of items that consist
255 of more than one character:
256
257 (a) /^(..)\1/ did not partially match "aba" because checking references was
258 done on an "all or nothing" basis. This also applied to repeated
259 references.
260
261 (b) \R did not give a hard partial match if \r was found at the end of the
262 subject.
263
264 (c) \X did not give a hard partial match after matching one or more
265 characters at the end of the subject.
266
267 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
268 a partial match for the string "\r".
269
270 (e) When newline was set to CRLF, the metacharacter "." did not recognize
271 a partial match for a CR character at the end of the subject string.
272
273 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
274 running pcretest, the text "(JIT)" added to the output whenever JIT is
275 actually used to run the match.
276
277 7. Individual JIT compile options can be set in pcretest by following -s+[+]
278 or /S+[+] with a digit between 1 and 7.
279
280 8. OP_NOT now supports any UTF character not just single-byte ones.
281
282 9. (*MARK) control verb is now supported by the JIT compiler.
283
284 10. The command "./RunTest list" lists the available tests without actually
285 running any of them. (Because I keep forgetting what they all are.)
286
287 11. Add PCRE_INFO_MAXLOOKBEHIND.
288
289 12. Applied a (slightly modified) user-supplied patch that improves performance
290 when the heap is used for recursion (compiled with --disable-stack-for-
291 recursion). Instead of malloc and free for each heap frame each time a
292 logical recursion happens, frames are retained on a chain and re-used where
293 possible. This sometimes gives as much as 30% improvement.
294
295 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
296 call.
297
298 14. As documented, (*COMMIT) is now confined to within a positive assertion.
299
300 15. It is now possible to link pcretest with libedit as an alternative to
301 libreadline.
302
303 16. (*COMMIT) control verb is now supported by the JIT compiler.
304
305 17. The Unicode data tables have been updated to Unicode 6.1.0.
306
307 18. Added --file-list option to pcregrep.
308
309 19. Added binary file support to pcregrep, including the -a, --binary-files,
310 -I, and --text options.
311
312 20. The madvise function is renamed for posix_madvise for QNX compatibility
313 reasons. Fixed by Giuseppe D'Angelo.
314
315 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
316
317 22. Changed the option for creating symbolic links for 16-bit man pages from
318 -s to -sf so that re-installing does not cause issues.
319
320 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
321
322 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
323 matches in certain environments (the workspace was not being correctly
324 retained). Also added to pcre_dfa_exec() a simple plausibility check on
325 some of the workspace data at the beginning of a restart.
326
327 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
328 was not doing so when it should - probably a typo introduced by SVN 528
329 (change 8.10/14).
330
331 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
332 \w+ when the character tables indicated that \x{c4} was a word character.
333 There were several related cases, all because the tests for doing a table
334 lookup were testing for characters less than 127 instead of 255.
335
336 27. If a pattern contains capturing parentheses that are not used in a match,
337 their slots in the ovector are set to -1. For those that are higher than
338 any matched groups, this happens at the end of processing. In the case when
339 there were back references that the ovector was too small to contain
340 (causing temporary malloc'd memory to be used during matching), and the
341 highest capturing number was not used, memory off the end of the ovector
342 was incorrectly being set to -1. (It was using the size of the temporary
343 memory instead of the true size.)
344
345 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
346 ovector size, it uses memory at the end of the block that it has got.
347
348 29. Check for an overlong MARK name and give an error at compile time. The
349 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
350
351 30. JIT compiler update.
352
353 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
354 Rill for the patch.
355
356 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
357
358 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
359
360 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
361 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
362 was enabled.
363
364 35. Improve JIT code generation for greedy plus quantifier.
365
366 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
367 1 to "aa" instead of to an empty string. The bug affected repeated groups
368 that could potentially match an empty string.
369
370 37. Optimizing single character iterators in JIT.
371
372 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
373 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
374 codepoints that are too big for the mode are faulted, and in a UTF mode,
375 disallowed codepoints are also faulted.
376
377 39. If PCRE was compiled with UTF support, in three places in the DFA
378 matcher there was code that should only have been obeyed in UTF mode, but
379 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
380 processing when bytes with values greater than 127 were present. In 16-bit
381 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
382 both cases the values are those that cannot be the first data item in a UTF
383 character. The three items that might have provoked this were recursions,
384 possessively repeated groups, and atomic groups.
385
386 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
387 and pcregrep, because some OS require shared objects to be explicitly
388 passed to ld, causing the link step to fail if they are not.
389
390 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
391 mode, patterns that started with \h* or \R* might be incorrectly matched.
392
393
394 Version 8.30 04-February-2012
395 -----------------------------
396
397 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
398 name is defined in ctype.h.
399
400 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
401 only in quite long subpatterns.
402
403 3. Removed the function pcre_info(), which has been obsolete and deprecated
404 since it was replaced by pcre_fullinfo() in February 2000.
405
406 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
407 match a (*MARK), and the match failed at the start of the subject, a
408 reference to memory before the start of the subject could occur. This bug
409 was introduced by fix 17 of release 8.21.
410
411 5. A reference to an unset group with zero minimum repetition was giving
412 totally wrong answers (in non-JavaScript-compatibility mode). For example,
413 /(another)?(\1?)test/ matched against "hello world test". This bug was
414 introduced in release 8.13.
415
416 6. Add support for 16-bit character strings (a large amount of work involving
417 many changes and refactorings).
418
419 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
420 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
421 from a file.
422
423 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
424 rounding is not applied in this particular case).
425
426 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
427 if they appear, or are escaped, in patterns.
428
429 10. Get rid of a number of -Wunused-but-set-variable warnings.
430
431 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
432 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
433 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
434 also returns the mark "x". This bug applied to capturing parentheses,
435 non-capturing parentheses, and atomic parentheses. It also applied to some
436 assertions.
437
438 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
439 information out of configure.ac instead of relying on pcre.h.generic, which
440 is not stored in the repository.
441
442 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
443 -lreadline.
444
445 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
446
447 15. Applied Graycode's patch to put the top-level frame on the stack rather
448 than the heap when not using the stack for recursion. This gives a
449 performance improvement in many cases when recursion is not deep.
450
451 16. Experimental code added to "pcretest -C" to output the stack frame size.
452
453
454 Version 8.21 12-Dec-2011
455 ------------------------
456
457 1. Updating the JIT compiler.
458
459 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
460 are added as well.
461
462 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
463 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
464 calling _pcre_jit_exec. Some extra comments are added.
465
466 4. (*MARK) settings inside atomic groups that do not contain any capturing
467 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
468 was introduced by change 18 for 8.20.
469
470 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
471 ECMA-262 standard.
472
473 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
474 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
475 This bug was probably introduced by change 9 of 8.13.
476
477 7. While fixing 6 above, I noticed that a number of other items were being
478 incorrectly rejected as "not fixed length". This arose partly because newer
479 opcodes had not been added to the fixed-length checking code. I have (a)
480 corrected the bug and added tests for these items, and (b) arranged for an
481 error to occur if an unknown opcode is encountered while checking for fixed
482 length instead of just assuming "not fixed length". The items that were
483 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
484 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
485 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
486
487 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
488 being incorrectly compiled and would have given unpredicatble results.
489
490 9. A possessively repeated subpattern with minimum repeat count greater than
491 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
492 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
493 the first (A) could occur when it should not.
494
495 10. Add a cast and remove a redundant test from the code.
496
497 11. JIT should use pcre_malloc/pcre_free for allocation.
498
499 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
500 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
501 is anything other than /usr, -L is still shown).
502
503 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
504
505 14. Perl does not support \N without a following name in a [] class; PCRE now
506 also gives an error.
507
508 15. If a forward reference was repeated with an upper limit of around 2000,
509 it caused the error "internal error: overran compiling workspace". The
510 maximum number of forward references (including repeats) was limited by the
511 internal workspace, and dependent on the LINK_SIZE. The code has been
512 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
513 the default depends on LINK_SIZE. There is a new upper limit (for safety)
514 of around 200,000 forward references. While doing this, I also speeded up
515 the filling in of repeated forward references.
516
517 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
518 incorrectly expecting the subject to contain another "a" after the start.
519
520 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
521 in the match, the SKIP should be ignored. This was not happening; instead
522 the SKIP was being treated as NOMATCH. For patterns such as
523 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
524 tested.
525
526 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
527 now much more compatible with Perl, in particular in cases where the result
528 is a non-match for a non-anchored pattern. For example, if
529 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
530 "m", where previously it did not return a name. A side effect of this
531 change is that for partial matches, the last encountered mark name is
532 returned, as for non matches. A number of tests that were previously not
533 Perl-compatible have been moved into the Perl-compatible test files. The
534 refactoring has had the pleasing side effect of removing one argument from
535 the match() function, thus reducing its stack requirements.
536
537 19. If the /S+ option was used in pcretest to study a pattern using JIT,
538 subsequent uses of /S (without +) incorrectly behaved like /S+.
539
540 21. Retrieve executable code size support for the JIT compiler and fixing
541 some warnings.
542
543 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
544 not work when the shorter character appeared right at the end of the
545 subject string.
546
547 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
548 systems.
549
550 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
551 output it when the /M option is used in pcretest.
552
553 25. The CheckMan script was not being included in the distribution. Also, added
554 an explicit "perl" to run Perl scripts from the PrepareRelease script
555 because this is reportedly needed in Windows.
556
557 26. If study data was being save in a file and studying had not found a set of
558 "starts with" bytes for the pattern, the data written to the file (though
559 never used) was taken from uninitialized memory and so caused valgrind to
560 complain.
561
562 27. Updated RunTest.bat as provided by Sheri Pierce.
563
564 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
565
566 29. Computation of memory usage for the table of capturing group names was
567 giving an unnecessarily large value.
568
569
570 Version 8.20 21-Oct-2011
571 ------------------------
572
573 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
574 a POSIX class. After further experiments with Perl, which convinced me that
575 Perl has bugs and confusions, a closing square bracket is no longer allowed
576 in a POSIX name. This bug also affected patterns with classes that started
577 with full stops.
578
579 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
580 captured substring, but while checking the failing first alternative,
581 substring 1 is temporarily captured. If the output vector supplied to
582 pcre_exec() was not big enough for this capture, the yield of the function
583 was still zero ("insufficient space for captured substrings"). This cannot
584 be totally fixed without adding another stack variable, which seems a lot
585 of expense for a edge case. However, I have improved the situation in cases
586 such as /(a)(b)x|abc/ matched against "abc", where the return code
587 indicates that fewer than the maximum number of slots in the ovector have
588 been set.
589
590 3. Related to (2) above: when there are more back references in a pattern than
591 slots in the output vector, pcre_exec() uses temporary memory during
592 matching, and copies in the captures as far as possible afterwards. It was
593 using the entire output vector, but this conflicts with the specification
594 that only 2/3 is used for passing back captured substrings. Now it uses
595 only the first 2/3, for compatibility. This is, of course, another edge
596 case.
597
598 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
599 main code base, and can be used by building with --enable-jit. When this is
600 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
601 runtime --no-jit option is given.
602
603 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
604 ovector, the return from the function was zero, implying that there were
605 other matches that did not fit. The correct "exactly full" value is now
606 returned.
607
608 6. If a subpattern that was called recursively or as a subroutine contained
609 (*PRUNE) or any other control that caused it to give a non-standard return,
610 invalid errors such as "Error -26 (nested recursion at the same subject
611 position)" or even infinite loops could occur.
612
613 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
614 computing the minimum length on reaching *ACCEPT, and so ended up with the
615 wrong value of 1 rather than 0. Further investigation indicates that
616 computing a minimum subject length in the presence of *ACCEPT is difficult
617 (think back references, subroutine calls), and so I have changed the code
618 so that no minimum is registered for a pattern that contains *ACCEPT.
619
620 8. If (*THEN) was present in the first (true) branch of a conditional group,
621 it was not handled as intended. [But see 16 below.]
622
623 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
624 Sheri Pierce.
625
626 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
627 the first byte in a match must be "a".
628
629 11. Change 17 for 8.13 increased the recursion depth for patterns like
630 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
631 pattern contains any instances of (*THEN). If it does not, the old
632 optimizations are restored. It would be nice to do this on a per-group
633 basis, but at the moment that is not feasible.
634
635 12. In some environments, the output of pcretest -C is CRLF terminated. This
636 broke RunTest's code that checks for the link size. A single white space
637 character after the value is now allowed for.
638
639 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
640 For "fr", it uses the Windows-specific input and output files.
641
642 14. If (*THEN) appeared in a group that was called recursively or as a
643 subroutine, it did not work as intended. [But see next item.]
644
645 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
646 pattern fragments (but not containing any | characters). If A and B are
647 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
648 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
649 D. In other words, Perl considers parentheses that do not contain any |
650 characters to be part of a surrounding alternative, whereas PCRE was
651 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
652 differently. PCRE now behaves in the same way as Perl, except in the case
653 of subroutine/recursion calls such as (?1) which have in any case always
654 been different (but PCRE had them first :-).
655
656 16. Related to 15 above: Perl does not treat the | in a conditional group as
657 creating alternatives. Such a group is treated in the same way as an
658 ordinary group without any | characters when processing (*THEN). PCRE has
659 been changed to match Perl's behaviour.
660
661 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
662 RunGrepTest script failed.
663
664 18. Change 22 for version 13 caused atomic groups to use more stack. This is
665 inevitable for groups that contain captures, but it can lead to a lot of
666 stack use in large patterns. The old behaviour has been restored for atomic
667 groups that do not contain any capturing parentheses.
668
669 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
670 suppress the check for a minimum subject length at run time. (If it was
671 given to pcre_exec() or pcre_dfa_exec() it did work.)
672
673 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
674 fail to work when decoding hex characters in data strings in EBCDIC
675 environments.
676
677 21. It appears that in at least one Mac OS environment, the isxdigit() function
678 is implemented as a macro that evaluates to its argument more than once,
679 contravening the C 90 Standard (I haven't checked a later standard). There
680 was an instance in pcretest which caused it to go wrong when processing
681 \x{...} escapes in subject strings. The has been rewritten to avoid using
682 things like p++ in the argument of isxdigit().
683
684
685 Version 8.13 16-Aug-2011
686 ------------------------
687
688 1. The Unicode data tables have been updated to Unicode 6.0.0.
689
690 2. Two minor typos in pcre_internal.h have been fixed.
691
692 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
693 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
694 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
695
696 4. There were a number of related bugs in the code for matching backrefences
697 caselessly in UTF-8 mode when codes for the characters concerned were
698 different numbers of bytes. For example, U+023A and U+2C65 are an upper
699 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
700 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
701 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
702 2-byte code at the end of the subject (it thought there wasn't enough data
703 left).
704
705 5. Comprehensive information about what went wrong is now returned by
706 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
707 as the output vector has at least 2 elements. The offset of the start of
708 the failing character and a reason code are placed in the vector.
709
710 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
711 now returned is for the first byte of the failing character, instead of the
712 last byte inspected. This is an incompatible change, but I hope it is small
713 enough not to be a problem. It makes the returned offset consistent with
714 pcre_exec() and pcre_dfa_exec().
715
716 7. pcretest now gives a text phrase as well as the error number when
717 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
718 failure, the offset and reason code are output.
719
720 8. When \R was used with a maximizing quantifier it failed to skip backwards
721 over a \r\n pair if the subsequent match failed. Instead, it just skipped
722 back over a single character (\n). This seems wrong (because it treated the
723 two characters as a single entity when going forwards), conflicts with the
724 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
725 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
726 has been changed.
727
728 9. Some internal refactoring has changed the processing so that the handling
729 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
730 time (the PCRE_DOTALL option was changed this way some time ago: version
731 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
732 which was always a bit of a fudge. It also means that there is one less
733 argument for the match() function, which reduces its stack requirements
734 slightly. This change also fixes an incompatibility with Perl: the pattern
735 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
736
737 10. More internal refactoring has drastically reduced the number of recursive
738 calls to match() for possessively repeated groups such as (abc)++ when
739 using pcre_exec().
740
741 11. While implementing 10, a number of bugs in the handling of groups were
742 discovered and fixed:
743
744 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
745 (a|)*(?1) gave a compile-time internal error.
746 ((a|)+)+ did not notice that the outer group could match an empty string.
747 (^a|^)+ was not marked as anchored.
748 (.*a|.*)+ was not marked as matching at start or after a newline.
749
750 12. Yet more internal refactoring has removed another argument from the match()
751 function. Special calls to this function are now indicated by setting a
752 value in a variable in the "match data" data block.
753
754 13. Be more explicit in pcre_study() instead of relying on "default" for
755 opcodes that mean there is no starting character; this means that when new
756 ones are added and accidentally left out of pcre_study(), testing should
757 pick them up.
758
759 14. The -s option of pcretest has been documented for ages as being an old
760 synonym of -m (show memory usage). I have changed it to mean "force study
761 for every regex", that is, assume /S for every regex. This is similar to -i
762 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
763 using it. It makes it easier to run collections of tests with and without
764 study enabled, and thereby test pcre_study() more easily. All the standard
765 tests are now run with and without -s (but some patterns can be marked as
766 "never study" - see 20 below).
767
768 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
769 restoration of the capturing data to the outer values was not happening
770 correctly.
771
772 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
773 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
774 pattern had matched an empty string, and so incorrectly returned a no
775 match.
776
777 17. There was optimizing code for the last branch of non-capturing parentheses,
778 and also for the obeyed branch of a conditional subexpression, which used
779 tail recursion to cut down on stack usage. Unfortunately, now that there is
780 the possibility of (*THEN) occurring in these branches, tail recursion is
781 no longer possible because the return has to be checked for (*THEN). These
782 two optimizations have therefore been removed. [But see 8.20/11 above.]
783
784 18. If a pattern containing \R was studied, it was assumed that \R always
785 matched two bytes, thus causing the minimum subject length to be
786 incorrectly computed because \R can also match just one byte.
787
788 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
789 was incorrectly computed.
790
791 20. If /S is present twice on a test pattern in pcretest input, it now
792 *disables* studying, thereby overriding the use of -s on the command line
793 (see 14 above). This is necessary for one or two tests to keep the output
794 identical in both cases.
795
796 21. When (*ACCEPT) was used in an assertion that matched an empty string and
797 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
798
799 22. When an atomic group that contained a capturing parenthesis was
800 successfully matched, but the branch in which it appeared failed, the
801 capturing was not being forgotten if a higher numbered group was later
802 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
803 group 1 to "a", when in fact it should be unset. This applied to multi-
804 branched capturing and non-capturing groups, repeated or not, and also to
805 positive assertions (capturing in negative assertions does not happen
806 in PCRE) and also to nested atomic groups.
807
808 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
809 subject after a captured substring, to make it easier to tell which of a
810 number of identical substrings has been captured.
811
812 24. The way atomic groups are processed by pcre_exec() has been changed so that
813 if they are repeated, backtracking one repetition now resets captured
814 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
815 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
816 "aaa". Previously, it would have been "a". As part of this code
817 refactoring, the way recursive calls are handled has also been changed.
818
819 25. If an assertion condition captured any substrings, they were not passed
820 back unless some other capturing happened later. For example, if
821 (?(?=(a))a) was matched against "a", no capturing was returned.
822
823 26. When studying a pattern that contained subroutine calls or assertions,
824 the code for finding the minimum length of a possible match was handling
825 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
826 group 1 called group 2 while simultaneously a separate group 2 called group
827 1). A stack overflow occurred in this case. I have fixed this by limiting
828 the recursion depth to 10.
829
830 27. Updated RunTest.bat in the distribution to the version supplied by Tom
831 Fortmann. This supports explicit test numbers on the command line, and has
832 argument validation and error reporting.
833
834 28. An instance of \X with an unlimited repeat could fail if at any point the
835 first character it looked at was a mark character.
836
837 29. Some minor code refactoring concerning Unicode properties and scripts
838 should reduce the stack requirement of match() slightly.
839
840 30. Added the '=' option to pcretest to check the setting of unused capturing
841 slots at the end of the pattern, which are documented as being -1, but are
842 not included in the return count.
843
844 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
845 compiled something random. Now it gives a compile-time error (as does
846 Perl).
847
848 32. A *MARK encountered during the processing of a positive assertion is now
849 recorded and passed back (compatible with Perl).
850
851 33. If --only-matching or --colour was set on a pcregrep call whose pattern
852 had alternative anchored branches, the search for a second match in a line
853 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
854 matched the line "0102" twice. The same bug affected patterns that started
855 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
856 twice.
857
858 34. Previously, PCRE did not allow quantification of assertions. However, Perl
859 does, and because of capturing effects, quantifying parenthesized
860 assertions may at times be useful. Quantifiers are now allowed for
861 parenthesized assertions.
862
863 35. A minor code tidy in pcre_compile() when checking options for \R usage.
864
865 36. \g was being checked for fancy things in a character class, when it should
866 just be a literal "g".
867
868 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
869 appearance of a nested POSIX class supersedes an apparent external class.
870 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
871 unescaped square brackets may also appear as part of class names. For
872 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
873 more like Perl. (But see 8.20/1 above.)
874
875 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
876 was because it thought it was \N{name}, which is not supported).
877
878 39. Add minix to OS list not supporting the -S option in pcretest.
879
880 40. PCRE tries to detect cases of infinite recursion at compile time, but it
881 cannot analyze patterns in sufficient detail to catch mutual recursions
882 such as ((?1))((?2)). There is now a runtime test that gives an error if a
883 subgroup is called recursively as a subpattern for a second time at the
884 same position in the subject string. In previous releases this might have
885 been caught by the recursion limit, or it might have run out of stack.
886
887 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
888 happen only once. PCRE was, however incorrectly giving a compile time error
889 "recursive call could loop indefinitely" because it cannot analyze the
890 pattern in sufficient detail. The compile time test no longer happens when
891 PCRE is compiling a conditional subpattern, but actual runaway loops are
892 now caught at runtime (see 40 above).
893
894 42. It seems that Perl allows any characters other than a closing parenthesis
895 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
896 has been changed to be the same.
897
898 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
899 as not to get warnings when autogen.sh is called. Also changed
900 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
901
902 44. To help people who use pcregrep to scan files containing exceedingly long
903 lines, the following changes have been made:
904
905 (a) The default value of the buffer size parameter has been increased from
906 8K to 20K. (The actual buffer used is three times this size.)
907
908 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
909 PCRE is built.
910
911 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
912 to be set at run time.
913
914 (d) Numerical values in pcregrep options can be followed by K or M, for
915 example --buffer-size=50K.
916
917 (e) If a line being scanned overflows pcregrep's buffer, an error is now
918 given and the return code is set to 2.
919
920 45. Add a pointer to the latest mark to the callout data block.
921
922 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
923 partial match of an empty string instead of no match. This was specific to
924 the use of ".".
925
926 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
927 complete match instead of a partial match. This bug was dependent on both
928 the PCRE_UTF8 and PCRE_DOTALL options being set.
929
930 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
931 starting byte set, because \b was not being ignored.
932
933
934 Version 8.12 15-Jan-2011
935 ------------------------
936
937 1. Fixed some typos in the markup of the man pages, and wrote a script that
938 checks for such things as part of the documentation building process.
939
940 2. On a big-endian 64-bit system, pcregrep did not correctly process the
941 --match-limit and --recursion-limit options (added for 8.11). In
942 particular, this made one of the standard tests fail. (The integer value
943 went into the wrong half of a long int.)
944
945 3. If the --colour option was given to pcregrep with -v (invert match), it
946 did strange things, either producing crazy output, or crashing. It should,
947 of course, ignore a request for colour when reporting lines that do not
948 match.
949
950 4. Another pcregrep bug caused similar problems if --colour was specified with
951 -M (multiline) and the pattern match finished with a line ending.
952
953 5. In pcregrep, when a pattern that ended with a literal newline sequence was
954 matched in multiline mode, the following line was shown as part of the
955 match. This seems wrong, so I have changed it.
956
957 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
958 the check for further matches in the same line (so they could be coloured)
959 to overrun the end of the current line. If another match was found, it was
960 incorrectly shown (and then shown again when found in the next line).
961
962 7. If pcregrep was compiled under Windows, there was a reference to the
963 function pcregrep_exit() before it was defined. I am assuming this was
964 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
965 reported by a user. I've moved the definition above the reference.
966
967
968 Version 8.11 10-Dec-2010
969 ------------------------
970
971 1. (*THEN) was not working properly if there were untried alternatives prior
972 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
973 backtracked to try for "b" instead of moving to the next alternative branch
974 at the same level (in this case, to look for "c"). The Perl documentation
975 is clear that when (*THEN) is backtracked onto, it goes to the "next
976 alternative in the innermost enclosing group".
977
978 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
979 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
980 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
981 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
982 (*THEN).
983
984 3. If \s appeared in a character class, it removed the VT character from
985 the class, even if it had been included by some previous item, for example
986 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
987 of \s, but is part of the POSIX "space" class.)
988
989 4. A partial match never returns an empty string (because you can always
990 match an empty string at the end of the subject); however the checking for
991 an empty string was starting at the "start of match" point. This has been
992 changed to the "earliest inspected character" point, because the returned
993 data for a partial match starts at this character. This means that, for
994 example, /(?<=abc)def/ gives a partial match for the subject "abc"
995 (previously it gave "no match").
996
997 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
998 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
999 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1000 has an implication that the given string is incomplete (because a partial
1001 match is preferred over a full match). For this reason, these items now
1002 give a partial match in this situation. [Aside: previously, the one case
1003 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1004 match rather than a full match, which was wrong by the old rules, but is
1005 now correct.]
1006
1007 6. There was a bug in the handling of #-introduced comments, recognized when
1008 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1009 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1010 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1011 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1012 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1013 places in pcre_compile().
1014
1015 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1016 comments when looking ahead for named forward references to subpatterns,
1017 the only newline sequence it recognized was NL. It now handles newlines
1018 according to the set newline convention.
1019
1020 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1021 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1022 cater for a lack of strerror(). These oversights have been fixed.
1023
1024 9. Added --match-limit and --recursion-limit to pcregrep.
1025
1026 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1027
1028 11. When the -o option was used, pcregrep was setting a return code of 1, even
1029 when matches were found, and --line-buffered was not being honoured.
1030
1031 12. Added an optional parentheses number to the -o and --only-matching options
1032 of pcregrep.
1033
1034 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1035 can match an empty string. The code to do it in pcretest and pcredemo
1036 needed fixing:
1037
1038 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1039 only one byte after an empty string match just before CRLF (this case
1040 just got forgotten; "any" and "anycrlf" were OK).
1041
1042 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1043 mode when an empty string match preceded an ASCII character followed by
1044 a non-ASCII character. (The code for advancing by one character rather
1045 than one byte was nonsense.)
1046
1047 (c) The pcredemo.c sample program did not have any code at all to handle
1048 the cases when CRLF is a valid newline sequence.
1049
1050 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1051 as a starting offset was within the subject string. There is now a new
1052 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1053 negative or greater than the length of the string. In order to test this,
1054 pcretest is extended to allow the setting of negative starting offsets.
1055
1056 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1057 starting offset points to the beginning of a UTF-8 character was
1058 unnecessarily clumsy. I tidied it up.
1059
1060 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1061 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1062
1063 17. Nobody had reported that the --include_dir option, which was added in
1064 release 7.7 should have been called --include-dir (hyphen, not underscore)
1065 for compatibility with GNU grep. I have changed it to --include-dir, but
1066 left --include_dir as an undocumented synonym, and the same for
1067 --exclude-dir, though that is not available in GNU grep, at least as of
1068 release 2.5.4.
1069
1070 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1071 characters from a string of bytes) have been redefined so as not to use
1072 loops, in order to improve performance in some environments. At the same
1073 time, I abstracted some of the common code into auxiliary macros to save
1074 repetition (this should not affect the compiled code).
1075
1076 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1077 compile-time error is now given if \c is not followed by an ASCII
1078 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1079 different, and any byte value is allowed.)
1080
1081 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1082 START_OPTIMIZE option, which is now allowed at compile time - but just
1083 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1084 to pcregrep and other applications that have no direct access to PCRE
1085 options. The new /Y option in pcretest sets this option when calling
1086 pcre_compile().
1087
1088 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1089 back references. Groups containing recursive back references were forced to
1090 be atomic by that change, but in the case of named groups, the amount of
1091 memory required was incorrectly computed, leading to "Failed: internal
1092 error: code overflow". This has been fixed.
1093
1094 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1095 pcretest.c, to avoid build problems in some Borland environments.
1096
1097
1098 Version 8.10 25-Jun-2010
1099 ------------------------
1100
1101 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1102 THEN.
1103
1104 2. (*ACCEPT) was not working when inside an atomic group.
1105
1106 3. Inside a character class, \B is treated as a literal by default, but
1107 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1108 causes the error). The code is unchanged, but I tidied the documentation.
1109
1110 4. Inside a character class, PCRE always treated \R and \X as literals,
1111 whereas Perl faults them if its -w option is set. I have changed PCRE so
1112 that it faults them when PCRE_EXTRA is set.
1113
1114 5. Added support for \N, which always matches any character other than
1115 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1116
1117 6. When compiling pcregrep with newer versions of gcc which may have
1118 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1119 declared with attribute warn_unused_result" were given. Just casting the
1120 result to (void) does not stop the warnings; a more elaborate fudge is
1121 needed. I've used a macro to implement this.
1122
1123 7. Minor change to pcretest.c to avoid a compiler warning.
1124
1125 8. Added four artifical Unicode properties to help with an option to make
1126 \s etc use properties (see next item). The new properties are: Xan
1127 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1128
1129 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1130 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1131 this option. Modified pcretest to add /W to test this facility. Added
1132 REG_UCP to make it available via the POSIX interface.
1133
1134 10. Added --line-buffered to pcregrep.
1135
1136 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1137 studied, and the match started with a letter with a code point greater than
1138 127 whose first byte was different to the first byte of the other case of
1139 the letter, the other case of this starting letter was not recognized
1140 (#976).
1141
1142 12. If a pattern that was studied started with a repeated Unicode property
1143 test, for example, \p{Nd}+, there was the theoretical possibility of
1144 setting up an incorrect bitmap of starting bytes, but fortunately it could
1145 not have actually happened in practice until change 8 above was made (it
1146 added property types that matched character-matching opcodes).
1147
1148 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1149 possible starting bytes for non-anchored patterns.
1150
1151 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1152 \R, and also a number of cases that involve Unicode properties, both
1153 explicit and implicit when PCRE_UCP is set.
1154
1155 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1156 input, it could crash or give wrong results if characters with values
1157 greater than 0xc0 were present in the subject string. (Detail: it assumed
1158 UTF-8 input when processing these items.)
1159
1160 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1161 size_t is 64-bit (#991).
1162
1163 17. Added a check for running out of memory when PCRE is compiled with
1164 --disable-stack-for-recursion (#990).
1165
1166 18. If the last data line in a file for pcretest does not have a newline on
1167 the end, a newline was missing in the output.
1168
1169 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1170 less than 128) in its various bitmaps. However, there is a facility for
1171 generating tables according to the current locale when PCRE is compiled. It
1172 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1173 characters, are recognized by isspace() and therefore were getting set in
1174 these tables, and indeed these tables seem to approximate to ISO 8859. This
1175 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1176 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1177 which of course cannot start UTF-8 characters. I have changed the code so
1178 that only real ASCII characters (less than 128) and the correct starting
1179 bytes for UTF-8 encodings are set for characters greater than 127 when in
1180 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1181 altogether.)
1182
1183 20. Added the /T option to pcretest so as to be able to run tests with non-
1184 standard character tables, thus making it possible to include the tests
1185 used for 19 above in the standard set of tests.
1186
1187 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1188 reference to a subpattern the other side of a comment that contains an
1189 opening parenthesis caused either an internal compiling error, or a
1190 reference to the wrong subpattern.
1191
1192
1193 Version 8.02 19-Mar-2010
1194 ------------------------
1195
1196 1. The Unicode data tables have been updated to Unicode 5.2.0.
1197
1198 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1199 configured.
1200
1201 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1202 original author of that file, following a query about its status.
1203
1204 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1205 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1206
1207 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1208 quantifier applied to a forward-referencing subroutine call, could compile
1209 incorrect code or give the error "internal error: previously-checked
1210 referenced subpattern not found".
1211
1212 6. Both MS Visual Studio and Symbian OS have problems with initializing
1213 variables to point to external functions. For these systems, therefore,
1214 pcre_malloc etc. are now initialized to local functions that call the
1215 relevant global functions.
1216
1217 7. There were two entries missing in the vectors called coptable and poptable
1218 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1219 I've fixed the data, and added a kludgy way of testing at compile time that
1220 the lengths are correct (equal to the number of opcodes).
1221
1222 8. Following on from 7, I added a similar kludge to check the length of the
1223 eint vector in pcreposix.c.
1224
1225 9. Error texts for pcre_compile() are held as one long string to avoid too
1226 much relocation at load time. To find a text, the string is searched,
1227 counting zeros. There was no check for running off the end of the string,
1228 which could happen if a new error number was added without updating the
1229 string.
1230
1231 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1232
1233 11. \K was not working if it appeared in an atomic group or in a group that
1234 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1235 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1236 the assertion is positive, but not if it is negative.
1237
1238 12. Change 11 fortuitously reduced the size of the stack frame used in the
1239 "match()" function of pcre_exec.c by one pointer. Forthcoming
1240 implementation of support for (*MARK) will need an extra pointer on the
1241 stack; I have reserved it now, so that the stack frame size does not
1242 decrease.
1243
1244 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1245 item in branch that calls a recursion is a subroutine call - as in the
1246 second branch in the above example - was incorrectly given the compile-
1247 time error "recursive call could loop indefinitely" because pcre_compile()
1248 was not correctly checking the subroutine for matching a non-empty string.
1249
1250 14. The checks for overrunning compiling workspace could trigger after an
1251 overrun had occurred. This is a "should never occur" error, but it can be
1252 triggered by pathological patterns such as hundreds of nested parentheses.
1253 The checks now trigger 100 bytes before the end of the workspace.
1254
1255 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1256
1257
1258 Version 8.01 19-Jan-2010
1259 ------------------------
1260
1261 1. If a pattern contained a conditional subpattern with only one branch (in
1262 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1263 computed the wrong minimum data length (which is of course zero for such
1264 subpatterns). This could cause incorrect "no match" results.
1265
1266 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1267 the pattern is reset in the first branch, pcre_compile() failed with
1268 "internal error: code overflow at offset...". This happened only when
1269 the reset was to the original external option setting. (An optimization
1270 abstracts leading options settings into an external setting, which was the
1271 cause of this.)
1272
1273 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1274 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1275 assertion pattern did not match (meaning that the assertion was true), it
1276 was incorrectly treated as false if the SKIP had been reached during the
1277 matching. This also applied to assertions used as conditions.
1278
1279 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1280 assertion subpattern, including such a pattern used as a condition,
1281 unpredictable results occurred, instead of the error return
1282 PCRE_ERROR_DFA_UITEM.
1283
1284 5. The C++ GlobalReplace function was not working like Perl for the special
1285 situation when an empty string is matched. It now does the fancy magic
1286 stuff that is necessary.
1287
1288 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1289 removed. (These were left over from very, very early versions of PCRE.)
1290
1291 7. Some cosmetic changes to the code to make life easier when compiling it
1292 as part of something else:
1293
1294 (a) Change DEBUG to PCRE_DEBUG.
1295
1296 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1297 called "current" as "current_branch", to prevent a collision with the
1298 Linux macro when compiled as a kernel module.
1299
1300 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1301 prevent a collision with the Linux macro when compiled as a kernel
1302 module.
1303
1304 8. In pcre_compile() there are some checks for integer overflows that used to
1305 cast potentially large values to (double). This has been changed to that
1306 when building, a check for int64_t is made, and if it is found, it is used
1307 instead, thus avoiding the use of floating point arithmetic. (There is no
1308 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1309 double.
1310
1311 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1312 2005 (difference between two addresses compared to an unsigned value).
1313
1314 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1315 custom one, because of the following reported problem in Windows:
1316
1317 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1318 under Win32.
1319 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1320 therefore missing the function definition.
1321 - The compiler thus generates a "C" signature for the test function.
1322 - The linker fails to find the "C" function.
1323 - PCRE fails to configure if asked to do so against libbz2.
1324
1325 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1326 messages were output:
1327
1328 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1329 rerunning libtoolize, to keep the correct libtool macros in-tree.
1330 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1331
1332 I have done both of these things.
1333
1334 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1335 most of the time, it *can* run out if it is given a pattern that contains a
1336 runaway infinite recursion. I updated the discussion in the pcrestack man
1337 page.
1338
1339 13. Now that we have gone to the x.xx style of version numbers, the minor
1340 version may start with zero. Using 08 or 09 is a bad idea because users
1341 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1342 interpreted as invalid octal numbers. I've updated the previous comment in
1343 configure.ac, and also added a check that gives an error if 08 or 09 are
1344 used.
1345
1346 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1347 causing partial matching to fail when the end of the subject matched \W
1348 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1349
1350 15. There were some discrepancies between the declarations in pcre_internal.h
1351 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1352 their definitions. The declarations used "const uschar *" and the
1353 definitions used USPTR. Even though USPTR is normally defined as "const
1354 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1355 reported that: "This difference in casting confuses some C++ compilers, for
1356 example, SunCC recognizes above declarations as different functions and
1357 generates broken code for hbpcre." I have changed the declarations to use
1358 USPTR.
1359
1360 16. GNU libtool is named differently on some systems. The autogen.sh script now
1361 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1362 (FreeBSD).
1363
1364 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1365 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1366 comment: "Figure out how to create a longlong from a string: strtoll and
1367 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1368 instance, but it only takes 2 args instead of 3!"
1369
1370 18. A subtle bug concerned with back references has been fixed by a change of
1371 specification, with a corresponding code fix. A pattern such as
1372 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1373 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1374 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1375 same bug. Such groups have to be quantified to be useful, or contained
1376 inside another quantified group. (If there's no repetition, the reference
1377 can never match.) The problem arises because, having left the group and
1378 moved on to the rest of the pattern, a later failure that backtracks into
1379 the group uses the captured value from the final iteration of the group
1380 rather than the correct earlier one. I have fixed this in PCRE by forcing
1381 any group that contains a reference to itself to be an atomic group; that
1382 is, there cannot be any backtracking into it once it has completed. This is
1383 similar to recursive and subroutine calls.
1384
1385
1386 Version 8.00 19-Oct-09
1387 ----------------------
1388
1389 1. The table for translating pcre_compile() error codes into POSIX error codes
1390 was out-of-date, and there was no check on the pcre_compile() error code
1391 being within the table. This could lead to an OK return being given in
1392 error.
1393
1394 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1395 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1396 in a Windows environment.
1397
1398 3. The pcregrep --count option prints the count for each file even when it is
1399 zero, as does GNU grep. However, pcregrep was also printing all files when
1400 --files-with-matches was added. Now, when both options are given, it prints
1401 counts only for those files that have at least one match. (GNU grep just
1402 prints the file name in this circumstance, but including the count seems
1403 more useful - otherwise, why use --count?) Also ensured that the
1404 combination -clh just lists non-zero counts, with no names.
1405
1406 4. The long form of the pcregrep -F option was incorrectly implemented as
1407 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1408 but it seems right to fix it, and I didn't think it was worth preserving
1409 the old behaviour.
1410
1411 5. The command line items --regex=pattern and --regexp=pattern were not
1412 recognized by pcregrep, which required --regex pattern or --regexp pattern
1413 (with a space rather than an '='). The man page documented the '=' forms,
1414 which are compatible with GNU grep; these now work.
1415
1416 6. No libpcreposix.pc file was created for pkg-config; there was just
1417 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1418
1419 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1420 when UCP support is not needed, by modifying the Python script that
1421 generates it from Unicode data files. This should not matter if the module
1422 is correctly used as a library, but I received one complaint about 50K of
1423 unwanted data. My guess is that the person linked everything into his
1424 program rather than using a library. Anyway, it does no harm.
1425
1426 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1427 was a minimum greater than 1 for a wide character in a possessive
1428 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1429 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1430 character. Chaos in the form of incorrect output or a compiling loop could
1431 result.
1432
1433 9. The restrictions on what a pattern can contain when partial matching is
1434 requested for pcre_exec() have been removed. All patterns can now be
1435 partially matched by this function. In addition, if there are at least two
1436 slots in the offset vector, the offset of the earliest inspected character
1437 for the match and the offset of the end of the subject are set in them when
1438 PCRE_ERROR_PARTIAL is returned.
1439
1440 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1441 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1442 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1443 and may be more useful for multi-segment matching.
1444
1445 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1446 used to be given if ever the end of the subject was reached; now it is
1447 given only if matching could not proceed because another character was
1448 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1449 string "Z", which now yields "no match" instead of "partial match". In the
1450 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1451 final character ended with (*FAIL).
1452
1453 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1454 if the pattern had a "must contain" character that was already found in the
1455 earlier partial match, unless partial matching was again requested. For
1456 example, with the pattern /dog.(body)?/, the "must contain" character is
1457 "g". If the first part-match was for the string "dog", restarting with
1458 "sbody" failed. This bug has been fixed.
1459
1460 13. The string returned by pcre_dfa_exec() after a partial match has been
1461 changed so that it starts at the first inspected character rather than the
1462 first character of the match. This makes a difference only if the pattern
1463 starts with a lookbehind assertion or \b or \B (\K is not supported by
1464 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1465 matching functions compatible, and I think it's the right thing to do.
1466
1467 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1468 so that the demonstration program is easily available in environments where
1469 PCRE has not been installed from source.
1470
1471 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1472 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1473 library.
1474
1475 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1476 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1477 is not the first non-POSIX option to be added. Clearly some people find
1478 these options useful.
1479
1480 17. If a caller to the POSIX matching function regexec() passes a non-zero
1481 value for nmatch with a NULL value for pmatch, the value of
1482 nmatch is forced to zero.
1483
1484 18. RunGrepTest did not have a test for the availability of the -u option of
1485 the diff command, as RunTest does. It now checks in the same way as
1486 RunTest, and also checks for the -b option.
1487
1488 19. If an odd number of negated classes containing just a single character
1489 interposed, within parentheses, between a forward reference to a named
1490 subpattern and the definition of the subpattern, compilation crashed with
1491 an internal error, complaining that it could not find the referenced
1492 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1493 [The bug was that it was starting one character too far in when skipping
1494 over the character class, thus treating the ] as data rather than
1495 terminating the class. This meant it could skip too much.]
1496
1497 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1498 /g option in pcretest when the pattern contains \K, which makes it possible
1499 to have an empty string match not at the start, even when the pattern is
1500 anchored. Updated pcretest and pcredemo to use this option.
1501
1502 21. If the maximum number of capturing subpatterns in a recursion was greater
1503 than the maximum at the outer level, the higher number was returned, but
1504 with unset values at the outer level. The correct (outer level) value is
1505 now given.
1506
1507 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1508 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1509 make it do so. The string so far is captured, making this feature
1510 compatible with Perl.
1511
1512 23. The tests have been re-organized, adding tests 11 and 12, to make it
1513 possible to check the Perl 5.10 features against Perl 5.10.
1514
1515 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1516 pattern matches a fixed length string. PCRE did not allow this; now it
1517 does. Neither allows recursion.
1518
1519 25. I finally figured out how to implement a request to provide the minimum
1520 length of subject string that was needed in order to match a given pattern.
1521 (It was back references and recursion that I had previously got hung up
1522 on.) This code has now been added to pcre_study(); it finds a lower bound
1523 to the length of subject needed. It is not necessarily the greatest lower
1524 bound, but using it to avoid searching strings that are too short does give
1525 some useful speed-ups. The value is available to calling programs via
1526 pcre_fullinfo().
1527
1528 26. While implementing 25, I discovered to my embarrassment that pcretest had
1529 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1530 study optimizations had never been tested with that matching function.
1531 Oops. What is worse, even when it was passed study data, there was a bug in
1532 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1533 were also very few tests of studied patterns with pcre_dfa_exec().
1534
1535 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1536 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1537 on the other side of the coin, they are no longer allowed to have different
1538 names, because these cannot be distinguished in PCRE, and this has caused
1539 confusion. (This is a difference from Perl.)
1540
1541 28. When duplicate subpattern names are present (necessarily with different
1542 numbers, as required by 27 above), and a test is made by name in a
1543 conditional pattern, either for a subpattern having been matched, or for
1544 recursion in such a pattern, all the associated numbered subpatterns are
1545 tested, and the overall condition is true if the condition is true for any
1546 one of them. This is the way Perl works, and is also more like the way
1547 testing by number works.
1548
1549
1550 Version 7.9 11-Apr-09
1551 ---------------------
1552
1553 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1554 (pcretest), all targets were linked against these libraries. This included
1555 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1556 libraries. This caused unwanted dependencies to be created. This problem
1557 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1558 pcretest is linked with readline.
1559
1560 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1561 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1562 moved outside it again, because FALSE and TRUE are already defined in AIX,
1563 but BOOL is not.
1564
1565 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1566 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1567
1568 4. The pcregrep documentation said spaces were inserted as well as colons (or
1569 hyphens) following file names and line numbers when outputting matching
1570 lines. This is not true; no spaces are inserted. I have also clarified the
1571 wording for the --colour (or --color) option.
1572
1573 5. In pcregrep, when --colour was used with -o, the list of matching strings
1574 was not coloured; this is different to GNU grep, so I have changed it to be
1575 the same.
1576
1577 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1578 each matching line was coloured. Now it goes on to look for further matches
1579 of any of the test patterns, which is the same behaviour as GNU grep.
1580
1581 7. A pattern that could match an empty string could cause pcregrep to loop; it
1582 doesn't make sense to accept an empty string match in pcregrep, so I have
1583 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1584 seems to be how GNU grep behaves.
1585
1586 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1587 start or after a newline", because the conditional assertion was not being
1588 correctly handled. The rule now is that both the assertion and what follows
1589 in the first alternative must satisfy the test.
1590
1591 9. If auto-callout was enabled in a pattern with a conditional group whose
1592 condition was an assertion, PCRE could crash during matching, both with
1593 pcre_exec() and pcre_dfa_exec().
1594
1595 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1596 used for matching.
1597
1598 11. Unicode property support in character classes was not working for
1599 characters (bytes) greater than 127 when not in UTF-8 mode.
1600
1601 12. Added the -M command line option to pcretest.
1602
1603 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1604
1605 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1606
1607 16. Added comments and documentation about mis-use of no_arg in the C++
1608 wrapper.
1609
1610 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1611 from Martin Jerabek that uses macro names for all relevant character and
1612 string constants.
1613
1614 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1615 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1616 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1617 these, but not everybody uses configure.
1618
1619 19. A conditional group that had only one branch was not being correctly
1620 recognized as an item that could match an empty string. This meant that an
1621 enclosing group might also not be so recognized, causing infinite looping
1622 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1623 with the subject "ab", where knowledge that the repeated group can match
1624 nothing is needed in order to break the loop.
1625
1626 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1627 exec(), but without supplying a callout function, matching went wrong.
1628
1629 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1630 leak if the size of the offset vector was greater than 30. When the vector
1631 is smaller, the saved offsets during recursion go onto a local stack
1632 vector, but for larger vectors malloc() is used. It was failing to free
1633 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1634 error, in fact).
1635
1636 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1637 heapframe that is used only when UTF-8 support is enabled. This caused no
1638 problem, but was untidy.
1639
1640 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1641 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1642 included within another project.
1643
1644 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1645 slightly modified by me:
1646
1647 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1648 not building pcregrep.
1649
1650 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1651 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1652
1653 25. Forward references, both numeric and by name, in patterns that made use of
1654 duplicate group numbers, could behave incorrectly or give incorrect errors,
1655 because when scanning forward to find the reference group, PCRE was not
1656 taking into account the duplicate group numbers. A pattern such as
1657 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1658
1659 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1660 the feature of a custom pointer more persuasive (as requested by a user).
1661
1662 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1663 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1664 pre-defined.
1665
1666 28. Added support for (*UTF8) at the start of a pattern.
1667
1668 29. Arrange for flags added by the "release type" setting in CMake to be shown
1669 in the configuration summary.
1670
1671
1672 Version 7.8 05-Sep-08
1673 ---------------------
1674
1675 1. Replaced UCP searching code with optimized version as implemented for Ad
1676 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1677 stage table and inline lookup instead of a function, giving speed ups of 2
1678 to 5 times on some simple patterns that I tested. Permission was given to
1679 distribute the MultiStage2.py script that generates the tables (it's not in
1680 the tarball, but is in the Subversion repository).
1681
1682 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1683 scripts.
1684
1685 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1686 a group with a zero qualifier. The result of the study could be incorrect,
1687 or the function might crash, depending on the pattern.
1688
1689 4. Caseless matching was not working for non-ASCII characters in back
1690 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1691 It now works when Unicode Property Support is available.
1692
1693 5. In pcretest, an escape such as \x{de} in the data was always generating
1694 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1695 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1696 truncation.
1697
1698 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1699
1700 7. Added two (int) casts to pcregrep when printing the difference of two
1701 pointers, in case they are 64-bit values.
1702
1703 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1704 test 2 if it fails.
1705
1706 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1707 and a #define of that name to empty if it is not externally set. This is to
1708 allow users of MSVC to set it if necessary.
1709
1710 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1711 the convenience functions in the pcre_get.c source file.
1712
1713 11. An option change at the start of a pattern that had top-level alternatives
1714 could cause overwriting and/or a crash. This command provoked a crash in
1715 some environments:
1716
1717 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1718
1719 This potential security problem was recorded as CVE-2008-2371.
1720
1721 12. For a pattern where the match had to start at the beginning or immediately
1722 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1723 pcre_dfa_exec() could read past the end of the passed subject if there was
1724 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1725 pcretest so that it places the subject at the end of its malloc-ed buffer.
1726
1727 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1728 exec() might read past the end of the data buffer in UTF-8 mode.
1729
1730 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1731 the data contained the byte 0x85 as part of a UTF-8 character within its
1732 first line. This applied both to normal and DFA matching.
1733
1734 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1735 /^[^d]*?$/8 failed to match "abc".
1736
1737 16. Added a missing copyright notice to pcrecpp_internal.h.
1738
1739 17. Make it more clear in the documentation that values returned from
1740 pcre_exec() in ovector are byte offsets, not character counts.
1741
1742 18. Tidied a few places to stop certain compilers from issuing warnings.
1743
1744 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1745 supplied by Stefan Weber. I made a further small update for 7.8 because
1746 there is a change of source arrangements: the pcre_searchfuncs.c module is
1747 replaced by pcre_ucd.c.
1748
1749
1750 Version 7.7 07-May-08
1751 ---------------------
1752
1753 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1754 a string to a long long, pretend we don't even have a long long." This is
1755 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1756
1757 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1758 pre-7.6 versions, which defined a global no_arg variable instead of putting
1759 it in the RE class. (See also #8 below.)
1760
1761 3. Remove a line of dead code, identified by coverity and reported by Nuno
1762 Lopes.
1763
1764 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1765
1766 (1) The include/exclude patterns were being applied to the whole pathnames
1767 of files, instead of just to the final components.
1768
1769 (2) If there was more than one level of directory, the subdirectories were
1770 skipped unless they satisfied the include/exclude conditions. This is
1771 inconsistent with GNU grep (and could even be seen as contrary to the
1772 pcregrep specification - which I improved to make it absolutely clear).
1773 The action now is always to scan all levels of directory, and just
1774 apply the include/exclude patterns to regular files.
1775
1776 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1777 --exclude_dir in the tests to avoid scanning .svn directories.
1778
1779 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1780 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1781 doesn't support NULs in patterns.
1782
1783 7. Added some missing "const"s to declarations of static tables in
1784 pcre_compile.c and pcre_dfa_exec.c.
1785
1786 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1787 caused by fix #2 above. (Subsequently also a second patch to fix the
1788 first patch. And a third patch - this was a messy problem.)
1789
1790 9. Applied Craig's patch to remove the use of push_back().
1791
1792 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1793 matching function regexec().
1794
1795 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1796 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1797 references. PCRE supports relative numbers with this syntax (I don't think
1798 Oniguruma does).
1799
1800 12. Previously, a group with a zero repeat such as (...){0} was completely
1801 omitted from the compiled regex. However, this means that if the group
1802 was called as a subroutine from elsewhere in the pattern, things went wrong
1803 (an internal error was given). Such groups are now left in the compiled
1804 pattern, with a new opcode that causes them to be skipped at execution
1805 time.
1806
1807 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1808 to the way PCRE behaves:
1809
1810 (a) A lone ] character is dis-allowed (Perl treats it as data).
1811
1812 (b) A back reference to an unmatched subpattern matches an empty string
1813 (Perl fails the current match path).
1814
1815 (c) A data ] in a character class must be notated as \] because if the
1816 first data character in a class is ], it defines an empty class. (In
1817 Perl it is not possible to have an empty class.) The empty class []
1818 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1819 The negative empty class [^] matches any one character, independently
1820 of the DOTALL setting.
1821
1822 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1823 non-existent subpattern following a character class starting with ']' and
1824 containing () gave an internal compiling error instead of "reference to
1825 non-existent subpattern". Fortunately, when the pattern did exist, the
1826 compiled code was correct. (When scanning forwards to check for the
1827 existencd of the subpattern, it was treating the data ']' as terminating
1828 the class, so got the count wrong. When actually compiling, the reference
1829 was subsequently set up correctly.)
1830
1831 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1832 it was being rejected as not supported by pcre_dfa_exec(), even though
1833 other assertions are supported. I have made pcre_dfa_exec() support
1834 (*FAIL).
1835
1836 16. The implementation of 13c above involved the invention of a new opcode,
1837 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1838 cannot be changed at match time, I realized I could make a small
1839 improvement to matching performance by compiling OP_ALLANY instead of
1840 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1841 on the OP_ANY path.
1842
1843 17. Compiling pcretest on Windows with readline support failed without the
1844 following two fixes: (1) Make the unistd.h include conditional on
1845 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1846
1847 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1848 ncurses library to be included for pcretest when ReadLine support is
1849 requested, but also to allow for it to be overridden. This patch came from
1850 Daniel Bergström.
1851
1852 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1853 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1854 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1855 spotting this.
1856
1857
1858 Version 7.6 28-Jan-08
1859 ---------------------
1860
1861 1. A character class containing a very large number of characters with
1862 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1863 overflow.
1864
1865 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1866 HAVE_LONG_LONG is not defined.
1867
1868 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1869 bring them up to date and include new features. This patch includes:
1870
1871 - Fixed PH's badly added libz and libbz2 support.
1872 - Fixed a problem with static linking.
1873 - Added pcredemo. [But later removed - see 7 below.]
1874 - Fixed dftables problem and added an option.
1875 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1876 HAVE_LONG_LONG.
1877 - Added readline support for pcretest.
1878 - Added an listing of the option settings after cmake has run.
1879
1880 4. A user submitted a patch to Makefile that makes it easy to create
1881 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1882 Makefile.am that cause it to include this special target, without
1883 affecting anything else. Note that the same mingw target plus all
1884 the other distribution libraries and programs are now supported
1885 when configuring with CMake (see 6 below) instead of with
1886 Configure/Make.
1887
1888 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1889 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1890 exported in the Windows port". It has not yet been confirmed that the patch
1891 solves the problem, but it does no harm.
1892
1893 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1894 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1895 with CMake, and also correct the comment about stack recursion.
1896
1897 7. Remove the automatic building of pcredemo from the ./configure system and
1898 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1899 of a program that users should build themselves after PCRE is installed, so
1900 building it automatically is not really right. What is more, it gave
1901 trouble in some build environments.
1902
1903 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1904
1905
1906 Version 7.5 10-Jan-08
1907 ---------------------
1908
1909 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1910 values in parens when parsing an RE using the C++ wrapper."
1911
1912 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1913 Characters greater than 255 were excluded from the class instead of being
1914 included.
1915
1916 3. The same bug as (2) above applied to negated POSIX classes such as
1917 [:^space:].
1918
1919 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1920 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1921 I have changed it.
1922
1923 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1924 first named subpattern) and a construct such as (?&a) would reference the
1925 first named subpattern whose name started with "a" (in other words, the
1926 length check was missing). Both these problems are fixed. "Subpattern name
1927 expected" is now given for (?&) (a zero-length name), and this patch also
1928 makes it give the same error for \k'' (previously it complained that that
1929 was a reference to a non-existent subpattern).
1930
1931 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1932 this is right because (?- can be followed by option settings as well as by
1933 digits. I have, however, made the messages clearer.
1934
1935 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1936 than the number used in the conditional) now cause a compile-time error.
1937 This is actually not compatible with Perl, which accepts such patterns, but
1938 treats the conditional as always being FALSE (as PCRE used to), but it
1939 seems to me that giving a diagnostic is better.
1940
1941 8. Change "alphameric" to the more common word "alphanumeric" in comments
1942 and messages.
1943
1944 9. Fix two occurrences of "backslash" in comments that should have been
1945 "backspace".
1946
1947 10. Remove two redundant lines of code that can never be obeyed (their function
1948 was moved elsewhere).
1949
1950 11. The program that makes PCRE's Unicode character property table had a bug
1951 which caused it to generate incorrect table entries for sequences of
1952 characters that have the same character type, but are in different scripts.
1953 It amalgamated them into a single range, with the script of the first of
1954 them. In other words, some characters were in the wrong script. There were
1955 thirteen such cases, affecting characters in the following ranges:
1956
1957 U+002b0 - U+002c1
1958 U+0060c - U+0060d
1959 U+0061e - U+00612
1960 U+0064b - U+0065e
1961 U+0074d - U+0076d
1962 U+01800 - U+01805
1963 U+01d00 - U+01d77
1964 U+01d9b - U+01dbf
1965 U+0200b - U+0200f
1966 U+030fc - U+030fe
1967 U+03260 - U+0327f
1968 U+0fb46 - U+0fbb1
1969 U+10450 - U+1049d
1970
1971 12. The -o option (show only the matching part of a line) for pcregrep was not
1972 compatible with GNU grep in that, if there was more than one match in a
1973 line, it showed only the first of them. It now behaves in the same way as
1974 GNU grep.
1975
1976 13. If the -o and -v options were combined for pcregrep, it printed a blank
1977 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1978 does the same. The return code can be used to tell if there were any
1979 non-matching lines.
1980
1981 14. Added --file-offsets and --line-offsets to pcregrep.
1982
1983 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1984 infinitely looping recursion. The bug was that positive lookaheads were not
1985 being skipped when checking for a possible empty match (negative lookaheads
1986 and both kinds of lookbehind were skipped).
1987
1988 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1989 inclusion of <windows.h> to before rather than after the definition of
1990 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1991
1992 17. Specifying a possessive quantifier with a specific limit for a Unicode
1993 character property caused pcre_compile() to compile bad code, which led at
1994 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1995 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1996 caused the error; without that there was no problem.
1997
1998 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1999
2000 19. Added --enable-pcretest-libreadline.
2001
2002 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2003 RE::GlobalReplace(). As a result, the number of replacements returned was
2004 double what it should be. I removed one of the increments, but Craig sent a
2005 later patch that removed the other one (the right fix) and added unit tests
2006 that check the return values (which was not done before).
2007
2008 21. Several CMake things:
2009
2010 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2011 the names libpcre and libpcreposix, not just pcre and pcreposix.
2012
2013 (2) The above change means that pcretest and pcregrep are now correctly
2014 linked with the newly-built libraries, not previously installed ones.
2015
2016 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2017
2018 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2019 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2020 UTF-8 newline character). The key issue is that the pattern starts .*;
2021 this means that the match must be either at the beginning, or after a
2022 newline. The bug was in the code for advancing after a failed match and
2023 checking that the new position followed a newline. It was not taking
2024 account of UTF-8 characters correctly.
2025
2026 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2027 character classes. PCRE was not treating the sequence [:...:] as a
2028 character class unless the ... were all letters. Perl, however, seems to
2029 allow any characters between [: and :], though of course it rejects as
2030 unknown any "names" that contain non-letters, because all the known class
2031 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2032 for example, whereas PCRE did not - it did not recognize a POSIX character
2033 class. This seemed a bit dangerous, so the code has been changed to be
2034 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2035 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2036 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2037 Perl does, and where it didn't before.
2038
2039 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2040 Windows environments %n is disabled by default.
2041
2042
2043 Version 7.4 21-Sep-07
2044 ---------------------
2045
2046 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2047 means that a class such as [\s] counted as "explicit reference to CR or
2048 LF". That isn't really right - the whole point of the change was to try to
2049 help when there was an actual mention of one of the two characters. So now
2050 the change happens only if \r or \n (or a literal CR or LF) character is
2051 encountered.
2052
2053 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2054 of both had grown to the point where there were only 3 bits left.
2055 Fortunately, there was spare space in the data structure, and so I have
2056 moved the internal flags into a new 16-bit field to free up more option
2057 bits.
2058
2059 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2060 but did not set the internal JCHANGED flag - either of these is enough to
2061 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2062 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2063 start sets both bits.
2064
2065 4. Added options (at build time, compile time, exec time) to change \R from
2066 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2067
2068 5. doc/pcresyntax.html was missing from the distribution.
2069
2070 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2071 compatibility, even though it is no longer used.
2072
2073 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2074 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2075 windows.h file is present (where different names are used). [This was
2076 reversed later after testing - see 16 below.]
2077
2078 8. Changed all #include <config.h> to #include "config.h". There were also
2079 some further <pcre.h> cases that I changed to "pcre.h".
2080
2081 9. When pcregrep was used with the --colour option, it missed the line ending
2082 sequence off the lines that it output.
2083
2084 10. It was pointed out to me that arrays of string pointers cause lots of
2085 relocations when a shared library is dynamically loaded. A technique of
2086 using a single long string with a table of offsets can drastically reduce
2087 these. I have refactored PCRE in four places to do this. The result is
2088 dramatic:
2089
2090 Originally: 290
2091 After changing UCP table: 187
2092 After changing error message table: 43
2093 After changing table of "verbs" 36
2094 After changing table of Posix names 22
2095
2096 Thanks to the folks working on Gregex for glib for this insight.
2097
2098 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2099 unicode-properties was also set.
2100
2101 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2102
2103 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2104 checked only for CRLF.
2105
2106 14. Added casts to pcretest.c to avoid compiler warnings.
2107
2108 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2109
2110 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2111 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2112 entirely. This removes changes made in 7 above.
2113
2114 17. The CMake files have been updated, and there is now more information about
2115 building with CMake in the NON-UNIX-USE document.
2116
2117
2118 Version 7.3 28-Aug-07
2119 ---------------------
2120
2121 1. In the rejigging of the build system that eventually resulted in 7.1, the
2122 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2123 brackets there is not right, since it causes compilers to look for an
2124 installed pcre.h, not the version that is in the source that is being
2125 compiled (which of course may be different). I have changed it back to:
2126
2127 #include "pcre.h"
2128
2129 I have a vague recollection that the change was concerned with compiling in
2130 different directories, but in the new build system, that is taken care of
2131 by the VPATH setting the Makefile.
2132
2133 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2134 when the subject happened to end in the byte 0x85 (e.g. if the last
2135 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2136 characters but of course it shouldn't be taken as a newline when it is part
2137 of another character. The bug was that, for an unlimited repeat of . in
2138 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2139 characters when looking for a newline.
2140
2141 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2142
2143 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2144 in debug output.
2145
2146 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2147 long printing in the pcrecpp unittest when running under MinGW.
2148
2149 6. ESC_K was left out of the EBCDIC table.
2150
2151 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2152 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2153 limit also applies to "virtual nesting" when a pattern is recursive, and in
2154 this case 1000 isn't so big. I have been able to remove this limit at the
2155 expense of backing off one optimization in certain circumstances. Normally,
2156 when pcre_exec() would call its internal match() function recursively and
2157 immediately return the result unconditionally, it uses a "tail recursion"
2158 feature to save stack. However, when a subpattern that can match an empty
2159 string has an unlimited repetition quantifier, it no longer makes this
2160 optimization. That gives it a stack frame in which to save the data for
2161 checking that an empty string has been matched. Previously this was taken
2162 from the 1000-entry workspace that had been reserved. So now there is no
2163 explicit limit, but more stack is used.
2164
2165 8. Applied Daniel's patches to solve problems with the import/export magic
2166 syntax that is required for Windows, and which was going wrong for the
2167 pcreposix and pcrecpp parts of the library. These were overlooked when this
2168 problem was solved for the main library.
2169
2170 9. There were some crude static tests to avoid integer overflow when computing
2171 the size of patterns that contain repeated groups with explicit upper
2172 limits. As the maximum quantifier is 65535, the maximum group length was
2173 set at 30,000 so that the product of these two numbers did not overflow a
2174 32-bit integer. However, it turns out that people want to use groups that
2175 are longer than 30,000 bytes (though not repeat them that many times).
2176 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2177 made it possible to implement the integer overflow checks in a much more
2178 dynamic way, which I have now done. The artificial limitation on group
2179 length has been removed - we now have only the limit on the total length of
2180 the compiled pattern, which depends on the LINK_SIZE setting.
2181
2182 10. Fixed a bug in the documentation for get/copy named substring when
2183 duplicate names are permitted. If none of the named substrings are set, the
2184 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2185 empty string.
2186
2187 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2188 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2189 because the ] is interpreted as the first data character and the
2190 terminating ] is not found. PCRE has been made compatible with Perl in this
2191 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2192 cause memory overwriting.
2193
2194 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2195 string has been matched (to stop an infinite loop). It was not recognizing
2196 a conditional subpattern that could match an empty string if that
2197 subpattern was within another subpattern. For example, it looped when
2198 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2199 condition was not nested. This bug has been fixed.
2200
2201 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2202 past the start of the subject in the presence of bytes with the top bit
2203 set, for example "\x8aBCD".
2204
2205 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2206 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2207
2208 14. Optimized (?!) to (*FAIL).
2209
2210 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2211 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2212 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2213 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2214 does: it's just the validity check that is more restrictive.
2215
2216 16. Inserted checks for integer overflows during escape sequence (backslash)
2217 processing, and also fixed erroneous offset values for syntax errors during
2218 backslash processing.
2219
2220 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2221 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2222
2223 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2224 caused an overrun.
2225
2226 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2227 something other than just ASCII characters) inside a group that had an
2228 unlimited repeat caused a loop at compile time (while checking to see
2229 whether the group could match an empty string).
2230
2231 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2232 [\P{Any}] did so. (Error in the code for printing property names.)
2233
2234 21. An orphan \E inside a character class could cause a crash.
2235
2236 22. A repeated capturing bracket such as (A)? could cause a wild memory
2237 reference during compilation.
2238
2239 23. There are several functions in pcre_compile() that scan along a compiled
2240 expression for various reasons (e.g. to see if it's fixed length for look
2241 behind). There were bugs in these functions when a repeated \p or \P was
2242 present in the pattern. These operators have additional parameters compared
2243 with \d, etc, and these were not being taken into account when moving along
2244 the compiled data. Specifically:
2245
2246 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2247 length.
2248
2249 (b) An item such as \pL+ within a repeated group could cause crashes or
2250 loops.
2251
2252 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2253 "reference to non-existent subpattern" error.
2254
2255 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2256
2257 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2258 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2259
2260 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2261
2262 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2263 character were causing crashes (broken optimization).
2264
2265 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2266 \p or \P) caused a compile-time loop.
2267
2268 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2269 break. For example, the unstudied pattern [\r\n]A does not match the string
2270 "\r\nA" because change 7.0/46 below moves the current point on by two
2271 characters after failing to match at the start. However, the pattern \nA
2272 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2273 the same is true. There doesn't seem any very clean way out of this, but
2274 what I have chosen to do makes the common cases work: PCRE now takes note
2275 of whether there can be an explicit match for \r or \n anywhere in the
2276 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2277 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2278 pattern has explicit CR or LF references.
2279
2280 29. Added (*CR) etc for changing newline setting at start of pattern.
2281
2282
2283 Version 7.2 19-Jun-07
2284 ---------------------
2285
2286 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2287 which is apparently normally available under Windows.
2288
2289 2. Re-jig the pcregrep tests with different newline settings in an attempt
2290 to make them independent of the local environment's newline setting.
2291
2292 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2293
2294 4. Some of the "internals" tests were previously cut out when the link size
2295 was not 2, because the output contained actual offsets. The recent new
2296 "Z" feature of pcretest means that these can be cut out, making the tests
2297 usable with all link sizes.
2298
2299 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2300 stack recursion. This gives a massive performance boost under BSD, but just
2301 a small improvement under Linux. However, it saves one field in the frame
2302 in all cases.
2303
2304 6. Added more features from the forthcoming Perl 5.10:
2305
2306 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2307 recursion call. It refers to the nth most recently opened parentheses.
2308
2309 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2310 to be opened parentheses.
2311
2312 (c) Conditions that refer to capturing parentheses can be specified
2313 relatively, for example, (?(-2)... or (?(+3)...
2314
2315 (d) \K resets the start of the current match so that everything before
2316 is not part of it.
2317
2318 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2319
2320 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2321 reference syntax.
2322
2323 (g) (?| introduces a group in which the numbering of parentheses in each
2324 alternative starts with the same number.
2325
2326 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2327
2328 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2329 PCRE_INFO_JCHANGED.
2330
2331 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2332 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2333 for detecting groups that can match an empty string.
2334
2335 9. A pattern with a very large number of alternatives (more than several
2336 hundred) was running out of internal workspace during the pre-compile
2337 phase, where pcre_compile() figures out how much memory will be needed. A
2338 bit of new cunning has reduced the workspace needed for groups with
2339 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2340 workspace instead of running out of the 4096 that are available.
2341
2342 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2343
2344 11. Applied patch from Google to remove an optimization that didn't quite work.
2345 The report of the bug said:
2346
2347 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2348 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2349 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2350
2351 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2352 it matched the wrong number of bytes.
2353
2354
2355 Version 7.1 24-Apr-07
2356 ---------------------
2357
2358 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2359 that is more "standard", making use of automake and other Autotools. There
2360 is some re-arrangement of the files and adjustment of comments consequent
2361 on this.
2362
2363 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2364 for recursive directory scanning broke on some systems because the files
2365 are not scanned in any specific order and on different systems the order
2366 was different. A call to "sort" has been inserted into RunGrepTest for the
2367 approprate test as a short-term fix. In the longer term there may be an
2368 alternative.
2369
2370 3. I had an email from Eric Raymond about problems translating some of PCRE's
2371 man pages to HTML (despite the fact that I distribute HTML pages, some
2372 people do their own conversions for various reasons). The problems
2373 concerned the use of low-level troff macros .br and .in. I have therefore
2374 removed all such uses from the man pages (some were redundant, some could
2375 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2376 HTML has been updated to handle .nf/.fi and to complain if it encounters
2377 .br or .in.
2378
2379 4. Updated comments in configure.ac that get placed in config.h.in and also
2380 arranged for config.h to be included in the distribution, with the name
2381 config.h.generic, for the benefit of those who have to compile without
2382 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2383
2384 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2385 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2386 makevp.bat for the current PCRE, using the additional files
2387 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2388
2389 6. A Windows user reported a minor discrepancy with test 2, which turned out
2390 to be caused by a trailing space on an input line that had got lost in his
2391 copy. The trailing space was an accident, so I've just removed it.
2392
2393 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2394 that is needed.
2395
2396 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2397 as "const" (a) because they are and (b) because it helps the PHP
2398 maintainers who have recently made a script to detect big data structures
2399 in the php code that should be moved to the .rodata section. I remembered
2400 to update Builducptable as well, so it won't revert if ucptable.h is ever
2401 re-created.
2402
2403 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2404 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2405 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2406 support is not required. This saves 1.5-2K of code, which is important in
2407 some applications.
2408
2409 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2410 so as not to refer to the tables, even though these functions will never be
2411 called when UTF-8 support is disabled. Otherwise there are problems with a
2412 shared library.
2413
2414 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2415
2416 (a) It was defining its arguments as char * instead of void *.
2417
2418 (b) It was assuming that all moves were upwards in memory; this was true
2419 a long time ago when I wrote it, but is no longer the case.
2420
2421 The emulated memove() is provided for those environments that have neither
2422 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2423 is clearly not the case, as these two bugs were recently reported.
2424
2425 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2426 and Detrail to create the HTML documentation, the .txt form of the man
2427 pages, and it removes trailing spaces from listed files. It also creates
2428 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2429 case, it wraps all the #defines with #ifndefs. This script should be run
2430 before "make dist".
2431
2432 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2433 with Unicode property support.
2434
2435 (a) For a maximizing quantifier, if the two different cases of the
2436 character were of different lengths in their UTF-8 codings (there are
2437 some cases like this - I found 11), and the matching function had to
2438 back up over a mixture of the two cases, it incorrectly assumed they
2439 were both the same length.
2440
2441 (b) When PCRE was configured to use the heap rather than the stack for
2442 recursion during matching, it was not correctly preserving the data for
2443 the other case of a UTF-8 character when checking ahead for a match
2444 while processing a minimizing repeat. If the check also involved
2445 matching a wide character, but failed, corruption could cause an
2446 erroneous result when trying to check for a repeat of the original
2447 character.
2448
2449 13. Some tidying changes to the testing mechanism:
2450
2451 (a) The RunTest script now detects the internal link size and whether there
2452 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2453 values substituted by "configure". (The RunGrepTest script already did
2454 this for UTF-8.) The configure.ac script no longer substitutes the
2455 relevant variables.
2456
2457 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2458 with length and offset values. This means that the output is different
2459 for different internal link sizes. Test 2 is skipped for link sizes
2460 other than 2 because of this, bypassing the problem. Unfortunately,
2461 there was also a test in test 3 (the locale tests) that used /B and
2462 failed for link sizes other than 2. Rather than cut the whole test out,
2463 I have added a new /Z option to pcretest that replaces the length and
2464 offset values with spaces. This is now used to make test 3 independent
2465 of link size. (Test 2 will be tidied up later.)
2466
2467 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2468 segmentation fault instead of returning the appropriate error message.
2469
2470 15. In multiline mode when the newline sequence was set to "any", the pattern
2471 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2472 This doesn't seem right; it now treats the CRLF combination as the line
2473 ending, and so does not match in that case. It's only a pattern such as ^$
2474 that would hit this one: something like ^ABC$ would have failed after \r
2475 and then tried again after \r\n.
2476
2477 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2478 in an attempt to make files that differ only in their line terminators
2479 compare equal. This works on Linux.
2480
2481 17. Under certain error circumstances pcregrep might try to free random memory
2482 as it exited. This is now fixed, thanks to valgrind.
2483
2484 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2485 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2486 was because its rules for how to advance for /g after matching an empty
2487 string at the end of a line did not allow for this case. They now check for
2488 it specially.
2489
2490 20. pcretest is supposed to handle patterns and data of any length, by
2491 extending its buffers when necessary. It was getting this wrong when the
2492 buffer for a data line had to be extended.
2493
2494 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2495 CRLF as a newline sequence.
2496
2497 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2498 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2499 I have nevertheless tidied it up.
2500
2501 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2502
2503 24. Added a man page for pcre-config.
2504
2505
2506 Version 7.0 19-Dec-06
2507 ---------------------
2508
2509 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2510 moving to gcc 4.1.1.
2511
2512 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2513 sys/time.h, which is documented as needed for this function. It doesn't
2514 seem to matter on Linux, but it showed up on some releases of OS X.
2515
2516 3. It seems that there are systems where bytes whose values are greater than
2517 127 match isprint() in the "C" locale. The "C" locale should be the
2518 default when a C program starts up. In most systems, only ASCII printing
2519 characters match isprint(). This difference caused the output from pcretest
2520 to vary, making some of the tests fail. I have changed pcretest so that:
2521
2522 (a) When it is outputting text in the compiled version of a pattern, bytes
2523 other than 32-126 are always shown as hex escapes.
2524
2525 (b) When it is outputting text that is a matched part of a subject string,
2526 it does the same, unless a different locale has been set for the match
2527 (using the /L modifier). In this case, it uses isprint() to decide.
2528
2529 4. Fixed a major bug that caused incorrect computation of the amount of memory
2530 required for a compiled pattern when options that changed within the
2531 pattern affected the logic of the preliminary scan that determines the
2532 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2533 that the computed length was too small. The symptoms of this bug were
2534 either the PCRE error "internal error: code overflow" from pcre_compile(),
2535 or a glibc crash with a message such as "pcretest: free(): invalid next
2536 size (fast)". Examples of patterns that provoked this bug (shown in
2537 pcretest format) are:
2538
2539 /(?-x: )/x
2540 /(?x)(?-x: \s*#\s*)/
2541 /((?i)[\x{c0}])/8
2542 /(?i:[\x{c0}])/8
2543
2544 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2545 is now done differently.
2546
2547 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2548 wrapper classes; (b) implement a new function in the C++ scanner that is
2549 more efficient than the old way of doing things because it avoids levels of
2550 recursion in the regex matching; (c) add a paragraph to the documentation
2551 for the FullMatch() function.
2552
2553 6. The escape sequence \n was being treated as whatever was defined as
2554 "newline". Not only was this contrary to the documentation, which states
2555 that \n is character 10 (hex 0A), but it also went horribly wrong when
2556 "newline" was defined as CRLF. This has been fixed.
2557
2558 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2559 was being set to -1 for the "end of line" case (supposedly a value that no
2560 character can have). Though this value is never used (the check for end of
2561 line is "zero bytes in current character"), it caused compiler complaints.
2562 I've changed it to 0xffffffff.
2563
2564 8. In pcre_version.c, the version string was being built by a sequence of
2565 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2566 string (as it is for production releases) called a macro with an empty
2567 argument. The C standard says the result of this is undefined. The gcc
2568 compiler treats it as an empty string (which was what was wanted) but it is
2569 reported that Visual C gives an error. The source has been hacked around to
2570 avoid this problem.
2571
2572 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2573 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2574 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2575 of them did).
2576
2577 10. Originally, pcretest opened its input and output without "b"; then I was
2578 told that "b" was needed in some environments, so it was added for release
2579 5.0 to both the input and output. (It makes no difference on Unix-like
2580 systems.) Later I was told that it is wrong for the input on Windows. I've
2581 now abstracted the modes into two macros, to make it easier to fiddle with
2582 them, and removed "b" from the input mode under Windows.
2583
2584 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2585
2586 12. Added -help and --help to pcretest as an official way of being reminded
2587 of the options.
2588
2589 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2590 and pcrecpp.cc because they annoy compilers at high warning levels.
2591
2592 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2593
2594 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2595 scripts are not C programs :-) and which was not noticed because it works
2596 on Linux.
2597
2598 16. pcretest is supposed to handle any length of pattern and data line (as one
2599 line or as a continued sequence of lines) by extending its input buffer if
2600 necessary. This feature was broken for very long pattern lines, leading to
2601 a string of junk being passed to pcre_compile() if the pattern was longer
2602 than about 50K.
2603
2604 17. I have done a major re-factoring of the way pcre_compile() computes the
2605 amount of memory needed for a compiled pattern. Previously, there was code
2606 that made a preliminary scan of the pattern in order to do this. That was
2607 OK when PCRE was new, but as the facilities have expanded, it has become
2608 harder and harder to keep it in step with the real compile phase, and there
2609 have been a number of bugs (see for example, 4 above). I have now found a
2610 cunning way of running the real compile function in a "fake" mode that
2611 enables it to compute how much memory it would need, while actually only
2612 ever using a few hundred bytes of working memory and without too many
2613 tests of the mode. This should make future maintenance and development
2614 easier. A side effect of this work is that the limit of 200 on the nesting
2615 depth of parentheses has been removed (though this was never a serious
2616 limitation, I suspect). However, there is a downside: pcre_compile() now
2617 runs more slowly than before (30% or more, depending on the pattern). I
2618 hope this isn't a big issue. There is no effect on runtime performance.
2619
2620 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2621 newline (only possible for the last line of a file) and it was a
2622 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2623
2624 19. Added additional timing features to pcretest. (1) The -tm option now times
2625 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2626 separate command line item, by a number that specifies the number of
2627 repeats to use when timing. The default is 50000; this gives better
2628 precision, but takes uncomfortably long for very large patterns.
2629
2630 20. Extended pcre_study() to be more clever in cases where a branch of a
2631 subpattern has no definite first character. For example, (a*|b*)[cd] would
2632 previously give no result from pcre_study(). Now it recognizes that the
2633 first character must be a, b, c, or d.
2634
2635 21. There was an incorrect error "recursive call could loop indefinitely" if
2636 a subpattern (or the entire pattern) that was being tested for matching an
2637 empty string contained only one non-empty item after a nested subpattern.
2638 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2639 incorrectly, because the \d was being skipped in the check.
2640
2641 22. The pcretest program now has a new pattern option /B and a command line
2642 option -b, which is equivalent to adding /B to every pattern. This causes
2643 it to show the compiled bytecode, without the additional information that
2644 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2645 is the same as /B/I).
2646
2647 23. A new optimization is now able automatically to treat some sequences such
2648 as a*b as a*+b. More specifically, if something simple (such as a character
2649 or a simple class like \d) has an unlimited quantifier, and is followed by
2650 something that cannot possibly match the quantified thing, the quantifier
2651 is automatically "possessified".
2652
2653 24. A recursive reference to a subpattern whose number was greater than 39
2654 went wrong under certain circumstances in UTF-8 mode. This bug could also
2655 have affected the operation of pcre_study().
2656
2657 25. Realized that a little bit of performance could be had by replacing
2658 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2659
2660 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2661
2662 27. Possessive quantifiers such as a++ were previously implemented by turning
2663 them into atomic groups such as ($>a+). Now they have their own opcodes,
2664 which improves performance. This includes the automatically created ones
2665 from 23 above.
2666
2667 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2668 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2669 the first matched character to be a colon. This applied both to named and
2670 numbered groups.
2671
2672 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2673
2674 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2675 building PCRE on Windows easier, so I have included it in the distribution.
2676
2677 31. There is now a check in pcretest against a ridiculously large number being
2678 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2679 loop, the loop is abandoned.
2680
2681 32. Forward references to subpatterns in conditions such as (?(2)...) where
2682 subpattern 2 is defined later cause pcre_compile() to search forwards in
2683 the pattern for the relevant set of parentheses. This search went wrong
2684 when there were unescaped parentheses in a character class, parentheses
2685 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2686
2687 33. "Subroutine" calls and backreferences were previously restricted to
2688 referencing subpatterns earlier in the regex. This restriction has now
2689 been removed.
2690
2691 34. Added a number of extra features that are going to be in Perl 5.10. On the
2692 whole, these are just syntactic alternatives for features that PCRE had
2693 previously implemented using the Python syntax or my own invention. The
2694 other formats are all retained for compatibility.
2695
2696 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2697 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2698 also .NET compatible.
2699
2700 (b) A recursion or subroutine call to a named group can now be defined as
2701 (?&name) as well as (?P>name).
2702
2703 (c) A backreference to a named group can now be defined as \k<name> or
2704 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2705 5.10, are also .NET compatible.
2706
2707 (d) A conditional reference to a named group can now use the syntax
2708 (?(<name>) or (?('name') as well as (?(name).
2709
2710 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2711 groups (named and numbered) that are never evaluated inline, but can be
2712 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2713 is always false. There may be only one alternative in such a group.
2714
2715 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2716 as the simple (?(R). The condition is true only if the most recent
2717 recursion is that of the given number or name. It does not search out
2718 through the entire recursion stack.
2719
2720 (g) The escape \gN or \g{N} has been added, where N is a positive or
2721 negative number, specifying an absolute or relative reference.
2722
2723 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2724 some "unreachable code" warnings.
2725
2726 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2727 things, this adds five new scripts.
2728
2729 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2730 There were also incompatibilities regarding the handling of \Q..\E inside
2731 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2732 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2733
2734 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2735 matches an empty string, and forcibly breaks the loop. There were bugs in
2736 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2737 against aaaa the result was just "a" rather than "aaaa", for example. Two
2738 separate and independent bugs (that affected different cases) have been
2739 fixed.
2740
2741 39. Refactored the code to abolish the use of different opcodes for small
2742 capturing bracket numbers. This is a tidy that I avoided doing when I
2743 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2744 The new approach is not only tidier, it makes it possible to reduce the
2745 memory needed to fix the previous bug (38).
2746
2747 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2748 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2749 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2750 mode.
2751
2752 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2753 report.
2754
2755 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2756 copy construction and assignment in the C++ wrapper.
2757
2758 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2759 couple of bugs that could have given wrong results in the "--newline=crlf"
2760 case.
2761
2762 44. Added a number of casts and did some reorganization of signed/unsigned int
2763 variables following suggestions from Dair Grant. Also renamed the variable
2764 "this" as "item" because it is a C++ keyword.
2765
2766 45. Arranged for dftables to add
2767
2768 #include "pcre_internal.h"
2769
2770 to pcre_chartables.c because without it, gcc 4.x may remove the array
2771 definition from the final binary if PCRE is built into a static library and
2772 dead code stripping is activated.
2773
2774 46. For an unanchored pattern, if a match attempt fails at the start of a
2775 newline sequence, and the newline setting is CRLF or ANY, and the next two
2776 characters are CRLF, advance by two characters instead of one.
2777
2778
2779 Version 6.7 04-Jul-06
2780 ---------------------
2781
2782 1. In order to handle tests when input lines are enormously long, pcretest has
2783 been re-factored so that it automatically extends its buffers when
2784 necessary. The code is crude, but this _is_ just a test program. The
2785 default size has been increased from 32K to 50K.
2786
2787 2. The code in pcre_study() was using the value of the re argument before
2788 testing it for NULL. (Of course, in any sensible call of the function, it
2789 won't be NULL.)
2790
2791 3. The memmove() emulation function in pcre_internal.h, which is used on
2792 systems that lack both memmove() and bcopy() - that is, hardly ever -
2793 was missing a "static" storage class specifier.
2794
2795 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2796 containing an extended class (one that cannot be represented by a bitmap
2797 because it contains high-valued characters or Unicode property items, e.g.
2798 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2799 pattern, but PCRE should not loop if you do not (it no longer does).
2800 [Detail: two cases were found: (a) a repeated subpattern containing an
2801 extended class; (b) a recursive reference to a subpattern that followed a
2802 previous extended class. It wasn't skipping over the extended class
2803 correctly when UTF-8 mode was not set.]
2804
2805 5. A negated single-character class was not being recognized as fixed-length
2806 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2807 compile error "lookbehind assertion is not fixed length".
2808
2809 6. The RunPerlTest auxiliary script was showing an unexpected difference
2810 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2811 write a Perl script that can interpret lines of an input file either as
2812 byte characters or as UTF-8, which is what "perltest" was being required to
2813 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2814 can't do is switch easily at run time between having the "use utf8;" pragma
2815 or not. In the end, I fudged it by using the RunPerlTest script to insert
2816 "use utf8;" explicitly for the UTF-8 tests.
2817
2818 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2819 the end of the subject string, contrary to the documentation and to what
2820 Perl does. This was true of both matching functions. Now it matches only at
2821 the start of the subject and immediately after *internal* newlines.
2822
2823 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2824 a pointer to an int instead of a pointer to an unsigned long int. This
2825 caused problems on 64-bit systems.
2826
2827 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2828 instance of the 'standard' template library not being so standard".
2829
2830 10. There was no check on the number of named subpatterns nor the maximum
2831 length of a subpattern name. The product of these values is used to compute
2832 the size of the memory block for a compiled pattern. By supplying a very
2833 long subpattern name and a large number of named subpatterns, the size
2834 computation could be caused to overflow. This is now prevented by limiting
2835 the length of names to 32 characters, and the number of named subpatterns
2836 to 10,000.
2837
2838 11. Subpatterns that are repeated with specific counts have to be replicated in
2839 the compiled pattern. The size of memory for this was computed from the
2840 length of the subpattern and the repeat count. The latter is limited to
2841 65535, but there was no limit on the former, meaning that integer overflow
2842 could in principle occur. The compiled length of a repeated subpattern is
2843 now limited to 30,000 bytes in order to prevent this.
2844
2845 12. Added the optional facility to have named substrings with the same name.
2846
2847 13. Added the ability to use a named substring as a condition, using the
2848 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2849 are numbers (not recommended). Forward references are permitted.
2850
2851 14. Added forward references in named backreferences (if you see what I mean).
2852
2853 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2854 pattern could run off the end of the subject. For example, the pattern
2855 "(?s)(.{1,5})"8 did this with the subject "ab".
2856
2857 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2858 PCRE_CASELESS was set when matching characters that were quantified with ?
2859 or *.
2860
2861 17. A character class other than a single negated character that had a minimum
2862 but no maximum quantifier - for example [ab]{6,} - was not handled
2863 correctly by pce_dfa_exec(). It would match only one character.
2864
2865 18. A valid (though odd) pattern that looked like a POSIX character
2866 class but used an invalid character after [ (for example [[,abc,]]) caused
2867 pcre_compile() to give the error "Failed: internal error: code overflow" or
2868 in some cases to crash with a glibc free() error. This could even happen if
2869 the pattern terminated after [[ but there just happened to be a sequence of
2870 letters, a binary zero, and a closing ] in the memory that followed.
2871
2872 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2873 over the years. Originally (before any Unicode support), just the bottom 8
2874 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2875 output from "man perlunicode" includes this:
2876
2877 The regular expression compiler produces polymorphic opcodes. That
2878 is, the pattern adapts to the data and automatically switches to
2879 the Unicode character scheme when presented with Unicode data--or
2880 instead uses a traditional byte scheme when presented with byte
2881 data.
2882
2883 Sadly, a wide octal escape does not cause a switch, and in a string with
2884 no other multibyte characters, these octal escapes are treated as before.
2885 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2886 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2887 Unicode string.
2888
2889 I have not perpetrated such confusion in PCRE. Up till now, it took just
2890 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2891 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2892 translate to the appropriate multibyte character.
2893
2894 29. Applied some refactoring to reduce the number of warnings from Microsoft
2895 and Borland compilers. This has included removing the fudge introduced
2896 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2897 a warning about an unused variable.
2898
2899 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2900 characters since release 4.0, because Perl (from release 5.004) does not.
2901 [Or at least, is documented not to: some releases seem to be in conflict
2902 with the documentation.] However, when a pattern was studied with
2903 pcre_study() and all its branches started with \s, PCRE still included VT
2904 as a possible starting character. Of course, this did no harm; it just
2905 caused an unnecessary match attempt.
2906
2907 22. Removed a now-redundant internal flag bit that recorded the fact that case
2908 dependency changed within the pattern. This was once needed for "required
2909 byte" processing, but is no longer used. This recovers a now-scarce options
2910 bit. Also moved the least significant internal flag bit to the most-
2911 significant bit of the word, which was not previously used (hangover from
2912 the days when it was an int rather than a uint) to free up another bit for
2913 the future.
2914
2915 23. Added support for CRLF line endings as well as CR and LF. As well as the
2916 default being selectable at build time, it can now be changed at runtime
2917 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2918 specify that it is scanning data with non-default line endings.
2919
2920 24. Changed the definition of CXXLINK to make it agree with the definition of
2921 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2922
2923 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2924 recursions. This makes a big different to stack usage for some patterns.
2925
2926 26. If a subpattern containing a named recursion or subroutine reference such
2927 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2928 the space required for the compiled pattern went wrong and gave too small a
2929 value. Depending on the environment, this could lead to "Failed: internal
2930 error: code overflow at offset 49" or "glibc detected double free or
2931 corruption" errors.
2932
2933 27. Applied patches from Google (a) to support the new newline modes and (b) to
2934 advance over multibyte UTF-8 characters in GlobalReplace.
2935
2936 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2937 difference for some implementation of PCRE in some Windows version.
2938
2939 29. Added some extra testing facilities to pcretest:
2940
2941 \q<number> in a data line sets the "match limit" value
2942 \Q<number> in a data line sets the "match recursion limt" value
2943 -S <number> sets the stack size, where <number> is in megabytes
2944
2945 The -S option isn't available for Windows.
2946
2947
2948 Version 6.6 06-Feb-06
2949 ---------------------
2950
2951 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2952 in pcreposix.h. I have copied the definition from pcre.h.
2953
2954 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2955 because pcre.h is no longer a built file.
2956
2957 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2958 not normally included in the compiled code.
2959
2960
2961 Version 6.5 01-Feb-06
2962 ---------------------
2963
2964 1. When using the partial match feature with pcre_dfa_exec(), it was not
2965 anchoring the second and subsequent partial matches at the new starting
2966 point. This could lead to incorrect results. For example, with the pattern
2967 /1234/, partially matching against "123" and then "a4" gave a match.
2968
2969 2. Changes to pcregrep:
2970
2971 (a) All non-match returns from pcre_exec() were being treated as failures
2972 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2973 error message is output. Some extra information is given for the
2974 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2975 probably the only errors that are likely to be caused by users (by
2976 specifying a regex that has nested indefinite repeats, for instance).
2977 If there are more than 20 of these errors, pcregrep is abandoned.
2978
2979 (b) A binary zero was treated as data while matching, but terminated the
2980 output line if it was written out. This has been fixed: binary zeroes
2981 are now no different to any other data bytes.
2982
2983 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2984 used to set a locale for matching. The --locale=xxxx long option has
2985 been added (no short equivalent) to specify a locale explicitly on the
2986 pcregrep command, overriding the environment variables.
2987
2988 (d) When -B was used with -n, some line numbers in the output were one less
2989 than they should have been.
2990
2991 (e) Added the -o (--only-matching) option.
2992
2993 (f) If -A or -C was used with -c (count only), some lines of context were
2994 accidentally printed for the final match.
2995
2996 (g) Added the -H (--with-filename) option.
2997
2998 (h) The combination of options -rh failed to suppress file names for files
2999 that were found from directory arguments.
3000
3001 (i) Added the -D (--devices) and -d (--directories) options.
3002
3003 (j) Added the -F (--fixed-strings) option.
3004
3005 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3006
3007 (l) Added the --colo(u)r option.
3008
3009 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3010 is not present by default.
3011
3012 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3013 items such as (?R) or (?1), when the recursion could match a number of
3014 alternatives. If it matched one of the alternatives, but subsequently,
3015 outside the recursion, there was a failure, the code tried to back up into
3016 the recursion. However, because of the way PCRE is implemented, this is not
3017 possible, and the result was an incorrect result from the match.
3018
3019 In order to prevent this happening, the specification of recursion has
3020 been changed so that all such subpatterns are automatically treated as
3021 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3022
3023 4. I had overlooked the fact that, in some locales, there are characters for
3024 which isalpha() is true but neither isupper() nor islower() are true. In
3025 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3026 and ordfeminine) are like this. This affected the treatment of \w and \W
3027 when they appeared in character classes, but not when they appeared outside
3028 a character class. The bit map for "word" characters is now created
3029 separately from the results of isalnum() instead of just taking it from the
3030 upper, lower, and digit maps. (Plus the underscore character, of course.)
3031
3032 5. The above bug also affected the handling of POSIX character classes such as
3033 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3034 permanent tables. Instead, the bit maps for such a class were previously
3035 created as the appropriate unions of the upper, lower, and digit bitmaps.
3036 Now they are created by subtraction from the [[:word:]] class, which has
3037 its own bitmap.
3038
3039 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3040 It is created by subtracting the vertical space characters (\x09, \x0a,
3041 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3042 subtraction was done in the overall bitmap for a character class, meaning
3043 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3044 be recognized. This bug has been fixed.
3045
3046 7. Patches from the folks at Google:
3047
3048 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3049 real life, but is still worth protecting against".
3050
3051 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3052 regular expressions".
3053
3054 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3055 have it.
3056
3057 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3058 "configure" and the latter not, in order to fix a problem somebody had
3059 with compiling the Arg class on HP-UX.
3060
3061 (e) Improve the error-handling of the C++ wrapper a little bit.
3062
3063 (f) New tests for checking recursion limiting.
3064
3065 8. The pcre_memmove() function, which is used only if the environment does not
3066 have a standard memmove() function (and is therefore rarely compiled),
3067 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3068 returning a result (though PCRE never actually uses the result).
3069
3070 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3071 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3072 returned instead of calling malloc() with an overflowing number that would
3073 most likely cause subsequent chaos.
3074
3075 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3076
3077 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3078 with this option is matched, the nmatch and pmatch options of regexec() are
3079 ignored.
3080
3081 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3082 provided in case anyone wants to the the POSIX interface with UTF-8
3083 strings.
3084
3085 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3086 C++ linking (needed for some HP-UX environments).
3087
3088 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3089 (unused parameter) and in the pcre_printint() function (omitted "default"
3090 switch label when the default is to do nothing).
3091
3092 15. Added some code to make it possible, when PCRE is compiled as a C++
3093 library, to replace subject pointers for pcre_exec() with a smart pointer
3094 class, thus making it possible to process discontinuous strings.
3095
3096 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3097 much the same function. They were added by different people who were trying
3098 to make PCRE easy to compile on non-Unix systems. It has been suggested
3099 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3100 for compiling on Windows systems. I have therefore replaced it with
3101 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3102 defaults to "extern" for C or "extern C" for C++, which works fine on
3103 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3104 SCOPE with something explicit in config.h. In addition:
3105
3106 (a) pcreposix.h still had just "extern" instead of either of these macros;
3107 I have replaced it with PCRE_DATA_SCOPE.
3108
3109 (b) Functions such as _pcre_xclass(), which are internal to the library,
3110 but external in the C sense, all had PCRE_EXPORT in their definitions.
3111 This is apparently wrong for the Windows case, so I have removed it.
3112 (It makes no difference on Unix-like systems.)
3113
3114 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3115 of recursive calls to match(). This is different to MATCH_LIMIT because
3116 that limits the total number of calls to match(), not all of which increase
3117 the depth of recursion. Limiting the recursion depth limits the amount of
3118 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3119 when PCRE is compiled, and changed at run time. A patch from Google adds
3120 this functionality to the C++ interface.
3121
3122 18. Changes to the handling of Unicode character properties:
3123
3124 (a) Updated the table to Unicode 4.1.0.
3125
3126 (b) Recognize characters that are not in the table as "Cn" (undefined).
3127
3128 (c) I revised the way the table is implemented to a much improved format
3129 which includes recognition of ranges. It now supports the ranges that
3130 are defined in UnicodeData.txt, and it also amalgamates other
3131 characters into ranges. This has reduced the number of entries in the
3132 table from around 16,000 to around 3,000, thus reducing its size
3133 considerably. I realized I did not need to use a tree structure after
3134 all - a binary chop search is just as efficient. Having reduced the
3135 number of entries, I extended their size from 6 bytes to 8 bytes to
3136 allow for more data.
3137
3138 (d) Added support for Unicode script names via properties such as \p{Han}.
3139
3140 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3141 matching that character.
3142
3143 20. When matching a repeated Unicode property with a minimum greater than zero,
3144 (for example \pL{2,}), PCRE could look past the end of the subject if it
3145 reached it while seeking the minimum number of characters. This could
3146 happen only if some of the characters were more than one byte long, because
3147 there is a check for at least the minimum number of bytes.
3148
3149 21. Refactored the implementation of \p and \P so as to be more general, to
3150 allow for more different types of property in future. This has changed the
3151 compiled form incompatibly. Anybody with saved compiled patterns that use
3152 \p or \P will have to recompile them.
3153
3154 22. Added "Any" and "L&" to the supported property types.
3155
3156 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3157 but give a compile time error if the value is greater than 0xff.
3158
3159 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3160 accidentally not being installed or uninstalled.
3161
3162 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3163 made were to insert the current release number. This seemed silly, because
3164 it made things harder for people building PCRE on systems that don't run
3165 "configure". I have turned pcre.h into a distributed file, no longer built
3166 by "configure", with the version identification directly included. There is
3167 no longer a pcre.h.in file.
3168
3169 However, this change necessitated a change to the pcre-config script as
3170 well. It is built from pcre-config.in, and one of the substitutions was the
3171 release number. I have updated configure.ac so that ./configure now finds
3172 the release number by grepping pcre.h.
3173
3174 26. Added the ability to run the tests under valgrind.
3175
3176
3177 Version 6.4 05-Sep-05
3178 ---------------------
3179
3180 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3181 "--" to be printed when multiple files were scanned, even when none of the
3182 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3183 consider it to be a bug, and have restored the previous behaviour.
3184
3185 2. A couple of code tidies to get rid of compiler warnings.
3186
3187 3. The pcretest program used to cheat by referring to symbols in the library
3188 whose names begin with _pcre_. These are internal symbols that are not
3189 really supposed to be visible externally, and in some environments it is
3190 possible to suppress them. The cheating is now confined to including
3191 certain files from the library's source, which is a bit cleaner.
3192
3193 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3194 file's purpose clearer.
3195
3196 5. Reorganized pcre_ucp_findchar().
3197
3198
3199 Version 6.3 15-Aug-05
3200 ---------------------
3201
3202 1. The file libpcre.pc.in did not have general read permission in the tarball.
3203
3204 2. There were some problems when building without C++ support:
3205
3206 (a) If C++ support was not built, "make install" and "make test" still
3207 tried to test it.
3208
3209 (b) There were problems when the value of CXX was explicitly set. Some
3210 changes have been made to try to fix these, and ...
3211
3212 (c) --disable-cpp can now be used to explicitly disable C++ support.
3213
3214 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3215 backslash in a target when C++ was disabled. This confuses some
3216 versions of "make", apparently. Using an intermediate variable solves
3217 this. (Same for CPP_LOBJ.)
3218
3219 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3220 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3221 necessary on certain architectures.
3222
3223 4. Added a setting of -export-symbols-regex to the link command to remove
3224 those symbols that are exported in the C sense, but actually are local
3225 within the library, and not documented. Their names all begin with
3226 "_pcre_". This is not a perfect job, because (a) we have to except some
3227 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3228 available (and never for static libraries). I have made a note to try to
3229 find a way round (a) in the future.
3230
3231
3232 Version 6.2 01-Aug-05
3233 ---------------------
3234
3235 1. There was no test for integer overflow of quantifier values. A construction
3236 such as {1111111111111111} would give undefined results. What is worse, if
3237 a minimum quantifier for a parenthesized subpattern overflowed and became
3238 negative, the calculation of the memory size went wrong. This could have
3239 led to memory overwriting.
3240
3241 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3242
3243 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3244 operating environments where this matters.
3245
3246 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3247 PCRE options from within the C++ wrapper.
3248
3249 5. Named capturing subpatterns were not being correctly counted when a pattern
3250 was compiled. This caused two problems: (a) If there were more than 100
3251 such subpatterns, the calculation of the memory needed for the whole
3252 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3253 back references of the form \12, where the number was greater than 9, were
3254 not recognized as back references, even though there were sufficient
3255 previous subpatterns.
3256
3257 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3258 versions of gcc, e.g. 2.95.4.
3259
3260
3261 Version 6.1 21-Jun-05
3262 ---------------------
3263
3264 1. There was one reference to the variable "posix" in pcretest.c that was not
3265 surrounded by "#if !defined NOPOSIX".
3266
3267 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3268 the cross-check on the old pcre_info() function, for the benefit of the
3269 cut-down version of PCRE that is currently imported into Exim.
3270
3271 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3272 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3273 patterns that start (?i) but I don't think that matters. The use of (?i) is
3274 just an example; this all applies to the other options as well.
3275
3276 4. Since libtool seems to echo the compile commands it is issuing, the output
3277 from "make" can be reduced a bit by putting "@" in front of each libtool
3278 compile command.
3279
3280 5. Patch from the folks at Google for configure.in to be a bit more thorough
3281 in checking for a suitable C++ installation before trying to compile the
3282 C++ stuff. This should fix a reported problem when a compiler was present,
3283 but no suitable headers.
3284
3285 6. The man pages all had just "PCRE" as their title. I have changed them to
3286 be the relevant file name. I have also arranged that these names are
3287 retained in the file doc/pcre.txt, which is a concatenation in text format
3288 of all the man pages except the little individual ones for each function.
3289
3290 7. The NON-UNIX-USE file had not been updated for the different set of source
3291 files that come with release 6. I also added a few comments about the C++
3292 wrapper.
3293
3294
3295 Version 6.0 07-Jun-05
3296 ---------------------
3297
3298 1. Some minor internal re-organization to help with my DFA experiments.
3299
3300 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3301 didn't matter for the library itself when fully configured, but did matter
3302 when compiling without UCP support, or within Exim, where the ucp files are
3303 not imported.
3304
3305 3. Refactoring of the library code to split up the various functions into
3306 different source modules. The addition of the new DFA matching code (see
3307 below) to a single monolithic source would have made it really too
3308 unwieldy, quite apart from causing all the code to be include in a
3309 statically linked application, when only some functions are used. This is
3310 relevant even without the DFA addition now that patterns can be compiled in
3311 one application and matched in another.
3312
3313 The downside of splitting up is that there have to be some external
3314 functions and data tables that are used internally in different modules of
3315 the library but which are not part of the API. These have all had their
3316 names changed to start with "_pcre_" so that they are unlikely to clash
3317 with other external names.
3318
3319 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3320 a different (DFA) algorithm. Although it is slower than the original
3321 function, it does have some advantages for certain types of matching
3322 problem.
3323
3324 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3325 including restarting after a partial match.
3326
3327 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3328 defined when compiling for Windows was sent to me. I have put it into the
3329 code, though I have no means of testing or verifying it.
3330
3331 7. Added the pcre_refcount() auxiliary function.
3332
3333 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3334 match before or at the first newline in the subject string. In pcretest,
3335 the /f option on a pattern can be used to set this.
3336
3337 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3338 would behave wrongly. This has been present in PCRE since release 4.0.
3339
3340 10. A number of changes to the pcregrep command:
3341
3342 (a) Refactored how -x works; insert ^(...)$ instead of setting
3343 PCRE_ANCHORED and checking the length, in preparation for adding
3344 something similar for -w.
3345
3346 (b) Added the -w (match as a word) option.
3347
3348 (c) Refactored the way lines are read and buffered so as to have more
3349 than one at a time available.
3350
3351 (d) Implemented a pcregrep test script.
3352
3353 (e) Added the -M (multiline match) option. This allows patterns to match
3354 over several lines of the subject. The buffering ensures that at least
3355 8K, or the rest of the document (whichever is the shorter) is available
3356 for matching (and similarly the previous 8K for lookbehind assertions).
3357
3358 (f) Changed the --help output so that it now says
3359
3360 -w, --word-regex(p)
3361
3362 instead of two lines, one with "regex" and the other with "regexp"
3363 because that confused at least one person since the short forms are the
3364 same. (This required a bit of code, as the output is generated
3365 automatically from a table. It wasn't just a text change.)
3366
3367 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3368 option but starts with a hyphen. Could be a pattern or a path name
3369 starting with a hyphen, for instance.
3370
3371 (h) "-" can be given as a file name to represent stdin.
3372
3373 (i) When file names are being printed, "(standard input)" is used for
3374 the standard input, for compatibility with GNU grep. Previously
3375 "<stdin>" was used.
3376
3377 (j) The option --label=xxx can be used to supply a name to be used for
3378 stdin when file names are being printed. There is no short form.
3379
3380 (k) Re-factored the options decoding logic because we are going to add
3381 two more options that take data. Such options can now be given in four
3382 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3383
3384 (l) Added the -A, -B, and -C options for requesting that lines of context
3385 around matches be printed.
3386
3387 (m) Added the -L option to print the names of files that do not contain
3388 any matching lines, that is, the complement of -l.
3389
3390 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3391 continue to scan other files.
3392
3393 (o) The -s option was incorrectly implemented. For compatibility with other
3394 greps, it now suppresses the error message for a non-existent or non-
3395 accessible file (but not the return code). There is a new option called
3396 -q that suppresses the output of matching lines, which was what -s was
3397 previously doing.
3398
3399 (p) Added --include and --exclude options to specify files for inclusion
3400 and exclusion when recursing.
3401
3402 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3403 Hopefully, it now does.
3404
3405 12. Missing cast in pcre_study().
3406
3407 13. Added an "uninstall" target to the makefile.
3408
3409 14. Replaced "extern" in the function prototypes in Makefile.in with
3410 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3411 world, but is set differently for Windows.
3412
3413 15. Added a second compiling function called pcre_compile2(). The only
3414 difference is that it has an extra argument, which is a pointer to an
3415 integer error code. When there is a compile-time failure, this is set
3416 non-zero, in addition to the error test pointer being set to point to an
3417 error message. The new argument may be NULL if no error number is required
3418 (but then you may as well call pcre_compile(), which is now just a
3419 wrapper). This facility is provided because some applications need a
3420 numeric error indication, but it has also enabled me to tidy up the way
3421 compile-time errors are handled in the POSIX wrapper.
3422
3423 16. Added VPATH=.libs to the makefile; this should help when building with one
3424 prefix path and installing with another. (Or so I'm told by someone who
3425 knows more about this stuff than I do.)
3426
3427 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3428 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3429 match everything, including newlines. This is not POSIX-compatible, but
3430 somebody wanted the feature. From pcretest it can be activated by using
3431 both the P and the s flags.
3432
3433 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3434
3435 19. libpcre.pc was being incorrectly installed as executable.
3436
3437 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3438 it now also looks for '\r' so that it will work unmodified on Windows.
3439
3440 21. Added Google's contributed C++ wrapper to the distribution.
3441
3442 22. Added some untidy missing memory free() calls in pcretest, to keep
3443 Electric Fence happy when testing.
3444
3445
3446
3447 Version 5.0 13-Sep-04
3448 ---------------------
3449
3450 1. Internal change: literal characters are no longer packed up into items
3451 containing multiple characters in a single byte-string. Each character
3452 is now matched using a separate opcode. However, there may be more than one
3453 byte in the character in UTF-8 mode.
3454
3455 2. The pcre_callout_block structure has two new fields: pattern_position and
3456 next_item_length. These contain the offset in the pattern to the next match
3457 item, and its length, respectively.
3458
3459 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3460 insertion of callouts before each pattern item. Added the /C option to
3461 pcretest to make use of this.
3462
3463 4. On the advice of a Windows user, the lines
3464
3465 #if defined(_WIN32) || defined(WIN32)
3466 _setmode( _fileno( stdout ), 0x8000 );
3467 #endif /* defined(_WIN32) || defined(WIN32) */
3468
3469 have been added to the source of pcretest. This apparently does useful
3470 magic in relation to line terminators.
3471
3472 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3473 for the benefit of those environments where the "b" makes a difference.
3474
3475 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3476 to know about it. I have put a hack into configure.in that adds in code
3477 to set GCC=yes if CC=icc. This seems to end up at a point in the
3478 generated configure script that is early enough to affect the setting of
3479 compiler options, which is what is needed, but I have no means of testing
3480 whether it really works. (The user who reported this had patched the
3481 generated configure script, which of course I cannot do.)
3482
3483 LATER: After change 22 below (new libtool files), the configure script
3484 seems to know about icc (and also ecc). Therefore, I have commented out
3485 this hack in configure.in.
3486
3487 7. Added support for pkg-config (2 patches were sent in).
3488
3489 8. Negated POSIX character classes that used a combination of internal tables
3490 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3491 [[:^ascii]]. Typically, they would match almost any characters. The other
3492 POSIX classes were not broken in this way.
3493
3494 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3495 to find the match, as PCRE was deluded into thinking that the match had to
3496 start at the start point or following a newline. The same bug applied to
3497 patterns with negative forward assertions or any backward assertions
3498 preceding ".*" at the start, unless the pattern required a fixed first
3499 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3500
3501 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3502 starting at the last subject character, bytes beyond the end of the subject
3503 string were read.
3504
3505 11. Renamed the variable "class" as "classbits" to make life easier for C++
3506 users. (Previously there was a macro definition, but it apparently wasn't
3507 enough.)
3508
3509 12. Added the new field "tables" to the extra data so that tables can be passed
3510 in at exec time, or the internal tables can be re-selected. This allows
3511 a compiled regex to be saved and re-used at a later time by a different
3512 program that might have everything at different addresses.
3513
3514 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3515 -R library as well as a -L library.
3516
3517 14. The debugging options of pcretest (-d on the command line or D on a
3518 pattern) showed incorrect output for anything following an extended class
3519 that contained multibyte characters and which was followed by a quantifier.
3520
3521 15. Added optional support for general category Unicode character properties
3522 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3523 support. It adds about 90K to the size of the library. The meanings of the
3524 inbuilt class escapes such as \d and \s have NOT been changed.
3525
3526 16. Updated pcredemo.c to include calls to free() to release the memory for the
3527 compiled pattern.
3528
3529 17. The generated file chartables.c was being created in the source directory
3530 instead of in the building directory. This caused the build to fail if the
3531 source directory was different from the building directory, and was
3532 read-only.
3533
3534 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3535 file. No doubt somebody will tell me if they don't make sense... Also added
3536 Dan Mooney's comments about building on OpenVMS.
3537
3538 19. Added support for partial matching via the PCRE_PARTIAL option for
3539 pcre_exec() and the \P data escape in pcretest.
3540
3541 20. Extended pcretest with 3 new pattern features:
3542
3543 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3544 write the compiled pattern to the file whose name is "rest-of-line".
3545 This is a straight binary dump of the data, with the saved pointer to
3546 the character tables forced to be NULL. The study data, if any, is
3547 written too. After writing, pcretest reads a new pattern.
3548
3549 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3550 compiled pattern from the given file. There must not be any
3551 occurrences of "<" in the file name (pretty unlikely); if there are,
3552 pcretest will instead treat the initial "<" as a pattern delimiter.
3553 After reading in the pattern, pcretest goes on to read data lines as
3554 usual.
3555
3556 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3557 and 16-bit fields in a compiled pattern, to simulate a pattern that
3558 was compiled on a host of opposite endianness.
3559
3560 21. The pcre-exec() function can now cope with patterns that were compiled on
3561 hosts of opposite endianness, with this restriction:
3562
3563 As for any compiled expression that is saved and used later, the tables
3564 pointer field cannot be preserved; the extra_data field in the arguments
3565 to pcre_exec() should be used to pass in a tables address if a value
3566 other than the default internal tables were used at compile time.
3567
3568 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3569 now diagnosed as an error. Previously, most of the time, a negative number
3570 would have been treated as zero, but if in addition "ovector" was passed as
3571 NULL, a crash could occur.
3572
3573 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3574 new versions from the libtool 1.5 distribution (the last one is a copy of
3575 a file called libtool.m4). This seems to have fixed the need to patch
3576 "configure" to support Darwin 1.3 (which I used to do). However, I still
3577 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3578 workstation).
3579
3580 24. Changed the PCRE licence to be the more standard "BSD" licence.
3581
3582
3583 Version 4.5 01-Dec-03
3584 ---------------------
3585
3586 1. There has been some re-arrangement of the code for the match() function so
3587 that it can be compiled in a version that does not call itself recursively.
3588 Instead, it keeps those local variables that need separate instances for
3589 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3590 needs to "recurse". Keeping track of where control must go is done by means
3591 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3592 hide most of the details from the main code, and operates only if
3593 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3594 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3595 operating.
3596
3597 To make it easier for callers to provide specially tailored get/free
3598 functions for this usage, two new functions, pcre_stack_malloc, and
3599 pcre_stack_free, are used. They are always called in strict stacking order,
3600 and the size of block requested is always the same.
3601
3602 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3603 PCRE has been compiled to use the stack or the heap for recursion. The
3604 -C option of pcretest uses this to show which version is compiled.
3605
3606 A new data escape \S, is added to pcretest; it causes the amounts of store
3607 obtained and freed by both kinds of malloc/free at match time to be added
3608 to the output.
3609
3610 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3611 what's available on my current Linux desktop machine.
3612
3613 3. When matching a UTF-8 string, the test for a valid string at the start has
3614 been extended. If start_offset is not zero, PCRE now checks that it points
3615 to a byte that is the start of a UTF-8 character. If not, it returns
3616 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3617 this is necessary because there may be backward assertions in the pattern.
3618 When matching the same subject several times, it may save resources to use
3619 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3620
3621 4. The code for checking the validity of UTF-8 strings has been tightened so
3622 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3623 containing "overlong sequences".
3624
3625 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3626 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3627 should have been "&", but it just so happened that all the cases this let
3628 through by mistake were picked up later in the function.
3629
3630 6. I had used a variable called "isblank" - this is a C99 function, causing
3631 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3632
3633 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3634 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3635 so the version is shown; (c) stop "make test" echoing "./RunTest".
3636
3637 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3638
3639 9. The prototype for memmove() for systems that don't have it was using
3640 size_t, but the inclusion of the header that defines size_t was later. I've
3641 moved the #includes for the C headers earlier to avoid this.
3642
3643 10. Added some adjustments to the code to make it easier to compiler on certain
3644 special systems:
3645
3646 (a) Some "const" qualifiers were missing.
3647 (b) Added the macro EXPORT before all exported functions; by default this
3648 is defined to be empty.
3649 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3650 that it reads its output file name as an argument instead of writing
3651 to the standard output and assuming this can be redirected.
3652
3653 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3654 class containing characters with values greater than 255, PCRE compilation
3655 went into a loop.
3656
3657 12. A recursive reference to a subpattern that was within another subpattern
3658 that had a minimum quantifier of zero caused PCRE to crash. For example,
3659 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3660 recursion. If the recursively-called subpattern itself had a zero repeat,
3661 that was OK.
3662
3663 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3664 buffer into which it was copied (for escape processing) was still set at
3665 1024, so long lines caused crashes.
3666
3667 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3668 "internal error: code overflow...". This applied to any character class
3669 that was followed by a possessive quantifier.
3670
3671 15. Modified the Makefile to add libpcre.la as a prerequisite for
3672 libpcreposix.la because I was told this is needed for a parallel build to
3673 work.
3674
3675 16. If a pattern that contained .* following optional items at the start was
3676 studied, the wrong optimizing data was generated, leading to matching
3677 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3678 matching string must start with a or b or c. The correct conclusion for
3679 this pattern is that a match can start with any character.
3680
3681
3682 Version 4.4 13-Aug-03
3683 ---------------------
3684
3685 1. In UTF-8 mode, a character class containing characters with values between
3686 127 and 255 was not handled correctly if the compiled pattern was studied.
3687 In fixing this, I have also improved the studying algorithm for such
3688 classes (slightly).
3689
3690 2. Three internal functions had redundant arguments passed to them. Removal
3691 might give a very teeny performance improvement.
3692
3693 3. Documentation bug: the value of the capture_top field in a callout is *one
3694 more than* the number of the hightest numbered captured substring.
3695
3696 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3697 in incorrectly linking with a previously installed version. They now link
3698 explicitly with libpcre.la.
3699
3700 5. configure.in no longer needs to recognize Cygwin specially.
3701
3702 6. A problem in pcre.in for Windows platforms is fixed.
3703
3704 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3705 pcretest, it used to include the size of the study block as part of its
3706 output. Unfortunately, the structure contains a field that has a different
3707 size on different hardware architectures. This meant that the tests that
3708 showed this size failed. As the block is currently always of a fixed size,
3709 this information isn't actually particularly useful in pcretest output, so
3710 I have just removed it.
3711
3712 8. Three pre-processor statements accidentally did not start in column 1.
3713 Sadly, there are *still* compilers around that complain, even though
3714 standard C has not required this for well over a decade. Sigh.
3715
3716 9. In pcretest, the code for checking callouts passed small integers in the
3717 callout_data field, which is a void * field. However, some picky compilers
3718 complained about the casts involved for this on 64-bit systems. Now
3719 pcretest passes the address of the small integer instead, which should get
3720 rid of the warnings.
3721
3722 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3723 both compile and run time, and gives an error if an invalid UTF-8 sequence
3724 is found. There is a option for disabling this check in cases where the
3725 string is known to be correct and/or the maximum performance is wanted.
3726
3727 11. In response to a bug report, I changed one line in Makefile.in from
3728
3729 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3730 to
3731 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3732
3733 to look similar to other lines, but I have no way of telling whether this
3734 is the right thing to do, as I do not use Windows. No doubt I'll get told
3735 if it's wrong...
3736
3737
3738 Version 4.3 21-May-03
3739 ---------------------
3740
3741 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3742 Makefile.
3743
3744 2. Some refactoring to improve the quality of the code:
3745
3746 (i) The utf8_table... variables are now declared "const".
3747
3748 (ii) The code for \cx, which used the "case flipping" table to upper case
3749 lower case letters, now just substracts 32. This is ASCII-specific,
3750 but the whole concept of \cx is ASCII-specific, so it seems
3751 reasonable.
3752
3753 (iii) PCRE was using its character types table to recognize decimal and
3754 hexadecimal digits in the pattern. This is silly, because it handles
3755 only 0-9, a-f, and A-F, but the character types table is locale-
3756 specific, which means strange things might happen. A private
3757 table is now used for this - though it costs 256 bytes, a table is
3758 much faster than multiple explicit tests. Of course, the standard
3759 character types table is still used for matching digits in subject
3760 strings against \d.
3761
3762 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3763 ending in _t are). So I've renamed it as ESC_tee.
3764
3765 3. The first argument for regexec() in the POSIX wrapper should have been
3766 defined as "const".
3767
3768 4. Changed pcretest to use malloc() for its buffers so that they can be
3769 Electric Fenced for debugging.
3770
3771 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3772 to read one or more bytes before the start of the subject string. Often this
3773 had no effect on PCRE's behaviour, but in some circumstances it could
3774 provoke a segmentation fault.
3775
3776 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3777 to try to read one or more bytes before the start of the subject string.
3778
3779 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3780 UTF-8 support could misbehave in various ways if the subject string
3781 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3782 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3783 back over UTF-8 characters.)
3784
3785
3786 Version 4.2 14-Apr-03
3787 ---------------------
3788
3789 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3790
3791 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3792 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3793 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3794 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3795 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3796 and BUILD_EXEEXT
3797 Note: automatic setting of the BUILD variables is not yet working
3798 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3799 compile-time but not at link-time
3800 [LINK]: use for linking executables only
3801 make different versions for Windows and non-Windows
3802 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3803 libraries
3804 [LINK_FOR_BUILD]: new variable
3805 [OBJEXT]: use throughout
3806 [EXEEXT]: use throughout
3807 <winshared>: new target
3808 <wininstall>: new target
3809 <dftables.o>: use native compiler
3810 <dftables>: use native linker
3811 <install>: handle Windows platform correctly
3812 <clean>: ditto
3813 <check>: ditto
3814 copy DLL to top builddir before testing
3815
3816 As part of these changes, -no-undefined was removed again. This was reported
3817 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3818 in any case.
3819
3820 3. Some tidies to get rid of compiler warnings:
3821
3822 . In the match_data structure, match_limit was an unsigned long int, whereas
3823 match_call_count was an int. I've made them both unsigned long ints.
3824
3825 . In pcretest the fact that a const uschar * doesn't automatically cast to
3826 a void * provoked a warning.
3827
3828 . Turning on some more compiler warnings threw up some "shadow" variables
3829 and a few more missing casts.
3830
3831 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3832 option, a class that contained a single character with a value between 128
3833 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3834
3835 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3836 option, a class that contained several characters, but with at least one
3837 whose value was between 128 and 255 caused PCRE to crash.
3838
3839
3840 Version 4.1 12-Mar-03
3841 ---------------------
3842
3843 1. Compiling with gcc -pedantic found a couple of places where casts were
3844 needed, and a string in dftables.c that was longer than standard compilers are
3845 required to support.
3846
3847 2. Compiling with Sun's compiler found a few more places where the code could
3848 be tidied up in order to avoid warnings.
3849
3850 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3851 first of these names is deprecated in the latest Autoconf in favour of the name
3852 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3853 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3854 analogy I have changed it to CFLAGS_FOR_BUILD.
3855
3856 4. Added -no-undefined to the linking command in the Makefile, because this is
3857 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3858 linking step for the pcreposix library.
3859
3860 5. PCRE was failing to diagnose the case of two named groups with the same
3861 name.
3862
3863 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3864 literal character that is needed in the subject for a match, and scans along to
3865 ensure that it is present before embarking on the full matching process. This
3866 saves time in cases of nested unlimited repeats that are never going to match.
3867 Problem: the scan can take a lot of time if the subject is very long (e.g.
3868 megabytes), thus penalizing straightforward matches. It is now done only if the
3869 amount of subject to be scanned is less than 1000 bytes.
3870
3871 7. A lesser problem with the same optimization is that it was recording the
3872 first character of an anchored pattern as "needed", thus provoking a search
3873 right along the subject, even when the first match of the pattern was going to
3874 fail. The "needed" character is now not set for anchored patterns, unless it
3875 follows something in the pattern that is of non-fixed length. Thus, it still
3876 fulfils its original purpose of finding quick non-matches in cases of nested
3877 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3878
3879
3880 Version 4.0 17-Feb-03
3881 ---------------------
3882
3883 1. If a comment in an extended regex that started immediately after a meta-item
3884 extended to the end of string, PCRE compiled incorrect data. This could lead to
3885 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3886
3887 2. Moved to autoconf 2.53 and libtool 1.4.2.
3888
3889 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3890 the special perltest8 script is no longer needed - all the tests can be run
3891 from a single perltest script.
3892
3893 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3894 by \s. It has now been removed in PCRE. This means it isn't recognized as
3895 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3896 class [:space:] *does* include VT, thereby creating a mess.
3897
3898 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3899 space and tab.
3900
3901 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3902 its new features into the main test script, reducing the number of scripts.
3903
3904 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3905 were backward compatible, and made the (?i) apply to the whole pattern, as if
3906 /i were given. Now it behaves more logically, and applies the option setting
3907 only to what follows. PCRE has been changed to follow suit. However, if it
3908 finds options settings right at the start of the pattern, it extracts them into
3909 the global options, as before. Thus, they show up in the info data.
3910
3911 8. Added support for the \Q...\E escape sequence. Characters in between are
3912 treated as literals. This is slightly different from Perl in that $ and @ are
3913 also handled as literals inside the quotes. In Perl, they will cause variable
3914 interpolation. Note the following examples:
3915
3916 Pattern PCRE matches Perl matches
3917
3918 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3919 \Qabc\$xyz\E abc\$xyz abc\$xyz
3920 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3921
3922 For compatibility with Perl, \Q...\E sequences are recognized inside character
3923 classes as well as outside them.
3924
3925 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3926 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3927 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3928 signed/unsigned warnings.
3929
3930 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3931 option for pcretest, so I've replaced it by a simple function that does just
3932 that job.
3933
3934 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3935 "pcregrep -".
3936
3937 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3938 Java package. This provides some syntactic sugar for simple cases of what my
3939 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3940 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3941 item, you can use this simplified notation. Note that only makes sense with
3942 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3943 greediness, whatever the setting of the PCRE_UNGREEDY option.
3944
3945 13. A change of greediness default within a pattern was not taking effect at
3946 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3947 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3948 was abstracted outside.
3949
3950 14. PCRE now supports the \G assertion. It is true when the current matching
3951 position is at the start point of the match. This differs from \A when the
3952 starting offset is non-zero. Used with the /g option of pcretest (or similar
3953 code), it works in the same way as it does for Perl's /g option. If all
3954 alternatives of a regex begin with \G, the expression is anchored to the start
3955 match position, and the "anchored" flag is set in the compiled expression.
3956
3957 15. Some bugs concerning the handling of certain option changes within patterns
3958 have been fixed. These applied to options other than (?ims). For example,
3959 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3960 been the other way round. Some of this was related to change 7 above.
3961
3962 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3963 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3964 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3965 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3966
3967 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3968 mode. Unlike ".", it always matches newline, whatever the setting of
3969 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3970 assertions. Perl allows it, but it doesn't (in general) work because it can't
3971 calculate the length of the lookbehind. At least, that's the case for Perl
3972 5.8.0 - I've been told they are going to document that it doesn't work in
3973 future.
3974
3975 18. Added an error diagnosis for escapes that PCRE does not support: these are
3976 \L, \l, \N, \P, \p, \U, \u, and \X.
3977
3978 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3979 reading past the end of the pattern in cases such as /[abcd/.
3980
3981 20. PCRE was getting more memory than necessary for patterns with classes that
3982 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3983
3984 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3985 compiling PCRE for use with Virtual Pascal.
3986
3987 22. Small fix to the Makefile to make it work properly if the build is done
3988 outside the source tree.
3989
3990 23. Added a new extension: a condition to go with recursion. If a conditional
3991 subpattern starts with (?(R) the "true" branch is used if recursion has
3992 happened, whereas the "false" branch is used only at the top level.
3993
3994 24. When there was a very long string of literal characters (over 255 bytes
3995 without UTF support, over 250 bytes with UTF support), the computation of how
3996 much memory was required could be incorrect, leading to segfaults or other
3997 strange effects.
3998
3999 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4000 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4001 there was a subsequent back reference to those brackets. This meant that, for
4002 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4003 possible to check for precisely this case. All we can do is abandon the
4004 optimization if .* occurs inside capturing brackets when there are any back
4005 references whatsoever. (See below for a better fix that came later.)
4006
4007 26. The handling of the optimization for finding the first character of a
4008 non-anchored pattern, and for finding a character that is required later in the
4009 match were failing in some cases. This didn't break the matching; it just
4010 failed to optimize when it could. The way this is done has been re-implemented.
4011
4012 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4013
4014 28. Added a new feature that provides some of the functionality that Perl
4015 provides with (?{...}). The facility is termed a "callout". The way it is done
4016 in PCRE is for the caller to provide an optional function, by setting
4017 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4018 global variable. By default it is unset, which disables all calling out. To get
4019 the function called, the regex must include (?C) at appropriate points. This
4020 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4021 This provides a means of identifying different callout points. When PCRE
4022 reaches such a point in the regex, if pcre_callout has been set, the external
4023 function is called. It is provided with data in a structure called
4024 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4025 matching continues; if it returns a non-zero value, the match at the current
4026 point fails. However, backtracking will occur if possible. [This was changed
4027 later and other features added - see item 49 below.]
4028
4029 29. pcretest is upgraded to test the callout functionality. It provides a
4030 callout function that displays information. By default, it shows the start of
4031 the match and the current position in the text. There are some new data escapes
4032 to vary what happens:
4033
4034 \C+ in addition, show current contents of captured substrings
4035 \C- do not supply a callout function
4036 \C!n return 1 when callout number n is reached
4037 \C!n!m return 1 when callout number n is reached for the mth time
4038
4039 30. If pcregrep was called with the -l option and just a single file name, it
4040 output "<stdin>" if a match was found, instead of the file name.
4041
4042 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4043 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4044 pcre_exec(). This saves a malloc/free per call. The default value of
4045 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4046 when configuring.
4047
4048 32. The default maximum size of a compiled pattern is 64K. There have been a
4049 few cases of people hitting this limit. The code now uses macros to handle the
4050 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4051 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4052 configuring. Tests 2 and 5 work only with 2-byte links because they output
4053 debugging information about compiled patterns.
4054
4055 33. Internal code re-arrangements:
4056
4057 (a) Moved the debugging function for printing out a compiled regex into
4058 its own source file (printint.c) and used #include to pull it into
4059 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4060 separate copies.
4061
4062 (b) Defined the list of op-code names for debugging as a macro in
4063 internal.h so that it is next to the definition of the opcodes.
4064
4065 (c) Defined a table of op-code lengths for simpler skipping along compiled
4066 code. This is again a macro in internal.h so that it is next to the
4067 definition of the opcodes.
4068
4069 34. Added support for recursive calls to individual subpatterns, along the
4070 lines of Robin Houston's patch (but implemented somewhat differently).
4071
4072 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4073 allow it to read and process whole directories in Win32. This code was
4074 contributed by Lionel Fourquaux; it has not been tested by me.
4075
4076 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4077 used to name a group. Names consist of alphanumerics and underscores, and must
4078 be unique. Back references use the syntax (?P=name) and recursive calls use
4079 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4080 numbers. The function pcre_fullinfo() can be used after compilation to extract
4081 a name/number map. There are three relevant calls:
4082
4083 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4084 PCRE_INFO_NAMECOUNT yields the number of entries
4085 PCRE_INFO_NAMETABLE yields a pointer to the map.
4086
4087 The map is a vector of fixed-size entries. The size of each entry depends on
4088 the length of the longest name used. The first two bytes of each entry are the
4089 group number, most significant byte first. There follows the corresponding
4090 name, zero terminated. The names are in alphabetical order.
4091
4092 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4093 case instead of 255. Making it the same both with and without UTF-8 support
4094 means that the same test output works with both.
4095
4096 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4097 calling malloc() with a zero argument.
4098
4099 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4100 optimization. I've improved things by keeping a bitmap of backreferences with
4101 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4102 fact referenced, the optimization can be applied. It is unlikely that a
4103 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4104 the match to follow \n) will appear inside brackets with a number greater than
4105 31, but if it does, any back reference > 31 suppresses the optimization.
4106
4107 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4108 of disabling numbered capturing parentheses. Any opening parenthesis that is
4109 not followed by ? behaves as if it were followed by ?: but named parentheses
4110 can still be used for capturing (and they will acquire numbers in the usual
4111 way).
4112
4113 41. Redesigned the return codes from the match() function into yes/no/error so
4114 that errors can be passed back from deep inside the nested calls. A malloc
4115 failure while inside a recursive subpattern call now causes the
4116 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4117
4118 42. It is now possible to set a limit on the number of times the match()
4119 function is called in a call to pcre_exec(). This facility makes it possible to
4120 limit the amount of recursion and backtracking, though not in a directly
4121 obvious way, because the match() function is used in a number of different
4122 circumstances. The count starts from zero for each position in the subject
4123 string (for non-anchored patterns). The default limit is, for compatibility, a
4124 large number, namely 10 000 000. You can change this in two ways:
4125
4126 (a) When configuring PCRE before making, you can use --with-match-limit=n
4127 to set a default value for the compiled library.
4128
4129 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4130 a different value is set. See 45 below.
4131
4132 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4133
4134 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4135 of things that can be changed at compile time. The first argument specifies
4136 what is wanted and the second points to where the information is to be placed.
4137 The current list of available information is:
4138
4139 PCRE_CONFIG_UTF8
4140
4141 The output is an integer that is set to one if UTF-8 support is available;
4142 otherwise it is set to zero.
4143
4144 PCRE_CONFIG_NEWLINE
4145
4146 The output is an integer that it set to the value of the code that is used for
4147 newline. It is either LF (10) or CR (13).
4148
4149 PCRE_CONFIG_LINK_SIZE
4150
4151 The output is an integer that contains the number of bytes used for internal
4152 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4153
4154 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4155
4156 The output is an integer that contains the threshold above which the POSIX
4157 interface uses malloc() for output vectors. See item 31 above.
4158
4159 PCRE_CONFIG_MATCH_LIMIT
4160
4161 The output is an unsigned integer that contains the default limit of the number
4162 of match() calls in a pcre_exec() execution. See 42 above.
4163
4164 44. pcretest has been upgraded by the addition of the -C option. This causes it
4165 to extract all the available output from the new pcre_config() function, and to
4166 output it. The program then exits immediately.
4167
4168 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4169 order to support additional features. One way would have been to define
4170 pcre_exec2() (for example) with extra arguments, but this would not have been
4171 extensible, and would also have required all calls to the original function to
4172 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4173 is used for passing in "extra" data from pcre_study().
4174
4175 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4176 contains the following fields:
4177
4178 flags a bitmap indicating which of the following fields are set
4179 study_data opaque data from pcre_study()
4180 match_limit a way of specifying a limit on match() calls for a specific
4181 call to pcre_exec()
4182 callout_data data for callouts (see 49 below)
4183
4184 The flag bits are also defined in pcre.h, and are
4185
4186 PCRE_EXTRA_STUDY_DATA
4187 PCRE_EXTRA_MATCH_LIMIT
4188 PCRE_EXTRA_CALLOUT_DATA
4189
4190 The pcre_study() function now returns one of these new pcre_extra blocks, with
4191 the actual study data pointed to by the study_data field, and the
4192 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4193 before. That is, this change is entirely upwards-compatible and requires no
4194 change to existing code.
4195
4196 If you want to pass in additional data to pcre_exec(), you can either place it
4197 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4198 block.
4199
4200 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4201 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4202 times with different match limits, until it finds the minimum value needed for
4203 pcre_exec() to complete. The value is then output. This can be instructive; for
4204 most simple matches the number is quite small, but for pathological cases it
4205 gets very large very quickly.
4206
4207 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4208 returns the size of the data block pointed to by the study_data field in a
4209 pcre_extra block, that is, the value that was passed as the argument to
4210 pcre_malloc() when PCRE was getting memory in which to place the information
4211 created by pcre_study(). The fourth argument should point to a size_t variable.
4212 pcretest has been extended so that this information is shown after a successful
4213 pcre_study() call when information about the compiled regex is being displayed.
4214
4215 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4216 because what follows is always an absolute path. (Later: it turns out that this
4217 is more than cosmetic for MinGW, because it doesn't like empty path
4218 components.)
4219
4220 49. Some changes have been made to the callout feature (see 28 above):
4221
4222 (i) A callout function now has three choices for what it returns:
4223
4224 0 => success, carry on matching
4225 > 0 => failure at this point, but backtrack if possible
4226 < 0 => serious error, return this value from pcre_exec()
4227
4228 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4229 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4230 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4231 use by callout functions. It will never be used by PCRE itself.
4232
4233 (ii) The pcre_extra structure (see 45 above) has a void * field called
4234 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4235 pcre_callout_block structure has a field of the same name. The contents of
4236 the field passed in the pcre_extra structure are passed to the callout
4237 function in the corresponding field in the callout block. This makes it
4238 easier to use the same callout-containing regex from multiple threads. For
4239 testing, the pcretest program has a new data escape
4240
4241 \C*n pass the number n (may be negative) as callout_data
4242
4243 If the callout function in pcretest receives a non-zero value as
4244 callout_data, it returns that value.
4245
4246 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4247 there were some redundant $(CFLAGS) in commands that are now specified as
4248 $(LINK), which already includes $(CFLAGS).
4249
4250 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4251 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4252 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4253 one-byte characters throughout. Note that case-insensitive matching applies
4254 only to characters whose values are less than 256. PCRE doesn't support the
4255 notion of cases for higher-valued characters.
4256
4257 (i) A character class whose characters are all within 0-255 is handled as
4258 a bit map, and the map is inverted for negative classes. Previously, a
4259 character > 255 always failed to match such a class; however it should
4260 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4261
4262 (ii) A negated character class with a single character < 255 is coded as
4263 "not this character" (OP_NOT). This wasn't working properly when the test
4264 character was multibyte, either singly or repeated.
4265
4266 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4267 mode, for example: \x{100}{2,3}.
4268
4269 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4270 singly or repeated) now correctly test multibyte characters. However,
4271 PCRE doesn't recognize any characters with values greater than 255 as
4272 digits, spaces, or word characters. Such characters always match \D, \S,
4273 and \W, and never match \d, \s, or \w.
4274
4275 (v) Classes may now contain characters and character ranges with values
4276 greater than 255. For example: [ab\x{100}-\x{400}].
4277
4278 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4279 PCRE in UTF-8 mode.
4280
4281 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4282 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4283 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4284 value.)
4285
4286 53. The single man page has become too large. I have therefore split it up into
4287 a number of separate man pages. These also give rise to individual HTML pages;
4288 these are now put in a separate directory, and there is an index.html page that
4289 lists them all. Some hyperlinking between the pages has been installed.
4290
4291 54. Added convenience functions for handling named capturing parentheses.
4292
4293 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4294 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4295 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4296 are faulted.
4297
4298 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4299 calling configure. These values are used when compiling the dftables.c program
4300 which is run to generate the source of the default character tables. They
4301 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4302 you will need to set these values.
4303
4304 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4305
4306
4307 Version 3.9 02-Jan-02
4308 ---------------------
4309
4310 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4311
4312 2. If --disable-static was given, the building process failed when trying to
4313 build pcretest and pcregrep. (For some reason it was using libtool to compile
4314 them, which is not right, as they aren't part of the library.)
4315
4316
4317 Version 3.8 18-Dec-01
4318 ---------------------
4319
4320 1. The experimental UTF-8 code was completely screwed up. It was packing the
4321 bytes in the wrong order. How dumb can you get?
4322
4323
4324 Version 3.7 29-Oct-01
4325 ---------------------
4326
4327 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4328 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4329 this didn't happen under Solaris 8, where I normally test things.
4330
4331 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4332 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4333 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4334 this fix an hour or so after the initial 3.7 release.)
4335
4336
4337 Version 3.6 23-Oct-01
4338 ---------------------
4339
4340 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4341 offsets passed as NULL with zero offset count.
4342
4343 2. The config.guess and config.sub files had not been updated when I moved to
4344 the latest autoconf.
4345
4346
4347 Version 3.5 15-Aug-01
4348 ---------------------
4349
4350 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4351 had been forgotten.
4352
4353 2. By using declared but undefined structures, we can avoid using "void"
4354 definitions in pcre.h while keeping the internal definitions of the structures
4355 private.
4356
4357 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4358 user point of view, this means that both static and shared libraries are built
4359 by default, but this can be individually controlled. More of the work of
4360 handling this static/shared cases is now inside libtool instead of PCRE's make
4361 file.
4362
4363 4. The pcretest utility is now installed along with pcregrep because it is
4364 useful for users (to test regexs) and by doing this, it automatically gets
4365 relinked by libtool. The documentation has been turned into a man page, so
4366 there are now .1, .txt, and .html versions in /doc.
4367
4368 5. Upgrades to pcregrep:
4369 (i) Added long-form option names like gnu grep.
4370 (ii) Added --help to list all options with an explanatory phrase.
4371 (iii) Added -r, --recursive to recurse into sub-directories.
4372 (iv) Added -f, --file to read patterns from a file.
4373
4374 6. pcre_exec() was referring to its "code" argument before testing that
4375 argument for NULL (and giving an error if it was NULL).
4376
4377 7. Upgraded Makefile.in to allow for compiling in a different directory from
4378 the source directory.
4379
4380 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4381 options bits, the pointer it was passed was to an int instead of to an unsigned
4382 long int. This mattered only on 64-bit systems.
4383
4384 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4385 generated) instead of pcre.in, which it its source. Also made the same change
4386 in several of the .c files.
4387
4388 10. A new release of gcc defines printf() as a macro, which broke pcretest
4389 because it had an ifdef in the middle of a string argument for printf(). Fixed
4390 by using separate calls to printf().
4391
4392 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4393 script, to force use of CR or LF instead of \n in the source. On non-Unix
4394 systems, the value can be set in config.h.
4395
4396 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4397 absolute limit. Changed the text of the error message to make this clear, and
4398 likewise updated the man page.
4399
4400 13. The limit of 99 on the number of capturing subpatterns has been removed.
4401 The new limit is 65535, which I hope will not be a "real" limit.
4402
4403
4404 Version 3.4 22-Aug-00
4405 ---------------------
4406
4407 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4408
4409 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4410
4411
4412 Version 3.3 01-Aug-00
4413 ---------------------
4414
4415 1. If an octal character was given, but the value was greater than \377, it
4416 was not getting masked to the least significant bits, as documented. This could
4417 lead to crashes in some systems.
4418
4419 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4420 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4421
4422 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4423 These just pass their arguments on to (pcre_free)(), but they are provided
4424 because some uses of PCRE bind it to non-C systems that can call its functions,
4425 but cannot call free() or pcre_free() directly.
4426
4427 4. Add "make test" as a synonym for "make check". Corrected some comments in
4428 the Makefile.
4429
4430 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4431 Makefile.
4432
4433 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4434 command called pgrep for grepping around the active processes.
4435
4436 7. Added the beginnings of support for UTF-8 character strings.
4437
4438 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4439 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4440 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4441 out for the ar command.)
4442
4443
4444 Version 3.2 12-May-00
4445 ---------------------
4446
4447 This is purely a bug fixing release.
4448
4449 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4450 of ZA. This was just one example of several cases that could provoke this bug,
4451 which was introduced by change 9 of version 2.00. The code for breaking
4452 infinite loops after an iteration that matches an empty string was't working
4453 correctly.
4454
4455 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4456 when matched against abbab (for example). After matching an empty string, it
4457 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4458 caused it to match further down the string than it should.
4459
4460 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4461 was there because it doesn't seem to be needed, and it causes trouble on some
4462 systems, as it is not a Standard C header. It has been removed.
4463
4464 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4465 were reported on the Macintosh. The changes were from
4466
4467 while ((c = *(++ptr)) != 0 && c != '\n');
4468 to
4469 while ((c = *(++ptr)) != 0 && c != '\n') ;
4470
4471 Totally extraordinary, but if that's what it takes...
4472
4473 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4474 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4475 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4476 assumes the way PCRE uses memmove() (always moving upwards).
4477
4478 6. PCRE is being used in one environment where strchr() is not available. There
4479 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4480 faster code anyway.
4481
4482
4483 Version 3.1 09-Feb-00
4484 ---------------------
4485
4486 The only change in this release is the fixing of some bugs in Makefile.in for
4487 the "install" target:
4488
4489 (1) It was failing to install pcreposix.h.
4490
4491 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4492
4493
4494 Version 3.0 01-Feb-00
4495 ---------------------
4496
4497 1. Add support for the /+ modifier to perltest (to output $` like it does in
4498 pcretest).
4499
4500 2. Add support for the /g modifier to perltest.
4501
4502 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4503 matches null strings.
4504
4505 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4506 pattern. Perl treats empty patterns specially - it reuses the most recent
4507 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4508 effect.
4509
4510 5. The POSIX interface was broken in that it was just handing over the POSIX
4511 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4512 required a bigger vector, with some working space on the end. This means that
4513 the POSIX wrapper now has to get and free some memory, and copy the results.
4514
4515 6. Added some simple autoconf support, placing the test data and the
4516 documentation in separate directories, re-organizing some of the
4517 information files, and making it build pcre-config (a GNU standard). Also added
4518 libtool support for building PCRE as a shared library, which is now the
4519 default.
4520
4521 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4522 09 are not valid octal constants. Single digits will be used for minor values
4523 less than 10.
4524
4525 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4526 existing programs that set these in the POSIX interface can use PCRE without
4527 modification.
4528
4529 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4530 return all that pcre_info() returns, plus additional data. The pcre_info()
4531 function is retained for compatibility, but is considered to be obsolete.
4532
4533 10. Added experimental recursion feature (?R) to handle one common case that
4534 Perl 5.6 will be able to do with (?p{...}).
4535
4536 11. Added support for POSIX character classes like [:alpha:], which Perl is
4537 adopting.
4538
4539
4540 Version 2.08 31-Aug-99
4541 ----------------------
4542
4543 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4544 trying to match at the startoffset position, but instead was moving forward to
4545 the next newline as if a previous match had failed.
4546
4547 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4548 and could get into a loop if a null string was matched other than at the start
4549 of the subject.
4550
4551 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4552 be distinguished at compile time, and for completeness also added PCRE_DATE.
4553
4554 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4555 in GnuWin32 environments.
4556
4557
4558 Version 2.07 29-Jul-99
4559 ----------------------
4560
4561 1. The documentation is now supplied in plain text form and HTML as well as in
4562 the form of man page sources.
4563
4564 2. C++ compilers don't like assigning (void *) values to other pointer types.
4565 In particular this affects malloc(). Although there is no problem in Standard
4566 C, I've put in casts to keep C++ compilers happy.
4567
4568 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4569 should be (const char *).
4570
4571 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4572 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4573 However, I haven't made this a standard facility. The documentation doesn't
4574 mention it, and the Makefile doesn't support it.
4575
4576 5. The Makefile now contains an "install" target, with editable destinations at
4577 the top of the file. The pcretest program is not installed.
4578
4579 6. pgrep -V now gives the PCRE version number and date.
4580
4581 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4582 causing the entire string to be ignored, instead of just the last character.
4583
4584 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4585 non-matching string, it can take a very, very long time, even for strings of
4586 quite modest length, because of the nested recursion. PCRE now does better in
4587 some of these cases. It does this by remembering the last required literal
4588 character in the pattern, and pre-searching the subject to ensure it is present
4589 before running the real match. In other words, it applies a heuristic to detect
4590 some types of certain failure quickly, and in the above example, if presented
4591 with a string that has no trailing " it gives "no match" very quickly.
4592
4593 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4594 other alternatives are tried instead.
4595
4596
4597 Version 2.06 09-Jun-99
4598 ----------------------
4599
4600 1. Change pcretest's output for amount of store used to show just the code
4601 space, because the remainder (the data block) varies in size between 32-bit and
4602 64-bit systems.
4603
4604 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4605 start matching at. This allows lookbehinds to work when searching for multiple
4606 occurrences in a string.
4607
4608 3. Added additional options to pcretest for testing multiple occurrences:
4609
4610 /+ outputs the rest of the string that follows a match
4611 /g loops for multiple occurrences, using the new startoffset argument
4612 /G loops for multiple occurrences by passing an incremented pointer
4613
4614 4. PCRE wasn't doing the "first character" optimization for patterns starting
4615 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4616 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4617 the letter 'x'. On long subject strings, this gives a significant speed-up.
4618
4619
4620 Version 2.05 21-Apr-99
4621 ----------------------
4622
4623 1. Changed the type of magic_number from int to long int so that it works
4624 properly on 16-bit systems.
4625
4626 2. Fixed a bug which caused patterns starting with .* not to work correctly
4627 when the subject string contained newline characters. PCRE was assuming
4628 anchoring for such patterns in all cases, which is not correct because .* will
4629 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4630 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4631 must be retried after every newline in the subject.
4632
4633
4634 Version 2.04 18-Feb-99
4635 ----------------------
4636
4637 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4638 computation of the store needed to hold the pattern was incorrect (too large).
4639 If such patterns were nested a few deep, this could multiply and become a real
4640 problem.
4641
4642 2. Added /M option to pcretest to show the memory requirement of a specific
4643 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4644
4645 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4646 compiled in such a way that the backtracking after subsequent failure was
4647 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4648 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4649
4650
4651 Version 2.03 02-Feb-99
4652 ----------------------
4653
4654 1. Fixed typo and small mistake in man page.
4655
4656 2. Added 4th condition (GPL supersedes if conflict) and created separate
4657 LICENCE file containing the conditions.
4658
4659 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4660 Perl, that is the internal \ allows the delimiter to be included in the
4661 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4662 the final delimiter, add \ to the end of the pattern (to test the error).
4663
4664 4. Added the convenience functions for extracting substrings after a successful
4665 match. Updated pcretest to make it able to test these functions.
4666
4667
4668 Version 2.02 14-Jan-99
4669 ----------------------
4670
4671 1. Initialized the working variables associated with each extraction so that
4672 their saving and restoring doesn't refer to uninitialized store.
4673
4674 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4675 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4676 fix the problem.
4677
4678 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4679 calls, and wasn't printing the correct value for compiling calls. Increased the
4680 default value of LOOPREPEAT, and the number of significant figures in the
4681 times.
4682
4683 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4684
4685 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4686 a building problem on Windows NT with a FAT file system.
4687
4688
4689 Version 2.01 21-Oct-98
4690 ----------------------
4691
4692 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4693 to character tables built by pcre_maketables() in the current locale. If NULL
4694 is passed, the default tables are used.
4695
4696
4697 Version 2.00 24-Sep-98
4698 ----------------------
4699
4700 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4701 it any more.
4702
4703 2. Allow quantification of (?>) groups, and make it work correctly.
4704
4705 3. The first character computation wasn't working for (?>) groups.
4706
4707 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4708 end of the subject) and add 5.005's \z, which really does match only at the
4709 very end of the subject.
4710
4711 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4712
4713 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4714 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4715 localized options. All options to pcre_study() were also removed.
4716
4717 7. Add other new features from 5.005:
4718
4719 $(?<= positive lookbehind
4720 $(?<! negative lookbehind
4721 (?imsx-imsx) added the unsetting capability
4722 such a setting is global if at outer level; local otherwise
4723 (?imsx-imsx:) non-capturing groups with option setting
4724 (?(cond)re|re) conditional pattern matching
4725
4726 A backreference to itself in a repeated group matches the previous
4727 captured string.
4728
4729 8. General tidying up of studying (both automatic and via "study")
4730 consequential on the addition of new assertions.
4731
4732 9. As in 5.005, unlimited repeated groups that could match an empty substring
4733 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4734 runtime if any iteration does actually match an empty substring.
4735
4736 10. Include the RunTest script in the distribution.
4737
4738 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4739 discrepancies, some of which were old and were also with respect to 5.004. They
4740 have now been fixed.
4741
4742
4743 Version 1.09 28-Apr-98
4744 ----------------------
4745
4746 1. A negated single character class followed by a quantifier with a minimum
4747 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4748 program crashes, or just wrong answers. This did not apply to negated classes
4749 containing more than one character, or to minima other than one.
4750
4751
4752 Version 1.08 27-Mar-98
4753 ----------------------
4754
4755 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4756
4757 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4758 latter must appear before anything that relies on it in the pattern.
4759
4760
4761 Version 1.07 16-Feb-98
4762 ----------------------
4763
4764 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4765 repeat of a potentially empty string).
4766
4767
4768 Version 1.06 23-Jan-98
4769 ----------------------
4770
4771 1. Added Markus Oberhumer's little patches for C++.
4772
4773 2. Literal strings longer than 255 characters were broken.
4774
4775
4776 Version 1.05 23-Dec-97
4777 ----------------------
4778
4779 1. Negated character classes containing more than one character were failing if
4780 PCRE_CASELESS was set at run time.
4781
4782
4783 Version 1.04 19-Dec-97
4784 ----------------------
4785
4786 1. Corrected the man page, where some "const" qualifiers had been omitted.
4787
4788 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4789 input syntax.
4790
4791 3. Fixed memory leak which occurred when a regex with back references was
4792 matched with an offsets vector that wasn't big enough. The temporary memory
4793 that is used in this case wasn't being freed if the match failed.
4794
4795 4. Tidied pcretest to ensure it frees memory that it gets.
4796
4797 5. Temporary memory was being obtained in the case where the passed offsets
4798 vector was exactly big enough.
4799
4800 6. Corrected definition of offsetof() from change 5 below.
4801
4802 7. I had screwed up change 6 below and broken the rules for the use of
4803 setjmp(). Now fixed.
4804
4805
4806 Version 1.03 18-Dec-97
4807 ----------------------
4808
4809 1. A erroneous regex with a missing opening parenthesis was correctly
4810 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4811 on some systems.
4812
4813 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4814 it was reported that one broken compiler failed on the former because "code" is
4815 also an independent variable.
4816
4817 3. The erroneous regex a[]b caused an array overrun reference.
4818
4819 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4820 fail on data ending with that character. (It was going on too far, and checking
4821 the next character, typically a binary zero.) This was specific to the
4822 optimized code for single-character negative classes.
4823
4824 5. Added a contributed patch from the TIN world which does the following:
4825
4826 + Add an undef for memmove, in case the the system defines a macro for it.
4827
4828 + Add a definition of offsetof(), in case there isn't one. (I don't know
4829 the reason behind this - offsetof() is part of the ANSI standard - but
4830 it does no harm).
4831
4832 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4833 most of the places where whitespace preceded '#'. I have given up and
4834 allowed the remaining 2 cases to be at the margin.
4835
4836 + Rename some variables in pcre to eliminate shadowing. This seems very
4837 pedantic, but does no harm, of course.
4838
4839 6. Moved the call to setjmp() into its own function, to get rid of warnings
4840 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4841
4842 7. Constructs such as \d{8,} were compiling into the equivalent of
4843 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4844 outcome, but in this particular case used more store than had been allocated,
4845 which caused the bug to be discovered because it threw up an internal error.
4846
4847 8. The debugging code in both pcre and pcretest for outputting the compiled
4848 form of a regex was going wrong in the case of back references followed by
4849 curly-bracketed repeats.
4850
4851
4852 Version 1.02 12-Dec-97
4853 ----------------------
4854
4855 1. Typos in pcre.3 and comments in the source fixed.
4856
4857 2. Applied a contributed patch to get rid of places where it used to remove
4858 'const' from variables, and fixed some signed/unsigned and uninitialized
4859 variable warnings.
4860
4861 3. Added the "runtest" target to Makefile.
4862
4863 4. Set default compiler flag to -O2 rather than just -O.
4864
4865
4866 Version 1.01 19-Nov-97
4867 ----------------------
4868
4869 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4870 like /([ab]*)*/, that is, for classes with more than one character in them.
4871
4872 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4873 as /((?>a*))*/ (a PCRE_EXTRA facility).
4874
4875
4876 Version 1.00 18-Nov-97
4877 ----------------------
4878
4879 1. Added compile-time macros to support systems such as SunOS4 which don't have
4880 memmove() or strerror() but have other things that can be used instead.
4881
4882 2. Arranged that "make clean" removes the executables.
4883
4884
4885 Version 0.99 27-Oct-97
4886 ----------------------
4887
4888 1. Fixed bug in code for optimizing classes with only one character. It was
4889 initializing a 32-byte map regardless, which could cause it to run off the end
4890 of the memory it had got.
4891
4892 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4893
4894
4895 Version 0.98 22-Oct-97
4896 ----------------------
4897
4898 1. Fixed bug in code for handling temporary memory usage when there are more
4899 back references than supplied space in the ovector. This could cause segfaults.
4900
4901
4902 Version 0.97 21-Oct-97
4903 ----------------------
4904
4905 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4906
4907 2. Optimized negated single characters not to use a bit map.
4908
4909 3. Brought error texts together as macro definitions; clarified some of them;
4910 fixed one that was wrong - it said "range out of order" when it meant "invalid
4911 escape sequence".
4912
4913 4. Changed some char * arguments to const char *.
4914
4915 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4916
4917 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4918 pcretest.
4919
4920
4921 Version 0.96 16-Oct-97
4922 ----------------------
4923
4924 1. Added a simple "pgrep" utility to the distribution.
4925
4926 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4927 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4928 where "ddd" means "one or more decimal digits".
4929
4930 3. Fixed serious bug. If a pattern had a back reference, but the call to
4931 pcre_exec() didn't supply a large enough ovector to record the related
4932 identifying subpattern, the match always failed. PCRE now remembers the number
4933 of the largest back reference, and gets some temporary memory in which to save
4934 the offsets during matching if necessary, in order to ensure that
4935 backreferences always work.
4936
4937 4. Increased the compatibility with Perl in a number of ways:
4938
4939 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4940 to request this handling. The option can be set at compile or exec time.
4941
4942 (b) $ matches before a terminating newline by default; an option
4943 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4944 mode). The option can be set at compile or exec time.
4945
4946 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4947 the same as Perl's. If the decimal number it represents is less than 10
4948 or there aren't that many previous left capturing parentheses, an octal
4949 escape is read. Inside a character class, it's always an octal escape,
4950 even if it is a single digit.
4951
4952 (d) An escaped but undefined alphabetic character is taken as a literal,
4953 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4954 escapes.
4955
4956 (e) {0} is now permitted. (The previous item is removed from the compiled
4957 pattern).
4958
4959 5. Changed all the names of code files so that the basic parts are no longer
4960 than 10 characters, and abolished the teeny "globals.c" file.
4961
4962 6. Changed the handling of character classes; they are now done with a 32-byte
4963 bit map always.
4964
4965 7. Added the -d and /D options to pcretest to make it possible to look at the
4966 internals of compilation without having to recompile pcre.
4967
4968
4969 Version 0.95 23-Sep-97
4970 ----------------------
4971
4972 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4973 \x20 at the start of a run of normal characters. These were being treated as
4974 real characters, instead of the source characters being re-checked.
4975
4976
4977 Version 0.94 18-Sep-97
4978 ----------------------
4979
4980 1. The functions are now thread-safe, with the caveat that the global variables
4981 containing pointers to malloc() and free() or alternative functions are the
4982 same for all threads.
4983
4984 2. Get pcre_study() to generate a bitmap of initial characters for non-
4985 anchored patterns when this is possible, and use it if passed to pcre_exec().
4986
4987
4988 Version 0.93 15-Sep-97
4989 ----------------------
4990
4991 1. /(b)|(:+)/ was computing an incorrect first character.
4992
4993 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4994 but not actually doing anything yet.
4995
4996 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4997 as Perl does (e.g. [-az] or [az-]).
4998
4999 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5000 all possible positions.
5001
5002 5. Split up into different modules to avoid including unneeded functions in a
5003 compiled binary. However, compile and exec are still in one module. The "study"
5004 function is split off.
5005
5006 6. The character tables are now in a separate module whose source is generated
5007 by an auxiliary program - but can then be edited by hand if required. There are
5008 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5009 toupper() in the code.
5010
5011 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5012 make them global. Abolish the function for setting them, as the caller can now
5013 set them directly.
5014
5015
5016 Version 0.92 11-Sep-97
5017 ----------------------
5018
5019 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5020 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5021
5022 2. Caseless matching was not working in character classes if the characters in
5023 the pattern were in upper case.
5024
5025 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5026
5027 4. Make PCRE_ANCHORED public and accept as a compile option.
5028
5029 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5030 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5031 pass them.
5032
5033 6. Give an error if bad option bits passed at compile or run time.
5034
5035 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5036 pcretest to cause it to pass that flag.
5037
5038 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5039 options, and the first character, if set.
5040
5041 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5042
5043
5044 Version 0.91 10-Sep-97
5045 ----------------------
5046
5047 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5048 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5049
5050 2. PCRE was looping on encountering an indefinitely repeated back reference to
5051 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5052 Perl does - treats the match as successful.
5053
5054 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5