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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1236 - (show annotations)
Sun Dec 30 17:09:59 2012 UTC (6 years, 10 months ago) by ph10
File size: 240815 byte(s)
Error occurred while calculating annotation data.
A minor change from uchar to unsigned int improves performance in some 
environments. Also correct a missing "unsigned".
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
26 Version 8.32 30-November-2012
27 -----------------------------
28
29 1. Improved JIT compiler optimizations for first character search and single
30 character iterators.
31
32 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
33 Patch by Daniel Richard G.
34
35 3. Single character iterator optimizations in the JIT compiler.
36
37 4. Improved JIT compiler optimizations for character ranges.
38
39 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
40 Reported by Giuseppe D'Angelo.
41
42 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
43 of a line, was being set incorrectly in cases where .* appeared inside
44 atomic brackets at the start of a pattern, or where there was a subsequent
45 *PRUNE or *SKIP.
46
47 7. Improved instruction cache flush for POWER/PowerPC.
48 Patch by Daniel Richard G.
49
50 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
51 grep:
52
53 (a) There is now no limit to the number of patterns to be matched.
54
55 (b) An error is given if a pattern is too long.
56
57 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
58 are now supported.
59
60 (d) --exclude-from and --include-from (multiple use) have been added.
61
62 (e) Exclusions and inclusions now apply to all files and directories, not
63 just to those obtained from scanning a directory recursively.
64
65 (f) Multiple uses of -f and --file-list are now supported.
66
67 (g) In a Windows environment, the default for -d has been changed from
68 "read" (the GNU grep default) to "skip", because otherwise the presence
69 of a directory in the file list provokes an error.
70
71 (h) The documentation has been revised and clarified in places.
72
73 9. Improve the matching speed of capturing brackets.
74
75 10. Changed the meaning of \X so that it now matches a Unicode extended
76 grapheme cluster.
77
78 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
79 out POSIX threads when JIT support is configured.
80
81 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
82
83 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
84 structure could go wrong in environments where size_t is not the same size
85 as int.
86
87 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
88 set.
89
90 15. The EBCDIC support had decayed; later updates to the code had included
91 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
92 general tidy up of EBCDIC-related issues, and the documentation was also
93 not quite right. There is now a test that can be run on ASCII systems to
94 check some of the EBCDIC-related things (but is it not a full test).
95
96 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
97 in a small tidy to the code.
98
99 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
100
101 18. If the --only-matching (-o) option in pcregrep is specified multiple
102 times, each one causes appropriate output. For example, -o1 -o2 outputs the
103 substrings matched by the 1st and 2nd capturing parentheses. A separating
104 string can be specified by --om-separator (default empty).
105
106 19. Improving the first n character searches.
107
108 20. Turn case lists for horizontal and vertical white space into macros so that
109 they are defined only once.
110
111 21. This set of changes together give more compatible Unicode case-folding
112 behaviour for characters that have more than one other case when UCP
113 support is available.
114
115 (a) The Unicode property table now has offsets into a new table of sets of
116 three or more characters that are case-equivalent. The MultiStage2.py
117 script that generates these tables (the pcre_ucd.c file) now scans
118 CaseFolding.txt instead of UnicodeData.txt for character case
119 information.
120
121 (b) The code for adding characters or ranges of characters to a character
122 class has been abstracted into a generalized function that also handles
123 case-independence. In UTF-mode with UCP support, this uses the new data
124 to handle characters with more than one other case.
125
126 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
127 whose other case is greater than 256 are now correctly matched
128 caselessly. Previously, the high codepoint matched the low one, but not
129 vice versa.
130
131 (d) The processing of \h, \H, \v, and \ in character classes now makes use
132 of the new class addition function, using character lists defined as
133 macros alongside the case definitions of 20 above.
134
135 (e) Caseless back references now work with characters that have more than
136 one other case.
137
138 (f) General caseless matching of characters with more than one other case
139 is supported.
140
141 22. Unicode character properties were updated from Unicode 6.2.0
142
143 23. Improved CMake support under Windows. Patch by Daniel Richard G.
144
145 24. Add support for 32-bit character strings, and UTF-32
146
147 25. Major JIT compiler update (code refactoring and bugfixing).
148 Experimental Sparc 32 support is added.
149
150 26. Applied a modified version of Daniel Richard G's patch to create
151 pcre.h.generic and config.h.generic by "make" instead of in the
152 PrepareRelease script.
153
154 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
155 pcre_compile.c when checking for a zero character.
156
157 28. Introducing a native interface for JIT. Through this interface, the compiled
158 machine code can be directly executed. The purpose of this interface is to
159 provide fast pattern matching, so several sanity checks are not performed.
160 However, feature tests are still performed. The new interface provides
161 1.4x speedup compared to the old one.
162
163 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
164 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
165 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
166
167 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
168 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
169 the "old" RFC 2279). Instead, it ended up passing a negative length to
170 pcre_exec().
171
172 31. Add support for GCC's visibility feature to hide internal functions.
173
174 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
175 "unknown -C option" after outputting 0 or 1.
176
177 33. There is now support for generating a code coverage report for the test
178 suite in environments where gcc is the compiler and lcov is installed. This
179 is mainly for the benefit of the developers.
180
181 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
182 unaddressable using valgrind annotations, allowing valgrind to detect
183 invalid memory accesses. This is mainly for the benefit of the developers.
184
185 25. (*UTF) can now be used to start a pattern in any of the three libraries.
186
187 26. Give configure error if --enable-cpp but no C++ compiler found.
188
189
190 Version 8.31 06-July-2012
191 -------------------------
192
193 1. Fixing a wrong JIT test case and some compiler warnings.
194
195 2. Removed a bashism from the RunTest script.
196
197 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
198 to unsigned type, result still unsigned" that was given by an MS compiler
199 on encountering the code "-sizeof(xxx)".
200
201 4. Partial matching support is added to the JIT compiler.
202
203 5. Fixed several bugs concerned with partial matching of items that consist
204 of more than one character:
205
206 (a) /^(..)\1/ did not partially match "aba" because checking references was
207 done on an "all or nothing" basis. This also applied to repeated
208 references.
209
210 (b) \R did not give a hard partial match if \r was found at the end of the
211 subject.
212
213 (c) \X did not give a hard partial match after matching one or more
214 characters at the end of the subject.
215
216 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
217 a partial match for the string "\r".
218
219 (e) When newline was set to CRLF, the metacharacter "." did not recognize
220 a partial match for a CR character at the end of the subject string.
221
222 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
223 running pcretest, the text "(JIT)" added to the output whenever JIT is
224 actually used to run the match.
225
226 7. Individual JIT compile options can be set in pcretest by following -s+[+]
227 or /S+[+] with a digit between 1 and 7.
228
229 8. OP_NOT now supports any UTF character not just single-byte ones.
230
231 9. (*MARK) control verb is now supported by the JIT compiler.
232
233 10. The command "./RunTest list" lists the available tests without actually
234 running any of them. (Because I keep forgetting what they all are.)
235
236 11. Add PCRE_INFO_MAXLOOKBEHIND.
237
238 12. Applied a (slightly modified) user-supplied patch that improves performance
239 when the heap is used for recursion (compiled with --disable-stack-for-
240 recursion). Instead of malloc and free for each heap frame each time a
241 logical recursion happens, frames are retained on a chain and re-used where
242 possible. This sometimes gives as much as 30% improvement.
243
244 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
245 call.
246
247 14. As documented, (*COMMIT) is now confined to within a positive assertion.
248
249 15. It is now possible to link pcretest with libedit as an alternative to
250 libreadline.
251
252 16. (*COMMIT) control verb is now supported by the JIT compiler.
253
254 17. The Unicode data tables have been updated to Unicode 6.1.0.
255
256 18. Added --file-list option to pcregrep.
257
258 19. Added binary file support to pcregrep, including the -a, --binary-files,
259 -I, and --text options.
260
261 20. The madvise function is renamed for posix_madvise for QNX compatibility
262 reasons. Fixed by Giuseppe D'Angelo.
263
264 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
265
266 22. Changed the option for creating symbolic links for 16-bit man pages from
267 -s to -sf so that re-installing does not cause issues.
268
269 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
270
271 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
272 matches in certain environments (the workspace was not being correctly
273 retained). Also added to pcre_dfa_exec() a simple plausibility check on
274 some of the workspace data at the beginning of a restart.
275
276 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
277 was not doing so when it should - probably a typo introduced by SVN 528
278 (change 8.10/14).
279
280 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
281 \w+ when the character tables indicated that \x{c4} was a word character.
282 There were several related cases, all because the tests for doing a table
283 lookup were testing for characters less than 127 instead of 255.
284
285 27. If a pattern contains capturing parentheses that are not used in a match,
286 their slots in the ovector are set to -1. For those that are higher than
287 any matched groups, this happens at the end of processing. In the case when
288 there were back references that the ovector was too small to contain
289 (causing temporary malloc'd memory to be used during matching), and the
290 highest capturing number was not used, memory off the end of the ovector
291 was incorrectly being set to -1. (It was using the size of the temporary
292 memory instead of the true size.)
293
294 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
295 ovector size, it uses memory at the end of the block that it has got.
296
297 29. Check for an overlong MARK name and give an error at compile time. The
298 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
299
300 30. JIT compiler update.
301
302 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
303 Rill for the patch.
304
305 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
306
307 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
308
309 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
310 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
311 was enabled.
312
313 35. Improve JIT code generation for greedy plus quantifier.
314
315 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
316 1 to "aa" instead of to an empty string. The bug affected repeated groups
317 that could potentially match an empty string.
318
319 37. Optimizing single character iterators in JIT.
320
321 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
322 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
323 codepoints that are too big for the mode are faulted, and in a UTF mode,
324 disallowed codepoints are also faulted.
325
326 39. If PCRE was compiled with UTF support, in three places in the DFA
327 matcher there was code that should only have been obeyed in UTF mode, but
328 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
329 processing when bytes with values greater than 127 were present. In 16-bit
330 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
331 both cases the values are those that cannot be the first data item in a UTF
332 character. The three items that might have provoked this were recursions,
333 possessively repeated groups, and atomic groups.
334
335 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
336 and pcregrep, because some OS require shared objects to be explicitly
337 passed to ld, causing the link step to fail if they are not.
338
339 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
340 mode, patterns that started with \h* or \R* might be incorrectly matched.
341
342
343 Version 8.30 04-February-2012
344 -----------------------------
345
346 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
347 name is defined in ctype.h.
348
349 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
350 only in quite long subpatterns.
351
352 3. Removed the function pcre_info(), which has been obsolete and deprecated
353 since it was replaced by pcre_fullinfo() in February 2000.
354
355 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
356 match a (*MARK), and the match failed at the start of the subject, a
357 reference to memory before the start of the subject could occur. This bug
358 was introduced by fix 17 of release 8.21.
359
360 5. A reference to an unset group with zero minimum repetition was giving
361 totally wrong answers (in non-JavaScript-compatibility mode). For example,
362 /(another)?(\1?)test/ matched against "hello world test". This bug was
363 introduced in release 8.13.
364
365 6. Add support for 16-bit character strings (a large amount of work involving
366 many changes and refactorings).
367
368 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
369 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
370 from a file.
371
372 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
373 rounding is not applied in this particular case).
374
375 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
376 if they appear, or are escaped, in patterns.
377
378 10. Get rid of a number of -Wunused-but-set-variable warnings.
379
380 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
381 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
382 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
383 also returns the mark "x". This bug applied to capturing parentheses,
384 non-capturing parentheses, and atomic parentheses. It also applied to some
385 assertions.
386
387 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
388 information out of configure.ac instead of relying on pcre.h.generic, which
389 is not stored in the repository.
390
391 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
392 -lreadline.
393
394 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
395
396 15. Applied Graycode's patch to put the top-level frame on the stack rather
397 than the heap when not using the stack for recursion. This gives a
398 performance improvement in many cases when recursion is not deep.
399
400 16. Experimental code added to "pcretest -C" to output the stack frame size.
401
402
403 Version 8.21 12-Dec-2011
404 ------------------------
405
406 1. Updating the JIT compiler.
407
408 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
409 are added as well.
410
411 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
412 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
413 calling _pcre_jit_exec. Some extra comments are added.
414
415 4. (*MARK) settings inside atomic groups that do not contain any capturing
416 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
417 was introduced by change 18 for 8.20.
418
419 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
420 ECMA-262 standard.
421
422 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
423 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
424 This bug was probably introduced by change 9 of 8.13.
425
426 7. While fixing 6 above, I noticed that a number of other items were being
427 incorrectly rejected as "not fixed length". This arose partly because newer
428 opcodes had not been added to the fixed-length checking code. I have (a)
429 corrected the bug and added tests for these items, and (b) arranged for an
430 error to occur if an unknown opcode is encountered while checking for fixed
431 length instead of just assuming "not fixed length". The items that were
432 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
433 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
434 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
435
436 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
437 being incorrectly compiled and would have given unpredicatble results.
438
439 9. A possessively repeated subpattern with minimum repeat count greater than
440 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
441 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
442 the first (A) could occur when it should not.
443
444 10. Add a cast and remove a redundant test from the code.
445
446 11. JIT should use pcre_malloc/pcre_free for allocation.
447
448 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
449 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
450 is anything other than /usr, -L is still shown).
451
452 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
453
454 14. Perl does not support \N without a following name in a [] class; PCRE now
455 also gives an error.
456
457 15. If a forward reference was repeated with an upper limit of around 2000,
458 it caused the error "internal error: overran compiling workspace". The
459 maximum number of forward references (including repeats) was limited by the
460 internal workspace, and dependent on the LINK_SIZE. The code has been
461 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
462 the default depends on LINK_SIZE. There is a new upper limit (for safety)
463 of around 200,000 forward references. While doing this, I also speeded up
464 the filling in of repeated forward references.
465
466 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
467 incorrectly expecting the subject to contain another "a" after the start.
468
469 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
470 in the match, the SKIP should be ignored. This was not happening; instead
471 the SKIP was being treated as NOMATCH. For patterns such as
472 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
473 tested.
474
475 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
476 now much more compatible with Perl, in particular in cases where the result
477 is a non-match for a non-anchored pattern. For example, if
478 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
479 "m", where previously it did not return a name. A side effect of this
480 change is that for partial matches, the last encountered mark name is
481 returned, as for non matches. A number of tests that were previously not
482 Perl-compatible have been moved into the Perl-compatible test files. The
483 refactoring has had the pleasing side effect of removing one argument from
484 the match() function, thus reducing its stack requirements.
485
486 19. If the /S+ option was used in pcretest to study a pattern using JIT,
487 subsequent uses of /S (without +) incorrectly behaved like /S+.
488
489 21. Retrieve executable code size support for the JIT compiler and fixing
490 some warnings.
491
492 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
493 not work when the shorter character appeared right at the end of the
494 subject string.
495
496 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
497 systems.
498
499 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
500 output it when the /M option is used in pcretest.
501
502 25. The CheckMan script was not being included in the distribution. Also, added
503 an explicit "perl" to run Perl scripts from the PrepareRelease script
504 because this is reportedly needed in Windows.
505
506 26. If study data was being save in a file and studying had not found a set of
507 "starts with" bytes for the pattern, the data written to the file (though
508 never used) was taken from uninitialized memory and so caused valgrind to
509 complain.
510
511 27. Updated RunTest.bat as provided by Sheri Pierce.
512
513 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
514
515 29. Computation of memory usage for the table of capturing group names was
516 giving an unnecessarily large value.
517
518
519 Version 8.20 21-Oct-2011
520 ------------------------
521
522 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
523 a POSIX class. After further experiments with Perl, which convinced me that
524 Perl has bugs and confusions, a closing square bracket is no longer allowed
525 in a POSIX name. This bug also affected patterns with classes that started
526 with full stops.
527
528 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
529 captured substring, but while checking the failing first alternative,
530 substring 1 is temporarily captured. If the output vector supplied to
531 pcre_exec() was not big enough for this capture, the yield of the function
532 was still zero ("insufficient space for captured substrings"). This cannot
533 be totally fixed without adding another stack variable, which seems a lot
534 of expense for a edge case. However, I have improved the situation in cases
535 such as /(a)(b)x|abc/ matched against "abc", where the return code
536 indicates that fewer than the maximum number of slots in the ovector have
537 been set.
538
539 3. Related to (2) above: when there are more back references in a pattern than
540 slots in the output vector, pcre_exec() uses temporary memory during
541 matching, and copies in the captures as far as possible afterwards. It was
542 using the entire output vector, but this conflicts with the specification
543 that only 2/3 is used for passing back captured substrings. Now it uses
544 only the first 2/3, for compatibility. This is, of course, another edge
545 case.
546
547 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
548 main code base, and can be used by building with --enable-jit. When this is
549 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
550 runtime --no-jit option is given.
551
552 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
553 ovector, the return from the function was zero, implying that there were
554 other matches that did not fit. The correct "exactly full" value is now
555 returned.
556
557 6. If a subpattern that was called recursively or as a subroutine contained
558 (*PRUNE) or any other control that caused it to give a non-standard return,
559 invalid errors such as "Error -26 (nested recursion at the same subject
560 position)" or even infinite loops could occur.
561
562 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
563 computing the minimum length on reaching *ACCEPT, and so ended up with the
564 wrong value of 1 rather than 0. Further investigation indicates that
565 computing a minimum subject length in the presence of *ACCEPT is difficult
566 (think back references, subroutine calls), and so I have changed the code
567 so that no minimum is registered for a pattern that contains *ACCEPT.
568
569 8. If (*THEN) was present in the first (true) branch of a conditional group,
570 it was not handled as intended. [But see 16 below.]
571
572 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
573 Sheri Pierce.
574
575 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
576 the first byte in a match must be "a".
577
578 11. Change 17 for 8.13 increased the recursion depth for patterns like
579 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
580 pattern contains any instances of (*THEN). If it does not, the old
581 optimizations are restored. It would be nice to do this on a per-group
582 basis, but at the moment that is not feasible.
583
584 12. In some environments, the output of pcretest -C is CRLF terminated. This
585 broke RunTest's code that checks for the link size. A single white space
586 character after the value is now allowed for.
587
588 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
589 For "fr", it uses the Windows-specific input and output files.
590
591 14. If (*THEN) appeared in a group that was called recursively or as a
592 subroutine, it did not work as intended. [But see next item.]
593
594 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
595 pattern fragments (but not containing any | characters). If A and B are
596 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
597 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
598 D. In other words, Perl considers parentheses that do not contain any |
599 characters to be part of a surrounding alternative, whereas PCRE was
600 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
601 differently. PCRE now behaves in the same way as Perl, except in the case
602 of subroutine/recursion calls such as (?1) which have in any case always
603 been different (but PCRE had them first :-).
604
605 16. Related to 15 above: Perl does not treat the | in a conditional group as
606 creating alternatives. Such a group is treated in the same way as an
607 ordinary group without any | characters when processing (*THEN). PCRE has
608 been changed to match Perl's behaviour.
609
610 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
611 RunGrepTest script failed.
612
613 18. Change 22 for version 13 caused atomic groups to use more stack. This is
614 inevitable for groups that contain captures, but it can lead to a lot of
615 stack use in large patterns. The old behaviour has been restored for atomic
616 groups that do not contain any capturing parentheses.
617
618 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
619 suppress the check for a minimum subject length at run time. (If it was
620 given to pcre_exec() or pcre_dfa_exec() it did work.)
621
622 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
623 fail to work when decoding hex characters in data strings in EBCDIC
624 environments.
625
626 21. It appears that in at least one Mac OS environment, the isxdigit() function
627 is implemented as a macro that evaluates to its argument more than once,
628 contravening the C 90 Standard (I haven't checked a later standard). There
629 was an instance in pcretest which caused it to go wrong when processing
630 \x{...} escapes in subject strings. The has been rewritten to avoid using
631 things like p++ in the argument of isxdigit().
632
633
634 Version 8.13 16-Aug-2011
635 ------------------------
636
637 1. The Unicode data tables have been updated to Unicode 6.0.0.
638
639 2. Two minor typos in pcre_internal.h have been fixed.
640
641 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
642 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
643 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
644
645 4. There were a number of related bugs in the code for matching backrefences
646 caselessly in UTF-8 mode when codes for the characters concerned were
647 different numbers of bytes. For example, U+023A and U+2C65 are an upper
648 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
649 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
650 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
651 2-byte code at the end of the subject (it thought there wasn't enough data
652 left).
653
654 5. Comprehensive information about what went wrong is now returned by
655 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
656 as the output vector has at least 2 elements. The offset of the start of
657 the failing character and a reason code are placed in the vector.
658
659 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
660 now returned is for the first byte of the failing character, instead of the
661 last byte inspected. This is an incompatible change, but I hope it is small
662 enough not to be a problem. It makes the returned offset consistent with
663 pcre_exec() and pcre_dfa_exec().
664
665 7. pcretest now gives a text phrase as well as the error number when
666 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
667 failure, the offset and reason code are output.
668
669 8. When \R was used with a maximizing quantifier it failed to skip backwards
670 over a \r\n pair if the subsequent match failed. Instead, it just skipped
671 back over a single character (\n). This seems wrong (because it treated the
672 two characters as a single entity when going forwards), conflicts with the
673 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
674 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
675 has been changed.
676
677 9. Some internal refactoring has changed the processing so that the handling
678 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
679 time (the PCRE_DOTALL option was changed this way some time ago: version
680 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
681 which was always a bit of a fudge. It also means that there is one less
682 argument for the match() function, which reduces its stack requirements
683 slightly. This change also fixes an incompatibility with Perl: the pattern
684 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
685
686 10. More internal refactoring has drastically reduced the number of recursive
687 calls to match() for possessively repeated groups such as (abc)++ when
688 using pcre_exec().
689
690 11. While implementing 10, a number of bugs in the handling of groups were
691 discovered and fixed:
692
693 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
694 (a|)*(?1) gave a compile-time internal error.
695 ((a|)+)+ did not notice that the outer group could match an empty string.
696 (^a|^)+ was not marked as anchored.
697 (.*a|.*)+ was not marked as matching at start or after a newline.
698
699 12. Yet more internal refactoring has removed another argument from the match()
700 function. Special calls to this function are now indicated by setting a
701 value in a variable in the "match data" data block.
702
703 13. Be more explicit in pcre_study() instead of relying on "default" for
704 opcodes that mean there is no starting character; this means that when new
705 ones are added and accidentally left out of pcre_study(), testing should
706 pick them up.
707
708 14. The -s option of pcretest has been documented for ages as being an old
709 synonym of -m (show memory usage). I have changed it to mean "force study
710 for every regex", that is, assume /S for every regex. This is similar to -i
711 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
712 using it. It makes it easier to run collections of tests with and without
713 study enabled, and thereby test pcre_study() more easily. All the standard
714 tests are now run with and without -s (but some patterns can be marked as
715 "never study" - see 20 below).
716
717 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
718 restoration of the capturing data to the outer values was not happening
719 correctly.
720
721 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
722 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
723 pattern had matched an empty string, and so incorrectly returned a no
724 match.
725
726 17. There was optimizing code for the last branch of non-capturing parentheses,
727 and also for the obeyed branch of a conditional subexpression, which used
728 tail recursion to cut down on stack usage. Unfortunately, now that there is
729 the possibility of (*THEN) occurring in these branches, tail recursion is
730 no longer possible because the return has to be checked for (*THEN). These
731 two optimizations have therefore been removed. [But see 8.20/11 above.]
732
733 18. If a pattern containing \R was studied, it was assumed that \R always
734 matched two bytes, thus causing the minimum subject length to be
735 incorrectly computed because \R can also match just one byte.
736
737 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
738 was incorrectly computed.
739
740 20. If /S is present twice on a test pattern in pcretest input, it now
741 *disables* studying, thereby overriding the use of -s on the command line
742 (see 14 above). This is necessary for one or two tests to keep the output
743 identical in both cases.
744
745 21. When (*ACCEPT) was used in an assertion that matched an empty string and
746 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
747
748 22. When an atomic group that contained a capturing parenthesis was
749 successfully matched, but the branch in which it appeared failed, the
750 capturing was not being forgotten if a higher numbered group was later
751 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
752 group 1 to "a", when in fact it should be unset. This applied to multi-
753 branched capturing and non-capturing groups, repeated or not, and also to
754 positive assertions (capturing in negative assertions does not happen
755 in PCRE) and also to nested atomic groups.
756
757 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
758 subject after a captured substring, to make it easier to tell which of a
759 number of identical substrings has been captured.
760
761 24. The way atomic groups are processed by pcre_exec() has been changed so that
762 if they are repeated, backtracking one repetition now resets captured
763 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
764 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
765 "aaa". Previously, it would have been "a". As part of this code
766 refactoring, the way recursive calls are handled has also been changed.
767
768 25. If an assertion condition captured any substrings, they were not passed
769 back unless some other capturing happened later. For example, if
770 (?(?=(a))a) was matched against "a", no capturing was returned.
771
772 26. When studying a pattern that contained subroutine calls or assertions,
773 the code for finding the minimum length of a possible match was handling
774 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
775 group 1 called group 2 while simultaneously a separate group 2 called group
776 1). A stack overflow occurred in this case. I have fixed this by limiting
777 the recursion depth to 10.
778
779 27. Updated RunTest.bat in the distribution to the version supplied by Tom
780 Fortmann. This supports explicit test numbers on the command line, and has
781 argument validation and error reporting.
782
783 28. An instance of \X with an unlimited repeat could fail if at any point the
784 first character it looked at was a mark character.
785
786 29. Some minor code refactoring concerning Unicode properties and scripts
787 should reduce the stack requirement of match() slightly.
788
789 30. Added the '=' option to pcretest to check the setting of unused capturing
790 slots at the end of the pattern, which are documented as being -1, but are
791 not included in the return count.
792
793 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
794 compiled something random. Now it gives a compile-time error (as does
795 Perl).
796
797 32. A *MARK encountered during the processing of a positive assertion is now
798 recorded and passed back (compatible with Perl).
799
800 33. If --only-matching or --colour was set on a pcregrep call whose pattern
801 had alternative anchored branches, the search for a second match in a line
802 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
803 matched the line "0102" twice. The same bug affected patterns that started
804 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
805 twice.
806
807 34. Previously, PCRE did not allow quantification of assertions. However, Perl
808 does, and because of capturing effects, quantifying parenthesized
809 assertions may at times be useful. Quantifiers are now allowed for
810 parenthesized assertions.
811
812 35. A minor code tidy in pcre_compile() when checking options for \R usage.
813
814 36. \g was being checked for fancy things in a character class, when it should
815 just be a literal "g".
816
817 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
818 appearance of a nested POSIX class supersedes an apparent external class.
819 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
820 unescaped square brackets may also appear as part of class names. For
821 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
822 more like Perl. (But see 8.20/1 above.)
823
824 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
825 was because it thought it was \N{name}, which is not supported).
826
827 39. Add minix to OS list not supporting the -S option in pcretest.
828
829 40. PCRE tries to detect cases of infinite recursion at compile time, but it
830 cannot analyze patterns in sufficient detail to catch mutual recursions
831 such as ((?1))((?2)). There is now a runtime test that gives an error if a
832 subgroup is called recursively as a subpattern for a second time at the
833 same position in the subject string. In previous releases this might have
834 been caught by the recursion limit, or it might have run out of stack.
835
836 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
837 happen only once. PCRE was, however incorrectly giving a compile time error
838 "recursive call could loop indefinitely" because it cannot analyze the
839 pattern in sufficient detail. The compile time test no longer happens when
840 PCRE is compiling a conditional subpattern, but actual runaway loops are
841 now caught at runtime (see 40 above).
842
843 42. It seems that Perl allows any characters other than a closing parenthesis
844 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
845 has been changed to be the same.
846
847 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
848 as not to get warnings when autogen.sh is called. Also changed
849 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
850
851 44. To help people who use pcregrep to scan files containing exceedingly long
852 lines, the following changes have been made:
853
854 (a) The default value of the buffer size parameter has been increased from
855 8K to 20K. (The actual buffer used is three times this size.)
856
857 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
858 PCRE is built.
859
860 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
861 to be set at run time.
862
863 (d) Numerical values in pcregrep options can be followed by K or M, for
864 example --buffer-size=50K.
865
866 (e) If a line being scanned overflows pcregrep's buffer, an error is now
867 given and the return code is set to 2.
868
869 45. Add a pointer to the latest mark to the callout data block.
870
871 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
872 partial match of an empty string instead of no match. This was specific to
873 the use of ".".
874
875 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
876 complete match instead of a partial match. This bug was dependent on both
877 the PCRE_UTF8 and PCRE_DOTALL options being set.
878
879 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
880 starting byte set, because \b was not being ignored.
881
882
883 Version 8.12 15-Jan-2011
884 ------------------------
885
886 1. Fixed some typos in the markup of the man pages, and wrote a script that
887 checks for such things as part of the documentation building process.
888
889 2. On a big-endian 64-bit system, pcregrep did not correctly process the
890 --match-limit and --recursion-limit options (added for 8.11). In
891 particular, this made one of the standard tests fail. (The integer value
892 went into the wrong half of a long int.)
893
894 3. If the --colour option was given to pcregrep with -v (invert match), it
895 did strange things, either producing crazy output, or crashing. It should,
896 of course, ignore a request for colour when reporting lines that do not
897 match.
898
899 4. Another pcregrep bug caused similar problems if --colour was specified with
900 -M (multiline) and the pattern match finished with a line ending.
901
902 5. In pcregrep, when a pattern that ended with a literal newline sequence was
903 matched in multiline mode, the following line was shown as part of the
904 match. This seems wrong, so I have changed it.
905
906 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
907 the check for further matches in the same line (so they could be coloured)
908 to overrun the end of the current line. If another match was found, it was
909 incorrectly shown (and then shown again when found in the next line).
910
911 7. If pcregrep was compiled under Windows, there was a reference to the
912 function pcregrep_exit() before it was defined. I am assuming this was
913 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
914 reported by a user. I've moved the definition above the reference.
915
916
917 Version 8.11 10-Dec-2010
918 ------------------------
919
920 1. (*THEN) was not working properly if there were untried alternatives prior
921 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
922 backtracked to try for "b" instead of moving to the next alternative branch
923 at the same level (in this case, to look for "c"). The Perl documentation
924 is clear that when (*THEN) is backtracked onto, it goes to the "next
925 alternative in the innermost enclosing group".
926
927 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
928 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
929 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
930 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
931 (*THEN).
932
933 3. If \s appeared in a character class, it removed the VT character from
934 the class, even if it had been included by some previous item, for example
935 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
936 of \s, but is part of the POSIX "space" class.)
937
938 4. A partial match never returns an empty string (because you can always
939 match an empty string at the end of the subject); however the checking for
940 an empty string was starting at the "start of match" point. This has been
941 changed to the "earliest inspected character" point, because the returned
942 data for a partial match starts at this character. This means that, for
943 example, /(?<=abc)def/ gives a partial match for the subject "abc"
944 (previously it gave "no match").
945
946 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
947 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
948 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
949 has an implication that the given string is incomplete (because a partial
950 match is preferred over a full match). For this reason, these items now
951 give a partial match in this situation. [Aside: previously, the one case
952 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
953 match rather than a full match, which was wrong by the old rules, but is
954 now correct.]
955
956 6. There was a bug in the handling of #-introduced comments, recognized when
957 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
958 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
959 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
960 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
961 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
962 places in pcre_compile().
963
964 7. Related to (6) above, when pcre_compile() was skipping #-introduced
965 comments when looking ahead for named forward references to subpatterns,
966 the only newline sequence it recognized was NL. It now handles newlines
967 according to the set newline convention.
968
969 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
970 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
971 cater for a lack of strerror(). These oversights have been fixed.
972
973 9. Added --match-limit and --recursion-limit to pcregrep.
974
975 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
976
977 11. When the -o option was used, pcregrep was setting a return code of 1, even
978 when matches were found, and --line-buffered was not being honoured.
979
980 12. Added an optional parentheses number to the -o and --only-matching options
981 of pcregrep.
982
983 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
984 can match an empty string. The code to do it in pcretest and pcredemo
985 needed fixing:
986
987 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
988 only one byte after an empty string match just before CRLF (this case
989 just got forgotten; "any" and "anycrlf" were OK).
990
991 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
992 mode when an empty string match preceded an ASCII character followed by
993 a non-ASCII character. (The code for advancing by one character rather
994 than one byte was nonsense.)
995
996 (c) The pcredemo.c sample program did not have any code at all to handle
997 the cases when CRLF is a valid newline sequence.
998
999 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1000 as a starting offset was within the subject string. There is now a new
1001 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1002 negative or greater than the length of the string. In order to test this,
1003 pcretest is extended to allow the setting of negative starting offsets.
1004
1005 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1006 starting offset points to the beginning of a UTF-8 character was
1007 unnecessarily clumsy. I tidied it up.
1008
1009 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1010 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1011
1012 17. Nobody had reported that the --include_dir option, which was added in
1013 release 7.7 should have been called --include-dir (hyphen, not underscore)
1014 for compatibility with GNU grep. I have changed it to --include-dir, but
1015 left --include_dir as an undocumented synonym, and the same for
1016 --exclude-dir, though that is not available in GNU grep, at least as of
1017 release 2.5.4.
1018
1019 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1020 characters from a string of bytes) have been redefined so as not to use
1021 loops, in order to improve performance in some environments. At the same
1022 time, I abstracted some of the common code into auxiliary macros to save
1023 repetition (this should not affect the compiled code).
1024
1025 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1026 compile-time error is now given if \c is not followed by an ASCII
1027 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1028 different, and any byte value is allowed.)
1029
1030 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1031 START_OPTIMIZE option, which is now allowed at compile time - but just
1032 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1033 to pcregrep and other applications that have no direct access to PCRE
1034 options. The new /Y option in pcretest sets this option when calling
1035 pcre_compile().
1036
1037 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1038 back references. Groups containing recursive back references were forced to
1039 be atomic by that change, but in the case of named groups, the amount of
1040 memory required was incorrectly computed, leading to "Failed: internal
1041 error: code overflow". This has been fixed.
1042
1043 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1044 pcretest.c, to avoid build problems in some Borland environments.
1045
1046
1047 Version 8.10 25-Jun-2010
1048 ------------------------
1049
1050 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1051 THEN.
1052
1053 2. (*ACCEPT) was not working when inside an atomic group.
1054
1055 3. Inside a character class, \B is treated as a literal by default, but
1056 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1057 causes the error). The code is unchanged, but I tidied the documentation.
1058
1059 4. Inside a character class, PCRE always treated \R and \X as literals,
1060 whereas Perl faults them if its -w option is set. I have changed PCRE so
1061 that it faults them when PCRE_EXTRA is set.
1062
1063 5. Added support for \N, which always matches any character other than
1064 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1065
1066 6. When compiling pcregrep with newer versions of gcc which may have
1067 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1068 declared with attribute warn_unused_result" were given. Just casting the
1069 result to (void) does not stop the warnings; a more elaborate fudge is
1070 needed. I've used a macro to implement this.
1071
1072 7. Minor change to pcretest.c to avoid a compiler warning.
1073
1074 8. Added four artifical Unicode properties to help with an option to make
1075 \s etc use properties (see next item). The new properties are: Xan
1076 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1077
1078 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1079 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1080 this option. Modified pcretest to add /W to test this facility. Added
1081 REG_UCP to make it available via the POSIX interface.
1082
1083 10. Added --line-buffered to pcregrep.
1084
1085 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1086 studied, and the match started with a letter with a code point greater than
1087 127 whose first byte was different to the first byte of the other case of
1088 the letter, the other case of this starting letter was not recognized
1089 (#976).
1090
1091 12. If a pattern that was studied started with a repeated Unicode property
1092 test, for example, \p{Nd}+, there was the theoretical possibility of
1093 setting up an incorrect bitmap of starting bytes, but fortunately it could
1094 not have actually happened in practice until change 8 above was made (it
1095 added property types that matched character-matching opcodes).
1096
1097 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1098 possible starting bytes for non-anchored patterns.
1099
1100 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1101 \R, and also a number of cases that involve Unicode properties, both
1102 explicit and implicit when PCRE_UCP is set.
1103
1104 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1105 input, it could crash or give wrong results if characters with values
1106 greater than 0xc0 were present in the subject string. (Detail: it assumed
1107 UTF-8 input when processing these items.)
1108
1109 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1110 size_t is 64-bit (#991).
1111
1112 17. Added a check for running out of memory when PCRE is compiled with
1113 --disable-stack-for-recursion (#990).
1114
1115 18. If the last data line in a file for pcretest does not have a newline on
1116 the end, a newline was missing in the output.
1117
1118 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1119 less than 128) in its various bitmaps. However, there is a facility for
1120 generating tables according to the current locale when PCRE is compiled. It
1121 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1122 characters, are recognized by isspace() and therefore were getting set in
1123 these tables, and indeed these tables seem to approximate to ISO 8859. This
1124 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1125 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1126 which of course cannot start UTF-8 characters. I have changed the code so
1127 that only real ASCII characters (less than 128) and the correct starting
1128 bytes for UTF-8 encodings are set for characters greater than 127 when in
1129 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1130 altogether.)
1131
1132 20. Added the /T option to pcretest so as to be able to run tests with non-
1133 standard character tables, thus making it possible to include the tests
1134 used for 19 above in the standard set of tests.
1135
1136 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1137 reference to a subpattern the other side of a comment that contains an
1138 opening parenthesis caused either an internal compiling error, or a
1139 reference to the wrong subpattern.
1140
1141
1142 Version 8.02 19-Mar-2010
1143 ------------------------
1144
1145 1. The Unicode data tables have been updated to Unicode 5.2.0.
1146
1147 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1148 configured.
1149
1150 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1151 original author of that file, following a query about its status.
1152
1153 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1154 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1155
1156 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1157 quantifier applied to a forward-referencing subroutine call, could compile
1158 incorrect code or give the error "internal error: previously-checked
1159 referenced subpattern not found".
1160
1161 6. Both MS Visual Studio and Symbian OS have problems with initializing
1162 variables to point to external functions. For these systems, therefore,
1163 pcre_malloc etc. are now initialized to local functions that call the
1164 relevant global functions.
1165
1166 7. There were two entries missing in the vectors called coptable and poptable
1167 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1168 I've fixed the data, and added a kludgy way of testing at compile time that
1169 the lengths are correct (equal to the number of opcodes).
1170
1171 8. Following on from 7, I added a similar kludge to check the length of the
1172 eint vector in pcreposix.c.
1173
1174 9. Error texts for pcre_compile() are held as one long string to avoid too
1175 much relocation at load time. To find a text, the string is searched,
1176 counting zeros. There was no check for running off the end of the string,
1177 which could happen if a new error number was added without updating the
1178 string.
1179
1180 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1181
1182 11. \K was not working if it appeared in an atomic group or in a group that
1183 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1184 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1185 the assertion is positive, but not if it is negative.
1186
1187 12. Change 11 fortuitously reduced the size of the stack frame used in the
1188 "match()" function of pcre_exec.c by one pointer. Forthcoming
1189 implementation of support for (*MARK) will need an extra pointer on the
1190 stack; I have reserved it now, so that the stack frame size does not
1191 decrease.
1192
1193 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1194 item in branch that calls a recursion is a subroutine call - as in the
1195 second branch in the above example - was incorrectly given the compile-
1196 time error "recursive call could loop indefinitely" because pcre_compile()
1197 was not correctly checking the subroutine for matching a non-empty string.
1198
1199 14. The checks for overrunning compiling workspace could trigger after an
1200 overrun had occurred. This is a "should never occur" error, but it can be
1201 triggered by pathological patterns such as hundreds of nested parentheses.
1202 The checks now trigger 100 bytes before the end of the workspace.
1203
1204 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1205
1206
1207 Version 8.01 19-Jan-2010
1208 ------------------------
1209
1210 1. If a pattern contained a conditional subpattern with only one branch (in
1211 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1212 computed the wrong minimum data length (which is of course zero for such
1213 subpatterns). This could cause incorrect "no match" results.
1214
1215 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1216 the pattern is reset in the first branch, pcre_compile() failed with
1217 "internal error: code overflow at offset...". This happened only when
1218 the reset was to the original external option setting. (An optimization
1219 abstracts leading options settings into an external setting, which was the
1220 cause of this.)
1221
1222 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1223 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1224 assertion pattern did not match (meaning that the assertion was true), it
1225 was incorrectly treated as false if the SKIP had been reached during the
1226 matching. This also applied to assertions used as conditions.
1227
1228 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1229 assertion subpattern, including such a pattern used as a condition,
1230 unpredictable results occurred, instead of the error return
1231 PCRE_ERROR_DFA_UITEM.
1232
1233 5. The C++ GlobalReplace function was not working like Perl for the special
1234 situation when an empty string is matched. It now does the fancy magic
1235 stuff that is necessary.
1236
1237 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1238 removed. (These were left over from very, very early versions of PCRE.)
1239
1240 7. Some cosmetic changes to the code to make life easier when compiling it
1241 as part of something else:
1242
1243 (a) Change DEBUG to PCRE_DEBUG.
1244
1245 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1246 called "current" as "current_branch", to prevent a collision with the
1247 Linux macro when compiled as a kernel module.
1248
1249 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1250 prevent a collision with the Linux macro when compiled as a kernel
1251 module.
1252
1253 8. In pcre_compile() there are some checks for integer overflows that used to
1254 cast potentially large values to (double). This has been changed to that
1255 when building, a check for int64_t is made, and if it is found, it is used
1256 instead, thus avoiding the use of floating point arithmetic. (There is no
1257 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1258 double.
1259
1260 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1261 2005 (difference between two addresses compared to an unsigned value).
1262
1263 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1264 custom one, because of the following reported problem in Windows:
1265
1266 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1267 under Win32.
1268 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1269 therefore missing the function definition.
1270 - The compiler thus generates a "C" signature for the test function.
1271 - The linker fails to find the "C" function.
1272 - PCRE fails to configure if asked to do so against libbz2.
1273
1274 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1275 messages were output:
1276
1277 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1278 rerunning libtoolize, to keep the correct libtool macros in-tree.
1279 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1280
1281 I have done both of these things.
1282
1283 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1284 most of the time, it *can* run out if it is given a pattern that contains a
1285 runaway infinite recursion. I updated the discussion in the pcrestack man
1286 page.
1287
1288 13. Now that we have gone to the x.xx style of version numbers, the minor
1289 version may start with zero. Using 08 or 09 is a bad idea because users
1290 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1291 interpreted as invalid octal numbers. I've updated the previous comment in
1292 configure.ac, and also added a check that gives an error if 08 or 09 are
1293 used.
1294
1295 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1296 causing partial matching to fail when the end of the subject matched \W
1297 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1298
1299 15. There were some discrepancies between the declarations in pcre_internal.h
1300 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1301 their definitions. The declarations used "const uschar *" and the
1302 definitions used USPTR. Even though USPTR is normally defined as "const
1303 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1304 reported that: "This difference in casting confuses some C++ compilers, for
1305 example, SunCC recognizes above declarations as different functions and
1306 generates broken code for hbpcre." I have changed the declarations to use
1307 USPTR.
1308
1309 16. GNU libtool is named differently on some systems. The autogen.sh script now
1310 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1311 (FreeBSD).
1312
1313 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1314 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1315 comment: "Figure out how to create a longlong from a string: strtoll and
1316 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1317 instance, but it only takes 2 args instead of 3!"
1318
1319 18. A subtle bug concerned with back references has been fixed by a change of
1320 specification, with a corresponding code fix. A pattern such as
1321 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1322 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1323 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1324 same bug. Such groups have to be quantified to be useful, or contained
1325 inside another quantified group. (If there's no repetition, the reference
1326 can never match.) The problem arises because, having left the group and
1327 moved on to the rest of the pattern, a later failure that backtracks into
1328 the group uses the captured value from the final iteration of the group
1329 rather than the correct earlier one. I have fixed this in PCRE by forcing
1330 any group that contains a reference to itself to be an atomic group; that
1331 is, there cannot be any backtracking into it once it has completed. This is
1332 similar to recursive and subroutine calls.
1333
1334
1335 Version 8.00 19-Oct-09
1336 ----------------------
1337
1338 1. The table for translating pcre_compile() error codes into POSIX error codes
1339 was out-of-date, and there was no check on the pcre_compile() error code
1340 being within the table. This could lead to an OK return being given in
1341 error.
1342
1343 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1344 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1345 in a Windows environment.
1346
1347 3. The pcregrep --count option prints the count for each file even when it is
1348 zero, as does GNU grep. However, pcregrep was also printing all files when
1349 --files-with-matches was added. Now, when both options are given, it prints
1350 counts only for those files that have at least one match. (GNU grep just
1351 prints the file name in this circumstance, but including the count seems
1352 more useful - otherwise, why use --count?) Also ensured that the
1353 combination -clh just lists non-zero counts, with no names.
1354
1355 4. The long form of the pcregrep -F option was incorrectly implemented as
1356 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1357 but it seems right to fix it, and I didn't think it was worth preserving
1358 the old behaviour.
1359
1360 5. The command line items --regex=pattern and --regexp=pattern were not
1361 recognized by pcregrep, which required --regex pattern or --regexp pattern
1362 (with a space rather than an '='). The man page documented the '=' forms,
1363 which are compatible with GNU grep; these now work.
1364
1365 6. No libpcreposix.pc file was created for pkg-config; there was just
1366 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1367
1368 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1369 when UCP support is not needed, by modifying the Python script that
1370 generates it from Unicode data files. This should not matter if the module
1371 is correctly used as a library, but I received one complaint about 50K of
1372 unwanted data. My guess is that the person linked everything into his
1373 program rather than using a library. Anyway, it does no harm.
1374
1375 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1376 was a minimum greater than 1 for a wide character in a possessive
1377 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1378 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1379 character. Chaos in the form of incorrect output or a compiling loop could
1380 result.
1381
1382 9. The restrictions on what a pattern can contain when partial matching is
1383 requested for pcre_exec() have been removed. All patterns can now be
1384 partially matched by this function. In addition, if there are at least two
1385 slots in the offset vector, the offset of the earliest inspected character
1386 for the match and the offset of the end of the subject are set in them when
1387 PCRE_ERROR_PARTIAL is returned.
1388
1389 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1390 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1391 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1392 and may be more useful for multi-segment matching.
1393
1394 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1395 used to be given if ever the end of the subject was reached; now it is
1396 given only if matching could not proceed because another character was
1397 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1398 string "Z", which now yields "no match" instead of "partial match". In the
1399 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1400 final character ended with (*FAIL).
1401
1402 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1403 if the pattern had a "must contain" character that was already found in the
1404 earlier partial match, unless partial matching was again requested. For
1405 example, with the pattern /dog.(body)?/, the "must contain" character is
1406 "g". If the first part-match was for the string "dog", restarting with
1407 "sbody" failed. This bug has been fixed.
1408
1409 13. The string returned by pcre_dfa_exec() after a partial match has been
1410 changed so that it starts at the first inspected character rather than the
1411 first character of the match. This makes a difference only if the pattern
1412 starts with a lookbehind assertion or \b or \B (\K is not supported by
1413 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1414 matching functions compatible, and I think it's the right thing to do.
1415
1416 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1417 so that the demonstration program is easily available in environments where
1418 PCRE has not been installed from source.
1419
1420 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1421 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1422 library.
1423
1424 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1425 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1426 is not the first non-POSIX option to be added. Clearly some people find
1427 these options useful.
1428
1429 17. If a caller to the POSIX matching function regexec() passes a non-zero
1430 value for nmatch with a NULL value for pmatch, the value of
1431 nmatch is forced to zero.
1432
1433 18. RunGrepTest did not have a test for the availability of the -u option of
1434 the diff command, as RunTest does. It now checks in the same way as
1435 RunTest, and also checks for the -b option.
1436
1437 19. If an odd number of negated classes containing just a single character
1438 interposed, within parentheses, between a forward reference to a named
1439 subpattern and the definition of the subpattern, compilation crashed with
1440 an internal error, complaining that it could not find the referenced
1441 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1442 [The bug was that it was starting one character too far in when skipping
1443 over the character class, thus treating the ] as data rather than
1444 terminating the class. This meant it could skip too much.]
1445
1446 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1447 /g option in pcretest when the pattern contains \K, which makes it possible
1448 to have an empty string match not at the start, even when the pattern is
1449 anchored. Updated pcretest and pcredemo to use this option.
1450
1451 21. If the maximum number of capturing subpatterns in a recursion was greater
1452 than the maximum at the outer level, the higher number was returned, but
1453 with unset values at the outer level. The correct (outer level) value is
1454 now given.
1455
1456 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1457 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1458 make it do so. The string so far is captured, making this feature
1459 compatible with Perl.
1460
1461 23. The tests have been re-organized, adding tests 11 and 12, to make it
1462 possible to check the Perl 5.10 features against Perl 5.10.
1463
1464 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1465 pattern matches a fixed length string. PCRE did not allow this; now it
1466 does. Neither allows recursion.
1467
1468 25. I finally figured out how to implement a request to provide the minimum
1469 length of subject string that was needed in order to match a given pattern.
1470 (It was back references and recursion that I had previously got hung up
1471 on.) This code has now been added to pcre_study(); it finds a lower bound
1472 to the length of subject needed. It is not necessarily the greatest lower
1473 bound, but using it to avoid searching strings that are too short does give
1474 some useful speed-ups. The value is available to calling programs via
1475 pcre_fullinfo().
1476
1477 26. While implementing 25, I discovered to my embarrassment that pcretest had
1478 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1479 study optimizations had never been tested with that matching function.
1480 Oops. What is worse, even when it was passed study data, there was a bug in
1481 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1482 were also very few tests of studied patterns with pcre_dfa_exec().
1483
1484 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1485 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1486 on the other side of the coin, they are no longer allowed to have different
1487 names, because these cannot be distinguished in PCRE, and this has caused
1488 confusion. (This is a difference from Perl.)
1489
1490 28. When duplicate subpattern names are present (necessarily with different
1491 numbers, as required by 27 above), and a test is made by name in a
1492 conditional pattern, either for a subpattern having been matched, or for
1493 recursion in such a pattern, all the associated numbered subpatterns are
1494 tested, and the overall condition is true if the condition is true for any
1495 one of them. This is the way Perl works, and is also more like the way
1496 testing by number works.
1497
1498
1499 Version 7.9 11-Apr-09
1500 ---------------------
1501
1502 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1503 (pcretest), all targets were linked against these libraries. This included
1504 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1505 libraries. This caused unwanted dependencies to be created. This problem
1506 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1507 pcretest is linked with readline.
1508
1509 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1510 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1511 moved outside it again, because FALSE and TRUE are already defined in AIX,
1512 but BOOL is not.
1513
1514 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1515 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1516
1517 4. The pcregrep documentation said spaces were inserted as well as colons (or
1518 hyphens) following file names and line numbers when outputting matching
1519 lines. This is not true; no spaces are inserted. I have also clarified the
1520 wording for the --colour (or --color) option.
1521
1522 5. In pcregrep, when --colour was used with -o, the list of matching strings
1523 was not coloured; this is different to GNU grep, so I have changed it to be
1524 the same.
1525
1526 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1527 each matching line was coloured. Now it goes on to look for further matches
1528 of any of the test patterns, which is the same behaviour as GNU grep.
1529
1530 7. A pattern that could match an empty string could cause pcregrep to loop; it
1531 doesn't make sense to accept an empty string match in pcregrep, so I have
1532 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1533 seems to be how GNU grep behaves.
1534
1535 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1536 start or after a newline", because the conditional assertion was not being
1537 correctly handled. The rule now is that both the assertion and what follows
1538 in the first alternative must satisfy the test.
1539
1540 9. If auto-callout was enabled in a pattern with a conditional group whose
1541 condition was an assertion, PCRE could crash during matching, both with
1542 pcre_exec() and pcre_dfa_exec().
1543
1544 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1545 used for matching.
1546
1547 11. Unicode property support in character classes was not working for
1548 characters (bytes) greater than 127 when not in UTF-8 mode.
1549
1550 12. Added the -M command line option to pcretest.
1551
1552 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1553
1554 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1555
1556 16. Added comments and documentation about mis-use of no_arg in the C++
1557 wrapper.
1558
1559 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1560 from Martin Jerabek that uses macro names for all relevant character and
1561 string constants.
1562
1563 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1564 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1565 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1566 these, but not everybody uses configure.
1567
1568 19. A conditional group that had only one branch was not being correctly
1569 recognized as an item that could match an empty string. This meant that an
1570 enclosing group might also not be so recognized, causing infinite looping
1571 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1572 with the subject "ab", where knowledge that the repeated group can match
1573 nothing is needed in order to break the loop.
1574
1575 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1576 exec(), but without supplying a callout function, matching went wrong.
1577
1578 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1579 leak if the size of the offset vector was greater than 30. When the vector
1580 is smaller, the saved offsets during recursion go onto a local stack
1581 vector, but for larger vectors malloc() is used. It was failing to free
1582 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1583 error, in fact).
1584
1585 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1586 heapframe that is used only when UTF-8 support is enabled. This caused no
1587 problem, but was untidy.
1588
1589 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1590 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1591 included within another project.
1592
1593 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1594 slightly modified by me:
1595
1596 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1597 not building pcregrep.
1598
1599 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1600 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1601
1602 25. Forward references, both numeric and by name, in patterns that made use of
1603 duplicate group numbers, could behave incorrectly or give incorrect errors,
1604 because when scanning forward to find the reference group, PCRE was not
1605 taking into account the duplicate group numbers. A pattern such as
1606 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1607
1608 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1609 the feature of a custom pointer more persuasive (as requested by a user).
1610
1611 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1612 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1613 pre-defined.
1614
1615 28. Added support for (*UTF8) at the start of a pattern.
1616
1617 29. Arrange for flags added by the "release type" setting in CMake to be shown
1618 in the configuration summary.
1619
1620
1621 Version 7.8 05-Sep-08
1622 ---------------------
1623
1624 1. Replaced UCP searching code with optimized version as implemented for Ad
1625 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1626 stage table and inline lookup instead of a function, giving speed ups of 2
1627 to 5 times on some simple patterns that I tested. Permission was given to
1628 distribute the MultiStage2.py script that generates the tables (it's not in
1629 the tarball, but is in the Subversion repository).
1630
1631 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1632 scripts.
1633
1634 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1635 a group with a zero qualifier. The result of the study could be incorrect,
1636 or the function might crash, depending on the pattern.
1637
1638 4. Caseless matching was not working for non-ASCII characters in back
1639 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1640 It now works when Unicode Property Support is available.
1641
1642 5. In pcretest, an escape such as \x{de} in the data was always generating
1643 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1644 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1645 truncation.
1646
1647 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1648
1649 7. Added two (int) casts to pcregrep when printing the difference of two
1650 pointers, in case they are 64-bit values.
1651
1652 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1653 test 2 if it fails.
1654
1655 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1656 and a #define of that name to empty if it is not externally set. This is to
1657 allow users of MSVC to set it if necessary.
1658
1659 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1660 the convenience functions in the pcre_get.c source file.
1661
1662 11. An option change at the start of a pattern that had top-level alternatives
1663 could cause overwriting and/or a crash. This command provoked a crash in
1664 some environments:
1665
1666 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1667
1668 This potential security problem was recorded as CVE-2008-2371.
1669
1670 12. For a pattern where the match had to start at the beginning or immediately
1671 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1672 pcre_dfa_exec() could read past the end of the passed subject if there was
1673 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1674 pcretest so that it places the subject at the end of its malloc-ed buffer.
1675
1676 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1677 exec() might read past the end of the data buffer in UTF-8 mode.
1678
1679 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1680 the data contained the byte 0x85 as part of a UTF-8 character within its
1681 first line. This applied both to normal and DFA matching.
1682
1683 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1684 /^[^d]*?$/8 failed to match "abc".
1685
1686 16. Added a missing copyright notice to pcrecpp_internal.h.
1687
1688 17. Make it more clear in the documentation that values returned from
1689 pcre_exec() in ovector are byte offsets, not character counts.
1690
1691 18. Tidied a few places to stop certain compilers from issuing warnings.
1692
1693 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1694 supplied by Stefan Weber. I made a further small update for 7.8 because
1695 there is a change of source arrangements: the pcre_searchfuncs.c module is
1696 replaced by pcre_ucd.c.
1697
1698
1699 Version 7.7 07-May-08
1700 ---------------------
1701
1702 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1703 a string to a long long, pretend we don't even have a long long." This is
1704 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1705
1706 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1707 pre-7.6 versions, which defined a global no_arg variable instead of putting
1708 it in the RE class. (See also #8 below.)
1709
1710 3. Remove a line of dead code, identified by coverity and reported by Nuno
1711 Lopes.
1712
1713 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1714
1715 (1) The include/exclude patterns were being applied to the whole pathnames
1716 of files, instead of just to the final components.
1717
1718 (2) If there was more than one level of directory, the subdirectories were
1719 skipped unless they satisfied the include/exclude conditions. This is
1720 inconsistent with GNU grep (and could even be seen as contrary to the
1721 pcregrep specification - which I improved to make it absolutely clear).
1722 The action now is always to scan all levels of directory, and just
1723 apply the include/exclude patterns to regular files.
1724
1725 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1726 --exclude_dir in the tests to avoid scanning .svn directories.
1727
1728 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1729 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1730 doesn't support NULs in patterns.
1731
1732 7. Added some missing "const"s to declarations of static tables in
1733 pcre_compile.c and pcre_dfa_exec.c.
1734
1735 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1736 caused by fix #2 above. (Subsequently also a second patch to fix the
1737 first patch. And a third patch - this was a messy problem.)
1738
1739 9. Applied Craig's patch to remove the use of push_back().
1740
1741 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1742 matching function regexec().
1743
1744 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1745 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1746 references. PCRE supports relative numbers with this syntax (I don't think
1747 Oniguruma does).
1748
1749 12. Previously, a group with a zero repeat such as (...){0} was completely
1750 omitted from the compiled regex. However, this means that if the group
1751 was called as a subroutine from elsewhere in the pattern, things went wrong
1752 (an internal error was given). Such groups are now left in the compiled
1753 pattern, with a new opcode that causes them to be skipped at execution
1754 time.
1755
1756 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1757 to the way PCRE behaves:
1758
1759 (a) A lone ] character is dis-allowed (Perl treats it as data).
1760
1761 (b) A back reference to an unmatched subpattern matches an empty string
1762 (Perl fails the current match path).
1763
1764 (c) A data ] in a character class must be notated as \] because if the
1765 first data character in a class is ], it defines an empty class. (In
1766 Perl it is not possible to have an empty class.) The empty class []
1767 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1768 The negative empty class [^] matches any one character, independently
1769 of the DOTALL setting.
1770
1771 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1772 non-existent subpattern following a character class starting with ']' and
1773 containing () gave an internal compiling error instead of "reference to
1774 non-existent subpattern". Fortunately, when the pattern did exist, the
1775 compiled code was correct. (When scanning forwards to check for the
1776 existencd of the subpattern, it was treating the data ']' as terminating
1777 the class, so got the count wrong. When actually compiling, the reference
1778 was subsequently set up correctly.)
1779
1780 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1781 it was being rejected as not supported by pcre_dfa_exec(), even though
1782 other assertions are supported. I have made pcre_dfa_exec() support
1783 (*FAIL).
1784
1785 16. The implementation of 13c above involved the invention of a new opcode,
1786 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1787 cannot be changed at match time, I realized I could make a small
1788 improvement to matching performance by compiling OP_ALLANY instead of
1789 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1790 on the OP_ANY path.
1791
1792 17. Compiling pcretest on Windows with readline support failed without the
1793 following two fixes: (1) Make the unistd.h include conditional on
1794 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1795
1796 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1797 ncurses library to be included for pcretest when ReadLine support is
1798 requested, but also to allow for it to be overridden. This patch came from
1799 Daniel Bergström.
1800
1801 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1802 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1803 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1804 spotting this.
1805
1806
1807 Version 7.6 28-Jan-08
1808 ---------------------
1809
1810 1. A character class containing a very large number of characters with
1811 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1812 overflow.
1813
1814 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1815 HAVE_LONG_LONG is not defined.
1816
1817 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1818 bring them up to date and include new features. This patch includes:
1819
1820 - Fixed PH's badly added libz and libbz2 support.
1821 - Fixed a problem with static linking.
1822 - Added pcredemo. [But later removed - see 7 below.]
1823 - Fixed dftables problem and added an option.
1824 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1825 HAVE_LONG_LONG.
1826 - Added readline support for pcretest.
1827 - Added an listing of the option settings after cmake has run.
1828
1829 4. A user submitted a patch to Makefile that makes it easy to create
1830 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1831 Makefile.am that cause it to include this special target, without
1832 affecting anything else. Note that the same mingw target plus all
1833 the other distribution libraries and programs are now supported
1834 when configuring with CMake (see 6 below) instead of with
1835 Configure/Make.
1836
1837 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1838 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1839 exported in the Windows port". It has not yet been confirmed that the patch
1840 solves the problem, but it does no harm.
1841
1842 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1843 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1844 with CMake, and also correct the comment about stack recursion.
1845
1846 7. Remove the automatic building of pcredemo from the ./configure system and
1847 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1848 of a program that users should build themselves after PCRE is installed, so
1849 building it automatically is not really right. What is more, it gave
1850 trouble in some build environments.
1851
1852 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1853
1854
1855 Version 7.5 10-Jan-08
1856 ---------------------
1857
1858 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1859 values in parens when parsing an RE using the C++ wrapper."
1860
1861 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1862 Characters greater than 255 were excluded from the class instead of being
1863 included.
1864
1865 3. The same bug as (2) above applied to negated POSIX classes such as
1866 [:^space:].
1867
1868 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1869 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1870 I have changed it.
1871
1872 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1873 first named subpattern) and a construct such as (?&a) would reference the
1874 first named subpattern whose name started with "a" (in other words, the
1875 length check was missing). Both these problems are fixed. "Subpattern name
1876 expected" is now given for (?&) (a zero-length name), and this patch also
1877 makes it give the same error for \k'' (previously it complained that that
1878 was a reference to a non-existent subpattern).
1879
1880 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1881 this is right because (?- can be followed by option settings as well as by
1882 digits. I have, however, made the messages clearer.
1883
1884 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1885 than the number used in the conditional) now cause a compile-time error.
1886 This is actually not compatible with Perl, which accepts such patterns, but
1887 treats the conditional as always being FALSE (as PCRE used to), but it
1888 seems to me that giving a diagnostic is better.
1889
1890 8. Change "alphameric" to the more common word "alphanumeric" in comments
1891 and messages.
1892
1893 9. Fix two occurrences of "backslash" in comments that should have been
1894 "backspace".
1895
1896 10. Remove two redundant lines of code that can never be obeyed (their function
1897 was moved elsewhere).
1898
1899 11. The program that makes PCRE's Unicode character property table had a bug
1900 which caused it to generate incorrect table entries for sequences of
1901 characters that have the same character type, but are in different scripts.
1902 It amalgamated them into a single range, with the script of the first of
1903 them. In other words, some characters were in the wrong script. There were
1904 thirteen such cases, affecting characters in the following ranges:
1905
1906 U+002b0 - U+002c1
1907 U+0060c - U+0060d
1908 U+0061e - U+00612
1909 U+0064b - U+0065e
1910 U+0074d - U+0076d
1911 U+01800 - U+01805
1912 U+01d00 - U+01d77
1913 U+01d9b - U+01dbf
1914 U+0200b - U+0200f
1915 U+030fc - U+030fe
1916 U+03260 - U+0327f
1917 U+0fb46 - U+0fbb1
1918 U+10450 - U+1049d
1919
1920 12. The -o option (show only the matching part of a line) for pcregrep was not
1921 compatible with GNU grep in that, if there was more than one match in a
1922 line, it showed only the first of them. It now behaves in the same way as
1923 GNU grep.
1924
1925 13. If the -o and -v options were combined for pcregrep, it printed a blank
1926 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1927 does the same. The return code can be used to tell if there were any
1928 non-matching lines.
1929
1930 14. Added --file-offsets and --line-offsets to pcregrep.
1931
1932 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1933 infinitely looping recursion. The bug was that positive lookaheads were not
1934 being skipped when checking for a possible empty match (negative lookaheads
1935 and both kinds of lookbehind were skipped).
1936
1937 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1938 inclusion of <windows.h> to before rather than after the definition of
1939 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1940
1941 17. Specifying a possessive quantifier with a specific limit for a Unicode
1942 character property caused pcre_compile() to compile bad code, which led at
1943 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1944 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1945 caused the error; without that there was no problem.
1946
1947 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1948
1949 19. Added --enable-pcretest-libreadline.
1950
1951 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1952 RE::GlobalReplace(). As a result, the number of replacements returned was
1953 double what it should be. I removed one of the increments, but Craig sent a
1954 later patch that removed the other one (the right fix) and added unit tests
1955 that check the return values (which was not done before).
1956
1957 21. Several CMake things:
1958
1959 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1960 the names libpcre and libpcreposix, not just pcre and pcreposix.
1961
1962 (2) The above change means that pcretest and pcregrep are now correctly
1963 linked with the newly-built libraries, not previously installed ones.
1964
1965 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1966
1967 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1968 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1969 UTF-8 newline character). The key issue is that the pattern starts .*;
1970 this means that the match must be either at the beginning, or after a
1971 newline. The bug was in the code for advancing after a failed match and
1972 checking that the new position followed a newline. It was not taking
1973 account of UTF-8 characters correctly.
1974
1975 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1976 character classes. PCRE was not treating the sequence [:...:] as a
1977 character class unless the ... were all letters. Perl, however, seems to
1978 allow any characters between [: and :], though of course it rejects as
1979 unknown any "names" that contain non-letters, because all the known class
1980 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1981 for example, whereas PCRE did not - it did not recognize a POSIX character
1982 class. This seemed a bit dangerous, so the code has been changed to be
1983 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1984 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1985 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1986 Perl does, and where it didn't before.
1987
1988 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1989 Windows environments %n is disabled by default.
1990
1991
1992 Version 7.4 21-Sep-07
1993 ---------------------
1994
1995 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1996 means that a class such as [\s] counted as "explicit reference to CR or
1997 LF". That isn't really right - the whole point of the change was to try to
1998 help when there was an actual mention of one of the two characters. So now
1999 the change happens only if \r or \n (or a literal CR or LF) character is
2000 encountered.
2001
2002 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2003 of both had grown to the point where there were only 3 bits left.
2004 Fortunately, there was spare space in the data structure, and so I have
2005 moved the internal flags into a new 16-bit field to free up more option
2006 bits.
2007
2008 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2009 but did not set the internal JCHANGED flag - either of these is enough to
2010 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2011 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2012 start sets both bits.
2013
2014 4. Added options (at build time, compile time, exec time) to change \R from
2015 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2016
2017 5. doc/pcresyntax.html was missing from the distribution.
2018
2019 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2020 compatibility, even though it is no longer used.
2021
2022 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2023 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2024 windows.h file is present (where different names are used). [This was
2025 reversed later after testing - see 16 below.]
2026
2027 8. Changed all #include <config.h> to #include "config.h". There were also
2028 some further <pcre.h> cases that I changed to "pcre.h".
2029
2030 9. When pcregrep was used with the --colour option, it missed the line ending
2031 sequence off the lines that it output.
2032
2033 10. It was pointed out to me that arrays of string pointers cause lots of
2034 relocations when a shared library is dynamically loaded. A technique of
2035 using a single long string with a table of offsets can drastically reduce
2036 these. I have refactored PCRE in four places to do this. The result is
2037 dramatic:
2038
2039 Originally: 290
2040 After changing UCP table: 187
2041 After changing error message table: 43
2042 After changing table of "verbs" 36
2043 After changing table of Posix names 22
2044
2045 Thanks to the folks working on Gregex for glib for this insight.
2046
2047 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2048 unicode-properties was also set.
2049
2050 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2051
2052 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2053 checked only for CRLF.
2054
2055 14. Added casts to pcretest.c to avoid compiler warnings.
2056
2057 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2058
2059 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2060 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2061 entirely. This removes changes made in 7 above.
2062
2063 17. The CMake files have been updated, and there is now more information about
2064 building with CMake in the NON-UNIX-USE document.
2065
2066
2067 Version 7.3 28-Aug-07
2068 ---------------------
2069
2070 1. In the rejigging of the build system that eventually resulted in 7.1, the
2071 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2072 brackets there is not right, since it causes compilers to look for an
2073 installed pcre.h, not the version that is in the source that is being
2074 compiled (which of course may be different). I have changed it back to:
2075
2076 #include "pcre.h"
2077
2078 I have a vague recollection that the change was concerned with compiling in
2079 different directories, but in the new build system, that is taken care of
2080 by the VPATH setting the Makefile.
2081
2082 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2083 when the subject happened to end in the byte 0x85 (e.g. if the last
2084 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2085 characters but of course it shouldn't be taken as a newline when it is part
2086 of another character. The bug was that, for an unlimited repeat of . in
2087 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2088 characters when looking for a newline.
2089
2090 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2091
2092 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2093 in debug output.
2094
2095 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2096 long printing in the pcrecpp unittest when running under MinGW.
2097
2098 6. ESC_K was left out of the EBCDIC table.
2099
2100 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2101 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2102 limit also applies to "virtual nesting" when a pattern is recursive, and in
2103 this case 1000 isn't so big. I have been able to remove this limit at the
2104 expense of backing off one optimization in certain circumstances. Normally,
2105 when pcre_exec() would call its internal match() function recursively and
2106 immediately return the result unconditionally, it uses a "tail recursion"
2107 feature to save stack. However, when a subpattern that can match an empty
2108 string has an unlimited repetition quantifier, it no longer makes this
2109 optimization. That gives it a stack frame in which to save the data for
2110 checking that an empty string has been matched. Previously this was taken
2111 from the 1000-entry workspace that had been reserved. So now there is no
2112 explicit limit, but more stack is used.
2113
2114 8. Applied Daniel's patches to solve problems with the import/export magic
2115 syntax that is required for Windows, and which was going wrong for the
2116 pcreposix and pcrecpp parts of the library. These were overlooked when this
2117 problem was solved for the main library.
2118
2119 9. There were some crude static tests to avoid integer overflow when computing
2120 the size of patterns that contain repeated groups with explicit upper
2121 limits. As the maximum quantifier is 65535, the maximum group length was
2122 set at 30,000 so that the product of these two numbers did not overflow a
2123 32-bit integer. However, it turns out that people want to use groups that
2124 are longer than 30,000 bytes (though not repeat them that many times).
2125 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2126 made it possible to implement the integer overflow checks in a much more
2127 dynamic way, which I have now done. The artificial limitation on group
2128 length has been removed - we now have only the limit on the total length of
2129 the compiled pattern, which depends on the LINK_SIZE setting.
2130
2131 10. Fixed a bug in the documentation for get/copy named substring when
2132 duplicate names are permitted. If none of the named substrings are set, the
2133 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2134 empty string.
2135
2136 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2137 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2138 because the ] is interpreted as the first data character and the
2139 terminating ] is not found. PCRE has been made compatible with Perl in this
2140 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2141 cause memory overwriting.
2142
2143 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2144 string has been matched (to stop an infinite loop). It was not recognizing
2145 a conditional subpattern that could match an empty string if that
2146 subpattern was within another subpattern. For example, it looped when
2147 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2148 condition was not nested. This bug has been fixed.
2149
2150 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2151 past the start of the subject in the presence of bytes with the top bit
2152 set, for example "\x8aBCD".
2153
2154 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2155 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2156
2157 14. Optimized (?!) to (*FAIL).
2158
2159 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2160 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2161 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2162 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2163 does: it's just the validity check that is more restrictive.
2164
2165 16. Inserted checks for integer overflows during escape sequence (backslash)
2166 processing, and also fixed erroneous offset values for syntax errors during
2167 backslash processing.
2168
2169 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2170 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2171
2172 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2173 caused an overrun.
2174
2175 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2176 something other than just ASCII characters) inside a group that had an
2177 unlimited repeat caused a loop at compile time (while checking to see
2178 whether the group could match an empty string).
2179
2180 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2181 [\P{Any}] did so. (Error in the code for printing property names.)
2182
2183 21. An orphan \E inside a character class could cause a crash.
2184
2185 22. A repeated capturing bracket such as (A)? could cause a wild memory
2186 reference during compilation.
2187
2188 23. There are several functions in pcre_compile() that scan along a compiled
2189 expression for various reasons (e.g. to see if it's fixed length for look
2190 behind). There were bugs in these functions when a repeated \p or \P was
2191 present in the pattern. These operators have additional parameters compared
2192 with \d, etc, and these were not being taken into account when moving along
2193 the compiled data. Specifically:
2194
2195 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2196 length.
2197
2198 (b) An item such as \pL+ within a repeated group could cause crashes or
2199 loops.
2200
2201 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2202 "reference to non-existent subpattern" error.
2203
2204 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2205
2206 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2207 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2208
2209 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2210
2211 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2212 character were causing crashes (broken optimization).
2213
2214 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2215 \p or \P) caused a compile-time loop.
2216
2217 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2218 break. For example, the unstudied pattern [\r\n]A does not match the string
2219 "\r\nA" because change 7.0/46 below moves the current point on by two
2220 characters after failing to match at the start. However, the pattern \nA
2221 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2222 the same is true. There doesn't seem any very clean way out of this, but
2223 what I have chosen to do makes the common cases work: PCRE now takes note
2224 of whether there can be an explicit match for \r or \n anywhere in the
2225 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2226 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2227 pattern has explicit CR or LF references.
2228
2229 29. Added (*CR) etc for changing newline setting at start of pattern.
2230
2231
2232 Version 7.2 19-Jun-07
2233 ---------------------
2234
2235 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2236 which is apparently normally available under Windows.
2237
2238 2. Re-jig the pcregrep tests with different newline settings in an attempt
2239 to make them independent of the local environment's newline setting.
2240
2241 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2242
2243 4. Some of the "internals" tests were previously cut out when the link size
2244 was not 2, because the output contained actual offsets. The recent new
2245 "Z" feature of pcretest means that these can be cut out, making the tests
2246 usable with all link sizes.
2247
2248 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2249 stack recursion. This gives a massive performance boost under BSD, but just
2250 a small improvement under Linux. However, it saves one field in the frame
2251 in all cases.
2252
2253 6. Added more features from the forthcoming Perl 5.10:
2254
2255 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2256 recursion call. It refers to the nth most recently opened parentheses.
2257
2258 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2259 to be opened parentheses.
2260
2261 (c) Conditions that refer to capturing parentheses can be specified
2262 relatively, for example, (?(-2)... or (?(+3)...
2263
2264 (d) \K resets the start of the current match so that everything before
2265 is not part of it.
2266
2267 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2268
2269 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2270 reference syntax.
2271
2272 (g) (?| introduces a group in which the numbering of parentheses in each
2273 alternative starts with the same number.
2274
2275 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2276
2277 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2278 PCRE_INFO_JCHANGED.
2279
2280 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2281 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2282 for detecting groups that can match an empty string.
2283
2284 9. A pattern with a very large number of alternatives (more than several
2285 hundred) was running out of internal workspace during the pre-compile
2286 phase, where pcre_compile() figures out how much memory will be needed. A
2287 bit of new cunning has reduced the workspace needed for groups with
2288 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2289 workspace instead of running out of the 4096 that are available.
2290
2291 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2292
2293 11. Applied patch from Google to remove an optimization that didn't quite work.
2294 The report of the bug said:
2295
2296 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2297 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2298 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2299
2300 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2301 it matched the wrong number of bytes.
2302
2303
2304 Version 7.1 24-Apr-07
2305 ---------------------
2306
2307 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2308 that is more "standard", making use of automake and other Autotools. There
2309 is some re-arrangement of the files and adjustment of comments consequent
2310 on this.
2311
2312 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2313 for recursive directory scanning broke on some systems because the files
2314 are not scanned in any specific order and on different systems the order
2315 was different. A call to "sort" has been inserted into RunGrepTest for the
2316 approprate test as a short-term fix. In the longer term there may be an
2317 alternative.
2318
2319 3. I had an email from Eric Raymond about problems translating some of PCRE's
2320 man pages to HTML (despite the fact that I distribute HTML pages, some
2321 people do their own conversions for various reasons). The problems
2322 concerned the use of low-level troff macros .br and .in. I have therefore
2323 removed all such uses from the man pages (some were redundant, some could
2324 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2325 HTML has been updated to handle .nf/.fi and to complain if it encounters
2326 .br or .in.
2327
2328 4. Updated comments in configure.ac that get placed in config.h.in and also
2329 arranged for config.h to be included in the distribution, with the name
2330 config.h.generic, for the benefit of those who have to compile without
2331 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2332
2333 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2334 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2335 makevp.bat for the current PCRE, using the additional files
2336 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2337
2338 6. A Windows user reported a minor discrepancy with test 2, which turned out
2339 to be caused by a trailing space on an input line that had got lost in his
2340 copy. The trailing space was an accident, so I've just removed it.
2341
2342 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2343 that is needed.
2344
2345 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2346 as "const" (a) because they are and (b) because it helps the PHP
2347 maintainers who have recently made a script to detect big data structures
2348 in the php code that should be moved to the .rodata section. I remembered
2349 to update Builducptable as well, so it won't revert if ucptable.h is ever
2350 re-created.
2351
2352 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2353 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2354 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2355 support is not required. This saves 1.5-2K of code, which is important in
2356 some applications.
2357
2358 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2359 so as not to refer to the tables, even though these functions will never be
2360 called when UTF-8 support is disabled. Otherwise there are problems with a
2361 shared library.
2362
2363 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2364
2365 (a) It was defining its arguments as char * instead of void *.
2366
2367 (b) It was assuming that all moves were upwards in memory; this was true
2368 a long time ago when I wrote it, but is no longer the case.
2369
2370 The emulated memove() is provided for those environments that have neither
2371 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2372 is clearly not the case, as these two bugs were recently reported.
2373
2374 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2375 and Detrail to create the HTML documentation, the .txt form of the man
2376 pages, and it removes trailing spaces from listed files. It also creates
2377 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2378 case, it wraps all the #defines with #ifndefs. This script should be run
2379 before "make dist".
2380
2381 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2382 with Unicode property support.
2383
2384 (a) For a maximizing quantifier, if the two different cases of the
2385 character were of different lengths in their UTF-8 codings (there are
2386 some cases like this - I found 11), and the matching function had to
2387 back up over a mixture of the two cases, it incorrectly assumed they
2388 were both the same length.
2389
2390 (b) When PCRE was configured to use the heap rather than the stack for
2391 recursion during matching, it was not correctly preserving the data for
2392 the other case of a UTF-8 character when checking ahead for a match
2393 while processing a minimizing repeat. If the check also involved
2394 matching a wide character, but failed, corruption could cause an
2395 erroneous result when trying to check for a repeat of the original
2396 character.
2397
2398 13. Some tidying changes to the testing mechanism:
2399
2400 (a) The RunTest script now detects the internal link size and whether there
2401 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2402 values substituted by "configure". (The RunGrepTest script already did
2403 this for UTF-8.) The configure.ac script no longer substitutes the
2404 relevant variables.
2405
2406 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2407 with length and offset values. This means that the output is different
2408 for different internal link sizes. Test 2 is skipped for link sizes
2409 other than 2 because of this, bypassing the problem. Unfortunately,
2410 there was also a test in test 3 (the locale tests) that used /B and
2411 failed for link sizes other than 2. Rather than cut the whole test out,
2412 I have added a new /Z option to pcretest that replaces the length and
2413 offset values with spaces. This is now used to make test 3 independent
2414 of link size. (Test 2 will be tidied up later.)
2415
2416 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2417 segmentation fault instead of returning the appropriate error message.
2418
2419 15. In multiline mode when the newline sequence was set to "any", the pattern
2420 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2421 This doesn't seem right; it now treats the CRLF combination as the line
2422 ending, and so does not match in that case. It's only a pattern such as ^$
2423 that would hit this one: something like ^ABC$ would have failed after \r
2424 and then tried again after \r\n.
2425
2426 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2427 in an attempt to make files that differ only in their line terminators
2428 compare equal. This works on Linux.
2429
2430 17. Under certain error circumstances pcregrep might try to free random memory
2431 as it exited. This is now fixed, thanks to valgrind.
2432
2433 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2434 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2435 was because its rules for how to advance for /g after matching an empty
2436 string at the end of a line did not allow for this case. They now check for
2437 it specially.
2438
2439 20. pcretest is supposed to handle patterns and data of any length, by
2440 extending its buffers when necessary. It was getting this wrong when the
2441 buffer for a data line had to be extended.
2442
2443 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2444 CRLF as a newline sequence.
2445
2446 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2447 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2448 I have nevertheless tidied it up.
2449
2450 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2451
2452 24. Added a man page for pcre-config.
2453
2454
2455 Version 7.0 19-Dec-06
2456 ---------------------
2457
2458 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2459 moving to gcc 4.1.1.
2460
2461 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2462 sys/time.h, which is documented as needed for this function. It doesn't
2463 seem to matter on Linux, but it showed up on some releases of OS X.
2464
2465 3. It seems that there are systems where bytes whose values are greater than
2466 127 match isprint() in the "C" locale. The "C" locale should be the
2467 default when a C program starts up. In most systems, only ASCII printing
2468 characters match isprint(). This difference caused the output from pcretest
2469 to vary, making some of the tests fail. I have changed pcretest so that:
2470
2471 (a) When it is outputting text in the compiled version of a pattern, bytes
2472 other than 32-126 are always shown as hex escapes.
2473
2474 (b) When it is outputting text that is a matched part of a subject string,
2475 it does the same, unless a different locale has been set for the match
2476 (using the /L modifier). In this case, it uses isprint() to decide.
2477
2478 4. Fixed a major bug that caused incorrect computation of the amount of memory
2479 required for a compiled pattern when options that changed within the
2480 pattern affected the logic of the preliminary scan that determines the
2481 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2482 that the computed length was too small. The symptoms of this bug were
2483 either the PCRE error "internal error: code overflow" from pcre_compile(),
2484 or a glibc crash with a message such as "pcretest: free(): invalid next
2485 size (fast)". Examples of patterns that provoked this bug (shown in
2486 pcretest format) are:
2487
2488 /(?-x: )/x
2489 /(?x)(?-x: \s*#\s*)/
2490 /((?i)[\x{c0}])/8
2491 /(?i:[\x{c0}])/8
2492
2493 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2494 is now done differently.
2495
2496 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2497 wrapper classes; (b) implement a new function in the C++ scanner that is
2498 more efficient than the old way of doing things because it avoids levels of
2499 recursion in the regex matching; (c) add a paragraph to the documentation
2500 for the FullMatch() function.
2501
2502 6. The escape sequence \n was being treated as whatever was defined as
2503 "newline". Not only was this contrary to the documentation, which states
2504 that \n is character 10 (hex 0A), but it also went horribly wrong when
2505 "newline" was defined as CRLF. This has been fixed.
2506
2507 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2508 was being set to -1 for the "end of line" case (supposedly a value that no
2509 character can have). Though this value is never used (the check for end of
2510 line is "zero bytes in current character"), it caused compiler complaints.
2511 I've changed it to 0xffffffff.
2512
2513 8. In pcre_version.c, the version string was being built by a sequence of
2514 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2515 string (as it is for production releases) called a macro with an empty
2516 argument. The C standard says the result of this is undefined. The gcc
2517 compiler treats it as an empty string (which was what was wanted) but it is
2518 reported that Visual C gives an error. The source has been hacked around to
2519 avoid this problem.
2520
2521 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2522 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2523 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2524 of them did).
2525
2526 10. Originally, pcretest opened its input and output without "b"; then I was
2527 told that "b" was needed in some environments, so it was added for release
2528 5.0 to both the input and output. (It makes no difference on Unix-like
2529 systems.) Later I was told that it is wrong for the input on Windows. I've
2530 now abstracted the modes into two macros, to make it easier to fiddle with
2531 them, and removed "b" from the input mode under Windows.
2532
2533 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2534
2535 12. Added -help and --help to pcretest as an official way of being reminded
2536 of the options.
2537
2538 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2539 and pcrecpp.cc because they annoy compilers at high warning levels.
2540
2541 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2542
2543 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2544 scripts are not C programs :-) and which was not noticed because it works
2545 on Linux.
2546
2547 16. pcretest is supposed to handle any length of pattern and data line (as one
2548 line or as a continued sequence of lines) by extending its input buffer if
2549 necessary. This feature was broken for very long pattern lines, leading to
2550 a string of junk being passed to pcre_compile() if the pattern was longer
2551 than about 50K.
2552
2553 17. I have done a major re-factoring of the way pcre_compile() computes the
2554 amount of memory needed for a compiled pattern. Previously, there was code
2555 that made a preliminary scan of the pattern in order to do this. That was
2556 OK when PCRE was new, but as the facilities have expanded, it has become
2557 harder and harder to keep it in step with the real compile phase, and there
2558 have been a number of bugs (see for example, 4 above). I have now found a
2559 cunning way of running the real compile function in a "fake" mode that
2560 enables it to compute how much memory it would need, while actually only
2561 ever using a few hundred bytes of working memory and without too many
2562 tests of the mode. This should make future maintenance and development
2563 easier. A side effect of this work is that the limit of 200 on the nesting
2564 depth of parentheses has been removed (though this was never a serious
2565 limitation, I suspect). However, there is a downside: pcre_compile() now
2566 runs more slowly than before (30% or more, depending on the pattern). I
2567 hope this isn't a big issue. There is no effect on runtime performance.
2568
2569 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2570 newline (only possible for the last line of a file) and it was a
2571 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2572
2573 19. Added additional timing features to pcretest. (1) The -tm option now times
2574 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2575 separate command line item, by a number that specifies the number of
2576 repeats to use when timing. The default is 50000; this gives better
2577 precision, but takes uncomfortably long for very large patterns.
2578
2579 20. Extended pcre_study() to be more clever in cases where a branch of a
2580 subpattern has no definite first character. For example, (a*|b*)[cd] would
2581 previously give no result from pcre_study(). Now it recognizes that the
2582 first character must be a, b, c, or d.
2583
2584 21. There was an incorrect error "recursive call could loop indefinitely" if
2585 a subpattern (or the entire pattern) that was being tested for matching an
2586 empty string contained only one non-empty item after a nested subpattern.
2587 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2588 incorrectly, because the \d was being skipped in the check.
2589
2590 22. The pcretest program now has a new pattern option /B and a command line
2591 option -b, which is equivalent to adding /B to every pattern. This causes
2592 it to show the compiled bytecode, without the additional information that
2593 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2594 is the same as /B/I).
2595
2596 23. A new optimization is now able automatically to treat some sequences such
2597 as a*b as a*+b. More specifically, if something simple (such as a character
2598 or a simple class like \d) has an unlimited quantifier, and is followed by
2599 something that cannot possibly match the quantified thing, the quantifier
2600 is automatically "possessified".
2601
2602 24. A recursive reference to a subpattern whose number was greater than 39
2603 went wrong under certain circumstances in UTF-8 mode. This bug could also
2604 have affected the operation of pcre_study().
2605
2606 25. Realized that a little bit of performance could be had by replacing
2607 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2608
2609 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2610
2611 27. Possessive quantifiers such as a++ were previously implemented by turning
2612 them into atomic groups such as ($>a+). Now they have their own opcodes,
2613 which improves performance. This includes the automatically created ones
2614 from 23 above.
2615
2616 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2617 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2618 the first matched character to be a colon. This applied both to named and
2619 numbered groups.
2620
2621 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2622
2623 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2624 building PCRE on Windows easier, so I have included it in the distribution.
2625
2626 31. There is now a check in pcretest against a ridiculously large number being
2627 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2628 loop, the loop is abandoned.
2629
2630 32. Forward references to subpatterns in conditions such as (?(2)...) where
2631 subpattern 2 is defined later cause pcre_compile() to search forwards in
2632 the pattern for the relevant set of parentheses. This search went wrong
2633 when there were unescaped parentheses in a character class, parentheses
2634 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2635
2636 33. "Subroutine" calls and backreferences were previously restricted to
2637 referencing subpatterns earlier in the regex. This restriction has now
2638 been removed.
2639
2640 34. Added a number of extra features that are going to be in Perl 5.10. On the
2641 whole, these are just syntactic alternatives for features that PCRE had
2642 previously implemented using the Python syntax or my own invention. The
2643 other formats are all retained for compatibility.
2644
2645 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2646 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2647 also .NET compatible.
2648
2649 (b) A recursion or subroutine call to a named group can now be defined as
2650 (?&name) as well as (?P>name).
2651
2652 (c) A backreference to a named group can now be defined as \k<name> or
2653 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2654 5.10, are also .NET compatible.
2655
2656 (d) A conditional reference to a named group can now use the syntax
2657 (?(<name>) or (?('name') as well as (?(name).
2658
2659 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2660 groups (named and numbered) that are never evaluated inline, but can be
2661 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2662 is always false. There may be only one alternative in such a group.
2663
2664 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2665 as the simple (?(R). The condition is true only if the most recent
2666 recursion is that of the given number or name. It does not search out
2667 through the entire recursion stack.
2668
2669 (g) The escape \gN or \g{N} has been added, where N is a positive or
2670 negative number, specifying an absolute or relative reference.
2671
2672 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2673 some "unreachable code" warnings.
2674
2675 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2676 things, this adds five new scripts.
2677
2678 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2679 There were also incompatibilities regarding the handling of \Q..\E inside
2680 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2681 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2682
2683 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2684 matches an empty string, and forcibly breaks the loop. There were bugs in
2685 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2686 against aaaa the result was just "a" rather than "aaaa", for example. Two
2687 separate and independent bugs (that affected different cases) have been
2688 fixed.
2689
2690 39. Refactored the code to abolish the use of different opcodes for small
2691 capturing bracket numbers. This is a tidy that I avoided doing when I
2692 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2693 The new approach is not only tidier, it makes it possible to reduce the
2694 memory needed to fix the previous bug (38).
2695
2696 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2697 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2698 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2699 mode.
2700
2701 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2702 report.
2703
2704 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2705 copy construction and assignment in the C++ wrapper.
2706
2707 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2708 couple of bugs that could have given wrong results in the "--newline=crlf"
2709 case.
2710
2711 44. Added a number of casts and did some reorganization of signed/unsigned int
2712 variables following suggestions from Dair Grant. Also renamed the variable
2713 "this" as "item" because it is a C++ keyword.
2714
2715 45. Arranged for dftables to add
2716
2717 #include "pcre_internal.h"
2718
2719 to pcre_chartables.c because without it, gcc 4.x may remove the array
2720 definition from the final binary if PCRE is built into a static library and
2721 dead code stripping is activated.
2722
2723 46. For an unanchored pattern, if a match attempt fails at the start of a
2724 newline sequence, and the newline setting is CRLF or ANY, and the next two
2725 characters are CRLF, advance by two characters instead of one.
2726
2727
2728 Version 6.7 04-Jul-06
2729 ---------------------
2730
2731 1. In order to handle tests when input lines are enormously long, pcretest has
2732 been re-factored so that it automatically extends its buffers when
2733 necessary. The code is crude, but this _is_ just a test program. The
2734 default size has been increased from 32K to 50K.
2735
2736 2. The code in pcre_study() was using the value of the re argument before
2737 testing it for NULL. (Of course, in any sensible call of the function, it
2738 won't be NULL.)
2739
2740 3. The memmove() emulation function in pcre_internal.h, which is used on
2741 systems that lack both memmove() and bcopy() - that is, hardly ever -
2742 was missing a "static" storage class specifier.
2743
2744 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2745 containing an extended class (one that cannot be represented by a bitmap
2746 because it contains high-valued characters or Unicode property items, e.g.
2747 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2748 pattern, but PCRE should not loop if you do not (it no longer does).
2749 [Detail: two cases were found: (a) a repeated subpattern containing an
2750 extended class; (b) a recursive reference to a subpattern that followed a
2751 previous extended class. It wasn't skipping over the extended class
2752 correctly when UTF-8 mode was not set.]
2753
2754 5. A negated single-character class was not being recognized as fixed-length
2755 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2756 compile error "lookbehind assertion is not fixed length".
2757
2758 6. The RunPerlTest auxiliary script was showing an unexpected difference
2759 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2760 write a Perl script that can interpret lines of an input file either as
2761 byte characters or as UTF-8, which is what "perltest" was being required to
2762 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2763 can't do is switch easily at run time between having the "use utf8;" pragma
2764 or not. In the end, I fudged it by using the RunPerlTest script to insert
2765 "use utf8;" explicitly for the UTF-8 tests.
2766
2767 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2768 the end of the subject string, contrary to the documentation and to what
2769 Perl does. This was true of both matching functions. Now it matches only at
2770 the start of the subject and immediately after *internal* newlines.
2771
2772 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2773 a pointer to an int instead of a pointer to an unsigned long int. This
2774 caused problems on 64-bit systems.
2775
2776 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2777 instance of the 'standard' template library not being so standard".
2778
2779 10. There was no check on the number of named subpatterns nor the maximum
2780 length of a subpattern name. The product of these values is used to compute
2781 the size of the memory block for a compiled pattern. By supplying a very
2782 long subpattern name and a large number of named subpatterns, the size
2783 computation could be caused to overflow. This is now prevented by limiting
2784 the length of names to 32 characters, and the number of named subpatterns
2785 to 10,000.
2786
2787 11. Subpatterns that are repeated with specific counts have to be replicated in
2788 the compiled pattern. The size of memory for this was computed from the
2789 length of the subpattern and the repeat count. The latter is limited to
2790 65535, but there was no limit on the former, meaning that integer overflow
2791 could in principle occur. The compiled length of a repeated subpattern is
2792 now limited to 30,000 bytes in order to prevent this.
2793
2794 12. Added the optional facility to have named substrings with the same name.
2795
2796 13. Added the ability to use a named substring as a condition, using the
2797 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2798 are numbers (not recommended). Forward references are permitted.
2799
2800 14. Added forward references in named backreferences (if you see what I mean).
2801
2802 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2803 pattern could run off the end of the subject. For example, the pattern
2804 "(?s)(.{1,5})"8 did this with the subject "ab".
2805
2806 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2807 PCRE_CASELESS was set when matching characters that were quantified with ?
2808 or *.
2809
2810 17. A character class other than a single negated character that had a minimum
2811 but no maximum quantifier - for example [ab]{6,} - was not handled
2812 correctly by pce_dfa_exec(). It would match only one character.
2813
2814 18. A valid (though odd) pattern that looked like a POSIX character
2815 class but used an invalid character after [ (for example [[,abc,]]) caused
2816 pcre_compile() to give the error "Failed: internal error: code overflow" or
2817 in some cases to crash with a glibc free() error. This could even happen if
2818 the pattern terminated after [[ but there just happened to be a sequence of
2819 letters, a binary zero, and a closing ] in the memory that followed.
2820
2821 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2822 over the years. Originally (before any Unicode support), just the bottom 8
2823 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2824 output from "man perlunicode" includes this:
2825
2826 The regular expression compiler produces polymorphic opcodes. That
2827 is, the pattern adapts to the data and automatically switches to
2828 the Unicode character scheme when presented with Unicode data--or
2829 instead uses a traditional byte scheme when presented with byte
2830 data.
2831
2832 Sadly, a wide octal escape does not cause a switch, and in a string with
2833 no other multibyte characters, these octal escapes are treated as before.
2834 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2835 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2836 Unicode string.
2837
2838 I have not perpetrated such confusion in PCRE. Up till now, it took just
2839 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2840 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2841 translate to the appropriate multibyte character.
2842
2843 29. Applied some refactoring to reduce the number of warnings from Microsoft
2844 and Borland compilers. This has included removing the fudge introduced
2845 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2846 a warning about an unused variable.
2847
2848 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2849 characters since release 4.0, because Perl (from release 5.004) does not.
2850 [Or at least, is documented not to: some releases seem to be in conflict
2851 with the documentation.] However, when a pattern was studied with
2852 pcre_study() and all its branches started with \s, PCRE still included VT
2853 as a possible starting character. Of course, this did no harm; it just
2854 caused an unnecessary match attempt.
2855
2856 22. Removed a now-redundant internal flag bit that recorded the fact that case
2857 dependency changed within the pattern. This was once needed for "required
2858 byte" processing, but is no longer used. This recovers a now-scarce options
2859 bit. Also moved the least significant internal flag bit to the most-
2860 significant bit of the word, which was not previously used (hangover from
2861 the days when it was an int rather than a uint) to free up another bit for
2862 the future.
2863
2864 23. Added support for CRLF line endings as well as CR and LF. As well as the
2865 default being selectable at build time, it can now be changed at runtime
2866 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2867 specify that it is scanning data with non-default line endings.
2868
2869 24. Changed the definition of CXXLINK to make it agree with the definition of
2870 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2871
2872 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2873 recursions. This makes a big different to stack usage for some patterns.
2874
2875 26. If a subpattern containing a named recursion or subroutine reference such
2876 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2877 the space required for the compiled pattern went wrong and gave too small a
2878 value. Depending on the environment, this could lead to "Failed: internal
2879 error: code overflow at offset 49" or "glibc detected double free or
2880 corruption" errors.
2881
2882 27. Applied patches from Google (a) to support the new newline modes and (b) to
2883 advance over multibyte UTF-8 characters in GlobalReplace.
2884
2885 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2886 difference for some implementation of PCRE in some Windows version.
2887
2888 29. Added some extra testing facilities to pcretest:
2889
2890 \q<number> in a data line sets the "match limit" value
2891 \Q<number> in a data line sets the "match recursion limt" value
2892 -S <number> sets the stack size, where <number> is in megabytes
2893
2894 The -S option isn't available for Windows.
2895
2896
2897 Version 6.6 06-Feb-06
2898 ---------------------
2899
2900 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2901 in pcreposix.h. I have copied the definition from pcre.h.
2902
2903 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2904 because pcre.h is no longer a built file.
2905
2906 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2907 not normally included in the compiled code.
2908
2909
2910 Version 6.5 01-Feb-06
2911 ---------------------
2912
2913 1. When using the partial match feature with pcre_dfa_exec(), it was not
2914 anchoring the second and subsequent partial matches at the new starting
2915 point. This could lead to incorrect results. For example, with the pattern
2916 /1234/, partially matching against "123" and then "a4" gave a match.
2917
2918 2. Changes to pcregrep:
2919
2920 (a) All non-match returns from pcre_exec() were being treated as failures
2921 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2922 error message is output. Some extra information is given for the
2923 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2924 probably the only errors that are likely to be caused by users (by
2925 specifying a regex that has nested indefinite repeats, for instance).
2926 If there are more than 20 of these errors, pcregrep is abandoned.
2927
2928 (b) A binary zero was treated as data while matching, but terminated the
2929 output line if it was written out. This has been fixed: binary zeroes
2930 are now no different to any other data bytes.
2931
2932 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2933 used to set a locale for matching. The --locale=xxxx long option has
2934 been added (no short equivalent) to specify a locale explicitly on the
2935 pcregrep command, overriding the environment variables.
2936
2937 (d) When -B was used with -n, some line numbers in the output were one less
2938 than they should have been.
2939
2940 (e) Added the -o (--only-matching) option.
2941
2942 (f) If -A or -C was used with -c (count only), some lines of context were
2943 accidentally printed for the final match.
2944
2945 (g) Added the -H (--with-filename) option.
2946
2947 (h) The combination of options -rh failed to suppress file names for files
2948 that were found from directory arguments.
2949
2950 (i) Added the -D (--devices) and -d (--directories) options.
2951
2952 (j) Added the -F (--fixed-strings) option.
2953
2954 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2955
2956 (l) Added the --colo(u)r option.
2957
2958 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2959 is not present by default.
2960
2961 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2962 items such as (?R) or (?1), when the recursion could match a number of
2963 alternatives. If it matched one of the alternatives, but subsequently,
2964 outside the recursion, there was a failure, the code tried to back up into
2965 the recursion. However, because of the way PCRE is implemented, this is not
2966 possible, and the result was an incorrect result from the match.
2967
2968 In order to prevent this happening, the specification of recursion has
2969 been changed so that all such subpatterns are automatically treated as
2970 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2971
2972 4. I had overlooked the fact that, in some locales, there are characters for
2973 which isalpha() is true but neither isupper() nor islower() are true. In
2974 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2975 and ordfeminine) are like this. This affected the treatment of \w and \W
2976 when they appeared in character classes, but not when they appeared outside
2977 a character class. The bit map for "word" characters is now created
2978 separately from the results of isalnum() instead of just taking it from the
2979 upper, lower, and digit maps. (Plus the underscore character, of course.)
2980
2981 5. The above bug also affected the handling of POSIX character classes such as
2982 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2983 permanent tables. Instead, the bit maps for such a class were previously
2984 created as the appropriate unions of the upper, lower, and digit bitmaps.
2985 Now they are created by subtraction from the [[:word:]] class, which has
2986 its own bitmap.
2987
2988 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2989 It is created by subtracting the vertical space characters (\x09, \x0a,
2990 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2991 subtraction was done in the overall bitmap for a character class, meaning
2992 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2993 be recognized. This bug has been fixed.
2994
2995 7. Patches from the folks at Google:
2996
2997 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2998 real life, but is still worth protecting against".
2999
3000 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3001 regular expressions".
3002
3003 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3004 have it.
3005
3006 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3007 "configure" and the latter not, in order to fix a problem somebody had
3008 with compiling the Arg class on HP-UX.
3009
3010 (e) Improve the error-handling of the C++ wrapper a little bit.
3011
3012 (f) New tests for checking recursion limiting.
3013
3014 8. The pcre_memmove() function, which is used only if the environment does not
3015 have a standard memmove() function (and is therefore rarely compiled),
3016 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3017 returning a result (though PCRE never actually uses the result).
3018
3019 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3020 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3021 returned instead of calling malloc() with an overflowing number that would
3022 most likely cause subsequent chaos.
3023
3024 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3025
3026 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3027 with this option is matched, the nmatch and pmatch options of regexec() are
3028 ignored.
3029
3030 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3031 provided in case anyone wants to the the POSIX interface with UTF-8
3032 strings.
3033
3034 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3035 C++ linking (needed for some HP-UX environments).
3036
3037 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3038 (unused parameter) and in the pcre_printint() function (omitted "default"
3039 switch label when the default is to do nothing).
3040
3041 15. Added some code to make it possible, when PCRE is compiled as a C++
3042 library, to replace subject pointers for pcre_exec() with a smart pointer
3043 class, thus making it possible to process discontinuous strings.
3044
3045 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3046 much the same function. They were added by different people who were trying
3047 to make PCRE easy to compile on non-Unix systems. It has been suggested
3048 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3049 for compiling on Windows systems. I have therefore replaced it with
3050 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3051 defaults to "extern" for C or "extern C" for C++, which works fine on
3052 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3053 SCOPE with something explicit in config.h. In addition:
3054
3055 (a) pcreposix.h still had just "extern" instead of either of these macros;
3056 I have replaced it with PCRE_DATA_SCOPE.
3057
3058 (b) Functions such as _pcre_xclass(), which are internal to the library,
3059 but external in the C sense, all had PCRE_EXPORT in their definitions.
3060 This is apparently wrong for the Windows case, so I have removed it.
3061 (It makes no difference on Unix-like systems.)
3062
3063 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3064 of recursive calls to match(). This is different to MATCH_LIMIT because
3065 that limits the total number of calls to match(), not all of which increase
3066 the depth of recursion. Limiting the recursion depth limits the amount of
3067 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3068 when PCRE is compiled, and changed at run time. A patch from Google adds
3069 this functionality to the C++ interface.
3070
3071 18. Changes to the handling of Unicode character properties:
3072
3073 (a) Updated the table to Unicode 4.1.0.
3074
3075 (b) Recognize characters that are not in the table as "Cn" (undefined).
3076
3077 (c) I revised the way the table is implemented to a much improved format
3078 which includes recognition of ranges. It now supports the ranges that
3079 are defined in UnicodeData.txt, and it also amalgamates other
3080 characters into ranges. This has reduced the number of entries in the
3081 table from around 16,000 to around 3,000, thus reducing its size
3082 considerably. I realized I did not need to use a tree structure after
3083 all - a binary chop search is just as efficient. Having reduced the
3084 number of entries, I extended their size from 6 bytes to 8 bytes to
3085 allow for more data.
3086
3087 (d) Added support for Unicode script names via properties such as \p{Han}.
3088
3089 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3090 matching that character.
3091
3092 20. When matching a repeated Unicode property with a minimum greater than zero,
3093 (for example \pL{2,}), PCRE could look past the end of the subject if it
3094 reached it while seeking the minimum number of characters. This could
3095 happen only if some of the characters were more than one byte long, because
3096 there is a check for at least the minimum number of bytes.
3097
3098 21. Refactored the implementation of \p and \P so as to be more general, to
3099 allow for more different types of property in future. This has changed the
3100 compiled form incompatibly. Anybody with saved compiled patterns that use
3101 \p or \P will have to recompile them.
3102
3103 22. Added "Any" and "L&" to the supported property types.
3104
3105 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3106 but give a compile time error if the value is greater than 0xff.
3107
3108 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3109 accidentally not being installed or uninstalled.
3110
3111 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3112 made were to insert the current release number. This seemed silly, because
3113 it made things harder for people building PCRE on systems that don't run
3114 "configure". I have turned pcre.h into a distributed file, no longer built
3115 by "configure", with the version identification directly included. There is
3116 no longer a pcre.h.in file.
3117
3118 However, this change necessitated a change to the pcre-config script as
3119 well. It is built from pcre-config.in, and one of the substitutions was the
3120 release number. I have updated configure.ac so that ./configure now finds
3121 the release number by grepping pcre.h.
3122
3123 26. Added the ability to run the tests under valgrind.
3124
3125
3126 Version 6.4 05-Sep-05
3127 ---------------------
3128
3129 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3130 "--" to be printed when multiple files were scanned, even when none of the
3131 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3132 consider it to be a bug, and have restored the previous behaviour.
3133
3134 2. A couple of code tidies to get rid of compiler warnings.
3135
3136 3. The pcretest program used to cheat by referring to symbols in the library
3137 whose names begin with _pcre_. These are internal symbols that are not
3138 really supposed to be visible externally, and in some environments it is
3139 possible to suppress them. The cheating is now confined to including
3140 certain files from the library's source, which is a bit cleaner.
3141
3142 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3143 file's purpose clearer.
3144
3145 5. Reorganized pcre_ucp_findchar().
3146
3147
3148 Version 6.3 15-Aug-05
3149 ---------------------
3150
3151 1. The file libpcre.pc.in did not have general read permission in the tarball.
3152
3153 2. There were some problems when building without C++ support:
3154
3155 (a) If C++ support was not built, "make install" and "make test" still
3156 tried to test it.
3157
3158 (b) There were problems when the value of CXX was explicitly set. Some
3159 changes have been made to try to fix these, and ...
3160
3161 (c) --disable-cpp can now be used to explicitly disable C++ support.
3162
3163 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3164 backslash in a target when C++ was disabled. This confuses some
3165 versions of "make", apparently. Using an intermediate variable solves
3166 this. (Same for CPP_LOBJ.)
3167
3168 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3169 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3170 necessary on certain architectures.
3171
3172 4. Added a setting of -export-symbols-regex to the link command to remove
3173 those symbols that are exported in the C sense, but actually are local
3174 within the library, and not documented. Their names all begin with
3175 "_pcre_". This is not a perfect job, because (a) we have to except some
3176 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3177 available (and never for static libraries). I have made a note to try to
3178 find a way round (a) in the future.
3179
3180
3181 Version 6.2 01-Aug-05
3182 ---------------------
3183
3184 1. There was no test for integer overflow of quantifier values. A construction
3185 such as {1111111111111111} would give undefined results. What is worse, if
3186 a minimum quantifier for a parenthesized subpattern overflowed and became
3187 negative, the calculation of the memory size went wrong. This could have
3188 led to memory overwriting.
3189
3190 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3191
3192 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3193 operating environments where this matters.
3194
3195 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3196 PCRE options from within the C++ wrapper.
3197
3198 5. Named capturing subpatterns were not being correctly counted when a pattern
3199 was compiled. This caused two problems: (a) If there were more than 100
3200 such subpatterns, the calculation of the memory needed for the whole
3201 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3202 back references of the form \12, where the number was greater than 9, were
3203 not recognized as back references, even though there were sufficient
3204 previous subpatterns.
3205
3206 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3207 versions of gcc, e.g. 2.95.4.
3208
3209
3210 Version 6.1 21-Jun-05
3211 ---------------------
3212
3213 1. There was one reference to the variable "posix" in pcretest.c that was not
3214 surrounded by "#if !defined NOPOSIX".
3215
3216 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3217 the cross-check on the old pcre_info() function, for the benefit of the
3218 cut-down version of PCRE that is currently imported into Exim.
3219
3220 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3221 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3222 patterns that start (?i) but I don't think that matters. The use of (?i) is
3223 just an example; this all applies to the other options as well.
3224
3225 4. Since libtool seems to echo the compile commands it is issuing, the output
3226 from "make" can be reduced a bit by putting "@" in front of each libtool
3227 compile command.
3228
3229 5. Patch from the folks at Google for configure.in to be a bit more thorough
3230 in checking for a suitable C++ installation before trying to compile the
3231 C++ stuff. This should fix a reported problem when a compiler was present,
3232 but no suitable headers.
3233
3234 6. The man pages all had just "PCRE" as their title. I have changed them to
3235 be the relevant file name. I have also arranged that these names are
3236 retained in the file doc/pcre.txt, which is a concatenation in text format
3237 of all the man pages except the little individual ones for each function.
3238
3239 7. The NON-UNIX-USE file had not been updated for the different set of source
3240 files that come with release 6. I also added a few comments about the C++
3241 wrapper.
3242
3243
3244 Version 6.0 07-Jun-05
3245 ---------------------
3246
3247 1. Some minor internal re-organization to help with my DFA experiments.
3248
3249 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3250 didn't matter for the library itself when fully configured, but did matter
3251 when compiling without UCP support, or within Exim, where the ucp files are
3252 not imported.
3253
3254 3. Refactoring of the library code to split up the various functions into
3255 different source modules. The addition of the new DFA matching code (see
3256 below) to a single monolithic source would have made it really too
3257 unwieldy, quite apart from causing all the code to be include in a
3258 statically linked application, when only some functions are used. This is
3259 relevant even without the DFA addition now that patterns can be compiled in
3260 one application and matched in another.
3261
3262 The downside of splitting up is that there have to be some external
3263 functions and data tables that are used internally in different modules of
3264 the library but which are not part of the API. These have all had their
3265 names changed to start with "_pcre_" so that they are unlikely to clash
3266 with other external names.
3267
3268 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3269 a different (DFA) algorithm. Although it is slower than the original
3270 function, it does have some advantages for certain types of matching
3271 problem.
3272
3273 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3274 including restarting after a partial match.
3275
3276 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3277 defined when compiling for Windows was sent to me. I have put it into the
3278 code, though I have no means of testing or verifying it.
3279
3280 7. Added the pcre_refcount() auxiliary function.
3281
3282 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3283 match before or at the first newline in the subject string. In pcretest,
3284 the /f option on a pattern can be used to set this.
3285
3286 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3287 would behave wrongly. This has been present in PCRE since release 4.0.
3288
3289 10. A number of changes to the pcregrep command:
3290
3291 (a) Refactored how -x works; insert ^(...)$ instead of setting
3292 PCRE_ANCHORED and checking the length, in preparation for adding
3293 something similar for -w.
3294
3295 (b) Added the -w (match as a word) option.
3296
3297 (c) Refactored the way lines are read and buffered so as to have more
3298 than one at a time available.
3299
3300 (d) Implemented a pcregrep test script.
3301
3302 (e) Added the -M (multiline match) option. This allows patterns to match
3303 over several lines of the subject. The buffering ensures that at least
3304 8K, or the rest of the document (whichever is the shorter) is available
3305 for matching (and similarly the previous 8K for lookbehind assertions).
3306
3307 (f) Changed the --help output so that it now says
3308
3309 -w, --word-regex(p)
3310
3311 instead of two lines, one with "regex" and the other with "regexp"
3312 because that confused at least one person since the short forms are the
3313 same. (This required a bit of code, as the output is generated
3314 automatically from a table. It wasn't just a text change.)
3315
3316 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3317 option but starts with a hyphen. Could be a pattern or a path name
3318 starting with a hyphen, for instance.
3319
3320 (h) "-" can be given as a file name to represent stdin.
3321
3322 (i) When file names are being printed, "(standard input)" is used for
3323 the standard input, for compatibility with GNU grep. Previously
3324 "<stdin>" was used.
3325
3326 (j) The option --label=xxx can be used to supply a name to be used for
3327 stdin when file names are being printed. There is no short form.
3328
3329 (k) Re-factored the options decoding logic because we are going to add
3330 two more options that take data. Such options can now be given in four
3331 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3332
3333 (l) Added the -A, -B, and -C options for requesting that lines of context
3334 around matches be printed.
3335
3336 (m) Added the -L option to print the names of files that do not contain
3337 any matching lines, that is, the complement of -l.
3338
3339 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3340 continue to scan other files.
3341
3342 (o) The -s option was incorrectly implemented. For compatibility with other
3343 greps, it now suppresses the error message for a non-existent or non-
3344 accessible file (but not the return code). There is a new option called
3345 -q that suppresses the output of matching lines, which was what -s was
3346 previously doing.
3347
3348 (p) Added --include and --exclude options to specify files for inclusion
3349 and exclusion when recursing.
3350
3351 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3352 Hopefully, it now does.
3353
3354 12. Missing cast in pcre_study().
3355
3356 13. Added an "uninstall" target to the makefile.
3357
3358 14. Replaced "extern" in the function prototypes in Makefile.in with
3359 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3360 world, but is set differently for Windows.
3361
3362 15. Added a second compiling function called pcre_compile2(). The only
3363 difference is that it has an extra argument, which is a pointer to an
3364 integer error code. When there is a compile-time failure, this is set
3365 non-zero, in addition to the error test pointer being set to point to an
3366 error message. The new argument may be NULL if no error number is required
3367 (but then you may as well call pcre_compile(), which is now just a
3368 wrapper). This facility is provided because some applications need a
3369 numeric error indication, but it has also enabled me to tidy up the way
3370 compile-time errors are handled in the POSIX wrapper.
3371
3372 16. Added VPATH=.libs to the makefile; this should help when building with one
3373 prefix path and installing with another. (Or so I'm told by someone who
3374 knows more about this stuff than I do.)
3375
3376 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3377 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3378 match everything, including newlines. This is not POSIX-compatible, but
3379 somebody wanted the feature. From pcretest it can be activated by using
3380 both the P and the s flags.
3381
3382 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3383
3384 19. libpcre.pc was being incorrectly installed as executable.
3385
3386 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3387 it now also looks for '\r' so that it will work unmodified on Windows.
3388
3389 21. Added Google's contributed C++ wrapper to the distribution.
3390
3391 22. Added some untidy missing memory free() calls in pcretest, to keep
3392 Electric Fence happy when testing.
3393
3394
3395
3396 Version 5.0 13-Sep-04
3397 ---------------------
3398
3399 1. Internal change: literal characters are no longer packed up into items
3400 containing multiple characters in a single byte-string. Each character
3401 is now matched using a separate opcode. However, there may be more than one
3402 byte in the character in UTF-8 mode.
3403
3404 2. The pcre_callout_block structure has two new fields: pattern_position and
3405 next_item_length. These contain the offset in the pattern to the next match
3406 item, and its length, respectively.
3407
3408 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3409 insertion of callouts before each pattern item. Added the /C option to
3410 pcretest to make use of this.
3411
3412 4. On the advice of a Windows user, the lines
3413
3414 #if defined(_WIN32) || defined(WIN32)
3415 _setmode( _fileno( stdout ), 0x8000 );
3416 #endif /* defined(_WIN32) || defined(WIN32) */
3417
3418 have been added to the source of pcretest. This apparently does useful
3419 magic in relation to line terminators.
3420
3421 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3422 for the benefit of those environments where the "b" makes a difference.
3423
3424 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3425 to know about it. I have put a hack into configure.in that adds in code
3426 to set GCC=yes if CC=icc. This seems to end up at a point in the
3427 generated configure script that is early enough to affect the setting of
3428 compiler options, which is what is needed, but I have no means of testing
3429 whether it really works. (The user who reported this had patched the
3430 generated configure script, which of course I cannot do.)
3431
3432 LATER: After change 22 below (new libtool files), the configure script
3433 seems to know about icc (and also ecc). Therefore, I have commented out
3434 this hack in configure.in.
3435
3436 7. Added support for pkg-config (2 patches were sent in).
3437
3438 8. Negated POSIX character classes that used a combination of internal tables
3439 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3440 [[:^ascii]]. Typically, they would match almost any characters. The other
3441 POSIX classes were not broken in this way.
3442
3443 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3444 to find the match, as PCRE was deluded into thinking that the match had to
3445 start at the start point or following a newline. The same bug applied to
3446 patterns with negative forward assertions or any backward assertions
3447 preceding ".*" at the start, unless the pattern required a fixed first
3448 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3449
3450 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3451 starting at the last subject character, bytes beyond the end of the subject
3452 string were read.
3453
3454 11. Renamed the variable "class" as "classbits" to make life easier for C++
3455 users. (Previously there was a macro definition, but it apparently wasn't
3456 enough.)
3457
3458 12. Added the new field "tables" to the extra data so that tables can be passed
3459 in at exec time, or the internal tables can be re-selected. This allows
3460 a compiled regex to be saved and re-used at a later time by a different
3461 program that might have everything at different addresses.
3462
3463 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3464 -R library as well as a -L library.
3465
3466 14. The debugging options of pcretest (-d on the command line or D on a
3467 pattern) showed incorrect output for anything following an extended class
3468 that contained multibyte characters and which was followed by a quantifier.
3469
3470 15. Added optional support for general category Unicode character properties
3471 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3472 support. It adds about 90K to the size of the library. The meanings of the
3473 inbuilt class escapes such as \d and \s have NOT been changed.
3474
3475 16. Updated pcredemo.c to include calls to free() to release the memory for the
3476 compiled pattern.
3477
3478 17. The generated file chartables.c was being created in the source directory
3479 instead of in the building directory. This caused the build to fail if the
3480 source directory was different from the building directory, and was
3481 read-only.
3482
3483 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3484 file. No doubt somebody will tell me if they don't make sense... Also added
3485 Dan Mooney's comments about building on OpenVMS.
3486
3487 19. Added support for partial matching via the PCRE_PARTIAL option for
3488 pcre_exec() and the \P data escape in pcretest.
3489
3490 20. Extended pcretest with 3 new pattern features:
3491
3492 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3493 write the compiled pattern to the file whose name is "rest-of-line".
3494 This is a straight binary dump of the data, with the saved pointer to
3495 the character tables forced to be NULL. The study data, if any, is
3496 written too. After writing, pcretest reads a new pattern.
3497
3498 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3499 compiled pattern from the given file. There must not be any
3500 occurrences of "<" in the file name (pretty unlikely); if there are,
3501 pcretest will instead treat the initial "<" as a pattern delimiter.
3502 After reading in the pattern, pcretest goes on to read data lines as
3503 usual.
3504
3505 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3506 and 16-bit fields in a compiled pattern, to simulate a pattern that
3507 was compiled on a host of opposite endianness.
3508
3509 21. The pcre-exec() function can now cope with patterns that were compiled on
3510 hosts of opposite endianness, with this restriction:
3511
3512 As for any compiled expression that is saved and used later, the tables
3513 pointer field cannot be preserved; the extra_data field in the arguments
3514 to pcre_exec() should be used to pass in a tables address if a value
3515 other than the default internal tables were used at compile time.
3516
3517 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3518 now diagnosed as an error. Previously, most of the time, a negative number
3519 would have been treated as zero, but if in addition "ovector" was passed as
3520 NULL, a crash could occur.
3521
3522 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3523 new versions from the libtool 1.5 distribution (the last one is a copy of
3524 a file called libtool.m4). This seems to have fixed the need to patch
3525 "configure" to support Darwin 1.3 (which I used to do). However, I still
3526 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3527 workstation).
3528
3529 24. Changed the PCRE licence to be the more standard "BSD" licence.
3530
3531
3532 Version 4.5 01-Dec-03
3533 ---------------------
3534
3535 1. There has been some re-arrangement of the code for the match() function so
3536 that it can be compiled in a version that does not call itself recursively.
3537 Instead, it keeps those local variables that need separate instances for
3538 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3539 needs to "recurse". Keeping track of where control must go is done by means
3540 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3541 hide most of the details from the main code, and operates only if
3542 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3543 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3544 operating.
3545
3546 To make it easier for callers to provide specially tailored get/free
3547 functions for this usage, two new functions, pcre_stack_malloc, and
3548 pcre_stack_free, are used. They are always called in strict stacking order,
3549 and the size of block requested is always the same.
3550
3551 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3552 PCRE has been compiled to use the stack or the heap for recursion. The
3553 -C option of pcretest uses this to show which version is compiled.
3554
3555 A new data escape \S, is added to pcretest; it causes the amounts of store
3556 obtained and freed by both kinds of malloc/free at match time to be added
3557 to the output.
3558
3559 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3560 what's available on my current Linux desktop machine.
3561
3562 3. When matching a UTF-8 string, the test for a valid string at the start has
3563 been extended. If start_offset is not zero, PCRE now checks that it points
3564 to a byte that is the start of a UTF-8 character. If not, it returns
3565 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3566 this is necessary because there may be backward assertions in the pattern.
3567 When matching the same subject several times, it may save resources to use
3568 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3569
3570 4. The code for checking the validity of UTF-8 strings has been tightened so
3571 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3572 containing "overlong sequences".
3573
3574 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3575 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3576 should have been "&", but it just so happened that all the cases this let
3577 through by mistake were picked up later in the function.
3578
3579 6. I had used a variable called "isblank" - this is a C99 function, causing
3580 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3581
3582 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3583 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3584 so the version is shown; (c) stop "make test" echoing "./RunTest".
3585
3586 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3587
3588 9. The prototype for memmove() for systems that don't have it was using
3589 size_t, but the inclusion of the header that defines size_t was later. I've
3590 moved the #includes for the C headers earlier to avoid this.
3591
3592 10. Added some adjustments to the code to make it easier to compiler on certain
3593 special systems:
3594
3595 (a) Some "const" qualifiers were missing.
3596 (b) Added the macro EXPORT before all exported functions; by default this
3597 is defined to be empty.
3598 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3599 that it reads its output file name as an argument instead of writing
3600 to the standard output and assuming this can be redirected.
3601
3602 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3603 class containing characters with values greater than 255, PCRE compilation
3604 went into a loop.
3605
3606 12. A recursive reference to a subpattern that was within another subpattern
3607 that had a minimum quantifier of zero caused PCRE to crash. For example,
3608 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3609 recursion. If the recursively-called subpattern itself had a zero repeat,
3610 that was OK.
3611
3612 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3613 buffer into which it was copied (for escape processing) was still set at
3614 1024, so long lines caused crashes.
3615
3616 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3617 "internal error: code overflow...". This applied to any character class
3618 that was followed by a possessive quantifier.
3619
3620 15. Modified the Makefile to add libpcre.la as a prerequisite for
3621 libpcreposix.la because I was told this is needed for a parallel build to
3622 work.
3623
3624 16. If a pattern that contained .* following optional items at the start was
3625 studied, the wrong optimizing data was generated, leading to matching
3626 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3627 matching string must start with a or b or c. The correct conclusion for
3628 this pattern is that a match can start with any character.
3629
3630
3631 Version 4.4 13-Aug-03
3632 ---------------------
3633
3634 1. In UTF-8 mode, a character class containing characters with values between
3635 127 and 255 was not handled correctly if the compiled pattern was studied.
3636 In fixing this, I have also improved the studying algorithm for such
3637 classes (slightly).
3638
3639 2. Three internal functions had redundant arguments passed to them. Removal
3640 might give a very teeny performance improvement.
3641
3642 3. Documentation bug: the value of the capture_top field in a callout is *one
3643 more than* the number of the hightest numbered captured substring.
3644
3645 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3646 in incorrectly linking with a previously installed version. They now link
3647 explicitly with libpcre.la.
3648
3649 5. configure.in no longer needs to recognize Cygwin specially.
3650
3651 6. A problem in pcre.in for Windows platforms is fixed.
3652
3653 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3654 pcretest, it used to include the size of the study block as part of its
3655 output. Unfortunately, the structure contains a field that has a different
3656 size on different hardware architectures. This meant that the tests that
3657 showed this size failed. As the block is currently always of a fixed size,
3658 this information isn't actually particularly useful in pcretest output, so
3659 I have just removed it.
3660
3661 8. Three pre-processor statements accidentally did not start in column 1.
3662 Sadly, there are *still* compilers around that complain, even though
3663 standard C has not required this for well over a decade. Sigh.
3664
3665 9. In pcretest, the code for checking callouts passed small integers in the
3666 callout_data field, which is a void * field. However, some picky compilers
3667 complained about the casts involved for this on 64-bit systems. Now
3668 pcretest passes the address of the small integer instead, which should get
3669 rid of the warnings.
3670
3671 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3672 both compile and run time, and gives an error if an invalid UTF-8 sequence
3673 is found. There is a option for disabling this check in cases where the
3674 string is known to be correct and/or the maximum performance is wanted.
3675
3676 11. In response to a bug report, I changed one line in Makefile.in from
3677
3678 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3679 to
3680 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3681
3682 to look similar to other lines, but I have no way of telling whether this
3683 is the right thing to do, as I do not use Windows. No doubt I'll get told
3684 if it's wrong...
3685
3686
3687 Version 4.3 21-May-03
3688 ---------------------
3689
3690 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3691 Makefile.
3692
3693 2. Some refactoring to improve the quality of the code:
3694
3695 (i) The utf8_table... variables are now declared "const".
3696
3697 (ii) The code for \cx, which used the "case flipping" table to upper case
3698 lower case letters, now just substracts 32. This is ASCII-specific,
3699 but the whole concept of \cx is ASCII-specific, so it seems
3700 reasonable.
3701
3702 (iii) PCRE was using its character types table to recognize decimal and
3703 hexadecimal digits in the pattern. This is silly, because it handles
3704 only 0-9, a-f, and A-F, but the character types table is locale-
3705 specific, which means strange things might happen. A private
3706 table is now used for this - though it costs 256 bytes, a table is
3707 much faster than multiple explicit tests. Of course, the standard
3708 character types table is still used for matching digits in subject
3709 strings against \d.
3710
3711 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3712 ending in _t are). So I've renamed it as ESC_tee.
3713
3714 3. The first argument for regexec() in the POSIX wrapper should have been
3715 defined as "const".
3716
3717 4. Changed pcretest to use malloc() for its buffers so that they can be
3718 Electric Fenced for debugging.
3719
3720 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3721 to read one or more bytes before the start of the subject string. Often this
3722 had no effect on PCRE's behaviour, but in some circumstances it could
3723 provoke a segmentation fault.
3724
3725 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3726 to try to read one or more bytes before the start of the subject string.
3727
3728 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3729 UTF-8 support could misbehave in various ways if the subject string
3730 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3731 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3732 back over UTF-8 characters.)
3733
3734
3735 Version 4.2 14-Apr-03
3736 ---------------------
3737
3738 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3739
3740 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3741 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3742 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3743 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3744 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3745 and BUILD_EXEEXT
3746 Note: automatic setting of the BUILD variables is not yet working
3747 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3748 compile-time but not at link-time
3749 [LINK]: use for linking executables only
3750 make different versions for Windows and non-Windows
3751 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3752 libraries
3753 [LINK_FOR_BUILD]: new variable
3754 [OBJEXT]: use throughout
3755 [EXEEXT]: use throughout
3756 <winshared>: new target
3757 <wininstall>: new target
3758 <dftables.o>: use native compiler
3759 <dftables>: use native linker
3760 <install>: handle Windows platform correctly
3761 <clean>: ditto
3762 <check>: ditto
3763 copy DLL to top builddir before testing
3764
3765 As part of these changes, -no-undefined was removed again. This was reported
3766 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3767 in any case.
3768
3769 3. Some tidies to get rid of compiler warnings:
3770
3771 . In the match_data structure, match_limit was an unsigned long int, whereas
3772 match_call_count was an int. I've made them both unsigned long ints.
3773
3774 . In pcretest the fact that a const uschar * doesn't automatically cast to
3775 a void * provoked a warning.
3776
3777 . Turning on some more compiler warnings threw up some "shadow" variables
3778 and a few more missing casts.
3779
3780 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3781 option, a class that contained a single character with a value between 128
3782 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3783
3784 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3785 option, a class that contained several characters, but with at least one
3786 whose value was between 128 and 255 caused PCRE to crash.
3787
3788
3789 Version 4.1 12-Mar-03
3790 ---------------------
3791
3792 1. Compiling with gcc -pedantic found a couple of places where casts were
3793 needed, and a string in dftables.c that was longer than standard compilers are
3794 required to support.
3795
3796 2. Compiling with Sun's compiler found a few more places where the code could
3797 be tidied up in order to avoid warnings.
3798
3799 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3800 first of these names is deprecated in the latest Autoconf in favour of the name
3801 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3802 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3803 analogy I have changed it to CFLAGS_FOR_BUILD.
3804
3805 4. Added -no-undefined to the linking command in the Makefile, because this is
3806 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3807 linking step for the pcreposix library.
3808
3809 5. PCRE was failing to diagnose the case of two named groups with the same
3810 name.
3811
3812 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3813 literal character that is needed in the subject for a match, and scans along to
3814 ensure that it is present before embarking on the full matching process. This
3815 saves time in cases of nested unlimited repeats that are never going to match.
3816 Problem: the scan can take a lot of time if the subject is very long (e.g.
3817 megabytes), thus penalizing straightforward matches. It is now done only if the
3818 amount of subject to be scanned is less than 1000 bytes.
3819
3820 7. A lesser problem with the same optimization is that it was recording the
3821 first character of an anchored pattern as "needed", thus provoking a search
3822 right along the subject, even when the first match of the pattern was going to
3823 fail. The "needed" character is now not set for anchored patterns, unless it
3824 follows something in the pattern that is of non-fixed length. Thus, it still
3825 fulfils its original purpose of finding quick non-matches in cases of nested
3826 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3827
3828
3829 Version 4.0 17-Feb-03
3830 ---------------------
3831
3832 1. If a comment in an extended regex that started immediately after a meta-item
3833 extended to the end of string, PCRE compiled incorrect data. This could lead to
3834 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3835
3836 2. Moved to autoconf 2.53 and libtool 1.4.2.
3837
3838 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3839 the special perltest8 script is no longer needed - all the tests can be run
3840 from a single perltest script.
3841
3842 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3843 by \s. It has now been removed in PCRE. This means it isn't recognized as
3844 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3845 class [:space:] *does* include VT, thereby creating a mess.
3846
3847 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3848 space and tab.
3849
3850 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3851 its new features into the main test script, reducing the number of scripts.
3852
3853 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3854 were backward compatible, and made the (?i) apply to the whole pattern, as if
3855 /i were given. Now it behaves more logically, and applies the option setting
3856 only to what follows. PCRE has been changed to follow suit. However, if it
3857 finds options settings right at the start of the pattern, it extracts them into
3858 the global options, as before. Thus, they show up in the info data.
3859
3860 8. Added support for the \Q...\E escape sequence. Characters in between are
3861 treated as literals. This is slightly different from Perl in that $ and @ are
3862 also handled as literals inside the quotes. In Perl, they will cause variable
3863 interpolation. Note the following examples:
3864
3865 Pattern PCRE matches Perl matches
3866
3867 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3868 \Qabc\$xyz\E abc\$xyz abc\$xyz
3869 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3870
3871 For compatibility with Perl, \Q...\E sequences are recognized inside character
3872 classes as well as outside them.
3873
3874 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3875 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3876 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3877 signed/unsigned warnings.
3878
3879 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3880 option for pcretest, so I've replaced it by a simple function that does just
3881 that job.
3882
3883 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3884 "pcregrep -".
3885
3886 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3887 Java package. This provides some syntactic sugar for simple cases of what my
3888 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3889 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3890 item, you can use this simplified notation. Note that only makes sense with
3891 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3892 greediness, whatever the setting of the PCRE_UNGREEDY option.
3893
3894 13. A change of greediness default within a pattern was not taking effect at
3895 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3896 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3897 was abstracted outside.
3898
3899 14. PCRE now supports the \G assertion. It is true when the current matching
3900 position is at the start point of the match. This differs from \A when the
3901 starting offset is non-zero. Used with the /g option of pcretest (or similar
3902 code), it works in the same way as it does for Perl's /g option. If all
3903 alternatives of a regex begin with \G, the expression is anchored to the start
3904 match position, and the "anchored" flag is set in the compiled expression.
3905
3906 15. Some bugs concerning the handling of certain option changes within patterns
3907 have been fixed. These applied to options other than (?ims). For example,
3908 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3909 been the other way round. Some of this was related to change 7 above.
3910
3911 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3912 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3913 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3914 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3915
3916 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3917 mode. Unlike ".", it always matches newline, whatever the setting of
3918 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3919 assertions. Perl allows it, but it doesn't (in general) work because it can't
3920 calculate the length of the lookbehind. At least, that's the case for Perl
3921 5.8.0 - I've been told they are going to document that it doesn't work in
3922 future.
3923
3924 18. Added an error diagnosis for escapes that PCRE does not support: these are
3925 \L, \l, \N, \P, \p, \U, \u, and \X.
3926
3927 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3928 reading past the end of the pattern in cases such as /[abcd/.
3929
3930 20. PCRE was getting more memory than necessary for patterns with classes that
3931 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3932
3933 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3934 compiling PCRE for use with Virtual Pascal.
3935
3936 22. Small fix to the Makefile to make it work properly if the build is done
3937 outside the source tree.
3938
3939 23. Added a new extension: a condition to go with recursion. If a conditional
3940 subpattern starts with (?(R) the "true" branch is used if recursion has
3941 happened, whereas the "false" branch is used only at the top level.
3942
3943 24. When there was a very long string of literal characters (over 255 bytes
3944 without UTF support, over 250 bytes with UTF support), the computation of how
3945 much memory was required could be incorrect, leading to segfaults or other
3946 strange effects.
3947
3948 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3949 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3950 there was a subsequent back reference to those brackets. This meant that, for
3951 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3952 possible to check for precisely this case. All we can do is abandon the
3953 optimization if .* occurs inside capturing brackets when there are any back
3954 references whatsoever. (See below for a better fix that came later.)
3955
3956 26. The handling of the optimization for finding the first character of a
3957 non-anchored pattern, and for finding a character that is required later in the
3958 match were failing in some cases. This didn't break the matching; it just
3959 failed to optimize when it could. The way this is done has been re-implemented.
3960
3961 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3962
3963 28. Added a new feature that provides some of the functionality that Perl
3964 provides with (?{...}). The facility is termed a "callout". The way it is done
3965 in PCRE is for the caller to provide an optional function, by setting
3966 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3967 global variable. By default it is unset, which disables all calling out. To get
3968 the function called, the regex must include (?C) at appropriate points. This
3969 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3970 This provides a means of identifying different callout points. When PCRE
3971 reaches such a point in the regex, if pcre_callout has been set, the external
3972 function is called. It is provided with data in a structure called
3973 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3974 matching continues; if it returns a non-zero value, the match at the current
3975 point fails. However, backtracking will occur if possible. [This was changed
3976 later and other features added - see item 49 below.]
3977
3978 29. pcretest is upgraded to test the callout functionality. It provides a
3979 callout function that displays information. By default, it shows the start of
3980 the match and the current position in the text. There are some new data escapes
3981 to vary what happens:
3982
3983 \C+ in addition, show current contents of captured substrings
3984 \C- do not supply a callout function
3985 \C!n return 1 when callout number n is reached
3986 \C!n!m return 1 when callout number n is reached for the mth time
3987
3988 30. If pcregrep was called with the -l option and just a single file name, it
3989 output "<stdin>" if a match was found, instead of the file name.
3990
3991 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3992 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3993 pcre_exec(). This saves a malloc/free per call. The default value of
3994 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3995 when configuring.
3996
3997 32. The default maximum size of a compiled pattern is 64K. There have been a
3998 few cases of people hitting this limit. The code now uses macros to handle the
3999 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4000 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4001 configuring. Tests 2 and 5 work only with 2-byte links because they output
4002 debugging information about compiled patterns.
4003
4004 33. Internal code re-arrangements:
4005
4006 (a) Moved the debugging function for printing out a compiled regex into
4007 its own source file (printint.c) and used #include to pull it into
4008 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4009 separate copies.
4010
4011 (b) Defined the list of op-code names for debugging as a macro in
4012 internal.h so that it is next to the definition of the opcodes.
4013
4014 (c) Defined a table of op-code lengths for simpler skipping along compiled
4015 code. This is again a macro in internal.h so that it is next to the
4016 definition of the opcodes.
4017
4018 34. Added support for recursive calls to individual subpatterns, along the
4019 lines of Robin Houston's patch (but implemented somewhat differently).
4020
4021 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4022 allow it to read and process whole directories in Win32. This code was
4023 contributed by Lionel Fourquaux; it has not been tested by me.
4024
4025 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4026 used to name a group. Names consist of alphanumerics and underscores, and must
4027 be unique. Back references use the syntax (?P=name) and recursive calls use
4028 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4029 numbers. The function pcre_fullinfo() can be used after compilation to extract
4030 a name/number map. There are three relevant calls:
4031
4032 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4033 PCRE_INFO_NAMECOUNT yields the number of entries
4034 PCRE_INFO_NAMETABLE yields a pointer to the map.
4035
4036 The map is a vector of fixed-size entries. The size of each entry depends on
4037 the length of the longest name used. The first two bytes of each entry are the
4038 group number, most significant byte first. There follows the corresponding
4039 name, zero terminated. The names are in alphabetical order.
4040
4041 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4042 case instead of 255. Making it the same both with and without UTF-8 support
4043 means that the same test output works with both.
4044
4045 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4046 calling malloc() with a zero argument.
4047
4048 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4049 optimization. I've improved things by keeping a bitmap of backreferences with
4050 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4051 fact referenced, the optimization can be applied. It is unlikely that a
4052 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4053 the match to follow \n) will appear inside brackets with a number greater than
4054 31, but if it does, any back reference > 31 suppresses the optimization.
4055
4056 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4057 of disabling numbered capturing parentheses. Any opening parenthesis that is
4058 not followed by ? behaves as if it were followed by ?: but named parentheses
4059 can still be used for capturing (and they will acquire numbers in the usual
4060 way).
4061
4062 41. Redesigned the return codes from the match() function into yes/no/error so
4063 that errors can be passed back from deep inside the nested calls. A malloc
4064 failure while inside a recursive subpattern call now causes the
4065 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4066
4067 42. It is now possible to set a limit on the number of times the match()
4068 function is called in a call to pcre_exec(). This facility makes it possible to
4069 limit the amount of recursion and backtracking, though not in a directly
4070 obvious way, because the match() function is used in a number of different
4071 circumstances. The count starts from zero for each position in the subject
4072 string (for non-anchored patterns). The default limit is, for compatibility, a
4073 large number, namely 10 000 000. You can change this in two ways:
4074
4075 (a) When configuring PCRE before making, you can use --with-match-limit=n
4076 to set a default value for the compiled library.
4077
4078 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4079 a different value is set. See 45 below.
4080
4081 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4082
4083 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4084 of things that can be changed at compile time. The first argument specifies
4085 what is wanted and the second points to where the information is to be placed.
4086 The current list of available information is:
4087
4088 PCRE_CONFIG_UTF8
4089
4090 The output is an integer that is set to one if UTF-8 support is available;
4091 otherwise it is set to zero.
4092
4093 PCRE_CONFIG_NEWLINE
4094
4095 The output is an integer that it set to the value of the code that is used for
4096 newline. It is either LF (10) or CR (13).
4097
4098 PCRE_CONFIG_LINK_SIZE
4099
4100 The output is an integer that contains the number of bytes used for internal
4101 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4102
4103 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4104
4105 The output is an integer that contains the threshold above which the POSIX
4106 interface uses malloc() for output vectors. See item 31 above.
4107
4108 PCRE_CONFIG_MATCH_LIMIT
4109
4110 The output is an unsigned integer that contains the default limit of the number
4111 of match() calls in a pcre_exec() execution. See 42 above.
4112
4113 44. pcretest has been upgraded by the addition of the -C option. This causes it
4114 to extract all the available output from the new pcre_config() function, and to
4115 output it. The program then exits immediately.
4116
4117 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4118 order to support additional features. One way would have been to define
4119 pcre_exec2() (for example) with extra arguments, but this would not have been
4120 extensible, and would also have required all calls to the original function to
4121 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4122 is used for passing in "extra" data from pcre_study().
4123
4124 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4125 contains the following fields:
4126
4127 flags a bitmap indicating which of the following fields are set
4128 study_data opaque data from pcre_study()
4129 match_limit a way of specifying a limit on match() calls for a specific
4130 call to pcre_exec()
4131 callout_data data for callouts (see 49 below)
4132
4133 The flag bits are also defined in pcre.h, and are
4134
4135 PCRE_EXTRA_STUDY_DATA
4136 PCRE_EXTRA_MATCH_LIMIT
4137 PCRE_EXTRA_CALLOUT_DATA
4138
4139 The pcre_study() function now returns one of these new pcre_extra blocks, with
4140 the actual study data pointed to by the study_data field, and the
4141 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4142 before. That is, this change is entirely upwards-compatible and requires no
4143 change to existing code.
4144
4145 If you want to pass in additional data to pcre_exec(), you can either place it
4146 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4147 block.
4148
4149 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4150 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4151 times with different match limits, until it finds the minimum value needed for
4152 pcre_exec() to complete. The value is then output. This can be instructive; for
4153 most simple matches the number is quite small, but for pathological cases it
4154 gets very large very quickly.
4155
4156 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4157 returns the size of the data block pointed to by the study_data field in a
4158 pcre_extra block, that is, the value that was passed as the argument to
4159 pcre_malloc() when PCRE was getting memory in which to place the information
4160 created by pcre_study(). The fourth argument should point to a size_t variable.
4161 pcretest has been extended so that this information is shown after a successful
4162 pcre_study() call when information about the compiled regex is being displayed.
4163
4164 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4165 because what follows is always an absolute path. (Later: it turns out that this
4166 is more than cosmetic for MinGW, because it doesn't like empty path
4167 components.)
4168
4169 49. Some changes have been made to the callout feature (see 28 above):
4170
4171 (i) A callout function now has three choices for what it returns:
4172
4173 0 => success, carry on matching
4174 > 0 => failure at this point, but backtrack if possible
4175 < 0 => serious error, return this value from pcre_exec()
4176
4177 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4178 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4179 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4180 use by callout functions. It will never be used by PCRE itself.
4181
4182 (ii) The pcre_extra structure (see 45 above) has a void * field called
4183 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4184 pcre_callout_block structure has a field of the same name. The contents of
4185 the field passed in the pcre_extra structure are passed to the callout
4186 function in the corresponding field in the callout block. This makes it
4187 easier to use the same callout-containing regex from multiple threads. For
4188 testing, the pcretest program has a new data escape
4189
4190 \C*n pass the number n (may be negative) as callout_data
4191
4192 If the callout function in pcretest receives a non-zero value as
4193 callout_data, it returns that value.
4194
4195 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4196 there were some redundant $(CFLAGS) in commands that are now specified as
4197 $(LINK), which already includes $(CFLAGS).
4198
4199 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4200 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4201 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4202 one-byte characters throughout. Note that case-insensitive matching applies
4203 only to characters whose values are less than 256. PCRE doesn't support the
4204 notion of cases for higher-valued characters.
4205
4206 (i) A character class whose characters are all within 0-255 is handled as
4207 a bit map, and the map is inverted for negative classes. Previously, a
4208 character > 255 always failed to match such a class; however it should
4209 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4210
4211 (ii) A negated character class with a single character < 255 is coded as
4212 "not this character" (OP_NOT). This wasn't working properly when the test
4213 character was multibyte, either singly or repeated.
4214
4215 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4216 mode, for example: \x{100}{2,3}.
4217
4218 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4219 singly or repeated) now correctly test multibyte characters. However,
4220 PCRE doesn't recognize any characters with values greater than 255 as
4221 digits, spaces, or word characters. Such characters always match \D, \S,
4222 and \W, and never match \d, \s, or \w.
4223
4224 (v) Classes may now contain characters and character ranges with values
4225 greater than 255. For example: [ab\x{100}-\x{400}].
4226
4227 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4228 PCRE in UTF-8 mode.
4229
4230 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4231 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4232 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4233 value.)
4234
4235 53. The single man page has become too large. I have therefore split it up into
4236 a number of separate man pages. These also give rise to individual HTML pages;
4237 these are now put in a separate directory, and there is an index.html page that
4238 lists them all. Some hyperlinking between the pages has been installed.
4239
4240 54. Added convenience functions for handling named capturing parentheses.
4241
4242 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4243 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4244 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4245 are faulted.
4246
4247 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4248 calling configure. These values are used when compiling the dftables.c program
4249 which is run to generate the source of the default character tables. They
4250 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4251 you will need to set these values.
4252
4253 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4254
4255
4256 Version 3.9 02-Jan-02
4257 ---------------------
4258
4259 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4260
4261 2. If --disable-static was given, the building process failed when trying to
4262 build pcretest and pcregrep. (For some reason it was using libtool to compile
4263 them, which is not right, as they aren't part of the library.)
4264
4265
4266 Version 3.8 18-Dec-01
4267 ---------------------
4268
4269 1. The experimental UTF-8 code was completely screwed up. It was packing the
4270 bytes in the wrong order. How dumb can you get?
4271
4272
4273 Version 3.7 29-Oct-01
4274 ---------------------
4275
4276 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4277 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4278 this didn't happen under Solaris 8, where I normally test things.
4279
4280 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4281 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4282 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4283 this fix an hour or so after the initial 3.7 release.)
4284
4285
4286 Version 3.6 23-Oct-01
4287 ---------------------
4288
4289 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4290 offsets passed as NULL with zero offset count.
4291
4292 2. The config.guess and config.sub files had not been updated when I moved to
4293 the latest autoconf.
4294
4295
4296 Version 3.5 15-Aug-01
4297 ---------------------
4298
4299 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4300 had been forgotten.
4301
4302 2. By using declared but undefined structures, we can avoid using "void"
4303 definitions in pcre.h while keeping the internal definitions of the structures
4304 private.
4305
4306 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4307 user point of view, this means that both static and shared libraries are built
4308 by default, but this can be individually controlled. More of the work of
4309 handling this static/shared cases is now inside libtool instead of PCRE's make
4310 file.
4311
4312 4. The pcretest utility is now installed along with pcregrep because it is
4313 useful for users (to test regexs) and by doing this, it automatically gets
4314 relinked by libtool. The documentation has been turned into a man page, so
4315 there are now .1, .txt, and .html versions in /doc.
4316
4317 5. Upgrades to pcregrep:
4318 (i) Added long-form option names like gnu grep.
4319 (ii) Added --help to list all options with an explanatory phrase.
4320 (iii) Added -r, --recursive to recurse into sub-directories.
4321 (iv) Added -f, --file to read patterns from a file.
4322
4323 6. pcre_exec() was referring to its "code" argument before testing that
4324 argument for NULL (and giving an error if it was NULL).
4325
4326 7. Upgraded Makefile.in to allow for compiling in a different directory from
4327 the source directory.
4328
4329 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4330 options bits, the pointer it was passed was to an int instead of to an unsigned
4331 long int. This mattered only on 64-bit systems.
4332
4333 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4334 generated) instead of pcre.in, which it its source. Also made the same change
4335 in several of the .c files.
4336
4337 10. A new release of gcc defines printf() as a macro, which broke pcretest
4338 because it had an ifdef in the middle of a string argument for printf(). Fixed
4339 by using separate calls to printf().
4340
4341 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4342 script, to force use of CR or LF instead of \n in the source. On non-Unix
4343 systems, the value can be set in config.h.
4344
4345 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4346 absolute limit. Changed the text of the error message to make this clear, and
4347 likewise updated the man page.
4348
4349 13. The limit of 99 on the number of capturing subpatterns has been removed.
4350 The new limit is 65535, which I hope will not be a "real" limit.
4351
4352
4353 Version 3.4 22-Aug-00
4354 ---------------------
4355
4356 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4357
4358 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4359
4360
4361 Version 3.3 01-Aug-00
4362 ---------------------
4363
4364 1. If an octal character was given, but the value was greater than \377, it
4365 was not getting masked to the least significant bits, as documented. This could
4366 lead to crashes in some systems.
4367
4368 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4369 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4370
4371 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4372 These just pass their arguments on to (pcre_free)(), but they are provided
4373 because some uses of PCRE bind it to non-C systems that can call its functions,
4374 but cannot call free() or pcre_free() directly.
4375
4376 4. Add "make test" as a synonym for "make check". Corrected some comments in
4377 the Makefile.
4378
4379 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4380 Makefile.
4381
4382 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4383 command called pgrep for grepping around the active processes.
4384
4385 7. Added the beginnings of support for UTF-8 character strings.
4386
4387 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4388 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4389 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4390 out for the ar command.)
4391
4392
4393 Version 3.2 12-May-00
4394 ---------------------
4395
4396 This is purely a bug fixing release.
4397
4398 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4399 of ZA. This was just one example of several cases that could provoke this bug,
4400 which was introduced by change 9 of version 2.00. The code for breaking
4401 infinite loops after an iteration that matches an empty string was't working
4402 correctly.
4403
4404 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4405 when matched against abbab (for example). After matching an empty string, it
4406 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4407 caused it to match further down the string than it should.
4408
4409 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4410 was there because it doesn't seem to be needed, and it causes trouble on some
4411 systems, as it is not a Standard C header. It has been removed.
4412
4413 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4414 were reported on the Macintosh. The changes were from
4415
4416 while ((c = *(++ptr)) != 0 && c != '\n');
4417 to
4418 while ((c = *(++ptr)) != 0 && c != '\n') ;
4419
4420 Totally extraordinary, but if that's what it takes...
4421
4422 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4423 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4424 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4425 assumes the way PCRE uses memmove() (always moving upwards).
4426
4427 6. PCRE is being used in one environment where strchr() is not available. There
4428 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4429 faster code anyway.
4430
4431
4432 Version 3.1 09-Feb-00
4433 ---------------------
4434
4435 The only change in this release is the fixing of some bugs in Makefile.in for
4436 the "install" target:
4437
4438 (1) It was failing to install pcreposix.h.
4439
4440 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4441
4442
4443 Version 3.0 01-Feb-00
4444 ---------------------
4445
4446 1. Add support for the /+ modifier to perltest (to output $` like it does in
4447 pcretest).
4448
4449 2. Add support for the /g modifier to perltest.
4450
4451 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4452 matches null strings.
4453
4454 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4455 pattern. Perl treats empty patterns specially - it reuses the most recent
4456 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4457 effect.
4458
4459 5. The POSIX interface was broken in that it was just handing over the POSIX
4460 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4461 required a bigger vector, with some working space on the end. This means that
4462 the POSIX wrapper now has to get and free some memory, and copy the results.
4463
4464 6. Added some simple autoconf support, placing the test data and the
4465 documentation in separate directories, re-organizing some of the
4466 information files, and making it build pcre-config (a GNU standard). Also added
4467 libtool support for building PCRE as a shared library, which is now the
4468 default.
4469
4470 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4471 09 are not valid octal constants. Single digits will be used for minor values
4472 less than 10.
4473
4474 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4475 existing programs that set these in the POSIX interface can use PCRE without
4476 modification.
4477
4478 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4479 return all that pcre_info() returns, plus additional data. The pcre_info()
4480 function is retained for compatibility, but is considered to be obsolete.
4481
4482 10. Added experimental recursion feature (?R) to handle one common case that
4483 Perl 5.6 will be able to do with (?p{...}).
4484
4485 11. Added support for POSIX character classes like [:alpha:], which Perl is
4486 adopting.
4487
4488
4489 Version 2.08 31-Aug-99
4490 ----------------------
4491
4492 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4493 trying to match at the startoffset position, but instead was moving forward to
4494 the next newline as if a previous match had failed.
4495
4496 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4497 and could get into a loop if a null string was matched other than at the start
4498 of the subject.
4499
4500 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4501 be distinguished at compile time, and for completeness also added PCRE_DATE.
4502
4503 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4504 in GnuWin32 environments.
4505
4506
4507 Version 2.07 29-Jul-99
4508 ----------------------
4509
4510 1. The documentation is now supplied in plain text form and HTML as well as in
4511 the form of man page sources.
4512
4513 2. C++ compilers don't like assigning (void *) values to other pointer types.
4514 In particular this affects malloc(). Although there is no problem in Standard
4515 C, I've put in casts to keep C++ compilers happy.
4516
4517 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4518 should be (const char *).
4519
4520 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4521 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4522 However, I haven't made this a standard facility. The documentation doesn't
4523 mention it, and the Makefile doesn't support it.
4524
4525 5. The Makefile now contains an "install" target, with editable destinations at
4526 the top of the file. The pcretest program is not installed.
4527
4528 6. pgrep -V now gives the PCRE version number and date.
4529
4530 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4531 causing the entire string to be ignored, instead of just the last character.
4532
4533 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4534 non-matching string, it can take a very, very long time, even for strings of
4535 quite modest length, because of the nested recursion. PCRE now does better in
4536 some of these cases. It does this by remembering the last required literal
4537 character in the pattern, and pre-searching the subject to ensure it is present
4538 before running the real match. In other words, it applies a heuristic to detect
4539 some types of certain failure quickly, and in the above example, if presented
4540 with a string that has no trailing " it gives "no match" very quickly.
4541
4542 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4543 other alternatives are tried instead.
4544
4545
4546 Version 2.06 09-Jun-99
4547 ----------------------
4548
4549 1. Change pcretest's output for amount of store used to show just the code
4550 space, because the remainder (the data block) varies in size between 32-bit and
4551 64-bit systems.
4552
4553 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4554 start matching at. This allows lookbehinds to work when searching for multiple
4555 occurrences in a string.
4556
4557 3. Added additional options to pcretest for testing multiple occurrences:
4558
4559 /+ outputs the rest of the string that follows a match
4560 /g loops for multiple occurrences, using the new startoffset argument
4561 /G loops for multiple occurrences by passing an incremented pointer
4562
4563 4. PCRE wasn't doing the "first character" optimization for patterns starting
4564 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4565 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4566 the letter 'x'. On long subject strings, this gives a significant speed-up.
4567
4568
4569 Version 2.05 21-Apr-99
4570 ----------------------
4571
4572 1. Changed the type of magic_number from int to long int so that it works
4573 properly on 16-bit systems.
4574
4575 2. Fixed a bug which caused patterns starting with .* not to work correctly
4576 when the subject string contained newline characters. PCRE was assuming
4577 anchoring for such patterns in all cases, which is not correct because .* will
4578 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4579 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4580 must be retried after every newline in the subject.
4581
4582
4583 Version 2.04 18-Feb-99
4584 ----------------------
4585
4586 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4587 computation of the store needed to hold the pattern was incorrect (too large).
4588 If such patterns were nested a few deep, this could multiply and become a real
4589 problem.
4590
4591 2. Added /M option to pcretest to show the memory requirement of a specific
4592 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4593
4594 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4595 compiled in such a way that the backtracking after subsequent failure was
4596 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4597 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4598
4599
4600 Version 2.03 02-Feb-99
4601 ----------------------
4602
4603 1. Fixed typo and small mistake in man page.
4604
4605 2. Added 4th condition (GPL supersedes if conflict) and created separate
4606 LICENCE file containing the conditions.
4607
4608 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4609 Perl, that is the internal \ allows the delimiter to be included in the
4610 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4611 the final delimiter, add \ to the end of the pattern (to test the error).
4612
4613 4. Added the convenience functions for extracting substrings after a successful
4614 match. Updated pcretest to make it able to test these functions.
4615
4616
4617 Version 2.02 14-Jan-99
4618 ----------------------
4619
4620 1. Initialized the working variables associated with each extraction so that
4621 their saving and restoring doesn't refer to uninitialized store.
4622
4623 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4624 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4625 fix the problem.
4626
4627 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4628 calls, and wasn't printing the correct value for compiling calls. Increased the
4629 default value of LOOPREPEAT, and the number of significant figures in the
4630 times.
4631
4632 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4633
4634 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4635 a building problem on Windows NT with a FAT file system.
4636
4637
4638 Version 2.01 21-Oct-98
4639 ----------------------
4640
4641 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4642 to character tables built by pcre_maketables() in the current locale. If NULL
4643 is passed, the default tables are used.
4644
4645
4646 Version 2.00 24-Sep-98
4647 ----------------------
4648
4649 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4650 it any more.
4651
4652 2. Allow quantification of (?>) groups, and make it work correctly.
4653
4654 3. The first character computation wasn't working for (?>) groups.
4655
4656 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4657 end of the subject) and add 5.005's \z, which really does match only at the
4658 very end of the subject.
4659
4660 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4661
4662 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4663 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4664 localized options. All options to pcre_study() were also removed.
4665
4666 7. Add other new features from 5.005:
4667
4668 $(?<= positive lookbehind
4669 $(?<! negative lookbehind
4670 (?imsx-imsx) added the unsetting capability
4671 such a setting is global if at outer level; local otherwise
4672 (?imsx-imsx:) non-capturing groups with option setting
4673 (?(cond)re|re) conditional pattern matching
4674
4675 A backreference to itself in a repeated group matches the previous
4676 captured string.
4677
4678 8. General tidying up of studying (both automatic and via "study")
4679 consequential on the addition of new assertions.
4680
4681 9. As in 5.005, unlimited repeated groups that could match an empty substring
4682 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4683 runtime if any iteration does actually match an empty substring.
4684
4685 10. Include the RunTest script in the distribution.
4686
4687 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4688 discrepancies, some of which were old and were also with respect to 5.004. They
4689 have now been fixed.
4690
4691
4692 Version 1.09 28-Apr-98
4693 ----------------------
4694
4695 1. A negated single character class followed by a quantifier with a minimum
4696 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4697 program crashes, or just wrong answers. This did not apply to negated classes
4698 containing more than one character, or to minima other than one.
4699
4700
4701 Version 1.08 27-Mar-98
4702 ----------------------
4703
4704 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4705
4706 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4707 latter must appear before anything that relies on it in the pattern.
4708
4709
4710 Version 1.07 16-Feb-98
4711 ----------------------
4712
4713 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4714 repeat of a potentially empty string).
4715
4716
4717 Version 1.06 23-Jan-98
4718 ----------------------
4719
4720 1. Added Markus Oberhumer's little patches for C++.
4721
4722 2. Literal strings longer than 255 characters were broken.
4723
4724
4725 Version 1.05 23-Dec-97
4726 ----------------------
4727
4728 1. Negated character classes containing more than one character were failing if
4729 PCRE_CASELESS was set at run time.
4730
4731
4732 Version 1.04 19-Dec-97
4733 ----------------------
4734
4735 1. Corrected the man page, where some "const" qualifiers had been omitted.
4736
4737 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4738 input syntax.
4739
4740 3. Fixed memory leak which occurred when a regex with back references was
4741 matched with an offsets vector that wasn't big enough. The temporary memory
4742 that is used in this case wasn't being freed if the match failed.
4743
4744 4. Tidied pcretest to ensure it frees memory that it gets.
4745
4746 5. Temporary memory was being obtained in the case where the passed offsets
4747 vector was exactly big enough.
4748
4749 6. Corrected definition of offsetof() from change 5 below.
4750
4751 7. I had screwed up change 6 below and broken the rules for the use of
4752 setjmp(). Now fixed.
4753
4754
4755 Version 1.03 18-Dec-97
4756 ----------------------
4757
4758 1. A erroneous regex with a missing opening parenthesis was correctly
4759 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4760 on some systems.
4761
4762 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4763 it was reported that one broken compiler failed on the former because "code" is
4764 also an independent variable.
4765
4766 3. The erroneous regex a[]b caused an array overrun reference.
4767
4768 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4769 fail on data ending with that character. (It was going on too far, and checking
4770 the next character, typically a binary zero.) This was specific to the
4771 optimized code for single-character negative classes.
4772
4773 5. Added a contributed patch from the TIN world which does the following:
4774
4775 + Add an undef for memmove, in case the the system defines a macro for it.
4776
4777 + Add a definition of offsetof(), in case there isn't one. (I don't know
4778 the reason behind this - offsetof() is part of the ANSI standard - but
4779 it does no harm).
4780
4781 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4782 most of the places where whitespace preceded '#'. I have given up and
4783 allowed the remaining 2 cases to be at the margin.
4784
4785 + Rename some variables in pcre to eliminate shadowing. This seems very
4786 pedantic, but does no harm, of course.
4787
4788 6. Moved the call to setjmp() into its own function, to get rid of warnings
4789 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4790
4791 7. Constructs such as \d{8,} were compiling into the equivalent of
4792 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4793 outcome, but in this particular case used more store than had been allocated,
4794 which caused the bug to be discovered because it threw up an internal error.
4795
4796 8. The debugging code in both pcre and pcretest for outputting the compiled
4797 form of a regex was going wrong in the case of back references followed by
4798 curly-bracketed repeats.
4799
4800
4801 Version 1.02 12-Dec-97
4802 ----------------------
4803
4804 1. Typos in pcre.3 and comments in the source fixed.
4805
4806 2. Applied a contributed patch to get rid of places where it used to remove
4807 'const' from variables, and fixed some signed/unsigned and uninitialized
4808 variable warnings.
4809
4810 3. Added the "runtest" target to Makefile.
4811
4812 4. Set default compiler flag to -O2 rather than just -O.
4813
4814
4815 Version 1.01 19-Nov-97
4816 ----------------------
4817
4818 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4819 like /([ab]*)*/, that is, for classes with more than one character in them.
4820
4821 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4822 as /((?>a*))*/ (a PCRE_EXTRA facility).
4823
4824
4825 Version 1.00 18-Nov-97
4826 ----------------------
4827
4828 1. Added compile-time macros to support systems such as SunOS4 which don't have
4829 memmove() or strerror() but have other things that can be used instead.
4830
4831 2. Arranged that "make clean" removes the executables.
4832
4833
4834 Version 0.99 27-Oct-97
4835 ----------------------
4836
4837 1. Fixed bug in code for optimizing classes with only one character. It was
4838 initializing a 32-byte map regardless, which could cause it to run off the end
4839 of the memory it had got.
4840
4841 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4842
4843
4844 Version 0.98 22-Oct-97
4845 ----------------------
4846
4847 1. Fixed bug in code for handling temporary memory usage when there are more
4848 back references than supplied space in the ovector. This could cause segfaults.
4849
4850
4851 Version 0.97 21-Oct-97
4852 ----------------------
4853
4854 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4855
4856 2. Optimized negated single characters not to use a bit map.
4857
4858 3. Brought error texts together as macro definitions; clarified some of them;
4859 fixed one that was wrong - it said "range out of order" when it meant "invalid
4860 escape sequence".
4861
4862 4. Changed some char * arguments to const char *.
4863
4864 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4865
4866 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4867 pcretest.
4868
4869
4870 Version 0.96 16-Oct-97
4871 ----------------------
4872
4873 1. Added a simple "pgrep" utility to the distribution.
4874
4875 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4876 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4877 where "ddd" means "one or more decimal digits".
4878
4879 3. Fixed serious bug. If a pattern had a back reference, but the call to
4880 pcre_exec() didn't supply a large enough ovector to record the related
4881 identifying subpattern, the match always failed. PCRE now remembers the number
4882 of the largest back reference, and gets some temporary memory in which to save
4883 the offsets during matching if necessary, in order to ensure that
4884 backreferences always work.
4885
4886 4. Increased the compatibility with Perl in a number of ways:
4887
4888 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4889 to request this handling. The option can be set at compile or exec time.
4890
4891 (b) $ matches before a terminating newline by default; an option
4892 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4893 mode). The option can be set at compile or exec time.
4894
4895 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4896 the same as Perl's. If the decimal number it represents is less than 10
4897 or there aren't that many previous left capturing parentheses, an octal
4898 escape is read. Inside a character class, it's always an octal escape,
4899 even if it is a single digit.
4900
4901 (d) An escaped but undefined alphabetic character is taken as a literal,
4902 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4903 escapes.
4904
4905 (e) {0} is now permitted. (The previous item is removed from the compiled
4906 pattern).
4907
4908 5. Changed all the names of code files so that the basic parts are no longer
4909 than 10 characters, and abolished the teeny "globals.c" file.
4910
4911 6. Changed the handling of character classes; they are now done with a 32-byte
4912 bit map always.
4913
4914 7. Added the -d and /D options to pcretest to make it possible to look at the
4915 internals of compilation without having to recompile pcre.
4916
4917
4918 Version 0.95 23-Sep-97
4919 ----------------------
4920
4921 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4922 \x20 at the start of a run of normal characters. These were being treated as
4923 real characters, instead of the source characters being re-checked.
4924
4925
4926 Version 0.94 18-Sep-97
4927 ----------------------
4928
4929 1. The functions are now thread-safe, with the caveat that the global variables
4930 containing pointers to malloc() and free() or alternative functions are the
4931 same for all threads.
4932
4933 2. Get pcre_study() to generate a bitmap of initial characters for non-
4934 anchored patterns when this is possible, and use it if passed to pcre_exec().
4935
4936
4937 Version 0.93 15-Sep-97
4938 ----------------------
4939
4940 1. /(b)|(:+)/ was computing an incorrect first character.
4941
4942 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4943 but not actually doing anything yet.
4944
4945 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4946 as Perl does (e.g. [-az] or [az-]).
4947
4948 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4949 all possible positions.
4950
4951 5. Split up into different modules to avoid including unneeded functions in a
4952 compiled binary. However, compile and exec are still in one module. The "study"
4953 function is split off.
4954
4955 6. The character tables are now in a separate module whose source is generated
4956 by an auxiliary program - but can then be edited by hand if required. There are
4957 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4958 toupper() in the code.
4959
4960 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4961 make them global. Abolish the function for setting them, as the caller can now
4962 set them directly.
4963
4964
4965 Version 0.92 11-Sep-97
4966 ----------------------
4967
4968 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4969 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4970
4971 2. Caseless matching was not working in character classes if the characters in
4972 the pattern were in upper case.
4973
4974 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4975
4976 4. Make PCRE_ANCHORED public and accept as a compile option.
4977
4978 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4979 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4980 pass them.
4981
4982 6. Give an error if bad option bits passed at compile or run time.
4983
4984 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4985 pcretest to cause it to pass that flag.
4986
4987 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4988 options, and the first character, if set.
4989
4990 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4991
4992
4993 Version 0.91 10-Sep-97
4994 ----------------------
4995
4996 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4997 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4998
4999 2. PCRE was looping on encountering an indefinitely repeated back reference to
5000 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5001 Perl does - treats the match as successful.
5002
5003 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5