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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1003 - (show annotations)
Wed Aug 15 16:45:36 2012 UTC (7 years, 2 months ago) by ph10
File size: 234540 byte(s)
Error occurred while calculating annotation data.
Fix a number of issues in pcregrep.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51
52 Version 8.31 06-July-2012
53 -------------------------
54
55 1. Fixing a wrong JIT test case and some compiler warnings.
56
57 2. Removed a bashism from the RunTest script.
58
59 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
60 to unsigned type, result still unsigned" that was given by an MS compiler
61 on encountering the code "-sizeof(xxx)".
62
63 4. Partial matching support is added to the JIT compiler.
64
65 5. Fixed several bugs concerned with partial matching of items that consist
66 of more than one character:
67
68 (a) /^(..)\1/ did not partially match "aba" because checking references was
69 done on an "all or nothing" basis. This also applied to repeated
70 references.
71
72 (b) \R did not give a hard partial match if \r was found at the end of the
73 subject.
74
75 (c) \X did not give a hard partial match after matching one or more
76 characters at the end of the subject.
77
78 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
79 a partial match for the string "\r".
80
81 (e) When newline was set to CRLF, the metacharacter "." did not recognize
82 a partial match for a CR character at the end of the subject string.
83
84 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
85 running pcretest, the text "(JIT)" added to the output whenever JIT is
86 actually used to run the match.
87
88 7. Individual JIT compile options can be set in pcretest by following -s+[+]
89 or /S+[+] with a digit between 1 and 7.
90
91 8. OP_NOT now supports any UTF character not just single-byte ones.
92
93 9. (*MARK) control verb is now supported by the JIT compiler.
94
95 10. The command "./RunTest list" lists the available tests without actually
96 running any of them. (Because I keep forgetting what they all are.)
97
98 11. Add PCRE_INFO_MAXLOOKBEHIND.
99
100 12. Applied a (slightly modified) user-supplied patch that improves performance
101 when the heap is used for recursion (compiled with --disable-stack-for-
102 recursion). Instead of malloc and free for each heap frame each time a
103 logical recursion happens, frames are retained on a chain and re-used where
104 possible. This sometimes gives as much as 30% improvement.
105
106 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
107 call.
108
109 14. As documented, (*COMMIT) is now confined to within a positive assertion.
110
111 15. It is now possible to link pcretest with libedit as an alternative to
112 libreadline.
113
114 16. (*COMMIT) control verb is now supported by the JIT compiler.
115
116 17. The Unicode data tables have been updated to Unicode 6.1.0.
117
118 18. Added --file-list option to pcregrep.
119
120 19. Added binary file support to pcregrep, including the -a, --binary-files,
121 -I, and --text options.
122
123 20. The madvise function is renamed for posix_madvise for QNX compatibility
124 reasons. Fixed by Giuseppe D'Angelo.
125
126 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
127
128 22. Changed the option for creating symbolic links for 16-bit man pages from
129 -s to -sf so that re-installing does not cause issues.
130
131 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
132
133 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
134 matches in certain environments (the workspace was not being correctly
135 retained). Also added to pcre_dfa_exec() a simple plausibility check on
136 some of the workspace data at the beginning of a restart.
137
138 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
139 was not doing so when it should - probably a typo introduced by SVN 528
140 (change 8.10/14).
141
142 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
143 \w+ when the character tables indicated that \x{c4} was a word character.
144 There were several related cases, all because the tests for doing a table
145 lookup were testing for characters less than 127 instead of 255.
146
147 27. If a pattern contains capturing parentheses that are not used in a match,
148 their slots in the ovector are set to -1. For those that are higher than
149 any matched groups, this happens at the end of processing. In the case when
150 there were back references that the ovector was too small to contain
151 (causing temporary malloc'd memory to be used during matching), and the
152 highest capturing number was not used, memory off the end of the ovector
153 was incorrectly being set to -1. (It was using the size of the temporary
154 memory instead of the true size.)
155
156 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
157 ovector size, it uses memory at the end of the block that it has got.
158
159 29. Check for an overlong MARK name and give an error at compile time. The
160 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
161
162 30. JIT compiler update.
163
164 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
165 Rill for the patch.
166
167 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
168
169 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
170
171 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
172 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
173 was enabled.
174
175 35. Improve JIT code generation for greedy plus quantifier.
176
177 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
178 1 to "aa" instead of to an empty string. The bug affected repeated groups
179 that could potentially match an empty string.
180
181 37. Optimizing single character iterators in JIT.
182
183 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
184 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
185 codepoints that are too big for the mode are faulted, and in a UTF mode,
186 disallowed codepoints are also faulted.
187
188 39. If PCRE was compiled with UTF support, in three places in the DFA
189 matcher there was code that should only have been obeyed in UTF mode, but
190 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
191 processing when bytes with values greater than 127 were present. In 16-bit
192 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
193 both cases the values are those that cannot be the first data item in a UTF
194 character. The three items that might have provoked this were recursions,
195 possessively repeated groups, and atomic groups.
196
197 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
198 and pcregrep, because some OS require shared objects to be explicitly
199 passed to ld, causing the link step to fail if they are not.
200
201 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
202 mode, patterns that started with \h* or \R* might be incorrectly matched.
203
204
205 Version 8.30 04-February-2012
206 -----------------------------
207
208 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
209 name is defined in ctype.h.
210
211 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
212 only in quite long subpatterns.
213
214 3. Removed the function pcre_info(), which has been obsolete and deprecated
215 since it was replaced by pcre_fullinfo() in February 2000.
216
217 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
218 match a (*MARK), and the match failed at the start of the subject, a
219 reference to memory before the start of the subject could occur. This bug
220 was introduced by fix 17 of release 8.21.
221
222 5. A reference to an unset group with zero minimum repetition was giving
223 totally wrong answers (in non-JavaScript-compatibility mode). For example,
224 /(another)?(\1?)test/ matched against "hello world test". This bug was
225 introduced in release 8.13.
226
227 6. Add support for 16-bit character strings (a large amount of work involving
228 many changes and refactorings).
229
230 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
231 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
232 from a file.
233
234 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
235 rounding is not applied in this particular case).
236
237 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
238 if they appear, or are escaped, in patterns.
239
240 10. Get rid of a number of -Wunused-but-set-variable warnings.
241
242 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
243 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
244 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
245 also returns the mark "x". This bug applied to capturing parentheses,
246 non-capturing parentheses, and atomic parentheses. It also applied to some
247 assertions.
248
249 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
250 information out of configure.ac instead of relying on pcre.h.generic, which
251 is not stored in the repository.
252
253 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
254 -lreadline.
255
256 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
257
258 15. Applied Graycode's patch to put the top-level frame on the stack rather
259 than the heap when not using the stack for recursion. This gives a
260 performance improvement in many cases when recursion is not deep.
261
262 16. Experimental code added to "pcretest -C" to output the stack frame size.
263
264
265 Version 8.21 12-Dec-2011
266 ------------------------
267
268 1. Updating the JIT compiler.
269
270 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
271 are added as well.
272
273 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
274 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
275 calling _pcre_jit_exec. Some extra comments are added.
276
277 4. (*MARK) settings inside atomic groups that do not contain any capturing
278 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
279 was introduced by change 18 for 8.20.
280
281 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
282 ECMA-262 standard.
283
284 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
285 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
286 This bug was probably introduced by change 9 of 8.13.
287
288 7. While fixing 6 above, I noticed that a number of other items were being
289 incorrectly rejected as "not fixed length". This arose partly because newer
290 opcodes had not been added to the fixed-length checking code. I have (a)
291 corrected the bug and added tests for these items, and (b) arranged for an
292 error to occur if an unknown opcode is encountered while checking for fixed
293 length instead of just assuming "not fixed length". The items that were
294 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
295 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
296 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
297
298 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
299 being incorrectly compiled and would have given unpredicatble results.
300
301 9. A possessively repeated subpattern with minimum repeat count greater than
302 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
303 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
304 the first (A) could occur when it should not.
305
306 10. Add a cast and remove a redundant test from the code.
307
308 11. JIT should use pcre_malloc/pcre_free for allocation.
309
310 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
311 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
312 is anything other than /usr, -L is still shown).
313
314 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
315
316 14. Perl does not support \N without a following name in a [] class; PCRE now
317 also gives an error.
318
319 15. If a forward reference was repeated with an upper limit of around 2000,
320 it caused the error "internal error: overran compiling workspace". The
321 maximum number of forward references (including repeats) was limited by the
322 internal workspace, and dependent on the LINK_SIZE. The code has been
323 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
324 the default depends on LINK_SIZE. There is a new upper limit (for safety)
325 of around 200,000 forward references. While doing this, I also speeded up
326 the filling in of repeated forward references.
327
328 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
329 incorrectly expecting the subject to contain another "a" after the start.
330
331 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
332 in the match, the SKIP should be ignored. This was not happening; instead
333 the SKIP was being treated as NOMATCH. For patterns such as
334 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
335 tested.
336
337 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
338 now much more compatible with Perl, in particular in cases where the result
339 is a non-match for a non-anchored pattern. For example, if
340 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
341 "m", where previously it did not return a name. A side effect of this
342 change is that for partial matches, the last encountered mark name is
343 returned, as for non matches. A number of tests that were previously not
344 Perl-compatible have been moved into the Perl-compatible test files. The
345 refactoring has had the pleasing side effect of removing one argument from
346 the match() function, thus reducing its stack requirements.
347
348 19. If the /S+ option was used in pcretest to study a pattern using JIT,
349 subsequent uses of /S (without +) incorrectly behaved like /S+.
350
351 21. Retrieve executable code size support for the JIT compiler and fixing
352 some warnings.
353
354 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
355 not work when the shorter character appeared right at the end of the
356 subject string.
357
358 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
359 systems.
360
361 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
362 output it when the /M option is used in pcretest.
363
364 25. The CheckMan script was not being included in the distribution. Also, added
365 an explicit "perl" to run Perl scripts from the PrepareRelease script
366 because this is reportedly needed in Windows.
367
368 26. If study data was being save in a file and studying had not found a set of
369 "starts with" bytes for the pattern, the data written to the file (though
370 never used) was taken from uninitialized memory and so caused valgrind to
371 complain.
372
373 27. Updated RunTest.bat as provided by Sheri Pierce.
374
375 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
376
377 29. Computation of memory usage for the table of capturing group names was
378 giving an unnecessarily large value.
379
380
381 Version 8.20 21-Oct-2011
382 ------------------------
383
384 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
385 a POSIX class. After further experiments with Perl, which convinced me that
386 Perl has bugs and confusions, a closing square bracket is no longer allowed
387 in a POSIX name. This bug also affected patterns with classes that started
388 with full stops.
389
390 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
391 captured substring, but while checking the failing first alternative,
392 substring 1 is temporarily captured. If the output vector supplied to
393 pcre_exec() was not big enough for this capture, the yield of the function
394 was still zero ("insufficient space for captured substrings"). This cannot
395 be totally fixed without adding another stack variable, which seems a lot
396 of expense for a edge case. However, I have improved the situation in cases
397 such as /(a)(b)x|abc/ matched against "abc", where the return code
398 indicates that fewer than the maximum number of slots in the ovector have
399 been set.
400
401 3. Related to (2) above: when there are more back references in a pattern than
402 slots in the output vector, pcre_exec() uses temporary memory during
403 matching, and copies in the captures as far as possible afterwards. It was
404 using the entire output vector, but this conflicts with the specification
405 that only 2/3 is used for passing back captured substrings. Now it uses
406 only the first 2/3, for compatibility. This is, of course, another edge
407 case.
408
409 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
410 main code base, and can be used by building with --enable-jit. When this is
411 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
412 runtime --no-jit option is given.
413
414 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
415 ovector, the return from the function was zero, implying that there were
416 other matches that did not fit. The correct "exactly full" value is now
417 returned.
418
419 6. If a subpattern that was called recursively or as a subroutine contained
420 (*PRUNE) or any other control that caused it to give a non-standard return,
421 invalid errors such as "Error -26 (nested recursion at the same subject
422 position)" or even infinite loops could occur.
423
424 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
425 computing the minimum length on reaching *ACCEPT, and so ended up with the
426 wrong value of 1 rather than 0. Further investigation indicates that
427 computing a minimum subject length in the presence of *ACCEPT is difficult
428 (think back references, subroutine calls), and so I have changed the code
429 so that no minimum is registered for a pattern that contains *ACCEPT.
430
431 8. If (*THEN) was present in the first (true) branch of a conditional group,
432 it was not handled as intended. [But see 16 below.]
433
434 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
435 Sheri Pierce.
436
437 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
438 the first byte in a match must be "a".
439
440 11. Change 17 for 8.13 increased the recursion depth for patterns like
441 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
442 pattern contains any instances of (*THEN). If it does not, the old
443 optimizations are restored. It would be nice to do this on a per-group
444 basis, but at the moment that is not feasible.
445
446 12. In some environments, the output of pcretest -C is CRLF terminated. This
447 broke RunTest's code that checks for the link size. A single white space
448 character after the value is now allowed for.
449
450 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
451 For "fr", it uses the Windows-specific input and output files.
452
453 14. If (*THEN) appeared in a group that was called recursively or as a
454 subroutine, it did not work as intended. [But see next item.]
455
456 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
457 pattern fragments (but not containing any | characters). If A and B are
458 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
459 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
460 D. In other words, Perl considers parentheses that do not contain any |
461 characters to be part of a surrounding alternative, whereas PCRE was
462 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
463 differently. PCRE now behaves in the same way as Perl, except in the case
464 of subroutine/recursion calls such as (?1) which have in any case always
465 been different (but PCRE had them first :-).
466
467 16. Related to 15 above: Perl does not treat the | in a conditional group as
468 creating alternatives. Such a group is treated in the same way as an
469 ordinary group without any | characters when processing (*THEN). PCRE has
470 been changed to match Perl's behaviour.
471
472 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
473 RunGrepTest script failed.
474
475 18. Change 22 for version 13 caused atomic groups to use more stack. This is
476 inevitable for groups that contain captures, but it can lead to a lot of
477 stack use in large patterns. The old behaviour has been restored for atomic
478 groups that do not contain any capturing parentheses.
479
480 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
481 suppress the check for a minimum subject length at run time. (If it was
482 given to pcre_exec() or pcre_dfa_exec() it did work.)
483
484 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
485 fail to work when decoding hex characters in data strings in EBCDIC
486 environments.
487
488 21. It appears that in at least one Mac OS environment, the isxdigit() function
489 is implemented as a macro that evaluates to its argument more than once,
490 contravening the C 90 Standard (I haven't checked a later standard). There
491 was an instance in pcretest which caused it to go wrong when processing
492 \x{...} escapes in subject strings. The has been rewritten to avoid using
493 things like p++ in the argument of isxdigit().
494
495
496 Version 8.13 16-Aug-2011
497 ------------------------
498
499 1. The Unicode data tables have been updated to Unicode 6.0.0.
500
501 2. Two minor typos in pcre_internal.h have been fixed.
502
503 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
504 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
505 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
506
507 4. There were a number of related bugs in the code for matching backrefences
508 caselessly in UTF-8 mode when codes for the characters concerned were
509 different numbers of bytes. For example, U+023A and U+2C65 are an upper
510 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
511 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
512 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
513 2-byte code at the end of the subject (it thought there wasn't enough data
514 left).
515
516 5. Comprehensive information about what went wrong is now returned by
517 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
518 as the output vector has at least 2 elements. The offset of the start of
519 the failing character and a reason code are placed in the vector.
520
521 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
522 now returned is for the first byte of the failing character, instead of the
523 last byte inspected. This is an incompatible change, but I hope it is small
524 enough not to be a problem. It makes the returned offset consistent with
525 pcre_exec() and pcre_dfa_exec().
526
527 7. pcretest now gives a text phrase as well as the error number when
528 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
529 failure, the offset and reason code are output.
530
531 8. When \R was used with a maximizing quantifier it failed to skip backwards
532 over a \r\n pair if the subsequent match failed. Instead, it just skipped
533 back over a single character (\n). This seems wrong (because it treated the
534 two characters as a single entity when going forwards), conflicts with the
535 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
536 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
537 has been changed.
538
539 9. Some internal refactoring has changed the processing so that the handling
540 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
541 time (the PCRE_DOTALL option was changed this way some time ago: version
542 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
543 which was always a bit of a fudge. It also means that there is one less
544 argument for the match() function, which reduces its stack requirements
545 slightly. This change also fixes an incompatibility with Perl: the pattern
546 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
547
548 10. More internal refactoring has drastically reduced the number of recursive
549 calls to match() for possessively repeated groups such as (abc)++ when
550 using pcre_exec().
551
552 11. While implementing 10, a number of bugs in the handling of groups were
553 discovered and fixed:
554
555 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
556 (a|)*(?1) gave a compile-time internal error.
557 ((a|)+)+ did not notice that the outer group could match an empty string.
558 (^a|^)+ was not marked as anchored.
559 (.*a|.*)+ was not marked as matching at start or after a newline.
560
561 12. Yet more internal refactoring has removed another argument from the match()
562 function. Special calls to this function are now indicated by setting a
563 value in a variable in the "match data" data block.
564
565 13. Be more explicit in pcre_study() instead of relying on "default" for
566 opcodes that mean there is no starting character; this means that when new
567 ones are added and accidentally left out of pcre_study(), testing should
568 pick them up.
569
570 14. The -s option of pcretest has been documented for ages as being an old
571 synonym of -m (show memory usage). I have changed it to mean "force study
572 for every regex", that is, assume /S for every regex. This is similar to -i
573 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
574 using it. It makes it easier to run collections of tests with and without
575 study enabled, and thereby test pcre_study() more easily. All the standard
576 tests are now run with and without -s (but some patterns can be marked as
577 "never study" - see 20 below).
578
579 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
580 restoration of the capturing data to the outer values was not happening
581 correctly.
582
583 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
584 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
585 pattern had matched an empty string, and so incorrectly returned a no
586 match.
587
588 17. There was optimizing code for the last branch of non-capturing parentheses,
589 and also for the obeyed branch of a conditional subexpression, which used
590 tail recursion to cut down on stack usage. Unfortunately, now that there is
591 the possibility of (*THEN) occurring in these branches, tail recursion is
592 no longer possible because the return has to be checked for (*THEN). These
593 two optimizations have therefore been removed. [But see 8.20/11 above.]
594
595 18. If a pattern containing \R was studied, it was assumed that \R always
596 matched two bytes, thus causing the minimum subject length to be
597 incorrectly computed because \R can also match just one byte.
598
599 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
600 was incorrectly computed.
601
602 20. If /S is present twice on a test pattern in pcretest input, it now
603 *disables* studying, thereby overriding the use of -s on the command line
604 (see 14 above). This is necessary for one or two tests to keep the output
605 identical in both cases.
606
607 21. When (*ACCEPT) was used in an assertion that matched an empty string and
608 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
609
610 22. When an atomic group that contained a capturing parenthesis was
611 successfully matched, but the branch in which it appeared failed, the
612 capturing was not being forgotten if a higher numbered group was later
613 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
614 group 1 to "a", when in fact it should be unset. This applied to multi-
615 branched capturing and non-capturing groups, repeated or not, and also to
616 positive assertions (capturing in negative assertions does not happen
617 in PCRE) and also to nested atomic groups.
618
619 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
620 subject after a captured substring, to make it easier to tell which of a
621 number of identical substrings has been captured.
622
623 24. The way atomic groups are processed by pcre_exec() has been changed so that
624 if they are repeated, backtracking one repetition now resets captured
625 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
626 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
627 "aaa". Previously, it would have been "a". As part of this code
628 refactoring, the way recursive calls are handled has also been changed.
629
630 25. If an assertion condition captured any substrings, they were not passed
631 back unless some other capturing happened later. For example, if
632 (?(?=(a))a) was matched against "a", no capturing was returned.
633
634 26. When studying a pattern that contained subroutine calls or assertions,
635 the code for finding the minimum length of a possible match was handling
636 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
637 group 1 called group 2 while simultaneously a separate group 2 called group
638 1). A stack overflow occurred in this case. I have fixed this by limiting
639 the recursion depth to 10.
640
641 27. Updated RunTest.bat in the distribution to the version supplied by Tom
642 Fortmann. This supports explicit test numbers on the command line, and has
643 argument validation and error reporting.
644
645 28. An instance of \X with an unlimited repeat could fail if at any point the
646 first character it looked at was a mark character.
647
648 29. Some minor code refactoring concerning Unicode properties and scripts
649 should reduce the stack requirement of match() slightly.
650
651 30. Added the '=' option to pcretest to check the setting of unused capturing
652 slots at the end of the pattern, which are documented as being -1, but are
653 not included in the return count.
654
655 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
656 compiled something random. Now it gives a compile-time error (as does
657 Perl).
658
659 32. A *MARK encountered during the processing of a positive assertion is now
660 recorded and passed back (compatible with Perl).
661
662 33. If --only-matching or --colour was set on a pcregrep call whose pattern
663 had alternative anchored branches, the search for a second match in a line
664 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
665 matched the line "0102" twice. The same bug affected patterns that started
666 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
667 twice.
668
669 34. Previously, PCRE did not allow quantification of assertions. However, Perl
670 does, and because of capturing effects, quantifying parenthesized
671 assertions may at times be useful. Quantifiers are now allowed for
672 parenthesized assertions.
673
674 35. A minor code tidy in pcre_compile() when checking options for \R usage.
675
676 36. \g was being checked for fancy things in a character class, when it should
677 just be a literal "g".
678
679 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
680 appearance of a nested POSIX class supersedes an apparent external class.
681 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
682 unescaped square brackets may also appear as part of class names. For
683 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
684 more like Perl. (But see 8.20/1 above.)
685
686 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
687 was because it thought it was \N{name}, which is not supported).
688
689 39. Add minix to OS list not supporting the -S option in pcretest.
690
691 40. PCRE tries to detect cases of infinite recursion at compile time, but it
692 cannot analyze patterns in sufficient detail to catch mutual recursions
693 such as ((?1))((?2)). There is now a runtime test that gives an error if a
694 subgroup is called recursively as a subpattern for a second time at the
695 same position in the subject string. In previous releases this might have
696 been caught by the recursion limit, or it might have run out of stack.
697
698 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
699 happen only once. PCRE was, however incorrectly giving a compile time error
700 "recursive call could loop indefinitely" because it cannot analyze the
701 pattern in sufficient detail. The compile time test no longer happens when
702 PCRE is compiling a conditional subpattern, but actual runaway loops are
703 now caught at runtime (see 40 above).
704
705 42. It seems that Perl allows any characters other than a closing parenthesis
706 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
707 has been changed to be the same.
708
709 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
710 as not to get warnings when autogen.sh is called. Also changed
711 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
712
713 44. To help people who use pcregrep to scan files containing exceedingly long
714 lines, the following changes have been made:
715
716 (a) The default value of the buffer size parameter has been increased from
717 8K to 20K. (The actual buffer used is three times this size.)
718
719 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
720 PCRE is built.
721
722 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
723 to be set at run time.
724
725 (d) Numerical values in pcregrep options can be followed by K or M, for
726 example --buffer-size=50K.
727
728 (e) If a line being scanned overflows pcregrep's buffer, an error is now
729 given and the return code is set to 2.
730
731 45. Add a pointer to the latest mark to the callout data block.
732
733 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
734 partial match of an empty string instead of no match. This was specific to
735 the use of ".".
736
737 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
738 complete match instead of a partial match. This bug was dependent on both
739 the PCRE_UTF8 and PCRE_DOTALL options being set.
740
741 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
742 starting byte set, because \b was not being ignored.
743
744
745 Version 8.12 15-Jan-2011
746 ------------------------
747
748 1. Fixed some typos in the markup of the man pages, and wrote a script that
749 checks for such things as part of the documentation building process.
750
751 2. On a big-endian 64-bit system, pcregrep did not correctly process the
752 --match-limit and --recursion-limit options (added for 8.11). In
753 particular, this made one of the standard tests fail. (The integer value
754 went into the wrong half of a long int.)
755
756 3. If the --colour option was given to pcregrep with -v (invert match), it
757 did strange things, either producing crazy output, or crashing. It should,
758 of course, ignore a request for colour when reporting lines that do not
759 match.
760
761 4. Another pcregrep bug caused similar problems if --colour was specified with
762 -M (multiline) and the pattern match finished with a line ending.
763
764 5. In pcregrep, when a pattern that ended with a literal newline sequence was
765 matched in multiline mode, the following line was shown as part of the
766 match. This seems wrong, so I have changed it.
767
768 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
769 the check for further matches in the same line (so they could be coloured)
770 to overrun the end of the current line. If another match was found, it was
771 incorrectly shown (and then shown again when found in the next line).
772
773 7. If pcregrep was compiled under Windows, there was a reference to the
774 function pcregrep_exit() before it was defined. I am assuming this was
775 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
776 reported by a user. I've moved the definition above the reference.
777
778
779 Version 8.11 10-Dec-2010
780 ------------------------
781
782 1. (*THEN) was not working properly if there were untried alternatives prior
783 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
784 backtracked to try for "b" instead of moving to the next alternative branch
785 at the same level (in this case, to look for "c"). The Perl documentation
786 is clear that when (*THEN) is backtracked onto, it goes to the "next
787 alternative in the innermost enclosing group".
788
789 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
790 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
791 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
792 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
793 (*THEN).
794
795 3. If \s appeared in a character class, it removed the VT character from
796 the class, even if it had been included by some previous item, for example
797 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
798 of \s, but is part of the POSIX "space" class.)
799
800 4. A partial match never returns an empty string (because you can always
801 match an empty string at the end of the subject); however the checking for
802 an empty string was starting at the "start of match" point. This has been
803 changed to the "earliest inspected character" point, because the returned
804 data for a partial match starts at this character. This means that, for
805 example, /(?<=abc)def/ gives a partial match for the subject "abc"
806 (previously it gave "no match").
807
808 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
809 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
810 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
811 has an implication that the given string is incomplete (because a partial
812 match is preferred over a full match). For this reason, these items now
813 give a partial match in this situation. [Aside: previously, the one case
814 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
815 match rather than a full match, which was wrong by the old rules, but is
816 now correct.]
817
818 6. There was a bug in the handling of #-introduced comments, recognized when
819 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
820 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
821 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
822 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
823 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
824 places in pcre_compile().
825
826 7. Related to (6) above, when pcre_compile() was skipping #-introduced
827 comments when looking ahead for named forward references to subpatterns,
828 the only newline sequence it recognized was NL. It now handles newlines
829 according to the set newline convention.
830
831 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
832 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
833 cater for a lack of strerror(). These oversights have been fixed.
834
835 9. Added --match-limit and --recursion-limit to pcregrep.
836
837 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
838
839 11. When the -o option was used, pcregrep was setting a return code of 1, even
840 when matches were found, and --line-buffered was not being honoured.
841
842 12. Added an optional parentheses number to the -o and --only-matching options
843 of pcregrep.
844
845 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
846 can match an empty string. The code to do it in pcretest and pcredemo
847 needed fixing:
848
849 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
850 only one byte after an empty string match just before CRLF (this case
851 just got forgotten; "any" and "anycrlf" were OK).
852
853 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
854 mode when an empty string match preceded an ASCII character followed by
855 a non-ASCII character. (The code for advancing by one character rather
856 than one byte was nonsense.)
857
858 (c) The pcredemo.c sample program did not have any code at all to handle
859 the cases when CRLF is a valid newline sequence.
860
861 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
862 as a starting offset was within the subject string. There is now a new
863 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
864 negative or greater than the length of the string. In order to test this,
865 pcretest is extended to allow the setting of negative starting offsets.
866
867 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
868 starting offset points to the beginning of a UTF-8 character was
869 unnecessarily clumsy. I tidied it up.
870
871 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
872 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
873
874 17. Nobody had reported that the --include_dir option, which was added in
875 release 7.7 should have been called --include-dir (hyphen, not underscore)
876 for compatibility with GNU grep. I have changed it to --include-dir, but
877 left --include_dir as an undocumented synonym, and the same for
878 --exclude-dir, though that is not available in GNU grep, at least as of
879 release 2.5.4.
880
881 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
882 characters from a string of bytes) have been redefined so as not to use
883 loops, in order to improve performance in some environments. At the same
884 time, I abstracted some of the common code into auxiliary macros to save
885 repetition (this should not affect the compiled code).
886
887 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
888 compile-time error is now given if \c is not followed by an ASCII
889 character, that is, a byte less than 128. (In EBCDIC mode, the code is
890 different, and any byte value is allowed.)
891
892 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
893 START_OPTIMIZE option, which is now allowed at compile time - but just
894 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
895 to pcregrep and other applications that have no direct access to PCRE
896 options. The new /Y option in pcretest sets this option when calling
897 pcre_compile().
898
899 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
900 back references. Groups containing recursive back references were forced to
901 be atomic by that change, but in the case of named groups, the amount of
902 memory required was incorrectly computed, leading to "Failed: internal
903 error: code overflow". This has been fixed.
904
905 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
906 pcretest.c, to avoid build problems in some Borland environments.
907
908
909 Version 8.10 25-Jun-2010
910 ------------------------
911
912 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
913 THEN.
914
915 2. (*ACCEPT) was not working when inside an atomic group.
916
917 3. Inside a character class, \B is treated as a literal by default, but
918 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
919 causes the error). The code is unchanged, but I tidied the documentation.
920
921 4. Inside a character class, PCRE always treated \R and \X as literals,
922 whereas Perl faults them if its -w option is set. I have changed PCRE so
923 that it faults them when PCRE_EXTRA is set.
924
925 5. Added support for \N, which always matches any character other than
926 newline. (It is the same as "." when PCRE_DOTALL is not set.)
927
928 6. When compiling pcregrep with newer versions of gcc which may have
929 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
930 declared with attribute warn_unused_result" were given. Just casting the
931 result to (void) does not stop the warnings; a more elaborate fudge is
932 needed. I've used a macro to implement this.
933
934 7. Minor change to pcretest.c to avoid a compiler warning.
935
936 8. Added four artifical Unicode properties to help with an option to make
937 \s etc use properties (see next item). The new properties are: Xan
938 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
939
940 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
941 use Unicode properties. (*UCP) at the start of a pattern can be used to set
942 this option. Modified pcretest to add /W to test this facility. Added
943 REG_UCP to make it available via the POSIX interface.
944
945 10. Added --line-buffered to pcregrep.
946
947 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
948 studied, and the match started with a letter with a code point greater than
949 127 whose first byte was different to the first byte of the other case of
950 the letter, the other case of this starting letter was not recognized
951 (#976).
952
953 12. If a pattern that was studied started with a repeated Unicode property
954 test, for example, \p{Nd}+, there was the theoretical possibility of
955 setting up an incorrect bitmap of starting bytes, but fortunately it could
956 not have actually happened in practice until change 8 above was made (it
957 added property types that matched character-matching opcodes).
958
959 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
960 possible starting bytes for non-anchored patterns.
961
962 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
963 \R, and also a number of cases that involve Unicode properties, both
964 explicit and implicit when PCRE_UCP is set.
965
966 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
967 input, it could crash or give wrong results if characters with values
968 greater than 0xc0 were present in the subject string. (Detail: it assumed
969 UTF-8 input when processing these items.)
970
971 16. Added a lot of (int) casts to avoid compiler warnings in systems where
972 size_t is 64-bit (#991).
973
974 17. Added a check for running out of memory when PCRE is compiled with
975 --disable-stack-for-recursion (#990).
976
977 18. If the last data line in a file for pcretest does not have a newline on
978 the end, a newline was missing in the output.
979
980 19. The default pcre_chartables.c file recognizes only ASCII characters (values
981 less than 128) in its various bitmaps. However, there is a facility for
982 generating tables according to the current locale when PCRE is compiled. It
983 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
984 characters, are recognized by isspace() and therefore were getting set in
985 these tables, and indeed these tables seem to approximate to ISO 8859. This
986 caused a problem in UTF-8 mode when pcre_study() was used to create a list
987 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
988 which of course cannot start UTF-8 characters. I have changed the code so
989 that only real ASCII characters (less than 128) and the correct starting
990 bytes for UTF-8 encodings are set for characters greater than 127 when in
991 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
992 altogether.)
993
994 20. Added the /T option to pcretest so as to be able to run tests with non-
995 standard character tables, thus making it possible to include the tests
996 used for 19 above in the standard set of tests.
997
998 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
999 reference to a subpattern the other side of a comment that contains an
1000 opening parenthesis caused either an internal compiling error, or a
1001 reference to the wrong subpattern.
1002
1003
1004 Version 8.02 19-Mar-2010
1005 ------------------------
1006
1007 1. The Unicode data tables have been updated to Unicode 5.2.0.
1008
1009 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1010 configured.
1011
1012 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1013 original author of that file, following a query about its status.
1014
1015 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1016 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1017
1018 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1019 quantifier applied to a forward-referencing subroutine call, could compile
1020 incorrect code or give the error "internal error: previously-checked
1021 referenced subpattern not found".
1022
1023 6. Both MS Visual Studio and Symbian OS have problems with initializing
1024 variables to point to external functions. For these systems, therefore,
1025 pcre_malloc etc. are now initialized to local functions that call the
1026 relevant global functions.
1027
1028 7. There were two entries missing in the vectors called coptable and poptable
1029 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1030 I've fixed the data, and added a kludgy way of testing at compile time that
1031 the lengths are correct (equal to the number of opcodes).
1032
1033 8. Following on from 7, I added a similar kludge to check the length of the
1034 eint vector in pcreposix.c.
1035
1036 9. Error texts for pcre_compile() are held as one long string to avoid too
1037 much relocation at load time. To find a text, the string is searched,
1038 counting zeros. There was no check for running off the end of the string,
1039 which could happen if a new error number was added without updating the
1040 string.
1041
1042 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1043
1044 11. \K was not working if it appeared in an atomic group or in a group that
1045 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1046 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1047 the assertion is positive, but not if it is negative.
1048
1049 12. Change 11 fortuitously reduced the size of the stack frame used in the
1050 "match()" function of pcre_exec.c by one pointer. Forthcoming
1051 implementation of support for (*MARK) will need an extra pointer on the
1052 stack; I have reserved it now, so that the stack frame size does not
1053 decrease.
1054
1055 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1056 item in branch that calls a recursion is a subroutine call - as in the
1057 second branch in the above example - was incorrectly given the compile-
1058 time error "recursive call could loop indefinitely" because pcre_compile()
1059 was not correctly checking the subroutine for matching a non-empty string.
1060
1061 14. The checks for overrunning compiling workspace could trigger after an
1062 overrun had occurred. This is a "should never occur" error, but it can be
1063 triggered by pathological patterns such as hundreds of nested parentheses.
1064 The checks now trigger 100 bytes before the end of the workspace.
1065
1066 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1067
1068
1069 Version 8.01 19-Jan-2010
1070 ------------------------
1071
1072 1. If a pattern contained a conditional subpattern with only one branch (in
1073 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1074 computed the wrong minimum data length (which is of course zero for such
1075 subpatterns). This could cause incorrect "no match" results.
1076
1077 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1078 the pattern is reset in the first branch, pcre_compile() failed with
1079 "internal error: code overflow at offset...". This happened only when
1080 the reset was to the original external option setting. (An optimization
1081 abstracts leading options settings into an external setting, which was the
1082 cause of this.)
1083
1084 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1085 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1086 assertion pattern did not match (meaning that the assertion was true), it
1087 was incorrectly treated as false if the SKIP had been reached during the
1088 matching. This also applied to assertions used as conditions.
1089
1090 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1091 assertion subpattern, including such a pattern used as a condition,
1092 unpredictable results occurred, instead of the error return
1093 PCRE_ERROR_DFA_UITEM.
1094
1095 5. The C++ GlobalReplace function was not working like Perl for the special
1096 situation when an empty string is matched. It now does the fancy magic
1097 stuff that is necessary.
1098
1099 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1100 removed. (These were left over from very, very early versions of PCRE.)
1101
1102 7. Some cosmetic changes to the code to make life easier when compiling it
1103 as part of something else:
1104
1105 (a) Change DEBUG to PCRE_DEBUG.
1106
1107 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1108 called "current" as "current_branch", to prevent a collision with the
1109 Linux macro when compiled as a kernel module.
1110
1111 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1112 prevent a collision with the Linux macro when compiled as a kernel
1113 module.
1114
1115 8. In pcre_compile() there are some checks for integer overflows that used to
1116 cast potentially large values to (double). This has been changed to that
1117 when building, a check for int64_t is made, and if it is found, it is used
1118 instead, thus avoiding the use of floating point arithmetic. (There is no
1119 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1120 double.
1121
1122 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1123 2005 (difference between two addresses compared to an unsigned value).
1124
1125 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1126 custom one, because of the following reported problem in Windows:
1127
1128 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1129 under Win32.
1130 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1131 therefore missing the function definition.
1132 - The compiler thus generates a "C" signature for the test function.
1133 - The linker fails to find the "C" function.
1134 - PCRE fails to configure if asked to do so against libbz2.
1135
1136 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1137 messages were output:
1138
1139 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1140 rerunning libtoolize, to keep the correct libtool macros in-tree.
1141 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1142
1143 I have done both of these things.
1144
1145 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1146 most of the time, it *can* run out if it is given a pattern that contains a
1147 runaway infinite recursion. I updated the discussion in the pcrestack man
1148 page.
1149
1150 13. Now that we have gone to the x.xx style of version numbers, the minor
1151 version may start with zero. Using 08 or 09 is a bad idea because users
1152 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1153 interpreted as invalid octal numbers. I've updated the previous comment in
1154 configure.ac, and also added a check that gives an error if 08 or 09 are
1155 used.
1156
1157 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1158 causing partial matching to fail when the end of the subject matched \W
1159 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1160
1161 15. There were some discrepancies between the declarations in pcre_internal.h
1162 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1163 their definitions. The declarations used "const uschar *" and the
1164 definitions used USPTR. Even though USPTR is normally defined as "const
1165 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1166 reported that: "This difference in casting confuses some C++ compilers, for
1167 example, SunCC recognizes above declarations as different functions and
1168 generates broken code for hbpcre." I have changed the declarations to use
1169 USPTR.
1170
1171 16. GNU libtool is named differently on some systems. The autogen.sh script now
1172 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1173 (FreeBSD).
1174
1175 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1176 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1177 comment: "Figure out how to create a longlong from a string: strtoll and
1178 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1179 instance, but it only takes 2 args instead of 3!"
1180
1181 18. A subtle bug concerned with back references has been fixed by a change of
1182 specification, with a corresponding code fix. A pattern such as
1183 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1184 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1185 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1186 same bug. Such groups have to be quantified to be useful, or contained
1187 inside another quantified group. (If there's no repetition, the reference
1188 can never match.) The problem arises because, having left the group and
1189 moved on to the rest of the pattern, a later failure that backtracks into
1190 the group uses the captured value from the final iteration of the group
1191 rather than the correct earlier one. I have fixed this in PCRE by forcing
1192 any group that contains a reference to itself to be an atomic group; that
1193 is, there cannot be any backtracking into it once it has completed. This is
1194 similar to recursive and subroutine calls.
1195
1196
1197 Version 8.00 19-Oct-09
1198 ----------------------
1199
1200 1. The table for translating pcre_compile() error codes into POSIX error codes
1201 was out-of-date, and there was no check on the pcre_compile() error code
1202 being within the table. This could lead to an OK return being given in
1203 error.
1204
1205 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1206 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1207 in a Windows environment.
1208
1209 3. The pcregrep --count option prints the count for each file even when it is
1210 zero, as does GNU grep. However, pcregrep was also printing all files when
1211 --files-with-matches was added. Now, when both options are given, it prints
1212 counts only for those files that have at least one match. (GNU grep just
1213 prints the file name in this circumstance, but including the count seems
1214 more useful - otherwise, why use --count?) Also ensured that the
1215 combination -clh just lists non-zero counts, with no names.
1216
1217 4. The long form of the pcregrep -F option was incorrectly implemented as
1218 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1219 but it seems right to fix it, and I didn't think it was worth preserving
1220 the old behaviour.
1221
1222 5. The command line items --regex=pattern and --regexp=pattern were not
1223 recognized by pcregrep, which required --regex pattern or --regexp pattern
1224 (with a space rather than an '='). The man page documented the '=' forms,
1225 which are compatible with GNU grep; these now work.
1226
1227 6. No libpcreposix.pc file was created for pkg-config; there was just
1228 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1229
1230 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1231 when UCP support is not needed, by modifying the Python script that
1232 generates it from Unicode data files. This should not matter if the module
1233 is correctly used as a library, but I received one complaint about 50K of
1234 unwanted data. My guess is that the person linked everything into his
1235 program rather than using a library. Anyway, it does no harm.
1236
1237 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1238 was a minimum greater than 1 for a wide character in a possessive
1239 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1240 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1241 character. Chaos in the form of incorrect output or a compiling loop could
1242 result.
1243
1244 9. The restrictions on what a pattern can contain when partial matching is
1245 requested for pcre_exec() have been removed. All patterns can now be
1246 partially matched by this function. In addition, if there are at least two
1247 slots in the offset vector, the offset of the earliest inspected character
1248 for the match and the offset of the end of the subject are set in them when
1249 PCRE_ERROR_PARTIAL is returned.
1250
1251 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1252 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1253 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1254 and may be more useful for multi-segment matching.
1255
1256 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1257 used to be given if ever the end of the subject was reached; now it is
1258 given only if matching could not proceed because another character was
1259 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1260 string "Z", which now yields "no match" instead of "partial match". In the
1261 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1262 final character ended with (*FAIL).
1263
1264 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1265 if the pattern had a "must contain" character that was already found in the
1266 earlier partial match, unless partial matching was again requested. For
1267 example, with the pattern /dog.(body)?/, the "must contain" character is
1268 "g". If the first part-match was for the string "dog", restarting with
1269 "sbody" failed. This bug has been fixed.
1270
1271 13. The string returned by pcre_dfa_exec() after a partial match has been
1272 changed so that it starts at the first inspected character rather than the
1273 first character of the match. This makes a difference only if the pattern
1274 starts with a lookbehind assertion or \b or \B (\K is not supported by
1275 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1276 matching functions compatible, and I think it's the right thing to do.
1277
1278 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1279 so that the demonstration program is easily available in environments where
1280 PCRE has not been installed from source.
1281
1282 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1283 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1284 library.
1285
1286 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1287 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1288 is not the first non-POSIX option to be added. Clearly some people find
1289 these options useful.
1290
1291 17. If a caller to the POSIX matching function regexec() passes a non-zero
1292 value for nmatch with a NULL value for pmatch, the value of
1293 nmatch is forced to zero.
1294
1295 18. RunGrepTest did not have a test for the availability of the -u option of
1296 the diff command, as RunTest does. It now checks in the same way as
1297 RunTest, and also checks for the -b option.
1298
1299 19. If an odd number of negated classes containing just a single character
1300 interposed, within parentheses, between a forward reference to a named
1301 subpattern and the definition of the subpattern, compilation crashed with
1302 an internal error, complaining that it could not find the referenced
1303 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1304 [The bug was that it was starting one character too far in when skipping
1305 over the character class, thus treating the ] as data rather than
1306 terminating the class. This meant it could skip too much.]
1307
1308 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1309 /g option in pcretest when the pattern contains \K, which makes it possible
1310 to have an empty string match not at the start, even when the pattern is
1311 anchored. Updated pcretest and pcredemo to use this option.
1312
1313 21. If the maximum number of capturing subpatterns in a recursion was greater
1314 than the maximum at the outer level, the higher number was returned, but
1315 with unset values at the outer level. The correct (outer level) value is
1316 now given.
1317
1318 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1319 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1320 make it do so. The string so far is captured, making this feature
1321 compatible with Perl.
1322
1323 23. The tests have been re-organized, adding tests 11 and 12, to make it
1324 possible to check the Perl 5.10 features against Perl 5.10.
1325
1326 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1327 pattern matches a fixed length string. PCRE did not allow this; now it
1328 does. Neither allows recursion.
1329
1330 25. I finally figured out how to implement a request to provide the minimum
1331 length of subject string that was needed in order to match a given pattern.
1332 (It was back references and recursion that I had previously got hung up
1333 on.) This code has now been added to pcre_study(); it finds a lower bound
1334 to the length of subject needed. It is not necessarily the greatest lower
1335 bound, but using it to avoid searching strings that are too short does give
1336 some useful speed-ups. The value is available to calling programs via
1337 pcre_fullinfo().
1338
1339 26. While implementing 25, I discovered to my embarrassment that pcretest had
1340 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1341 study optimizations had never been tested with that matching function.
1342 Oops. What is worse, even when it was passed study data, there was a bug in
1343 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1344 were also very few tests of studied patterns with pcre_dfa_exec().
1345
1346 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1347 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1348 on the other side of the coin, they are no longer allowed to have different
1349 names, because these cannot be distinguished in PCRE, and this has caused
1350 confusion. (This is a difference from Perl.)
1351
1352 28. When duplicate subpattern names are present (necessarily with different
1353 numbers, as required by 27 above), and a test is made by name in a
1354 conditional pattern, either for a subpattern having been matched, or for
1355 recursion in such a pattern, all the associated numbered subpatterns are
1356 tested, and the overall condition is true if the condition is true for any
1357 one of them. This is the way Perl works, and is also more like the way
1358 testing by number works.
1359
1360
1361 Version 7.9 11-Apr-09
1362 ---------------------
1363
1364 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1365 (pcretest), all targets were linked against these libraries. This included
1366 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1367 libraries. This caused unwanted dependencies to be created. This problem
1368 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1369 pcretest is linked with readline.
1370
1371 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1372 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1373 moved outside it again, because FALSE and TRUE are already defined in AIX,
1374 but BOOL is not.
1375
1376 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1377 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1378
1379 4. The pcregrep documentation said spaces were inserted as well as colons (or
1380 hyphens) following file names and line numbers when outputting matching
1381 lines. This is not true; no spaces are inserted. I have also clarified the
1382 wording for the --colour (or --color) option.
1383
1384 5. In pcregrep, when --colour was used with -o, the list of matching strings
1385 was not coloured; this is different to GNU grep, so I have changed it to be
1386 the same.
1387
1388 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1389 each matching line was coloured. Now it goes on to look for further matches
1390 of any of the test patterns, which is the same behaviour as GNU grep.
1391
1392 7. A pattern that could match an empty string could cause pcregrep to loop; it
1393 doesn't make sense to accept an empty string match in pcregrep, so I have
1394 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1395 seems to be how GNU grep behaves.
1396
1397 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1398 start or after a newline", because the conditional assertion was not being
1399 correctly handled. The rule now is that both the assertion and what follows
1400 in the first alternative must satisfy the test.
1401
1402 9. If auto-callout was enabled in a pattern with a conditional group whose
1403 condition was an assertion, PCRE could crash during matching, both with
1404 pcre_exec() and pcre_dfa_exec().
1405
1406 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1407 used for matching.
1408
1409 11. Unicode property support in character classes was not working for
1410 characters (bytes) greater than 127 when not in UTF-8 mode.
1411
1412 12. Added the -M command line option to pcretest.
1413
1414 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1415
1416 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1417
1418 16. Added comments and documentation about mis-use of no_arg in the C++
1419 wrapper.
1420
1421 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1422 from Martin Jerabek that uses macro names for all relevant character and
1423 string constants.
1424
1425 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1426 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1427 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1428 these, but not everybody uses configure.
1429
1430 19. A conditional group that had only one branch was not being correctly
1431 recognized as an item that could match an empty string. This meant that an
1432 enclosing group might also not be so recognized, causing infinite looping
1433 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1434 with the subject "ab", where knowledge that the repeated group can match
1435 nothing is needed in order to break the loop.
1436
1437 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1438 exec(), but without supplying a callout function, matching went wrong.
1439
1440 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1441 leak if the size of the offset vector was greater than 30. When the vector
1442 is smaller, the saved offsets during recursion go onto a local stack
1443 vector, but for larger vectors malloc() is used. It was failing to free
1444 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1445 error, in fact).
1446
1447 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1448 heapframe that is used only when UTF-8 support is enabled. This caused no
1449 problem, but was untidy.
1450
1451 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1452 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1453 included within another project.
1454
1455 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1456 slightly modified by me:
1457
1458 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1459 not building pcregrep.
1460
1461 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1462 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1463
1464 25. Forward references, both numeric and by name, in patterns that made use of
1465 duplicate group numbers, could behave incorrectly or give incorrect errors,
1466 because when scanning forward to find the reference group, PCRE was not
1467 taking into account the duplicate group numbers. A pattern such as
1468 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1469
1470 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1471 the feature of a custom pointer more persuasive (as requested by a user).
1472
1473 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1474 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1475 pre-defined.
1476
1477 28. Added support for (*UTF8) at the start of a pattern.
1478
1479 29. Arrange for flags added by the "release type" setting in CMake to be shown
1480 in the configuration summary.
1481
1482
1483 Version 7.8 05-Sep-08
1484 ---------------------
1485
1486 1. Replaced UCP searching code with optimized version as implemented for Ad
1487 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1488 stage table and inline lookup instead of a function, giving speed ups of 2
1489 to 5 times on some simple patterns that I tested. Permission was given to
1490 distribute the MultiStage2.py script that generates the tables (it's not in
1491 the tarball, but is in the Subversion repository).
1492
1493 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1494 scripts.
1495
1496 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1497 a group with a zero qualifier. The result of the study could be incorrect,
1498 or the function might crash, depending on the pattern.
1499
1500 4. Caseless matching was not working for non-ASCII characters in back
1501 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1502 It now works when Unicode Property Support is available.
1503
1504 5. In pcretest, an escape such as \x{de} in the data was always generating
1505 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1506 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1507 truncation.
1508
1509 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1510
1511 7. Added two (int) casts to pcregrep when printing the difference of two
1512 pointers, in case they are 64-bit values.
1513
1514 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1515 test 2 if it fails.
1516
1517 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1518 and a #define of that name to empty if it is not externally set. This is to
1519 allow users of MSVC to set it if necessary.
1520
1521 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1522 the convenience functions in the pcre_get.c source file.
1523
1524 11. An option change at the start of a pattern that had top-level alternatives
1525 could cause overwriting and/or a crash. This command provoked a crash in
1526 some environments:
1527
1528 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1529
1530 This potential security problem was recorded as CVE-2008-2371.
1531
1532 12. For a pattern where the match had to start at the beginning or immediately
1533 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1534 pcre_dfa_exec() could read past the end of the passed subject if there was
1535 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1536 pcretest so that it places the subject at the end of its malloc-ed buffer.
1537
1538 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1539 exec() might read past the end of the data buffer in UTF-8 mode.
1540
1541 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1542 the data contained the byte 0x85 as part of a UTF-8 character within its
1543 first line. This applied both to normal and DFA matching.
1544
1545 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1546 /^[^d]*?$/8 failed to match "abc".
1547
1548 16. Added a missing copyright notice to pcrecpp_internal.h.
1549
1550 17. Make it more clear in the documentation that values returned from
1551 pcre_exec() in ovector are byte offsets, not character counts.
1552
1553 18. Tidied a few places to stop certain compilers from issuing warnings.
1554
1555 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1556 supplied by Stefan Weber. I made a further small update for 7.8 because
1557 there is a change of source arrangements: the pcre_searchfuncs.c module is
1558 replaced by pcre_ucd.c.
1559
1560
1561 Version 7.7 07-May-08
1562 ---------------------
1563
1564 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1565 a string to a long long, pretend we don't even have a long long." This is
1566 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1567
1568 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1569 pre-7.6 versions, which defined a global no_arg variable instead of putting
1570 it in the RE class. (See also #8 below.)
1571
1572 3. Remove a line of dead code, identified by coverity and reported by Nuno
1573 Lopes.
1574
1575 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1576
1577 (1) The include/exclude patterns were being applied to the whole pathnames
1578 of files, instead of just to the final components.
1579
1580 (2) If there was more than one level of directory, the subdirectories were
1581 skipped unless they satisfied the include/exclude conditions. This is
1582 inconsistent with GNU grep (and could even be seen as contrary to the
1583 pcregrep specification - which I improved to make it absolutely clear).
1584 The action now is always to scan all levels of directory, and just
1585 apply the include/exclude patterns to regular files.
1586
1587 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1588 --exclude_dir in the tests to avoid scanning .svn directories.
1589
1590 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1591 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1592 doesn't support NULs in patterns.
1593
1594 7. Added some missing "const"s to declarations of static tables in
1595 pcre_compile.c and pcre_dfa_exec.c.
1596
1597 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1598 caused by fix #2 above. (Subsequently also a second patch to fix the
1599 first patch. And a third patch - this was a messy problem.)
1600
1601 9. Applied Craig's patch to remove the use of push_back().
1602
1603 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1604 matching function regexec().
1605
1606 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1607 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1608 references. PCRE supports relative numbers with this syntax (I don't think
1609 Oniguruma does).
1610
1611 12. Previously, a group with a zero repeat such as (...){0} was completely
1612 omitted from the compiled regex. However, this means that if the group
1613 was called as a subroutine from elsewhere in the pattern, things went wrong
1614 (an internal error was given). Such groups are now left in the compiled
1615 pattern, with a new opcode that causes them to be skipped at execution
1616 time.
1617
1618 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1619 to the way PCRE behaves:
1620
1621 (a) A lone ] character is dis-allowed (Perl treats it as data).
1622
1623 (b) A back reference to an unmatched subpattern matches an empty string
1624 (Perl fails the current match path).
1625
1626 (c) A data ] in a character class must be notated as \] because if the
1627 first data character in a class is ], it defines an empty class. (In
1628 Perl it is not possible to have an empty class.) The empty class []
1629 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1630 The negative empty class [^] matches any one character, independently
1631 of the DOTALL setting.
1632
1633 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1634 non-existent subpattern following a character class starting with ']' and
1635 containing () gave an internal compiling error instead of "reference to
1636 non-existent subpattern". Fortunately, when the pattern did exist, the
1637 compiled code was correct. (When scanning forwards to check for the
1638 existencd of the subpattern, it was treating the data ']' as terminating
1639 the class, so got the count wrong. When actually compiling, the reference
1640 was subsequently set up correctly.)
1641
1642 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1643 it was being rejected as not supported by pcre_dfa_exec(), even though
1644 other assertions are supported. I have made pcre_dfa_exec() support
1645 (*FAIL).
1646
1647 16. The implementation of 13c above involved the invention of a new opcode,
1648 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1649 cannot be changed at match time, I realized I could make a small
1650 improvement to matching performance by compiling OP_ALLANY instead of
1651 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1652 on the OP_ANY path.
1653
1654 17. Compiling pcretest on Windows with readline support failed without the
1655 following two fixes: (1) Make the unistd.h include conditional on
1656 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1657
1658 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1659 ncurses library to be included for pcretest when ReadLine support is
1660 requested, but also to allow for it to be overridden. This patch came from
1661 Daniel Bergström.
1662
1663 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1664 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1665 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1666 spotting this.
1667
1668
1669 Version 7.6 28-Jan-08
1670 ---------------------
1671
1672 1. A character class containing a very large number of characters with
1673 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1674 overflow.
1675
1676 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1677 HAVE_LONG_LONG is not defined.
1678
1679 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1680 bring them up to date and include new features. This patch includes:
1681
1682 - Fixed PH's badly added libz and libbz2 support.
1683 - Fixed a problem with static linking.
1684 - Added pcredemo. [But later removed - see 7 below.]
1685 - Fixed dftables problem and added an option.
1686 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1687 HAVE_LONG_LONG.
1688 - Added readline support for pcretest.
1689 - Added an listing of the option settings after cmake has run.
1690
1691 4. A user submitted a patch to Makefile that makes it easy to create
1692 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1693 Makefile.am that cause it to include this special target, without
1694 affecting anything else. Note that the same mingw target plus all
1695 the other distribution libraries and programs are now supported
1696 when configuring with CMake (see 6 below) instead of with
1697 Configure/Make.
1698
1699 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1700 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1701 exported in the Windows port". It has not yet been confirmed that the patch
1702 solves the problem, but it does no harm.
1703
1704 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1705 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1706 with CMake, and also correct the comment about stack recursion.
1707
1708 7. Remove the automatic building of pcredemo from the ./configure system and
1709 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1710 of a program that users should build themselves after PCRE is installed, so
1711 building it automatically is not really right. What is more, it gave
1712 trouble in some build environments.
1713
1714 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1715
1716
1717 Version 7.5 10-Jan-08
1718 ---------------------
1719
1720 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1721 values in parens when parsing an RE using the C++ wrapper."
1722
1723 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1724 Characters greater than 255 were excluded from the class instead of being
1725 included.
1726
1727 3. The same bug as (2) above applied to negated POSIX classes such as
1728 [:^space:].
1729
1730 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1731 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1732 I have changed it.
1733
1734 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1735 first named subpattern) and a construct such as (?&a) would reference the
1736 first named subpattern whose name started with "a" (in other words, the
1737 length check was missing). Both these problems are fixed. "Subpattern name
1738 expected" is now given for (?&) (a zero-length name), and this patch also
1739 makes it give the same error for \k'' (previously it complained that that
1740 was a reference to a non-existent subpattern).
1741
1742 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1743 this is right because (?- can be followed by option settings as well as by
1744 digits. I have, however, made the messages clearer.
1745
1746 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1747 than the number used in the conditional) now cause a compile-time error.
1748 This is actually not compatible with Perl, which accepts such patterns, but
1749 treats the conditional as always being FALSE (as PCRE used to), but it
1750 seems to me that giving a diagnostic is better.
1751
1752 8. Change "alphameric" to the more common word "alphanumeric" in comments
1753 and messages.
1754
1755 9. Fix two occurrences of "backslash" in comments that should have been
1756 "backspace".
1757
1758 10. Remove two redundant lines of code that can never be obeyed (their function
1759 was moved elsewhere).
1760
1761 11. The program that makes PCRE's Unicode character property table had a bug
1762 which caused it to generate incorrect table entries for sequences of
1763 characters that have the same character type, but are in different scripts.
1764 It amalgamated them into a single range, with the script of the first of
1765 them. In other words, some characters were in the wrong script. There were
1766 thirteen such cases, affecting characters in the following ranges:
1767
1768 U+002b0 - U+002c1
1769 U+0060c - U+0060d
1770 U+0061e - U+00612
1771 U+0064b - U+0065e
1772 U+0074d - U+0076d
1773 U+01800 - U+01805
1774 U+01d00 - U+01d77
1775 U+01d9b - U+01dbf
1776 U+0200b - U+0200f
1777 U+030fc - U+030fe
1778 U+03260 - U+0327f
1779 U+0fb46 - U+0fbb1
1780 U+10450 - U+1049d
1781
1782 12. The -o option (show only the matching part of a line) for pcregrep was not
1783 compatible with GNU grep in that, if there was more than one match in a
1784 line, it showed only the first of them. It now behaves in the same way as
1785 GNU grep.
1786
1787 13. If the -o and -v options were combined for pcregrep, it printed a blank
1788 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1789 does the same. The return code can be used to tell if there were any
1790 non-matching lines.
1791
1792 14. Added --file-offsets and --line-offsets to pcregrep.
1793
1794 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1795 infinitely looping recursion. The bug was that positive lookaheads were not
1796 being skipped when checking for a possible empty match (negative lookaheads
1797 and both kinds of lookbehind were skipped).
1798
1799 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1800 inclusion of <windows.h> to before rather than after the definition of
1801 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1802
1803 17. Specifying a possessive quantifier with a specific limit for a Unicode
1804 character property caused pcre_compile() to compile bad code, which led at
1805 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1806 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1807 caused the error; without that there was no problem.
1808
1809 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1810
1811 19. Added --enable-pcretest-libreadline.
1812
1813 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1814 RE::GlobalReplace(). As a result, the number of replacements returned was
1815 double what it should be. I removed one of the increments, but Craig sent a
1816 later patch that removed the other one (the right fix) and added unit tests
1817 that check the return values (which was not done before).
1818
1819 21. Several CMake things:
1820
1821 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1822 the names libpcre and libpcreposix, not just pcre and pcreposix.
1823
1824 (2) The above change means that pcretest and pcregrep are now correctly
1825 linked with the newly-built libraries, not previously installed ones.
1826
1827 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1828
1829 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1830 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1831 UTF-8 newline character). The key issue is that the pattern starts .*;
1832 this means that the match must be either at the beginning, or after a
1833 newline. The bug was in the code for advancing after a failed match and
1834 checking that the new position followed a newline. It was not taking
1835 account of UTF-8 characters correctly.
1836
1837 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1838 character classes. PCRE was not treating the sequence [:...:] as a
1839 character class unless the ... were all letters. Perl, however, seems to
1840 allow any characters between [: and :], though of course it rejects as
1841 unknown any "names" that contain non-letters, because all the known class
1842 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1843 for example, whereas PCRE did not - it did not recognize a POSIX character
1844 class. This seemed a bit dangerous, so the code has been changed to be
1845 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1846 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1847 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1848 Perl does, and where it didn't before.
1849
1850 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1851 Windows environments %n is disabled by default.
1852
1853
1854 Version 7.4 21-Sep-07
1855 ---------------------
1856
1857 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1858 means that a class such as [\s] counted as "explicit reference to CR or
1859 LF". That isn't really right - the whole point of the change was to try to
1860 help when there was an actual mention of one of the two characters. So now
1861 the change happens only if \r or \n (or a literal CR or LF) character is
1862 encountered.
1863
1864 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1865 of both had grown to the point where there were only 3 bits left.
1866 Fortunately, there was spare space in the data structure, and so I have
1867 moved the internal flags into a new 16-bit field to free up more option
1868 bits.
1869
1870 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1871 but did not set the internal JCHANGED flag - either of these is enough to
1872 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1873 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1874 start sets both bits.
1875
1876 4. Added options (at build time, compile time, exec time) to change \R from
1877 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1878
1879 5. doc/pcresyntax.html was missing from the distribution.
1880
1881 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1882 compatibility, even though it is no longer used.
1883
1884 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1885 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1886 windows.h file is present (where different names are used). [This was
1887 reversed later after testing - see 16 below.]
1888
1889 8. Changed all #include <config.h> to #include "config.h". There were also
1890 some further <pcre.h> cases that I changed to "pcre.h".
1891
1892 9. When pcregrep was used with the --colour option, it missed the line ending
1893 sequence off the lines that it output.
1894
1895 10. It was pointed out to me that arrays of string pointers cause lots of
1896 relocations when a shared library is dynamically loaded. A technique of
1897 using a single long string with a table of offsets can drastically reduce
1898 these. I have refactored PCRE in four places to do this. The result is
1899 dramatic:
1900
1901 Originally: 290
1902 After changing UCP table: 187
1903 After changing error message table: 43
1904 After changing table of "verbs" 36
1905 After changing table of Posix names 22
1906
1907 Thanks to the folks working on Gregex for glib for this insight.
1908
1909 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1910 unicode-properties was also set.
1911
1912 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1913
1914 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1915 checked only for CRLF.
1916
1917 14. Added casts to pcretest.c to avoid compiler warnings.
1918
1919 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1920
1921 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1922 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1923 entirely. This removes changes made in 7 above.
1924
1925 17. The CMake files have been updated, and there is now more information about
1926 building with CMake in the NON-UNIX-USE document.
1927
1928
1929 Version 7.3 28-Aug-07
1930 ---------------------
1931
1932 1. In the rejigging of the build system that eventually resulted in 7.1, the
1933 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1934 brackets there is not right, since it causes compilers to look for an
1935 installed pcre.h, not the version that is in the source that is being
1936 compiled (which of course may be different). I have changed it back to:
1937
1938 #include "pcre.h"
1939
1940 I have a vague recollection that the change was concerned with compiling in
1941 different directories, but in the new build system, that is taken care of
1942 by the VPATH setting the Makefile.
1943
1944 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1945 when the subject happened to end in the byte 0x85 (e.g. if the last
1946 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1947 characters but of course it shouldn't be taken as a newline when it is part
1948 of another character. The bug was that, for an unlimited repeat of . in
1949 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1950 characters when looking for a newline.
1951
1952 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1953
1954 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1955 in debug output.
1956
1957 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1958 long printing in the pcrecpp unittest when running under MinGW.
1959
1960 6. ESC_K was left out of the EBCDIC table.
1961
1962 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1963 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1964 limit also applies to "virtual nesting" when a pattern is recursive, and in
1965 this case 1000 isn't so big. I have been able to remove this limit at the
1966 expense of backing off one optimization in certain circumstances. Normally,
1967 when pcre_exec() would call its internal match() function recursively and
1968 immediately return the result unconditionally, it uses a "tail recursion"
1969 feature to save stack. However, when a subpattern that can match an empty
1970 string has an unlimited repetition quantifier, it no longer makes this
1971 optimization. That gives it a stack frame in which to save the data for
1972 checking that an empty string has been matched. Previously this was taken
1973 from the 1000-entry workspace that had been reserved. So now there is no
1974 explicit limit, but more stack is used.
1975
1976 8. Applied Daniel's patches to solve problems with the import/export magic
1977 syntax that is required for Windows, and which was going wrong for the
1978 pcreposix and pcrecpp parts of the library. These were overlooked when this
1979 problem was solved for the main library.
1980
1981 9. There were some crude static tests to avoid integer overflow when computing
1982 the size of patterns that contain repeated groups with explicit upper
1983 limits. As the maximum quantifier is 65535, the maximum group length was
1984 set at 30,000 so that the product of these two numbers did not overflow a
1985 32-bit integer. However, it turns out that people want to use groups that
1986 are longer than 30,000 bytes (though not repeat them that many times).
1987 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1988 made it possible to implement the integer overflow checks in a much more
1989 dynamic way, which I have now done. The artificial limitation on group
1990 length has been removed - we now have only the limit on the total length of
1991 the compiled pattern, which depends on the LINK_SIZE setting.
1992
1993 10. Fixed a bug in the documentation for get/copy named substring when
1994 duplicate names are permitted. If none of the named substrings are set, the
1995 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1996 empty string.
1997
1998 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1999 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2000 because the ] is interpreted as the first data character and the
2001 terminating ] is not found. PCRE has been made compatible with Perl in this
2002 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2003 cause memory overwriting.
2004
2005 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2006 string has been matched (to stop an infinite loop). It was not recognizing
2007 a conditional subpattern that could match an empty string if that
2008 subpattern was within another subpattern. For example, it looped when
2009 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2010 condition was not nested. This bug has been fixed.
2011
2012 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2013 past the start of the subject in the presence of bytes with the top bit
2014 set, for example "\x8aBCD".
2015
2016 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2017 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2018
2019 14. Optimized (?!) to (*FAIL).
2020
2021 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2022 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2023 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2024 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2025 does: it's just the validity check that is more restrictive.
2026
2027 16. Inserted checks for integer overflows during escape sequence (backslash)
2028 processing, and also fixed erroneous offset values for syntax errors during
2029 backslash processing.
2030
2031 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2032 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2033
2034 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2035 caused an overrun.
2036
2037 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2038 something other than just ASCII characters) inside a group that had an
2039 unlimited repeat caused a loop at compile time (while checking to see
2040 whether the group could match an empty string).
2041
2042 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2043 [\P{Any}] did so. (Error in the code for printing property names.)
2044
2045 21. An orphan \E inside a character class could cause a crash.
2046
2047 22. A repeated capturing bracket such as (A)? could cause a wild memory
2048 reference during compilation.
2049
2050 23. There are several functions in pcre_compile() that scan along a compiled
2051 expression for various reasons (e.g. to see if it's fixed length for look
2052 behind). There were bugs in these functions when a repeated \p or \P was
2053 present in the pattern. These operators have additional parameters compared
2054 with \d, etc, and these were not being taken into account when moving along
2055 the compiled data. Specifically:
2056
2057 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2058 length.
2059
2060 (b) An item such as \pL+ within a repeated group could cause crashes or
2061 loops.
2062
2063 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2064 "reference to non-existent subpattern" error.
2065
2066 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2067
2068 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2069 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2070
2071 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2072
2073 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2074 character were causing crashes (broken optimization).
2075
2076 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2077 \p or \P) caused a compile-time loop.
2078
2079 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2080 break. For example, the unstudied pattern [\r\n]A does not match the string
2081 "\r\nA" because change 7.0/46 below moves the current point on by two
2082 characters after failing to match at the start. However, the pattern \nA
2083 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2084 the same is true. There doesn't seem any very clean way out of this, but
2085 what I have chosen to do makes the common cases work: PCRE now takes note
2086 of whether there can be an explicit match for \r or \n anywhere in the
2087 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2088 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2089 pattern has explicit CR or LF references.
2090
2091 29. Added (*CR) etc for changing newline setting at start of pattern.
2092
2093
2094 Version 7.2 19-Jun-07
2095 ---------------------
2096
2097 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2098 which is apparently normally available under Windows.
2099
2100 2. Re-jig the pcregrep tests with different newline settings in an attempt
2101 to make them independent of the local environment's newline setting.
2102
2103 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2104
2105 4. Some of the "internals" tests were previously cut out when the link size
2106 was not 2, because the output contained actual offsets. The recent new
2107 "Z" feature of pcretest means that these can be cut out, making the tests
2108 usable with all link sizes.
2109
2110 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2111 stack recursion. This gives a massive performance boost under BSD, but just
2112 a small improvement under Linux. However, it saves one field in the frame
2113 in all cases.
2114
2115 6. Added more features from the forthcoming Perl 5.10:
2116
2117 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2118 recursion call. It refers to the nth most recently opened parentheses.
2119
2120 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2121 to be opened parentheses.
2122
2123 (c) Conditions that refer to capturing parentheses can be specified
2124 relatively, for example, (?(-2)... or (?(+3)...
2125
2126 (d) \K resets the start of the current match so that everything before
2127 is not part of it.
2128
2129 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2130
2131 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2132 reference syntax.
2133
2134 (g) (?| introduces a group in which the numbering of parentheses in each
2135 alternative starts with the same number.
2136
2137 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2138
2139 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2140 PCRE_INFO_JCHANGED.
2141
2142 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2143 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2144 for detecting groups that can match an empty string.
2145
2146 9. A pattern with a very large number of alternatives (more than several
2147 hundred) was running out of internal workspace during the pre-compile
2148 phase, where pcre_compile() figures out how much memory will be needed. A
2149 bit of new cunning has reduced the workspace needed for groups with
2150 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2151 workspace instead of running out of the 4096 that are available.
2152
2153 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2154
2155 11. Applied patch from Google to remove an optimization that didn't quite work.
2156 The report of the bug said:
2157
2158 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2159 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2160 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2161
2162 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2163 it matched the wrong number of bytes.
2164
2165
2166 Version 7.1 24-Apr-07
2167 ---------------------
2168
2169 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2170 that is more "standard", making use of automake and other Autotools. There
2171 is some re-arrangement of the files and adjustment of comments consequent
2172 on this.
2173
2174 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2175 for recursive directory scanning broke on some systems because the files
2176 are not scanned in any specific order and on different systems the order
2177 was different. A call to "sort" has been inserted into RunGrepTest for the
2178 approprate test as a short-term fix. In the longer term there may be an
2179 alternative.
2180
2181 3. I had an email from Eric Raymond about problems translating some of PCRE's
2182 man pages to HTML (despite the fact that I distribute HTML pages, some
2183 people do their own conversions for various reasons). The problems
2184 concerned the use of low-level troff macros .br and .in. I have therefore
2185 removed all such uses from the man pages (some were redundant, some could
2186 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2187 HTML has been updated to handle .nf/.fi and to complain if it encounters
2188 .br or .in.
2189
2190 4. Updated comments in configure.ac that get placed in config.h.in and also
2191 arranged for config.h to be included in the distribution, with the name
2192 config.h.generic, for the benefit of those who have to compile without
2193 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2194
2195 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2196 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2197 makevp.bat for the current PCRE, using the additional files
2198 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2199
2200 6. A Windows user reported a minor discrepancy with test 2, which turned out
2201 to be caused by a trailing space on an input line that had got lost in his
2202 copy. The trailing space was an accident, so I've just removed it.
2203
2204 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2205 that is needed.
2206
2207 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2208 as "const" (a) because they are and (b) because it helps the PHP
2209 maintainers who have recently made a script to detect big data structures
2210 in the php code that should be moved to the .rodata section. I remembered
2211 to update Builducptable as well, so it won't revert if ucptable.h is ever
2212 re-created.
2213
2214 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2215 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2216 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2217 support is not required. This saves 1.5-2K of code, which is important in
2218 some applications.
2219
2220 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2221 so as not to refer to the tables, even though these functions will never be
2222 called when UTF-8 support is disabled. Otherwise there are problems with a
2223 shared library.
2224
2225 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2226
2227 (a) It was defining its arguments as char * instead of void *.
2228
2229 (b) It was assuming that all moves were upwards in memory; this was true
2230 a long time ago when I wrote it, but is no longer the case.
2231
2232 The emulated memove() is provided for those environments that have neither
2233 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2234 is clearly not the case, as these two bugs were recently reported.
2235
2236 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2237 and Detrail to create the HTML documentation, the .txt form of the man
2238 pages, and it removes trailing spaces from listed files. It also creates
2239 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2240 case, it wraps all the #defines with #ifndefs. This script should be run
2241 before "make dist".
2242
2243 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2244 with Unicode property support.
2245
2246 (a) For a maximizing quantifier, if the two different cases of the
2247 character were of different lengths in their UTF-8 codings (there are
2248 some cases like this - I found 11), and the matching function had to
2249 back up over a mixture of the two cases, it incorrectly assumed they
2250 were both the same length.
2251
2252 (b) When PCRE was configured to use the heap rather than the stack for
2253 recursion during matching, it was not correctly preserving the data for
2254 the other case of a UTF-8 character when checking ahead for a match
2255 while processing a minimizing repeat. If the check also involved
2256 matching a wide character, but failed, corruption could cause an
2257 erroneous result when trying to check for a repeat of the original
2258 character.
2259
2260 13. Some tidying changes to the testing mechanism:
2261
2262 (a) The RunTest script now detects the internal link size and whether there
2263 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2264 values substituted by "configure". (The RunGrepTest script already did
2265 this for UTF-8.) The configure.ac script no longer substitutes the
2266 relevant variables.
2267
2268 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2269 with length and offset values. This means that the output is different
2270 for different internal link sizes. Test 2 is skipped for link sizes
2271 other than 2 because of this, bypassing the problem. Unfortunately,
2272 there was also a test in test 3 (the locale tests) that used /B and
2273 failed for link sizes other than 2. Rather than cut the whole test out,
2274 I have added a new /Z option to pcretest that replaces the length and
2275 offset values with spaces. This is now used to make test 3 independent
2276 of link size. (Test 2 will be tidied up later.)
2277
2278 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2279 segmentation fault instead of returning the appropriate error message.
2280
2281 15. In multiline mode when the newline sequence was set to "any", the pattern
2282 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2283 This doesn't seem right; it now treats the CRLF combination as the line
2284 ending, and so does not match in that case. It's only a pattern such as ^$
2285 that would hit this one: something like ^ABC$ would have failed after \r
2286 and then tried again after \r\n.
2287
2288 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2289 in an attempt to make files that differ only in their line terminators
2290 compare equal. This works on Linux.
2291
2292 17. Under certain error circumstances pcregrep might try to free random memory
2293 as it exited. This is now fixed, thanks to valgrind.
2294
2295 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2296 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2297 was because its rules for how to advance for /g after matching an empty
2298 string at the end of a line did not allow for this case. They now check for
2299 it specially.
2300
2301 20. pcretest is supposed to handle patterns and data of any length, by
2302 extending its buffers when necessary. It was getting this wrong when the
2303 buffer for a data line had to be extended.
2304
2305 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2306 CRLF as a newline sequence.
2307
2308 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2309 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2310 I have nevertheless tidied it up.
2311
2312 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2313
2314 24. Added a man page for pcre-config.
2315
2316
2317 Version 7.0 19-Dec-06
2318 ---------------------
2319
2320 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2321 moving to gcc 4.1.1.
2322
2323 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2324 sys/time.h, which is documented as needed for this function. It doesn't
2325 seem to matter on Linux, but it showed up on some releases of OS X.
2326
2327 3. It seems that there are systems where bytes whose values are greater than
2328 127 match isprint() in the "C" locale. The "C" locale should be the
2329 default when a C program starts up. In most systems, only ASCII printing
2330 characters match isprint(). This difference caused the output from pcretest
2331 to vary, making some of the tests fail. I have changed pcretest so that:
2332
2333 (a) When it is outputting text in the compiled version of a pattern, bytes
2334 other than 32-126 are always shown as hex escapes.
2335
2336 (b) When it is outputting text that is a matched part of a subject string,
2337 it does the same, unless a different locale has been set for the match
2338 (using the /L modifier). In this case, it uses isprint() to decide.
2339
2340 4. Fixed a major bug that caused incorrect computation of the amount of memory
2341 required for a compiled pattern when options that changed within the
2342 pattern affected the logic of the preliminary scan that determines the
2343 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2344 that the computed length was too small. The symptoms of this bug were
2345 either the PCRE error "internal error: code overflow" from pcre_compile(),
2346 or a glibc crash with a message such as "pcretest: free(): invalid next
2347 size (fast)". Examples of patterns that provoked this bug (shown in
2348 pcretest format) are:
2349
2350 /(?-x: )/x
2351 /(?x)(?-x: \s*#\s*)/
2352 /((?i)[\x{c0}])/8
2353 /(?i:[\x{c0}])/8
2354
2355 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2356 is now done differently.
2357
2358 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2359 wrapper classes; (b) implement a new function in the C++ scanner that is
2360 more efficient than the old way of doing things because it avoids levels of
2361 recursion in the regex matching; (c) add a paragraph to the documentation
2362 for the FullMatch() function.
2363
2364 6. The escape sequence \n was being treated as whatever was defined as
2365 "newline". Not only was this contrary to the documentation, which states
2366 that \n is character 10 (hex 0A), but it also went horribly wrong when
2367 "newline" was defined as CRLF. This has been fixed.
2368
2369 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2370 was being set to -1 for the "end of line" case (supposedly a value that no
2371 character can have). Though this value is never used (the check for end of
2372 line is "zero bytes in current character"), it caused compiler complaints.
2373 I've changed it to 0xffffffff.
2374
2375 8. In pcre_version.c, the version string was being built by a sequence of
2376 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2377 string (as it is for production releases) called a macro with an empty
2378 argument. The C standard says the result of this is undefined. The gcc
2379 compiler treats it as an empty string (which was what was wanted) but it is
2380 reported that Visual C gives an error. The source has been hacked around to
2381 avoid this problem.
2382
2383 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2384 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2385 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2386 of them did).
2387
2388 10. Originally, pcretest opened its input and output without "b"; then I was
2389 told that "b" was needed in some environments, so it was added for release
2390 5.0 to both the input and output. (It makes no difference on Unix-like
2391 systems.) Later I was told that it is wrong for the input on Windows. I've
2392 now abstracted the modes into two macros, to make it easier to fiddle with
2393 them, and removed "b" from the input mode under Windows.
2394
2395 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2396
2397 12. Added -help and --help to pcretest as an official way of being reminded
2398 of the options.
2399
2400 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2401 and pcrecpp.cc because they annoy compilers at high warning levels.
2402
2403 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2404
2405 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2406 scripts are not C programs :-) and which was not noticed because it works
2407 on Linux.
2408
2409 16. pcretest is supposed to handle any length of pattern and data line (as one
2410 line or as a continued sequence of lines) by extending its input buffer if
2411 necessary. This feature was broken for very long pattern lines, leading to
2412 a string of junk being passed to pcre_compile() if the pattern was longer
2413 than about 50K.
2414
2415 17. I have done a major re-factoring of the way pcre_compile() computes the
2416 amount of memory needed for a compiled pattern. Previously, there was code
2417 that made a preliminary scan of the pattern in order to do this. That was
2418 OK when PCRE was new, but as the facilities have expanded, it has become
2419 harder and harder to keep it in step with the real compile phase, and there
2420 have been a number of bugs (see for example, 4 above). I have now found a
2421 cunning way of running the real compile function in a "fake" mode that
2422 enables it to compute how much memory it would need, while actually only
2423 ever using a few hundred bytes of working memory and without too many
2424 tests of the mode. This should make future maintenance and development
2425 easier. A side effect of this work is that the limit of 200 on the nesting
2426 depth of parentheses has been removed (though this was never a serious
2427 limitation, I suspect). However, there is a downside: pcre_compile() now
2428 runs more slowly than before (30% or more, depending on the pattern). I
2429 hope this isn't a big issue. There is no effect on runtime performance.
2430
2431 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2432 newline (only possible for the last line of a file) and it was a
2433 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2434
2435 19. Added additional timing features to pcretest. (1) The -tm option now times
2436 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2437 separate command line item, by a number that specifies the number of
2438 repeats to use when timing. The default is 50000; this gives better
2439 precision, but takes uncomfortably long for very large patterns.
2440
2441 20. Extended pcre_study() to be more clever in cases where a branch of a
2442 subpattern has no definite first character. For example, (a*|b*)[cd] would
2443 previously give no result from pcre_study(). Now it recognizes that the
2444 first character must be a, b, c, or d.
2445
2446 21. There was an incorrect error "recursive call could loop indefinitely" if
2447 a subpattern (or the entire pattern) that was being tested for matching an
2448 empty string contained only one non-empty item after a nested subpattern.
2449 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2450 incorrectly, because the \d was being skipped in the check.
2451
2452 22. The pcretest program now has a new pattern option /B and a command line
2453 option -b, which is equivalent to adding /B to every pattern. This causes
2454 it to show the compiled bytecode, without the additional information that
2455 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2456 is the same as /B/I).
2457
2458 23. A new optimization is now able automatically to treat some sequences such
2459 as a*b as a*+b. More specifically, if something simple (such as a character
2460 or a simple class like \d) has an unlimited quantifier, and is followed by
2461 something that cannot possibly match the quantified thing, the quantifier
2462 is automatically "possessified".
2463
2464 24. A recursive reference to a subpattern whose number was greater than 39
2465 went wrong under certain circumstances in UTF-8 mode. This bug could also
2466 have affected the operation of pcre_study().
2467
2468 25. Realized that a little bit of performance could be had by replacing
2469 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2470
2471 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2472
2473 27. Possessive quantifiers such as a++ were previously implemented by turning
2474 them into atomic groups such as ($>a+). Now they have their own opcodes,
2475 which improves performance. This includes the automatically created ones
2476 from 23 above.
2477
2478 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2479 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2480 the first matched character to be a colon. This applied both to named and
2481 numbered groups.
2482
2483 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2484
2485 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2486 building PCRE on Windows easier, so I have included it in the distribution.
2487
2488 31. There is now a check in pcretest against a ridiculously large number being
2489 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2490 loop, the loop is abandoned.
2491
2492 32. Forward references to subpatterns in conditions such as (?(2)...) where
2493 subpattern 2 is defined later cause pcre_compile() to search forwards in
2494 the pattern for the relevant set of parentheses. This search went wrong
2495 when there were unescaped parentheses in a character class, parentheses
2496 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2497
2498 33. "Subroutine" calls and backreferences were previously restricted to
2499 referencing subpatterns earlier in the regex. This restriction has now
2500 been removed.
2501
2502 34. Added a number of extra features that are going to be in Perl 5.10. On the
2503 whole, these are just syntactic alternatives for features that PCRE had
2504 previously implemented using the Python syntax or my own invention. The
2505 other formats are all retained for compatibility.
2506
2507 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2508 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2509 also .NET compatible.
2510
2511 (b) A recursion or subroutine call to a named group can now be defined as
2512 (?&name) as well as (?P>name).
2513
2514 (c) A backreference to a named group can now be defined as \k<name> or
2515 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2516 5.10, are also .NET compatible.
2517
2518 (d) A conditional reference to a named group can now use the syntax
2519 (?(<name>) or (?('name') as well as (?(name).
2520
2521 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2522 groups (named and numbered) that are never evaluated inline, but can be
2523 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2524 is always false. There may be only one alternative in such a group.
2525
2526 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2527 as the simple (?(R). The condition is true only if the most recent
2528 recursion is that of the given number or name. It does not search out
2529 through the entire recursion stack.
2530
2531 (g) The escape \gN or \g{N} has been added, where N is a positive or
2532 negative number, specifying an absolute or relative reference.
2533
2534 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2535 some "unreachable code" warnings.
2536
2537 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2538 things, this adds five new scripts.
2539
2540 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2541 There were also incompatibilities regarding the handling of \Q..\E inside
2542 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2543 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2544
2545 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2546 matches an empty string, and forcibly breaks the loop. There were bugs in
2547 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2548 against aaaa the result was just "a" rather than "aaaa", for example. Two
2549 separate and independent bugs (that affected different cases) have been
2550 fixed.
2551
2552 39. Refactored the code to abolish the use of different opcodes for small
2553 capturing bracket numbers. This is a tidy that I avoided doing when I
2554 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2555 The new approach is not only tidier, it makes it possible to reduce the
2556 memory needed to fix the previous bug (38).
2557
2558 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2559 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2560 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2561 mode.
2562
2563 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2564 report.
2565
2566 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2567 copy construction and assignment in the C++ wrapper.
2568
2569 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2570 couple of bugs that could have given wrong results in the "--newline=crlf"
2571 case.
2572
2573 44. Added a number of casts and did some reorganization of signed/unsigned int
2574 variables following suggestions from Dair Grant. Also renamed the variable
2575 "this" as "item" because it is a C++ keyword.
2576
2577 45. Arranged for dftables to add
2578
2579 #include "pcre_internal.h"
2580
2581 to pcre_chartables.c because without it, gcc 4.x may remove the array
2582 definition from the final binary if PCRE is built into a static library and
2583 dead code stripping is activated.
2584
2585 46. For an unanchored pattern, if a match attempt fails at the start of a
2586 newline sequence, and the newline setting is CRLF or ANY, and the next two
2587 characters are CRLF, advance by two characters instead of one.
2588
2589
2590 Version 6.7 04-Jul-06
2591 ---------------------
2592
2593 1. In order to handle tests when input lines are enormously long, pcretest has
2594 been re-factored so that it automatically extends its buffers when
2595 necessary. The code is crude, but this _is_ just a test program. The
2596 default size has been increased from 32K to 50K.
2597
2598 2. The code in pcre_study() was using the value of the re argument before
2599 testing it for NULL. (Of course, in any sensible call of the function, it
2600 won't be NULL.)
2601
2602 3. The memmove() emulation function in pcre_internal.h, which is used on
2603 systems that lack both memmove() and bcopy() - that is, hardly ever -
2604 was missing a "static" storage class specifier.
2605
2606 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2607 containing an extended class (one that cannot be represented by a bitmap
2608 because it contains high-valued characters or Unicode property items, e.g.
2609 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2610 pattern, but PCRE should not loop if you do not (it no longer does).
2611 [Detail: two cases were found: (a) a repeated subpattern containing an
2612 extended class; (b) a recursive reference to a subpattern that followed a
2613 previous extended class. It wasn't skipping over the extended class
2614 correctly when UTF-8 mode was not set.]
2615
2616 5. A negated single-character class was not being recognized as fixed-length
2617 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2618 compile error "lookbehind assertion is not fixed length".
2619
2620 6. The RunPerlTest auxiliary script was showing an unexpected difference
2621 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2622 write a Perl script that can interpret lines of an input file either as
2623 byte characters or as UTF-8, which is what "perltest" was being required to
2624 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2625 can't do is switch easily at run time between having the "use utf8;" pragma
2626 or not. In the end, I fudged it by using the RunPerlTest script to insert
2627 "use utf8;" explicitly for the UTF-8 tests.
2628
2629 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2630 the end of the subject string, contrary to the documentation and to what
2631 Perl does. This was true of both matching functions. Now it matches only at
2632 the start of the subject and immediately after *internal* newlines.
2633
2634 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2635 a pointer to an int instead of a pointer to an unsigned long int. This
2636 caused problems on 64-bit systems.
2637
2638 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2639 instance of the 'standard' template library not being so standard".
2640
2641 10. There was no check on the number of named subpatterns nor the maximum
2642 length of a subpattern name. The product of these values is used to compute
2643 the size of the memory block for a compiled pattern. By supplying a very
2644 long subpattern name and a large number of named subpatterns, the size
2645 computation could be caused to overflow. This is now prevented by limiting
2646 the length of names to 32 characters, and the number of named subpatterns
2647 to 10,000.
2648
2649 11. Subpatterns that are repeated with specific counts have to be replicated in
2650 the compiled pattern. The size of memory for this was computed from the
2651 length of the subpattern and the repeat count. The latter is limited to
2652 65535, but there was no limit on the former, meaning that integer overflow
2653 could in principle occur. The compiled length of a repeated subpattern is
2654 now limited to 30,000 bytes in order to prevent this.
2655
2656 12. Added the optional facility to have named substrings with the same name.
2657
2658 13. Added the ability to use a named substring as a condition, using the
2659 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2660 are numbers (not recommended). Forward references are permitted.
2661
2662 14. Added forward references in named backreferences (if you see what I mean).
2663
2664 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2665 pattern could run off the end of the subject. For example, the pattern
2666 "(?s)(.{1,5})"8 did this with the subject "ab".
2667
2668 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2669 PCRE_CASELESS was set when matching characters that were quantified with ?
2670 or *.
2671
2672 17. A character class other than a single negated character that had a minimum
2673 but no maximum quantifier - for example [ab]{6,} - was not handled
2674 correctly by pce_dfa_exec(). It would match only one character.
2675
2676 18. A valid (though odd) pattern that looked like a POSIX character
2677 class but used an invalid character after [ (for example [[,abc,]]) caused
2678 pcre_compile() to give the error "Failed: internal error: code overflow" or
2679 in some cases to crash with a glibc free() error. This could even happen if
2680 the pattern terminated after [[ but there just happened to be a sequence of
2681 letters, a binary zero, and a closing ] in the memory that followed.
2682
2683 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2684 over the years. Originally (before any Unicode support), just the bottom 8
2685 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2686 output from "man perlunicode" includes this:
2687
2688 The regular expression compiler produces polymorphic opcodes. That
2689 is, the pattern adapts to the data and automatically switches to
2690 the Unicode character scheme when presented with Unicode data--or
2691 instead uses a traditional byte scheme when presented with byte
2692 data.
2693
2694 Sadly, a wide octal escape does not cause a switch, and in a string with
2695 no other multibyte characters, these octal escapes are treated as before.
2696 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2697 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2698 Unicode string.
2699
2700 I have not perpetrated such confusion in PCRE. Up till now, it took just
2701 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2702 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2703 translate to the appropriate multibyte character.
2704
2705 29. Applied some refactoring to reduce the number of warnings from Microsoft
2706 and Borland compilers. This has included removing the fudge introduced
2707 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2708 a warning about an unused variable.
2709
2710 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2711 characters since release 4.0, because Perl (from release 5.004) does not.
2712 [Or at least, is documented not to: some releases seem to be in conflict
2713 with the documentation.] However, when a pattern was studied with
2714 pcre_study() and all its branches started with \s, PCRE still included VT
2715 as a possible starting character. Of course, this did no harm; it just
2716 caused an unnecessary match attempt.
2717
2718 22. Removed a now-redundant internal flag bit that recorded the fact that case
2719 dependency changed within the pattern. This was once needed for "required
2720 byte" processing, but is no longer used. This recovers a now-scarce options
2721 bit. Also moved the least significant internal flag bit to the most-
2722 significant bit of the word, which was not previously used (hangover from
2723 the days when it was an int rather than a uint) to free up another bit for
2724 the future.
2725
2726 23. Added support for CRLF line endings as well as CR and LF. As well as the
2727 default being selectable at build time, it can now be changed at runtime
2728 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2729 specify that it is scanning data with non-default line endings.
2730
2731 24. Changed the definition of CXXLINK to make it agree with the definition of
2732 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2733
2734 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2735 recursions. This makes a big different to stack usage for some patterns.
2736
2737 26. If a subpattern containing a named recursion or subroutine reference such
2738 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2739 the space required for the compiled pattern went wrong and gave too small a
2740 value. Depending on the environment, this could lead to "Failed: internal
2741 error: code overflow at offset 49" or "glibc detected double free or
2742 corruption" errors.
2743
2744 27. Applied patches from Google (a) to support the new newline modes and (b) to
2745 advance over multibyte UTF-8 characters in GlobalReplace.
2746
2747 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2748 difference for some implementation of PCRE in some Windows version.
2749
2750 29. Added some extra testing facilities to pcretest:
2751
2752 \q<number> in a data line sets the "match limit" value
2753 \Q<number> in a data line sets the "match recursion limt" value
2754 -S <number> sets the stack size, where <number> is in megabytes
2755
2756 The -S option isn't available for Windows.
2757
2758
2759 Version 6.6 06-Feb-06
2760 ---------------------
2761
2762 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2763 in pcreposix.h. I have copied the definition from pcre.h.
2764
2765 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2766 because pcre.h is no longer a built file.
2767
2768 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2769 not normally included in the compiled code.
2770
2771
2772 Version 6.5 01-Feb-06
2773 ---------------------
2774
2775 1. When using the partial match feature with pcre_dfa_exec(), it was not
2776 anchoring the second and subsequent partial matches at the new starting
2777 point. This could lead to incorrect results. For example, with the pattern
2778 /1234/, partially matching against "123" and then "a4" gave a match.
2779
2780 2. Changes to pcregrep:
2781
2782 (a) All non-match returns from pcre_exec() were being treated as failures
2783 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2784 error message is output. Some extra information is given for the
2785 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2786 probably the only errors that are likely to be caused by users (by
2787 specifying a regex that has nested indefinite repeats, for instance).
2788 If there are more than 20 of these errors, pcregrep is abandoned.
2789
2790 (b) A binary zero was treated as data while matching, but terminated the
2791 output line if it was written out. This has been fixed: binary zeroes
2792 are now no different to any other data bytes.
2793
2794 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2795 used to set a locale for matching. The --locale=xxxx long option has
2796 been added (no short equivalent) to specify a locale explicitly on the
2797 pcregrep command, overriding the environment variables.
2798
2799 (d) When -B was used with -n, some line numbers in the output were one less
2800 than they should have been.
2801
2802 (e) Added the -o (--only-matching) option.
2803
2804 (f) If -A or -C was used with -c (count only), some lines of context were
2805 accidentally printed for the final match.
2806
2807 (g) Added the -H (--with-filename) option.
2808
2809 (h) The combination of options -rh failed to suppress file names for files
2810 that were found from directory arguments.
2811
2812 (i) Added the -D (--devices) and -d (--directories) options.
2813
2814 (j) Added the -F (--fixed-strings) option.
2815
2816 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2817
2818 (l) Added the --colo(u)r option.
2819
2820 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2821 is not present by default.
2822
2823 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2824 items such as (?R) or (?1), when the recursion could match a number of
2825 alternatives. If it matched one of the alternatives, but subsequently,
2826 outside the recursion, there was a failure, the code tried to back up into
2827 the recursion. However, because of the way PCRE is implemented, this is not
2828 possible, and the result was an incorrect result from the match.
2829
2830 In order to prevent this happening, the specification of recursion has
2831 been changed so that all such subpatterns are automatically treated as
2832 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2833
2834 4. I had overlooked the fact that, in some locales, there are characters for
2835 which isalpha() is true but neither isupper() nor islower() are true. In
2836 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2837 and ordfeminine) are like this. This affected the treatment of \w and \W
2838 when they appeared in character classes, but not when they appeared outside
2839 a character class. The bit map for "word" characters is now created
2840 separately from the results of isalnum() instead of just taking it from the
2841 upper, lower, and digit maps. (Plus the underscore character, of course.)
2842
2843 5. The above bug also affected the handling of POSIX character classes such as
2844 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2845 permanent tables. Instead, the bit maps for such a class were previously
2846 created as the appropriate unions of the upper, lower, and digit bitmaps.
2847 Now they are created by subtraction from the [[:word:]] class, which has
2848 its own bitmap.
2849
2850 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2851 It is created by subtracting the vertical space characters (\x09, \x0a,
2852 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2853 subtraction was done in the overall bitmap for a character class, meaning
2854 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2855 be recognized. This bug has been fixed.
2856
2857 7. Patches from the folks at Google:
2858
2859 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2860 real life, but is still worth protecting against".
2861
2862 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2863 regular expressions".
2864
2865 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2866 have it.
2867
2868 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2869 "configure" and the latter not, in order to fix a problem somebody had
2870 with compiling the Arg class on HP-UX.
2871
2872 (e) Improve the error-handling of the C++ wrapper a little bit.
2873
2874 (f) New tests for checking recursion limiting.
2875
2876 8. The pcre_memmove() function, which is used only if the environment does not
2877 have a standard memmove() function (and is therefore rarely compiled),
2878 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2879 returning a result (though PCRE never actually uses the result).
2880
2881 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2882 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2883 returned instead of calling malloc() with an overflowing number that would
2884 most likely cause subsequent chaos.
2885
2886 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2887
2888 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2889 with this option is matched, the nmatch and pmatch options of regexec() are
2890 ignored.
2891
2892 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2893 provided in case anyone wants to the the POSIX interface with UTF-8
2894 strings.
2895
2896 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2897 C++ linking (needed for some HP-UX environments).
2898
2899 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2900 (unused parameter) and in the pcre_printint() function (omitted "default"
2901 switch label when the default is to do nothing).
2902
2903 15. Added some code to make it possible, when PCRE is compiled as a C++
2904 library, to replace subject pointers for pcre_exec() with a smart pointer
2905 class, thus making it possible to process discontinuous strings.
2906
2907 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2908 much the same function. They were added by different people who were trying
2909 to make PCRE easy to compile on non-Unix systems. It has been suggested
2910 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2911 for compiling on Windows systems. I have therefore replaced it with
2912 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2913 defaults to "extern" for C or "extern C" for C++, which works fine on
2914 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2915 SCOPE with something explicit in config.h. In addition:
2916
2917 (a) pcreposix.h still had just "extern" instead of either of these macros;
2918 I have replaced it with PCRE_DATA_SCOPE.
2919
2920 (b) Functions such as _pcre_xclass(), which are internal to the library,
2921 but external in the C sense, all had PCRE_EXPORT in their definitions.
2922 This is apparently wrong for the Windows case, so I have removed it.
2923 (It makes no difference on Unix-like systems.)
2924
2925 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2926 of recursive calls to match(). This is different to MATCH_LIMIT because
2927 that limits the total number of calls to match(), not all of which increase
2928 the depth of recursion. Limiting the recursion depth limits the amount of
2929 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2930 when PCRE is compiled, and changed at run time. A patch from Google adds
2931 this functionality to the C++ interface.
2932
2933 18. Changes to the handling of Unicode character properties:
2934
2935 (a) Updated the table to Unicode 4.1.0.
2936
2937 (b) Recognize characters that are not in the table as "Cn" (undefined).
2938
2939 (c) I revised the way the table is implemented to a much improved format
2940 which includes recognition of ranges. It now supports the ranges that
2941 are defined in UnicodeData.txt, and it also amalgamates other
2942 characters into ranges. This has reduced the number of entries in the
2943 table from around 16,000 to around 3,000, thus reducing its size
2944 considerably. I realized I did not need to use a tree structure after
2945 all - a binary chop search is just as efficient. Having reduced the
2946 number of entries, I extended their size from 6 bytes to 8 bytes to
2947 allow for more data.
2948
2949 (d) Added support for Unicode script names via properties such as \p{Han}.
2950
2951 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2952 matching that character.
2953
2954 20. When matching a repeated Unicode property with a minimum greater than zero,
2955 (for example \pL{2,}), PCRE could look past the end of the subject if it
2956 reached it while seeking the minimum number of characters. This could
2957 happen only if some of the characters were more than one byte long, because
2958 there is a check for at least the minimum number of bytes.
2959
2960 21. Refactored the implementation of \p and \P so as to be more general, to
2961 allow for more different types of property in future. This has changed the
2962 compiled form incompatibly. Anybody with saved compiled patterns that use
2963 \p or \P will have to recompile them.
2964
2965 22. Added "Any" and "L&" to the supported property types.
2966
2967 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2968 but give a compile time error if the value is greater than 0xff.
2969
2970 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2971 accidentally not being installed or uninstalled.
2972
2973 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2974 made were to insert the current release number. This seemed silly, because
2975 it made things harder for people building PCRE on systems that don't run
2976 "configure". I have turned pcre.h into a distributed file, no longer built
2977 by "configure", with the version identification directly included. There is
2978 no longer a pcre.h.in file.
2979
2980 However, this change necessitated a change to the pcre-config script as
2981 well. It is built from pcre-config.in, and one of the substitutions was the
2982 release number. I have updated configure.ac so that ./configure now finds
2983 the release number by grepping pcre.h.
2984
2985 26. Added the ability to run the tests under valgrind.
2986
2987
2988 Version 6.4 05-Sep-05
2989 ---------------------
2990
2991 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2992 "--" to be printed when multiple files were scanned, even when none of the
2993 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2994 consider it to be a bug, and have restored the previous behaviour.
2995
2996 2. A couple of code tidies to get rid of compiler warnings.
2997
2998 3. The pcretest program used to cheat by referring to symbols in the library
2999 whose names begin with _pcre_. These are internal symbols that are not
3000 really supposed to be visible externally, and in some environments it is
3001 possible to suppress them. The cheating is now confined to including
3002 certain files from the library's source, which is a bit cleaner.
3003
3004 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3005 file's purpose clearer.
3006
3007 5. Reorganized pcre_ucp_findchar().
3008
3009
3010 Version 6.3 15-Aug-05
3011 ---------------------
3012
3013 1. The file libpcre.pc.in did not have general read permission in the tarball.
3014
3015 2. There were some problems when building without C++ support:
3016
3017 (a) If C++ support was not built, "make install" and "make test" still
3018 tried to test it.
3019
3020 (b) There were problems when the value of CXX was explicitly set. Some
3021 changes have been made to try to fix these, and ...
3022
3023 (c) --disable-cpp can now be used to explicitly disable C++ support.
3024
3025 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3026 backslash in a target when C++ was disabled. This confuses some
3027 versions of "make", apparently. Using an intermediate variable solves
3028 this. (Same for CPP_LOBJ.)
3029
3030 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3031 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3032 necessary on certain architectures.
3033
3034 4. Added a setting of -export-symbols-regex to the link command to remove
3035 those symbols that are exported in the C sense, but actually are local
3036 within the library, and not documented. Their names all begin with
3037 "_pcre_". This is not a perfect job, because (a) we have to except some
3038 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3039 available (and never for static libraries). I have made a note to try to
3040 find a way round (a) in the future.
3041
3042
3043 Version 6.2 01-Aug-05
3044 ---------------------
3045
3046 1. There was no test for integer overflow of quantifier values. A construction
3047 such as {1111111111111111} would give undefined results. What is worse, if
3048 a minimum quantifier for a parenthesized subpattern overflowed and became
3049 negative, the calculation of the memory size went wrong. This could have
3050 led to memory overwriting.
3051
3052 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3053
3054 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3055 operating environments where this matters.
3056
3057 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3058 PCRE options from within the C++ wrapper.
3059
3060 5. Named capturing subpatterns were not being correctly counted when a pattern
3061 was compiled. This caused two problems: (a) If there were more than 100
3062 such subpatterns, the calculation of the memory needed for the whole
3063 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3064 back references of the form \12, where the number was greater than 9, were
3065 not recognized as back references, even though there were sufficient
3066 previous subpatterns.
3067
3068 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3069 versions of gcc, e.g. 2.95.4.
3070
3071
3072 Version 6.1 21-Jun-05
3073 ---------------------
3074
3075 1. There was one reference to the variable "posix" in pcretest.c that was not
3076 surrounded by "#if !defined NOPOSIX".
3077
3078 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3079 the cross-check on the old pcre_info() function, for the benefit of the
3080 cut-down version of PCRE that is currently imported into Exim.
3081
3082 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3083 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3084 patterns that start (?i) but I don't think that matters. The use of (?i) is
3085 just an example; this all applies to the other options as well.
3086
3087 4. Since libtool seems to echo the compile commands it is issuing, the output
3088 from "make" can be reduced a bit by putting "@" in front of each libtool
3089 compile command.
3090
3091 5. Patch from the folks at Google for configure.in to be a bit more thorough
3092 in checking for a suitable C++ installation before trying to compile the
3093 C++ stuff. This should fix a reported problem when a compiler was present,
3094 but no suitable headers.
3095
3096 6. The man pages all had just "PCRE" as their title. I have changed them to
3097 be the relevant file name. I have also arranged that these names are
3098 retained in the file doc/pcre.txt, which is a concatenation in text format
3099 of all the man pages except the little individual ones for each function.
3100
3101 7. The NON-UNIX-USE file had not been updated for the different set of source
3102 files that come with release 6. I also added a few comments about the C++
3103 wrapper.
3104
3105
3106 Version 6.0 07-Jun-05
3107 ---------------------
3108
3109 1. Some minor internal re-organization to help with my DFA experiments.
3110
3111 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3112 didn't matter for the library itself when fully configured, but did matter
3113 when compiling without UCP support, or within Exim, where the ucp files are
3114 not imported.
3115
3116 3. Refactoring of the library code to split up the various functions into
3117 different source modules. The addition of the new DFA matching code (see
3118 below) to a single monolithic source would have made it really too
3119 unwieldy, quite apart from causing all the code to be include in a
3120 statically linked application, when only some functions are used. This is
3121 relevant even without the DFA addition now that patterns can be compiled in
3122 one application and matched in another.
3123
3124 The downside of splitting up is that there have to be some external
3125 functions and data tables that are used internally in different modules of
3126 the library but which are not part of the API. These have all had their
3127 names changed to start with "_pcre_" so that they are unlikely to clash
3128 with other external names.
3129
3130 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3131 a different (DFA) algorithm. Although it is slower than the original
3132 function, it does have some advantages for certain types of matching
3133 problem.
3134
3135 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3136 including restarting after a partial match.
3137
3138 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3139 defined when compiling for Windows was sent to me. I have put it into the
3140 code, though I have no means of testing or verifying it.
3141
3142 7. Added the pcre_refcount() auxiliary function.
3143
3144 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3145 match before or at the first newline in the subject string. In pcretest,
3146 the /f option on a pattern can be used to set this.
3147
3148 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3149 would behave wrongly. This has been present in PCRE since release 4.0.
3150
3151 10. A number of changes to the pcregrep command:
3152
3153 (a) Refactored how -x works; insert ^(...)$ instead of setting
3154 PCRE_ANCHORED and checking the length, in preparation for adding
3155 something similar for -w.
3156
3157 (b) Added the -w (match as a word) option.
3158
3159 (c) Refactored the way lines are read and buffered so as to have more
3160 than one at a time available.
3161
3162 (d) Implemented a pcregrep test script.
3163
3164 (e) Added the -M (multiline match) option. This allows patterns to match
3165 over several lines of the subject. The buffering ensures that at least
3166 8K, or the rest of the document (whichever is the shorter) is available
3167 for matching (and similarly the previous 8K for lookbehind assertions).
3168
3169 (f) Changed the --help output so that it now says
3170
3171 -w, --word-regex(p)
3172
3173 instead of two lines, one with "regex" and the other with "regexp"
3174 because that confused at least one person since the short forms are the
3175 same. (This required a bit of code, as the output is generated
3176 automatically from a table. It wasn't just a text change.)
3177
3178 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3179 option but starts with a hyphen. Could be a pattern or a path name
3180 starting with a hyphen, for instance.
3181
3182 (h) "-" can be given as a file name to represent stdin.
3183
3184 (i) When file names are being printed, "(standard input)" is used for
3185 the standard input, for compatibility with GNU grep. Previously
3186 "<stdin>" was used.
3187
3188 (j) The option --label=xxx can be used to supply a name to be used for
3189 stdin when file names are being printed. There is no short form.
3190
3191 (k) Re-factored the options decoding logic because we are going to add
3192 two more options that take data. Such options can now be given in four
3193 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3194
3195 (l) Added the -A, -B, and -C options for requesting that lines of context
3196 around matches be printed.
3197
3198 (m) Added the -L option to print the names of files that do not contain
3199 any matching lines, that is, the complement of -l.
3200
3201 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3202 continue to scan other files.
3203
3204 (o) The -s option was incorrectly implemented. For compatibility with other
3205 greps, it now suppresses the error message for a non-existent or non-
3206 accessible file (but not the return code). There is a new option called
3207 -q that suppresses the output of matching lines, which was what -s was
3208 previously doing.
3209
3210 (p) Added --include and --exclude options to specify files for inclusion
3211 and exclusion when recursing.
3212
3213 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3214 Hopefully, it now does.
3215
3216 12. Missing cast in pcre_study().
3217
3218 13. Added an "uninstall" target to the makefile.
3219
3220 14. Replaced "extern" in the function prototypes in Makefile.in with
3221 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3222 world, but is set differently for Windows.
3223
3224 15. Added a second compiling function called pcre_compile2(). The only
3225 difference is that it has an extra argument, which is a pointer to an
3226 integer error code. When there is a compile-time failure, this is set
3227 non-zero, in addition to the error test pointer being set to point to an
3228 error message. The new argument may be NULL if no error number is required
3229 (but then you may as well call pcre_compile(), which is now just a
3230 wrapper). This facility is provided because some applications need a
3231 numeric error indication, but it has also enabled me to tidy up the way
3232 compile-time errors are handled in the POSIX wrapper.
3233
3234 16. Added VPATH=.libs to the makefile; this should help when building with one
3235 prefix path and installing with another. (Or so I'm told by someone who
3236 knows more about this stuff than I do.)
3237
3238 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3239 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3240 match everything, including newlines. This is not POSIX-compatible, but
3241 somebody wanted the feature. From pcretest it can be activated by using
3242 both the P and the s flags.
3243
3244 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3245
3246 19. libpcre.pc was being incorrectly installed as executable.
3247
3248 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3249 it now also looks for '\r' so that it will work unmodified on Windows.
3250
3251 21. Added Google's contributed C++ wrapper to the distribution.
3252
3253 22. Added some untidy missing memory free() calls in pcretest, to keep
3254 Electric Fence happy when testing.
3255
3256
3257
3258 Version 5.0 13-Sep-04
3259 ---------------------
3260
3261 1. Internal change: literal characters are no longer packed up into items
3262 containing multiple characters in a single byte-string. Each character
3263 is now matched using a separate opcode. However, there may be more than one
3264 byte in the character in UTF-8 mode.
3265
3266 2. The pcre_callout_block structure has two new fields: pattern_position and
3267 next_item_length. These contain the offset in the pattern to the next match
3268 item, and its length, respectively.
3269
3270 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3271 insertion of callouts before each pattern item. Added the /C option to
3272 pcretest to make use of this.
3273
3274 4. On the advice of a Windows user, the lines
3275
3276 #if defined(_WIN32) || defined(WIN32)
3277 _setmode( _fileno( stdout ), 0x8000 );
3278 #endif /* defined(_WIN32) || defined(WIN32) */
3279
3280 have been added to the source of pcretest. This apparently does useful
3281 magic in relation to line terminators.
3282
3283 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3284 for the benefit of those environments where the "b" makes a difference.
3285
3286 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3287 to know about it. I have put a hack into configure.in that adds in code
3288 to set GCC=yes if CC=icc. This seems to end up at a point in the
3289 generated configure script that is early enough to affect the setting of
3290 compiler options, which is what is needed, but I have no means of testing
3291 whether it really works. (The user who reported this had patched the
3292 generated configure script, which of course I cannot do.)
3293
3294 LATER: After change 22 below (new libtool files), the configure script
3295 seems to know about icc (and also ecc). Therefore, I have commented out
3296 this hack in configure.in.
3297
3298 7. Added support for pkg-config (2 patches were sent in).
3299
3300 8. Negated POSIX character classes that used a combination of internal tables
3301 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3302 [[:^ascii]]. Typically, they would match almost any characters. The other
3303 POSIX classes were not broken in this way.
3304
3305 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3306 to find the match, as PCRE was deluded into thinking that the match had to
3307 start at the start point or following a newline. The same bug applied to
3308 patterns with negative forward assertions or any backward assertions
3309 preceding ".*" at the start, unless the pattern required a fixed first
3310 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3311
3312 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3313 starting at the last subject character, bytes beyond the end of the subject
3314 string were read.
3315
3316 11. Renamed the variable "class" as "classbits" to make life easier for C++
3317 users. (Previously there was a macro definition, but it apparently wasn't
3318 enough.)
3319
3320 12. Added the new field "tables" to the extra data so that tables can be passed
3321 in at exec time, or the internal tables can be re-selected. This allows
3322 a compiled regex to be saved and re-used at a later time by a different
3323 program that might have everything at different addresses.
3324
3325 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3326 -R library as well as a -L library.
3327
3328 14. The debugging options of pcretest (-d on the command line or D on a
3329 pattern) showed incorrect output for anything following an extended class
3330 that contained multibyte characters and which was followed by a quantifier.
3331
3332 15. Added optional support for general category Unicode character properties
3333 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3334 support. It adds about 90K to the size of the library. The meanings of the
3335 inbuilt class escapes such as \d and \s have NOT been changed.
3336
3337 16. Updated pcredemo.c to include calls to free() to release the memory for the
3338 compiled pattern.
3339
3340 17. The generated file chartables.c was being created in the source directory
3341 instead of in the building directory. This caused the build to fail if the
3342 source directory was different from the building directory, and was
3343 read-only.
3344
3345 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3346 file. No doubt somebody will tell me if they don't make sense... Also added
3347 Dan Mooney's comments about building on OpenVMS.
3348
3349 19. Added support for partial matching via the PCRE_PARTIAL option for
3350 pcre_exec() and the \P data escape in pcretest.
3351
3352 20. Extended pcretest with 3 new pattern features:
3353
3354 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3355 write the compiled pattern to the file whose name is "rest-of-line".
3356 This is a straight binary dump of the data, with the saved pointer to
3357 the character tables forced to be NULL. The study data, if any, is
3358 written too. After writing, pcretest reads a new pattern.
3359
3360 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3361 compiled pattern from the given file. There must not be any
3362 occurrences of "<" in the file name (pretty unlikely); if there are,
3363 pcretest will instead treat the initial "<" as a pattern delimiter.
3364 After reading in the pattern, pcretest goes on to read data lines as
3365 usual.
3366
3367 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3368 and 16-bit fields in a compiled pattern, to simulate a pattern that
3369 was compiled on a host of opposite endianness.
3370
3371 21. The pcre-exec() function can now cope with patterns that were compiled on
3372 hosts of opposite endianness, with this restriction:
3373
3374 As for any compiled expression that is saved and used later, the tables
3375 pointer field cannot be preserved; the extra_data field in the arguments
3376 to pcre_exec() should be used to pass in a tables address if a value
3377 other than the default internal tables were used at compile time.
3378
3379 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3380 now diagnosed as an error. Previously, most of the time, a negative number
3381 would have been treated as zero, but if in addition "ovector" was passed as
3382 NULL, a crash could occur.
3383
3384 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3385 new versions from the libtool 1.5 distribution (the last one is a copy of
3386 a file called libtool.m4). This seems to have fixed the need to patch
3387 "configure" to support Darwin 1.3 (which I used to do). However, I still
3388 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3389 workstation).
3390
3391 24. Changed the PCRE licence to be the more standard "BSD" licence.
3392
3393
3394 Version 4.5 01-Dec-03
3395 ---------------------
3396
3397 1. There has been some re-arrangement of the code for the match() function so
3398 that it can be compiled in a version that does not call itself recursively.
3399 Instead, it keeps those local variables that need separate instances for
3400 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3401 needs to "recurse". Keeping track of where control must go is done by means
3402 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3403 hide most of the details from the main code, and operates only if
3404 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3405 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3406 operating.
3407
3408 To make it easier for callers to provide specially tailored get/free
3409 functions for this usage, two new functions, pcre_stack_malloc, and
3410 pcre_stack_free, are used. They are always called in strict stacking order,
3411 and the size of block requested is always the same.
3412
3413 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3414 PCRE has been compiled to use the stack or the heap for recursion. The
3415 -C option of pcretest uses this to show which version is compiled.
3416
3417 A new data escape \S, is added to pcretest; it causes the amounts of store
3418 obtained and freed by both kinds of malloc/free at match time to be added
3419 to the output.
3420
3421 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3422 what's available on my current Linux desktop machine.
3423
3424 3. When matching a UTF-8 string, the test for a valid string at the start has
3425 been extended. If start_offset is not zero, PCRE now checks that it points
3426 to a byte that is the start of a UTF-8 character. If not, it returns
3427 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3428 this is necessary because there may be backward assertions in the pattern.
3429 When matching the same subject several times, it may save resources to use
3430 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3431
3432 4. The code for checking the validity of UTF-8 strings has been tightened so
3433 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3434 containing "overlong sequences".
3435
3436 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3437 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3438 should have been "&", but it just so happened that all the cases this let
3439 through by mistake were picked up later in the function.
3440
3441 6. I had used a variable called "isblank" - this is a C99 function, causing
3442 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3443
3444 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3445 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3446 so the version is shown; (c) stop "make test" echoing "./RunTest".
3447
3448 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3449
3450 9. The prototype for memmove() for systems that don't have it was using
3451 size_t, but the inclusion of the header that defines size_t was later. I've
3452 moved the #includes for the C headers earlier to avoid this.
3453
3454 10. Added some adjustments to the code to make it easier to compiler on certain
3455 special systems:
3456
3457 (a) Some "const" qualifiers were missing.
3458 (b) Added the macro EXPORT before all exported functions; by default this
3459 is defined to be empty.
3460 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3461 that it reads its output file name as an argument instead of writing
3462 to the standard output and assuming this can be redirected.
3463
3464 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3465 class containing characters with values greater than 255, PCRE compilation
3466 went into a loop.
3467
3468 12. A recursive reference to a subpattern that was within another subpattern
3469 that had a minimum quantifier of zero caused PCRE to crash. For example,
3470 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3471 recursion. If the recursively-called subpattern itself had a zero repeat,
3472 that was OK.
3473
3474 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3475 buffer into which it was copied (for escape processing) was still set at
3476 1024, so long lines caused crashes.
3477
3478 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3479 "internal error: code overflow...". This applied to any character class
3480 that was followed by a possessive quantifier.
3481
3482 15. Modified the Makefile to add libpcre.la as a prerequisite for
3483 libpcreposix.la because I was told this is needed for a parallel build to
3484 work.
3485
3486 16. If a pattern that contained .* following optional items at the start was
3487 studied, the wrong optimizing data was generated, leading to matching
3488 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3489 matching string must start with a or b or c. The correct conclusion for
3490 this pattern is that a match can start with any character.
3491
3492
3493 Version 4.4 13-Aug-03
3494 ---------------------
3495
3496 1. In UTF-8 mode, a character class containing characters with values between
3497 127 and 255 was not handled correctly if the compiled pattern was studied.
3498 In fixing this, I have also improved the studying algorithm for such
3499 classes (slightly).
3500
3501 2. Three internal functions had redundant arguments passed to them. Removal
3502 might give a very teeny performance improvement.
3503
3504 3. Documentation bug: the value of the capture_top field in a callout is *one
3505 more than* the number of the hightest numbered captured substring.
3506
3507 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3508 in incorrectly linking with a previously installed version. They now link
3509 explicitly with libpcre.la.
3510
3511 5. configure.in no longer needs to recognize Cygwin specially.
3512
3513 6. A problem in pcre.in for Windows platforms is fixed.
3514
3515 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3516 pcretest, it used to include the size of the study block as part of its
3517 output. Unfortunately, the structure contains a field that has a different
3518 size on different hardware architectures. This meant that the tests that
3519 showed this size failed. As the block is currently always of a fixed size,
3520 this information isn't actually particularly useful in pcretest output, so
3521 I have just removed it.
3522
3523 8. Three pre-processor statements accidentally did not start in column 1.
3524 Sadly, there are *still* compilers around that complain, even though
3525 standard C has not required this for well over a decade. Sigh.
3526
3527 9. In pcretest, the code for checking callouts passed small integers in the
3528 callout_data field, which is a void * field. However, some picky compilers
3529 complained about the casts involved for this on 64-bit systems. Now
3530 pcretest passes the address of the small integer instead, which should get
3531 rid of the warnings.
3532
3533 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3534 both compile and run time, and gives an error if an invalid UTF-8 sequence
3535 is found. There is a option for disabling this check in cases where the
3536 string is known to be correct and/or the maximum performance is wanted.
3537
3538 11. In response to a bug report, I changed one line in Makefile.in from
3539
3540 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3541 to
3542 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3543
3544 to look similar to other lines, but I have no way of telling whether this
3545 is the right thing to do, as I do not use Windows. No doubt I'll get told
3546 if it's wrong...
3547
3548
3549 Version 4.3 21-May-03
3550 ---------------------
3551
3552 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3553 Makefile.
3554
3555 2. Some refactoring to improve the quality of the code:
3556
3557 (i) The utf8_table... variables are now declared "const".
3558
3559 (ii) The code for \cx, which used the "case flipping" table to upper case
3560 lower case letters, now just substracts 32. This is ASCII-specific,
3561 but the whole concept of \cx is ASCII-specific, so it seems
3562 reasonable.
3563
3564 (iii) PCRE was using its character types table to recognize decimal and
3565 hexadecimal digits in the pattern. This is silly, because it handles
3566 only 0-9, a-f, and A-F, but the character types table is locale-
3567 specific, which means strange things might happen. A private
3568 table is now used for this - though it costs 256 bytes, a table is
3569 much faster than multiple explicit tests. Of course, the standard
3570 character types table is still used for matching digits in subject
3571 strings against \d.
3572
3573 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3574 ending in _t are). So I've renamed it as ESC_tee.
3575
3576 3. The first argument for regexec() in the POSIX wrapper should have been
3577 defined as "const".
3578
3579 4. Changed pcretest to use malloc() for its buffers so that they can be
3580 Electric Fenced for debugging.
3581
3582 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3583 to read one or more bytes before the start of the subject string. Often this
3584 had no effect on PCRE's behaviour, but in some circumstances it could
3585 provoke a segmentation fault.
3586
3587 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3588 to try to read one or more bytes before the start of the subject string.
3589
3590 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3591 UTF-8 support could misbehave in various ways if the subject string
3592 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3593 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3594 back over UTF-8 characters.)
3595
3596
3597 Version 4.2 14-Apr-03
3598 ---------------------
3599
3600 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3601
3602 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3603 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3604 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3605 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3606 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3607 and BUILD_EXEEXT
3608 Note: automatic setting of the BUILD variables is not yet working
3609 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3610 compile-time but not at link-time
3611 [LINK]: use for linking executables only
3612 make different versions for Windows and non-Windows
3613 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3614 libraries
3615 [LINK_FOR_BUILD]: new variable
3616 [OBJEXT]: use throughout
3617 [EXEEXT]: use throughout
3618 <winshared>: new target
3619 <wininstall>: new target
3620 <dftables.o>: use native compiler
3621 <dftables>: use native linker
3622 <install>: handle Windows platform correctly
3623 <clean>: ditto
3624 <check>: ditto
3625 copy DLL to top builddir before testing
3626
3627 As part of these changes, -no-undefined was removed again. This was reported
3628 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3629 in any case.
3630
3631 3. Some tidies to get rid of compiler warnings:
3632
3633 . In the match_data structure, match_limit was an unsigned long int, whereas
3634 match_call_count was an int. I've made them both unsigned long ints.
3635
3636 . In pcretest the fact that a const uschar * doesn't automatically cast to
3637 a void * provoked a warning.
3638
3639 . Turning on some more compiler warnings threw up some "shadow" variables
3640 and a few more missing casts.
3641
3642 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3643 option, a class that contained a single character with a value between 128
3644 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3645
3646 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3647 option, a class that contained several characters, but with at least one
3648 whose value was between 128 and 255 caused PCRE to crash.
3649
3650
3651 Version 4.1 12-Mar-03
3652 ---------------------
3653
3654 1. Compiling with gcc -pedantic found a couple of places where casts were
3655 needed, and a string in dftables.c that was longer than standard compilers are
3656 required to support.
3657
3658 2. Compiling with Sun's compiler found a few more places where the code could
3659 be tidied up in order to avoid warnings.
3660
3661 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3662 first of these names is deprecated in the latest Autoconf in favour of the name
3663 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3664 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3665 analogy I have changed it to CFLAGS_FOR_BUILD.
3666
3667 4. Added -no-undefined to the linking command in the Makefile, because this is
3668 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3669 linking step for the pcreposix library.
3670
3671 5. PCRE was failing to diagnose the case of two named groups with the same
3672 name.
3673
3674 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3675 literal character that is needed in the subject for a match, and scans along to
3676 ensure that it is present before embarking on the full matching process. This
3677 saves time in cases of nested unlimited repeats that are never going to match.
3678 Problem: the scan can take a lot of time if the subject is very long (e.g.
3679 megabytes), thus penalizing straightforward matches. It is now done only if the
3680 amount of subject to be scanned is less than 1000 bytes.
3681
3682 7. A lesser problem with the same optimization is that it was recording the
3683 first character of an anchored pattern as "needed", thus provoking a search
3684 right along the subject, even when the first match of the pattern was going to
3685 fail. The "needed" character is now not set for anchored patterns, unless it
3686 follows something in the pattern that is of non-fixed length. Thus, it still
3687 fulfils its original purpose of finding quick non-matches in cases of nested
3688 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3689
3690
3691 Version 4.0 17-Feb-03
3692 ---------------------
3693
3694 1. If a comment in an extended regex that started immediately after a meta-item
3695 extended to the end of string, PCRE compiled incorrect data. This could lead to
3696 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3697
3698 2. Moved to autoconf 2.53 and libtool 1.4.2.
3699
3700 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3701 the special perltest8 script is no longer needed - all the tests can be run
3702 from a single perltest script.
3703
3704 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3705 by \s. It has now been removed in PCRE. This means it isn't recognized as
3706 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3707 class [:space:] *does* include VT, thereby creating a mess.
3708
3709 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3710 space and tab.
3711
3712 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3713 its new features into the main test script, reducing the number of scripts.
3714
3715 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3716 were backward compatible, and made the (?i) apply to the whole pattern, as if
3717 /i were given. Now it behaves more logically, and applies the option setting
3718 only to what follows. PCRE has been changed to follow suit. However, if it
3719 finds options settings right at the start of the pattern, it extracts them into
3720 the global options, as before. Thus, they show up in the info data.
3721
3722 8. Added support for the \Q...\E escape sequence. Characters in between are
3723 treated as literals. This is slightly different from Perl in that $ and @ are
3724 also handled as literals inside the quotes. In Perl, they will cause variable
3725 interpolation. Note the following examples:
3726
3727 Pattern PCRE matches Perl matches
3728
3729 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3730 \Qabc\$xyz\E abc\$xyz abc\$xyz
3731 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3732
3733 For compatibility with Perl, \Q...\E sequences are recognized inside character
3734 classes as well as outside them.
3735
3736 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3737 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3738 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3739 signed/unsigned warnings.
3740
3741 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3742 option for pcretest, so I've replaced it by a simple function that does just
3743 that job.
3744
3745 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3746 "pcregrep -".
3747
3748 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3749 Java package. This provides some syntactic sugar for simple cases of what my
3750 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3751 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3752 item, you can use this simplified notation. Note that only makes sense with
3753 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3754 greediness, whatever the setting of the PCRE_UNGREEDY option.
3755
3756 13. A change of greediness default within a pattern was not taking effect at
3757 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3758 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3759 was abstracted outside.
3760
3761 14. PCRE now supports the \G assertion. It is true when the current matching
3762 position is at the start point of the match. This differs from \A when the
3763 starting offset is non-zero. Used with the /g option of pcretest (or similar
3764 code), it works in the same way as it does for Perl's /g option. If all
3765 alternatives of a regex begin with \G, the expression is anchored to the start
3766 match position, and the "anchored" flag is set in the compiled expression.
3767
3768 15. Some bugs concerning the handling of certain option changes within patterns
3769 have been fixed. These applied to options other than (?ims). For example,
3770 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3771 been the other way round. Some of this was related to change 7 above.
3772
3773 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3774 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3775 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3776 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3777
3778 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3779 mode. Unlike ".", it always matches newline, whatever the setting of
3780 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3781 assertions. Perl allows it, but it doesn't (in general) work because it can't
3782 calculate the length of the lookbehind. At least, that's the case for Perl
3783 5.8.0 - I've been told they are going to document that it doesn't work in
3784 future.
3785
3786 18. Added an error diagnosis for escapes that PCRE does not support: these are
3787 \L, \l, \N, \P, \p, \U, \u, and \X.
3788
3789 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3790 reading past the end of the pattern in cases such as /[abcd/.
3791
3792 20. PCRE was getting more memory than necessary for patterns with classes that
3793 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3794
3795 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3796 compiling PCRE for use with Virtual Pascal.
3797
3798 22. Small fix to the Makefile to make it work properly if the build is done
3799 outside the source tree.
3800
3801 23. Added a new extension: a condition to go with recursion. If a conditional
3802 subpattern starts with (?(R) the "true" branch is used if recursion has
3803 happened, whereas the "false" branch is used only at the top level.
3804
3805 24. When there was a very long string of literal characters (over 255 bytes
3806 without UTF support, over 250 bytes with UTF support), the computation of how
3807 much memory was required could be incorrect, leading to segfaults or other
3808 strange effects.
3809
3810 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3811 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3812 there was a subsequent back reference to those brackets. This meant that, for
3813 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3814 possible to check for precisely this case. All we can do is abandon the
3815 optimization if .* occurs inside capturing brackets when there are any back
3816 references whatsoever. (See below for a better fix that came later.)
3817
3818 26. The handling of the optimization for finding the first character of a
3819 non-anchored pattern, and for finding a character that is required later in the
3820 match were failing in some cases. This didn't break the matching; it just
3821 failed to optimize when it could. The way this is done has been re-implemented.
3822
3823 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3824
3825 28. Added a new feature that provides some of the functionality that Perl
3826 provides with (?{...}). The facility is termed a "callout". The way it is done
3827 in PCRE is for the caller to provide an optional function, by setting
3828 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3829 global variable. By default it is unset, which disables all calling out. To get
3830 the function called, the regex must include (?C) at appropriate points. This
3831 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3832 This provides a means of identifying different callout points. When PCRE
3833 reaches such a point in the regex, if pcre_callout has been set, the external
3834 function is called. It is provided with data in a structure called
3835 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3836 matching continues; if it returns a non-zero value, the match at the current
3837 point fails. However, backtracking will occur if possible. [This was changed
3838 later and other features added - see item 49 below.]
3839
3840 29. pcretest is upgraded to test the callout functionality. It provides a
3841 callout function that displays information. By default, it shows the start of
3842 the match and the current position in the text. There are some new data escapes
3843 to vary what happens:
3844
3845 \C+ in addition, show current contents of captured substrings
3846 \C- do not supply a callout function
3847 \C!n return 1 when callout number n is reached
3848 \C!n!m return 1 when callout number n is reached for the mth time
3849
3850 30. If pcregrep was called with the -l option and just a single file name, it
3851 output "<stdin>" if a match was found, instead of the file name.
3852
3853 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3854 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3855 pcre_exec(). This saves a malloc/free per call. The default value of
3856 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3857 when configuring.
3858
3859 32. The default maximum size of a compiled pattern is 64K. There have been a
3860 few cases of people hitting this limit. The code now uses macros to handle the
3861 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3862 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3863 configuring. Tests 2 and 5 work only with 2-byte links because they output
3864 debugging information about compiled patterns.
3865
3866 33. Internal code re-arrangements:
3867
3868 (a) Moved the debugging function for printing out a compiled regex into
3869 its own source file (printint.c) and used #include to pull it into
3870 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3871 separate copies.
3872
3873 (b) Defined the list of op-code names for debugging as a macro in
3874 internal.h so that it is next to the definition of the opcodes.
3875
3876 (c) Defined a table of op-code lengths for simpler skipping along compiled
3877 code. This is again a macro in internal.h so that it is next to the
3878 definition of the opcodes.
3879
3880 34. Added support for recursive calls to individual subpatterns, along the
3881 lines of Robin Houston's patch (but implemented somewhat differently).
3882
3883 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3884 allow it to read and process whole directories in Win32. This code was
3885 contributed by Lionel Fourquaux; it has not been tested by me.
3886
3887 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3888 used to name a group. Names consist of alphanumerics and underscores, and must
3889 be unique. Back references use the syntax (?P=name) and recursive calls use
3890 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3891 numbers. The function pcre_fullinfo() can be used after compilation to extract
3892 a name/number map. There are three relevant calls:
3893
3894 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3895 PCRE_INFO_NAMECOUNT yields the number of entries
3896 PCRE_INFO_NAMETABLE yields a pointer to the map.
3897
3898 The map is a vector of fixed-size entries. The size of each entry depends on
3899 the length of the longest name used. The first two bytes of each entry are the
3900 group number, most significant byte first. There follows the corresponding
3901 name, zero terminated. The names are in alphabetical order.
3902
3903 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3904 case instead of 255. Making it the same both with and without UTF-8 support
3905 means that the same test output works with both.
3906
3907 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3908 calling malloc() with a zero argument.
3909
3910 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3911 optimization. I've improved things by keeping a bitmap of backreferences with
3912 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3913 fact referenced, the optimization can be applied. It is unlikely that a
3914 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3915 the match to follow \n) will appear inside brackets with a number greater than
3916 31, but if it does, any back reference > 31 suppresses the optimization.
3917
3918 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3919 of disabling numbered capturing parentheses. Any opening parenthesis that is
3920 not followed by ? behaves as if it were followed by ?: but named parentheses
3921 can still be used for capturing (and they will acquire numbers in the usual
3922 way).
3923
3924 41. Redesigned the return codes from the match() function into yes/no/error so
3925 that errors can be passed back from deep inside the nested calls. A malloc
3926 failure while inside a recursive subpattern call now causes the
3927 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3928
3929 42. It is now possible to set a limit on the number of times the match()
3930 function is called in a call to pcre_exec(). This facility makes it possible to
3931 limit the amount of recursion and backtracking, though not in a directly
3932 obvious way, because the match() function is used in a number of different
3933 circumstances. The count starts from zero for each position in the subject
3934 string (for non-anchored patterns). The default limit is, for compatibility, a
3935 large number, namely 10 000 000. You can change this in two ways:
3936
3937 (a) When configuring PCRE before making, you can use --with-match-limit=n
3938 to set a default value for the compiled library.
3939
3940 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3941 a different value is set. See 45 below.
3942
3943 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3944
3945 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3946 of things that can be changed at compile time. The first argument specifies
3947 what is wanted and the second points to where the information is to be placed.
3948 The current list of available information is:
3949
3950 PCRE_CONFIG_UTF8
3951
3952 The output is an integer that is set to one if UTF-8 support is available;
3953 otherwise it is set to zero.
3954
3955 PCRE_CONFIG_NEWLINE
3956
3957 The output is an integer that it set to the value of the code that is used for
3958 newline. It is either LF (10) or CR (13).
3959
3960 PCRE_CONFIG_LINK_SIZE
3961
3962 The output is an integer that contains the number of bytes used for internal
3963 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3964
3965 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3966
3967 The output is an integer that contains the threshold above which the POSIX
3968 interface uses malloc() for output vectors. See item 31 above.
3969
3970 PCRE_CONFIG_MATCH_LIMIT
3971
3972 The output is an unsigned integer that contains the default limit of the number
3973 of match() calls in a pcre_exec() execution. See 42 above.
3974
3975 44. pcretest has been upgraded by the addition of the -C option. This causes it
3976 to extract all the available output from the new pcre_config() function, and to
3977 output it. The program then exits immediately.
3978
3979 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3980 order to support additional features. One way would have been to define
3981 pcre_exec2() (for example) with extra arguments, but this would not have been
3982 extensible, and would also have required all calls to the original function to
3983 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3984 is used for passing in "extra" data from pcre_study().
3985
3986 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3987 contains the following fields:
3988
3989 flags a bitmap indicating which of the following fields are set
3990 study_data opaque data from pcre_study()
3991 match_limit a way of specifying a limit on match() calls for a specific
3992 call to pcre_exec()
3993 callout_data data for callouts (see 49 below)
3994
3995 The flag bits are also defined in pcre.h, and are
3996
3997 PCRE_EXTRA_STUDY_DATA
3998 PCRE_EXTRA_MATCH_LIMIT
3999 PCRE_EXTRA_CALLOUT_DATA
4000
4001 The pcre_study() function now returns one of these new pcre_extra blocks, with
4002 the actual study data pointed to by the study_data field, and the
4003 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4004 before. That is, this change is entirely upwards-compatible and requires no
4005 change to existing code.
4006
4007 If you want to pass in additional data to pcre_exec(), you can either place it
4008 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4009 block.
4010
4011 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4012 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4013 times with different match limits, until it finds the minimum value needed for
4014 pcre_exec() to complete. The value is then output. This can be instructive; for
4015 most simple matches the number is quite small, but for pathological cases it
4016 gets very large very quickly.
4017
4018 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4019 returns the size of the data block pointed to by the study_data field in a
4020 pcre_extra block, that is, the value that was passed as the argument to
4021 pcre_malloc() when PCRE was getting memory in which to place the information
4022 created by pcre_study(). The fourth argument should point to a size_t variable.
4023 pcretest has been extended so that this information is shown after a successful
4024 pcre_study() call when information about the compiled regex is being displayed.
4025
4026 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4027 because what follows is always an absolute path. (Later: it turns out that this
4028 is more than cosmetic for MinGW, because it doesn't like empty path
4029 components.)
4030
4031 49. Some changes have been made to the callout feature (see 28 above):
4032
4033 (i) A callout function now has three choices for what it returns:
4034
4035 0 => success, carry on matching
4036 > 0 => failure at this point, but backtrack if possible
4037 < 0 => serious error, return this value from pcre_exec()
4038
4039 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4040 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4041 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4042 use by callout functions. It will never be used by PCRE itself.
4043
4044 (ii) The pcre_extra structure (see 45 above) has a void * field called
4045 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4046 pcre_callout_block structure has a field of the same name. The contents of
4047 the field passed in the pcre_extra structure are passed to the callout
4048 function in the corresponding field in the callout block. This makes it
4049 easier to use the same callout-containing regex from multiple threads. For
4050 testing, the pcretest program has a new data escape
4051
4052 \C*n pass the number n (may be negative) as callout_data
4053
4054 If the callout function in pcretest receives a non-zero value as
4055 callout_data, it returns that value.
4056
4057 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4058 there were some redundant $(CFLAGS) in commands that are now specified as
4059 $(LINK), which already includes $(CFLAGS).
4060
4061 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4062 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4063 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4064 one-byte characters throughout. Note that case-insensitive matching applies
4065 only to characters whose values are less than 256. PCRE doesn't support the
4066 notion of cases for higher-valued characters.
4067
4068 (i) A character class whose characters are all within 0-255 is handled as
4069 a bit map, and the map is inverted for negative classes. Previously, a
4070 character > 255 always failed to match such a class; however it should
4071 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4072
4073 (ii) A negated character class with a single character < 255 is coded as
4074 "not this character" (OP_NOT). This wasn't working properly when the test
4075 character was multibyte, either singly or repeated.
4076
4077 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4078 mode, for example: \x{100}{2,3}.
4079
4080 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4081 singly or repeated) now correctly test multibyte characters. However,
4082 PCRE doesn't recognize any characters with values greater than 255 as
4083 digits, spaces, or word characters. Such characters always match \D, \S,
4084 and \W, and never match \d, \s, or \w.
4085
4086 (v) Classes may now contain characters and character ranges with values
4087 greater than 255. For example: [ab\x{100}-\x{400}].
4088
4089 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4090 PCRE in UTF-8 mode.
4091
4092 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4093 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4094 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4095 value.)
4096
4097 53. The single man page has become too large. I have therefore split it up into
4098 a number of separate man pages. These also give rise to individual HTML pages;
4099 these are now put in a separate directory, and there is an index.html page that
4100 lists them all. Some hyperlinking between the pages has been installed.
4101
4102 54. Added convenience functions for handling named capturing parentheses.
4103
4104 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4105 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4106 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4107 are faulted.
4108
4109 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4110 calling configure. These values are used when compiling the dftables.c program
4111 which is run to generate the source of the default character tables. They
4112 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4113 you will need to set these values.
4114
4115 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4116
4117
4118 Version 3.9 02-Jan-02
4119 ---------------------
4120
4121 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4122
4123 2. If --disable-static was given, the building process failed when trying to
4124 build pcretest and pcregrep. (For some reason it was using libtool to compile
4125 them, which is not right, as they aren't part of the library.)
4126
4127
4128 Version 3.8 18-Dec-01
4129 ---------------------
4130
4131 1. The experimental UTF-8 code was completely screwed up. It was packing the
4132 bytes in the wrong order. How dumb can you get?
4133
4134
4135 Version 3.7 29-Oct-01
4136 ---------------------
4137
4138 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4139 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4140 this didn't happen under Solaris 8, where I normally test things.
4141
4142 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4143 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4144 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4145 this fix an hour or so after the initial 3.7 release.)
4146
4147
4148 Version 3.6 23-Oct-01
4149 ---------------------
4150
4151 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4152 offsets passed as NULL with zero offset count.
4153
4154 2. The config.guess and config.sub files had not been updated when I moved to
4155 the latest autoconf.
4156
4157
4158 Version 3.5 15-Aug-01
4159 ---------------------
4160
4161 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4162 had been forgotten.
4163
4164 2. By using declared but undefined structures, we can avoid using "void"
4165 definitions in pcre.h while keeping the internal definitions of the structures
4166 private.
4167
4168 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4169 user point of view, this means that both static and shared libraries are built
4170 by default, but this can be individually controlled. More of the work of
4171 handling this static/shared cases is now inside libtool instead of PCRE's make
4172 file.
4173
4174 4. The pcretest utility is now installed along with pcregrep because it is
4175 useful for users (to test regexs) and by doing this, it automatically gets
4176 relinked by libtool. The documentation has been turned into a man page, so
4177 there are now .1, .txt, and .html versions in /doc.
4178
4179 5. Upgrades to pcregrep:
4180 (i) Added long-form option names like gnu grep.
4181 (ii) Added --help to list all options with an explanatory phrase.
4182 (iii) Added -r, --recursive to recurse into sub-directories.
4183 (iv) Added -f, --file to read patterns from a file.
4184
4185 6. pcre_exec() was referring to its "code" argument before testing that
4186 argument for NULL (and giving an error if it was NULL).
4187
4188 7. Upgraded Makefile.in to allow for compiling in a different directory from
4189 the source directory.
4190
4191 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4192 options bits, the pointer it was passed was to an int instead of to an unsigned
4193 long int. This mattered only on 64-bit systems.
4194
4195 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4196 generated) instead of pcre.in, which it its source. Also made the same change
4197 in several of the .c files.
4198
4199 10. A new release of gcc defines printf() as a macro, which broke pcretest
4200 because it had an ifdef in the middle of a string argument for printf(). Fixed
4201 by using separate calls to printf().
4202
4203 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4204 script, to force use of CR or LF instead of \n in the source. On non-Unix
4205 systems, the value can be set in config.h.
4206
4207 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4208 absolute limit. Changed the text of the error message to make this clear, and
4209 likewise updated the man page.
4210
4211 13. The limit of 99 on the number of capturing subpatterns has been removed.
4212 The new limit is 65535, which I hope will not be a "real" limit.
4213
4214
4215 Version 3.4 22-Aug-00
4216 ---------------------
4217
4218 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4219
4220 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4221
4222
4223 Version 3.3 01-Aug-00
4224 ---------------------
4225
4226 1. If an octal character was given, but the value was greater than \377, it
4227 was not getting masked to the least significant bits, as documented. This could
4228 lead to crashes in some systems.
4229
4230 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4231 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4232
4233 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4234 These just pass their arguments on to (pcre_free)(), but they are provided
4235 because some uses of PCRE bind it to non-C systems that can call its functions,
4236 but cannot call free() or pcre_free() directly.
4237
4238 4. Add "make test" as a synonym for "make check". Corrected some comments in
4239 the Makefile.
4240
4241 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4242 Makefile.
4243
4244 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4245 command called pgrep for grepping around the active processes.
4246
4247 7. Added the beginnings of support for UTF-8 character strings.
4248
4249 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4250 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4251 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4252 out for the ar command.)
4253
4254
4255 Version 3.2 12-May-00
4256 ---------------------
4257
4258 This is purely a bug fixing release.
4259
4260 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4261 of ZA. This was just one example of several cases that could provoke this bug,
4262 which was introduced by change 9 of version 2.00. The code for breaking
4263 infinite loops after an iteration that matches an empty string was't working
4264 correctly.
4265
4266 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4267 when matched against abbab (for example). After matching an empty string, it
4268 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4269 caused it to match further down the string than it should.
4270
4271 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4272 was there because it doesn't seem to be needed, and it causes trouble on some
4273 systems, as it is not a Standard C header. It has been removed.
4274
4275 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4276 were reported on the Macintosh. The changes were from
4277
4278 while ((c = *(++ptr)) != 0 && c != '\n');
4279 to
4280 while ((c = *(++ptr)) != 0 && c != '\n') ;
4281
4282 Totally extraordinary, but if that's what it takes...
4283
4284 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4285 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4286 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4287 assumes the way PCRE uses memmove() (always moving upwards).
4288
4289 6. PCRE is being used in one environment where strchr() is not available. There
4290 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4291 faster code anyway.
4292
4293
4294 Version 3.1 09-Feb-00
4295 ---------------------
4296
4297 The only change in this release is the fixing of some bugs in Makefile.in for
4298 the "install" target:
4299
4300 (1) It was failing to install pcreposix.h.
4301
4302 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4303
4304
4305 Version 3.0 01-Feb-00
4306 ---------------------
4307
4308 1. Add support for the /+ modifier to perltest (to output $` like it does in
4309 pcretest).
4310
4311 2. Add support for the /g modifier to perltest.
4312
4313 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4314 matches null strings.
4315
4316 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4317 pattern. Perl treats empty patterns specially - it reuses the most recent
4318 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4319 effect.
4320
4321 5. The POSIX interface was broken in that it was just handing over the POSIX
4322 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4323 required a bigger vector, with some working space on the end. This means that
4324 the POSIX wrapper now has to get and free some memory, and copy the results.
4325
4326 6. Added some simple autoconf support, placing the test data and the
4327 documentation in separate directories, re-organizing some of the
4328 information files, and making it build pcre-config (a GNU standard). Also added
4329 libtool support for building PCRE as a shared library, which is now the
4330 default.
4331
4332 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4333 09 are not valid octal constants. Single digits will be used for minor values
4334 less than 10.
4335
4336 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4337 existing programs that set these in the POSIX interface can use PCRE without
4338 modification.
4339
4340 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4341 return all that pcre_info() returns, plus additional data. The pcre_info()
4342 function is retained for compatibility, but is considered to be obsolete.
4343
4344 10. Added experimental recursion feature (?R) to handle one common case that
4345 Perl 5.6 will be able to do with (?p{...}).
4346
4347 11. Added support for POSIX character classes like [:alpha:], which Perl is
4348 adopting.
4349
4350
4351 Version 2.08 31-Aug-99
4352 ----------------------
4353
4354 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4355 trying to match at the startoffset position, but instead was moving forward to
4356 the next newline as if a previous match had failed.
4357
4358 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4359 and could get into a loop if a null string was matched other than at the start
4360 of the subject.
4361
4362 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4363 be distinguished at compile time, and for completeness also added PCRE_DATE.
4364
4365 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4366 in GnuWin32 environments.
4367
4368
4369 Version 2.07 29-Jul-99
4370 ----------------------
4371
4372 1. The documentation is now supplied in plain text form and HTML as well as in
4373 the form of man page sources.
4374
4375 2. C++ compilers don't like assigning (void *) values to other pointer types.
4376 In particular this affects malloc(). Although there is no problem in Standard
4377 C, I've put in casts to keep C++ compilers happy.
4378
4379 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4380 should be (const char *).
4381
4382 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4383 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4384 However, I haven't made this a standard facility. The documentation doesn't
4385 mention it, and the Makefile doesn't support it.
4386
4387 5. The Makefile now contains an "install" target, with editable destinations at
4388 the top of the file. The pcretest program is not installed.
4389
4390 6. pgrep -V now gives the PCRE version number and date.
4391
4392 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4393 causing the entire string to be ignored, instead of just the last character.
4394
4395 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4396 non-matching string, it can take a very, very long time, even for strings of
4397 quite modest length, because of the nested recursion. PCRE now does better in
4398 some of these cases. It does this by remembering the last required literal
4399 character in the pattern, and pre-searching the subject to ensure it is present
4400 before running the real match. In other words, it applies a heuristic to detect
4401 some types of certain failure quickly, and in the above example, if presented
4402 with a string that has no trailing " it gives "no match" very quickly.
4403
4404 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4405 other alternatives are tried instead.
4406
4407
4408 Version 2.06 09-Jun-99
4409 ----------------------
4410
4411 1. Change pcretest's output for amount of store used to show just the code
4412 space, because the remainder (the data block) varies in size between 32-bit and
4413 64-bit systems.
4414
4415 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4416 start matching at. This allows lookbehinds to work when searching for multiple
4417 occurrences in a string.
4418
4419 3. Added additional options to pcretest for testing multiple occurrences:
4420
4421 /+ outputs the rest of the string that follows a match
4422 /g loops for multiple occurrences, using the new startoffset argument
4423 /G loops for multiple occurrences by passing an incremented pointer
4424
4425 4. PCRE wasn't doing the "first character" optimization for patterns starting
4426 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4427 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4428 the letter 'x'. On long subject strings, this gives a significant speed-up.
4429
4430
4431 Version 2.05 21-Apr-99
4432 ----------------------
4433
4434 1. Changed the type of magic_number from int to long int so that it works
4435 properly on 16-bit systems.
4436
4437 2. Fixed a bug which caused patterns starting with .* not to work correctly
4438 when the subject string contained newline characters. PCRE was assuming
4439 anchoring for such patterns in all cases, which is not correct because .* will
4440 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4441 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4442 must be retried after every newline in the subject.
4443
4444
4445 Version 2.04 18-Feb-99
4446 ----------------------
4447
4448 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4449 computation of the store needed to hold the pattern was incorrect (too large).
4450 If such patterns were nested a few deep, this could multiply and become a real
4451 problem.
4452
4453 2. Added /M option to pcretest to show the memory requirement of a specific
4454 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4455
4456 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4457 compiled in such a way that the backtracking after subsequent failure was
4458 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4459 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4460
4461
4462 Version 2.03 02-Feb-99
4463 ----------------------
4464
4465 1. Fixed typo and small mistake in man page.
4466
4467 2. Added 4th condition (GPL supersedes if conflict) and created separate
4468 LICENCE file containing the conditions.
4469
4470 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4471 Perl, that is the internal \ allows the delimiter to be included in the
4472 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4473 the final delimiter, add \ to the end of the pattern (to test the error).
4474
4475 4. Added the convenience functions for extracting substrings after a successful
4476 match. Updated pcretest to make it able to test these functions.
4477
4478
4479 Version 2.02 14-Jan-99
4480 ----------------------
4481
4482 1. Initialized the working variables associated with each extraction so that
4483 their saving and restoring doesn't refer to uninitialized store.
4484
4485 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4486 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4487 fix the problem.
4488
4489 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4490 calls, and wasn't printing the correct value for compiling calls. Increased the
4491 default value of LOOPREPEAT, and the number of significant figures in the
4492 times.
4493
4494 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4495
4496 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4497 a building problem on Windows NT with a FAT file system.
4498
4499
4500 Version 2.01 21-Oct-98
4501 ----------------------
4502
4503 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4504 to character tables built by pcre_maketables() in the current locale. If NULL
4505 is passed, the default tables are used.
4506
4507
4508 Version 2.00 24-Sep-98
4509 ----------------------
4510
4511 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4512 it any more.
4513
4514 2. Allow quantification of (?>) groups, and make it work correctly.
4515
4516 3. The first character computation wasn't working for (?>) groups.
4517
4518 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4519 end of the subject) and add 5.005's \z, which really does match only at the
4520 very end of the subject.
4521
4522 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4523
4524 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4525 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4526 localized options. All options to pcre_study() were also removed.
4527
4528 7. Add other new features from 5.005:
4529
4530 $(?<= positive lookbehind
4531 $(?<! negative lookbehind
4532 (?imsx-imsx) added the unsetting capability
4533 such a setting is global if at outer level; local otherwise
4534 (?imsx-imsx:) non-capturing groups with option setting
4535 (?(cond)re|re) conditional pattern matching
4536
4537 A backreference to itself in a repeated group matches the previous
4538 captured string.
4539
4540 8. General tidying up of studying (both automatic and via "study")
4541 consequential on the addition of new assertions.
4542
4543 9. As in 5.005, unlimited repeated groups that could match an empty substring
4544 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4545 runtime if any iteration does actually match an empty substring.
4546
4547 10. Include the RunTest script in the distribution.
4548
4549 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4550 discrepancies, some of which were old and were also with respect to 5.004. They
4551 have now been fixed.
4552
4553
4554 Version 1.09 28-Apr-98
4555 ----------------------
4556
4557 1. A negated single character class followed by a quantifier with a minimum
4558 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4559 program crashes, or just wrong answers. This did not apply to negated classes
4560 containing more than one character, or to minima other than one.
4561
4562
4563 Version 1.08 27-Mar-98
4564 ----------------------
4565
4566 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4567
4568 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4569 latter must appear before anything that relies on it in the pattern.
4570
4571
4572 Version 1.07 16-Feb-98
4573 ----------------------
4574
4575 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4576 repeat of a potentially empty string).
4577
4578
4579 Version 1.06 23-Jan-98
4580 ----------------------
4581
4582 1. Added Markus Oberhumer's little patches for C++.
4583
4584 2. Literal strings longer than 255 characters were broken.
4585
4586
4587 Version 1.05 23-Dec-97
4588 ----------------------
4589
4590 1. Negated character classes containing more than one character were failing if
4591 PCRE_CASELESS was set at run time.
4592
4593
4594 Version 1.04 19-Dec-97
4595 ----------------------
4596
4597 1. Corrected the man page, where some "const" qualifiers had been omitted.
4598
4599 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4600 input syntax.
4601
4602 3. Fixed memory leak which occurred when a regex with back references was
4603 matched with an offsets vector that wasn't big enough. The temporary memory
4604 that is used in this case wasn't being freed if the match failed.
4605
4606 4. Tidied pcretest to ensure it frees memory that it gets.
4607
4608 5. Temporary memory was being obtained in the case where the passed offsets
4609 vector was exactly big enough.
4610
4611 6. Corrected definition of offsetof() from change 5 below.
4612
4613 7. I had screwed up change 6 below and broken the rules for the use of
4614 setjmp(). Now fixed.
4615
4616
4617 Version 1.03 18-Dec-97
4618 ----------------------
4619
4620 1. A erroneous regex with a missing opening parenthesis was correctly
4621 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4622 on some systems.
4623
4624 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4625 it was reported that one broken compiler failed on the former because "code" is
4626 also an independent variable.
4627
4628 3. The erroneous regex a[]b caused an array overrun reference.
4629
4630 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4631 fail on data ending with that character. (It was going on too far, and checking
4632 the next character, typically a binary zero.) This was specific to the
4633 optimized code for single-character negative classes.
4634
4635 5. Added a contributed patch from the TIN world which does the following:
4636
4637 + Add an undef for memmove, in case the the system defines a macro for it.
4638
4639 + Add a definition of offsetof(), in case there isn't one. (I don't know
4640 the reason behind this - offsetof() is part of the ANSI standard - but
4641 it does no harm).
4642
4643 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4644 most of the places where whitespace preceded '#'. I have given up and
4645 allowed the remaining 2 cases to be at the margin.
4646
4647 + Rename some variables in pcre to eliminate shadowing. This seems very
4648 pedantic, but does no harm, of course.
4649
4650 6. Moved the call to setjmp() into its own function, to get rid of warnings
4651 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4652
4653 7. Constructs such as \d{8,} were compiling into the equivalent of
4654 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4655 outcome, but in this particular case used more store than had been allocated,
4656 which caused the bug to be discovered because it threw up an internal error.
4657
4658 8. The debugging code in both pcre and pcretest for outputting the compiled
4659 form of a regex was going wrong in the case of back references followed by
4660 curly-bracketed repeats.
4661
4662
4663 Version 1.02 12-Dec-97
4664 ----------------------
4665
4666 1. Typos in pcre.3 and comments in the source fixed.
4667
4668 2. Applied a contributed patch to get rid of places where it used to remove
4669 'const' from variables, and fixed some signed/unsigned and uninitialized
4670 variable warnings.
4671
4672 3. Added the "runtest" target to Makefile.
4673
4674 4. Set default compiler flag to -O2 rather than just -O.
4675
4676
4677 Version 1.01 19-Nov-97
4678 ----------------------
4679
4680 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4681 like /([ab]*)*/, that is, for classes with more than one character in them.
4682
4683 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4684 as /((?>a*))*/ (a PCRE_EXTRA facility).
4685
4686
4687 Version 1.00 18-Nov-97
4688 ----------------------
4689
4690 1. Added compile-time macros to support systems such as SunOS4 which don't have
4691 memmove() or strerror() but have other things that can be used instead.
4692
4693 2. Arranged that "make clean" removes the executables.
4694
4695
4696 Version 0.99 27-Oct-97
4697 ----------------------
4698
4699 1. Fixed bug in code for optimizing classes with only one character. It was
4700 initializing a 32-byte map regardless, which could cause it to run off the end
4701 of the memory it had got.
4702
4703 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4704
4705
4706 Version 0.98 22-Oct-97
4707 ----------------------
4708
4709 1. Fixed bug in code for handling temporary memory usage when there are more
4710 back references than supplied space in the ovector. This could cause segfaults.
4711
4712
4713 Version 0.97 21-Oct-97
4714 ----------------------
4715
4716 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4717
4718 2. Optimized negated single characters not to use a bit map.
4719
4720 3. Brought error texts together as macro definitions; clarified some of them;
4721 fixed one that was wrong - it said "range out of order" when it meant "invalid
4722 escape sequence".
4723
4724 4. Changed some char * arguments to const char *.
4725
4726 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4727
4728 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4729 pcretest.
4730
4731
4732 Version 0.96 16-Oct-97
4733 ----------------------
4734
4735 1. Added a simple "pgrep" utility to the distribution.
4736
4737 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4738 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4739 where "ddd" means "one or more decimal digits".
4740
4741 3. Fixed serious bug. If a pattern had a back reference, but the call to
4742 pcre_exec() didn't supply a large enough ovector to record the related
4743 identifying subpattern, the match always failed. PCRE now remembers the number
4744 of the largest back reference, and gets some temporary memory in which to save
4745 the offsets during matching if necessary, in order to ensure that
4746 backreferences always work.
4747
4748 4. Increased the compatibility with Perl in a number of ways:
4749
4750 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4751 to request this handling. The option can be set at compile or exec time.
4752
4753 (b) $ matches before a terminating newline by default; an option
4754 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4755 mode). The option can be set at compile or exec time.
4756
4757 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4758 the same as Perl's. If the decimal number it represents is less than 10
4759 or there aren't that many previous left capturing parentheses, an octal
4760 escape is read. Inside a character class, it's always an octal escape,
4761 even if it is a single digit.
4762
4763 (d) An escaped but undefined alphabetic character is taken as a literal,
4764 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4765 escapes.
4766
4767 (e) {0} is now permitted. (The previous item is removed from the compiled
4768 pattern).
4769
4770 5. Changed all the names of code files so that the basic parts are no longer
4771 than 10 characters, and abolished the teeny "globals.c" file.
4772
4773 6. Changed the handling of character classes; they are now done with a 32-byte
4774 bit map always.
4775
4776 7. Added the -d and /D options to pcretest to make it possible to look at the
4777 internals of compilation without having to recompile pcre.
4778
4779
4780 Version 0.95 23-Sep-97
4781 ----------------------
4782
4783 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4784 \x20 at the start of a run of normal characters. These were being treated as
4785 real characters, instead of the source characters being re-checked.
4786
4787
4788 Version 0.94 18-Sep-97
4789 ----------------------
4790
4791 1. The functions are now thread-safe, with the caveat that the global variables
4792 containing pointers to malloc() and free() or alternative functions are the
4793 same for all threads.
4794
4795 2. Get pcre_study() to generate a bitmap of initial characters for non-
4796 anchored patterns when this is possible, and use it if passed to pcre_exec().
4797
4798
4799 Version 0.93 15-Sep-97
4800 ----------------------
4801
4802 1. /(b)|(:+)/ was computing an incorrect first character.
4803
4804 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4805 but not actually doing anything yet.
4806
4807 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4808 as Perl does (e.g. [-az] or [az-]).
4809
4810 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4811 all possible positions.
4812
4813 5. Split up into different modules to avoid including unneeded functions in a
4814 compiled binary. However, compile and exec are still in one module. The "study"
4815 function is split off.
4816
4817 6. The character tables are now in a separate module whose source is generated
4818 by an auxiliary program - but can then be edited by hand if required. There are
4819 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4820 toupper() in the code.
4821
4822 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4823 make them global. Abolish the function for setting them, as the caller can now
4824 set them directly.
4825
4826
4827 Version 0.92 11-Sep-97
4828 ----------------------
4829
4830 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4831 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4832
4833 2. Caseless matching was not working in character classes if the characters in
4834 the pattern were in upper case.
4835
4836 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4837
4838 4. Make PCRE_ANCHORED public and accept as a compile option.
4839
4840 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4841 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4842 pass them.
4843
4844 6. Give an error if bad option bits passed at compile or run time.
4845
4846 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4847 pcretest to cause it to pass that flag.
4848
4849 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4850 options, and the first character, if set.
4851
4852 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4853
4854
4855 Version 0.91 10-Sep-97
4856 ----------------------
4857
4858 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4859 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4860
4861 2. PCRE was looping on encountering an indefinitely repeated back reference to
4862 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4863 Perl does - treats the match as successful.
4864
4865 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5