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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 987 - (show annotations) (download)
Sat Jul 7 04:11:29 2012 UTC (2 years, 1 month ago) by zherczeg
File size: 232995 byte(s)
Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
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
14 Version 8.31 06-July-2012
15 -------------------------
16
17 1. Fixing a wrong JIT test case and some compiler warnings.
18
19 2. Removed a bashism from the RunTest script.
20
21 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
22 to unsigned type, result still unsigned" that was given by an MS compiler
23 on encountering the code "-sizeof(xxx)".
24
25 4. Partial matching support is added to the JIT compiler.
26
27 5. Fixed several bugs concerned with partial matching of items that consist
28 of more than one character:
29
30 (a) /^(..)\1/ did not partially match "aba" because checking references was
31 done on an "all or nothing" basis. This also applied to repeated
32 references.
33
34 (b) \R did not give a hard partial match if \r was found at the end of the
35 subject.
36
37 (c) \X did not give a hard partial match after matching one or more
38 characters at the end of the subject.
39
40 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
41 a partial match for the string "\r".
42
43 (e) When newline was set to CRLF, the metacharacter "." did not recognize
44 a partial match for a CR character at the end of the subject string.
45
46 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
47 running pcretest, the text "(JIT)" added to the output whenever JIT is
48 actually used to run the match.
49
50 7. Individual JIT compile options can be set in pcretest by following -s+[+]
51 or /S+[+] with a digit between 1 and 7.
52
53 8. OP_NOT now supports any UTF character not just single-byte ones.
54
55 9. (*MARK) control verb is now supported by the JIT compiler.
56
57 10. The command "./RunTest list" lists the available tests without actually
58 running any of them. (Because I keep forgetting what they all are.)
59
60 11. Add PCRE_INFO_MAXLOOKBEHIND.
61
62 12. Applied a (slightly modified) user-supplied patch that improves performance
63 when the heap is used for recursion (compiled with --disable-stack-for-
64 recursion). Instead of malloc and free for each heap frame each time a
65 logical recursion happens, frames are retained on a chain and re-used where
66 possible. This sometimes gives as much as 30% improvement.
67
68 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
69 call.
70
71 14. As documented, (*COMMIT) is now confined to within a positive assertion.
72
73 15. It is now possible to link pcretest with libedit as an alternative to
74 libreadline.
75
76 16. (*COMMIT) control verb is now supported by the JIT compiler.
77
78 17. The Unicode data tables have been updated to Unicode 6.1.0.
79
80 18. Added --file-list option to pcregrep.
81
82 19. Added binary file support to pcregrep, including the -a, --binary-files,
83 -I, and --text options.
84
85 20. The madvise function is renamed for posix_madvise for QNX compatibility
86 reasons. Fixed by Giuseppe D'Angelo.
87
88 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
89
90 22. Changed the option for creating symbolic links for 16-bit man pages from
91 -s to -sf so that re-installing does not cause issues.
92
93 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
94
95 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
96 matches in certain environments (the workspace was not being correctly
97 retained). Also added to pcre_dfa_exec() a simple plausibility check on
98 some of the workspace data at the beginning of a restart.
99
100 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
101 was not doing so when it should - probably a typo introduced by SVN 528
102 (change 8.10/14).
103
104 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
105 \w+ when the character tables indicated that \x{c4} was a word character.
106 There were several related cases, all because the tests for doing a table
107 lookup were testing for characters less than 127 instead of 255.
108
109 27. If a pattern contains capturing parentheses that are not used in a match,
110 their slots in the ovector are set to -1. For those that are higher than
111 any matched groups, this happens at the end of processing. In the case when
112 there were back references that the ovector was too small to contain
113 (causing temporary malloc'd memory to be used during matching), and the
114 highest capturing number was not used, memory off the end of the ovector
115 was incorrectly being set to -1. (It was using the size of the temporary
116 memory instead of the true size.)
117
118 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
119 ovector size, it uses memory at the end of the block that it has got.
120
121 29. Check for an overlong MARK name and give an error at compile time. The
122 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
123
124 30. JIT compiler update.
125
126 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
127 Rill for the patch.
128
129 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
130
131 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
132
133 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
134 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
135 was enabled.
136
137 35. Improve JIT code generation for greedy plus quantifier.
138
139 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
140 1 to "aa" instead of to an empty string. The bug affected repeated groups
141 that could potentially match an empty string.
142
143 37. Optimizing single character iterators in JIT.
144
145 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
146 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
147 codepoints that are too big for the mode are faulted, and in a UTF mode,
148 disallowed codepoints are also faulted.
149
150 39. If PCRE was compiled with UTF support, in three places in the DFA
151 matcher there was code that should only have been obeyed in UTF mode, but
152 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
153 processing when bytes with values greater than 127 were present. In 16-bit
154 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
155 both cases the values are those that cannot be the first data item in a UTF
156 character. The three items that might have provoked this were recursions,
157 possessively repeated groups, and atomic groups.
158
159 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
160 and pcregrep, because some OS require shared objects to be explicitly
161 passed to ld, causing the link step to fail if they are not.
162
163 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
164 mode, patterns that started with \h* or \R* might be incorrectly matched.
165
166
167 Version 8.30 04-February-2012
168 -----------------------------
169
170 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
171 name is defined in ctype.h.
172
173 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
174 only in quite long subpatterns.
175
176 3. Removed the function pcre_info(), which has been obsolete and deprecated
177 since it was replaced by pcre_fullinfo() in February 2000.
178
179 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
180 match a (*MARK), and the match failed at the start of the subject, a
181 reference to memory before the start of the subject could occur. This bug
182 was introduced by fix 17 of release 8.21.
183
184 5. A reference to an unset group with zero minimum repetition was giving
185 totally wrong answers (in non-JavaScript-compatibility mode). For example,
186 /(another)?(\1?)test/ matched against "hello world test". This bug was
187 introduced in release 8.13.
188
189 6. Add support for 16-bit character strings (a large amount of work involving
190 many changes and refactorings).
191
192 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
193 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
194 from a file.
195
196 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
197 rounding is not applied in this particular case).
198
199 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
200 if they appear, or are escaped, in patterns.
201
202 10. Get rid of a number of -Wunused-but-set-variable warnings.
203
204 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
205 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
206 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
207 also returns the mark "x". This bug applied to capturing parentheses,
208 non-capturing parentheses, and atomic parentheses. It also applied to some
209 assertions.
210
211 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
212 information out of configure.ac instead of relying on pcre.h.generic, which
213 is not stored in the repository.
214
215 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
216 -lreadline.
217
218 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
219
220 15. Applied Graycode's patch to put the top-level frame on the stack rather
221 than the heap when not using the stack for recursion. This gives a
222 performance improvement in many cases when recursion is not deep.
223
224 16. Experimental code added to "pcretest -C" to output the stack frame size.
225
226
227 Version 8.21 12-Dec-2011
228 ------------------------
229
230 1. Updating the JIT compiler.
231
232 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
233 are added as well.
234
235 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
236 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
237 calling _pcre_jit_exec. Some extra comments are added.
238
239 4. (*MARK) settings inside atomic groups that do not contain any capturing
240 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
241 was introduced by change 18 for 8.20.
242
243 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
244 ECMA-262 standard.
245
246 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
247 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
248 This bug was probably introduced by change 9 of 8.13.
249
250 7. While fixing 6 above, I noticed that a number of other items were being
251 incorrectly rejected as "not fixed length". This arose partly because newer
252 opcodes had not been added to the fixed-length checking code. I have (a)
253 corrected the bug and added tests for these items, and (b) arranged for an
254 error to occur if an unknown opcode is encountered while checking for fixed
255 length instead of just assuming "not fixed length". The items that were
256 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
257 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
258 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
259
260 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
261 being incorrectly compiled and would have given unpredicatble results.
262
263 9. A possessively repeated subpattern with minimum repeat count greater than
264 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
265 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
266 the first (A) could occur when it should not.
267
268 10. Add a cast and remove a redundant test from the code.
269
270 11. JIT should use pcre_malloc/pcre_free for allocation.
271
272 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
273 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
274 is anything other than /usr, -L is still shown).
275
276 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
277
278 14. Perl does not support \N without a following name in a [] class; PCRE now
279 also gives an error.
280
281 15. If a forward reference was repeated with an upper limit of around 2000,
282 it caused the error "internal error: overran compiling workspace". The
283 maximum number of forward references (including repeats) was limited by the
284 internal workspace, and dependent on the LINK_SIZE. The code has been
285 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
286 the default depends on LINK_SIZE. There is a new upper limit (for safety)
287 of around 200,000 forward references. While doing this, I also speeded up
288 the filling in of repeated forward references.
289
290 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
291 incorrectly expecting the subject to contain another "a" after the start.
292
293 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
294 in the match, the SKIP should be ignored. This was not happening; instead
295 the SKIP was being treated as NOMATCH. For patterns such as
296 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
297 tested.
298
299 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
300 now much more compatible with Perl, in particular in cases where the result
301 is a non-match for a non-anchored pattern. For example, if
302 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
303 "m", where previously it did not return a name. A side effect of this
304 change is that for partial matches, the last encountered mark name is
305 returned, as for non matches. A number of tests that were previously not
306 Perl-compatible have been moved into the Perl-compatible test files. The
307 refactoring has had the pleasing side effect of removing one argument from
308 the match() function, thus reducing its stack requirements.
309
310 19. If the /S+ option was used in pcretest to study a pattern using JIT,
311 subsequent uses of /S (without +) incorrectly behaved like /S+.
312
313 21. Retrieve executable code size support for the JIT compiler and fixing
314 some warnings.
315
316 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
317 not work when the shorter character appeared right at the end of the
318 subject string.
319
320 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
321 systems.
322
323 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
324 output it when the /M option is used in pcretest.
325
326 25. The CheckMan script was not being included in the distribution. Also, added
327 an explicit "perl" to run Perl scripts from the PrepareRelease script
328 because this is reportedly needed in Windows.
329
330 26. If study data was being save in a file and studying had not found a set of
331 "starts with" bytes for the pattern, the data written to the file (though
332 never used) was taken from uninitialized memory and so caused valgrind to
333 complain.
334
335 27. Updated RunTest.bat as provided by Sheri Pierce.
336
337 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
338
339 29. Computation of memory usage for the table of capturing group names was
340 giving an unnecessarily large value.
341
342
343 Version 8.20 21-Oct-2011
344 ------------------------
345
346 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
347 a POSIX class. After further experiments with Perl, which convinced me that
348 Perl has bugs and confusions, a closing square bracket is no longer allowed
349 in a POSIX name. This bug also affected patterns with classes that started
350 with full stops.
351
352 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
353 captured substring, but while checking the failing first alternative,
354 substring 1 is temporarily captured. If the output vector supplied to
355 pcre_exec() was not big enough for this capture, the yield of the function
356 was still zero ("insufficient space for captured substrings"). This cannot
357 be totally fixed without adding another stack variable, which seems a lot
358 of expense for a edge case. However, I have improved the situation in cases
359 such as /(a)(b)x|abc/ matched against "abc", where the return code
360 indicates that fewer than the maximum number of slots in the ovector have
361 been set.
362
363 3. Related to (2) above: when there are more back references in a pattern than
364 slots in the output vector, pcre_exec() uses temporary memory during
365 matching, and copies in the captures as far as possible afterwards. It was
366 using the entire output vector, but this conflicts with the specification
367 that only 2/3 is used for passing back captured substrings. Now it uses
368 only the first 2/3, for compatibility. This is, of course, another edge
369 case.
370
371 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
372 main code base, and can be used by building with --enable-jit. When this is
373 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
374 runtime --no-jit option is given.
375
376 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
377 ovector, the return from the function was zero, implying that there were
378 other matches that did not fit. The correct "exactly full" value is now
379 returned.
380
381 6. If a subpattern that was called recursively or as a subroutine contained
382 (*PRUNE) or any other control that caused it to give a non-standard return,
383 invalid errors such as "Error -26 (nested recursion at the same subject
384 position)" or even infinite loops could occur.
385
386 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
387 computing the minimum length on reaching *ACCEPT, and so ended up with the
388 wrong value of 1 rather than 0. Further investigation indicates that
389 computing a minimum subject length in the presence of *ACCEPT is difficult
390 (think back references, subroutine calls), and so I have changed the code
391 so that no minimum is registered for a pattern that contains *ACCEPT.
392
393 8. If (*THEN) was present in the first (true) branch of a conditional group,
394 it was not handled as intended. [But see 16 below.]
395
396 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
397 Sheri Pierce.
398
399 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
400 the first byte in a match must be "a".
401
402 11. Change 17 for 8.13 increased the recursion depth for patterns like
403 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
404 pattern contains any instances of (*THEN). If it does not, the old
405 optimizations are restored. It would be nice to do this on a per-group
406 basis, but at the moment that is not feasible.
407
408 12. In some environments, the output of pcretest -C is CRLF terminated. This
409 broke RunTest's code that checks for the link size. A single white space
410 character after the value is now allowed for.
411
412 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
413 For "fr", it uses the Windows-specific input and output files.
414
415 14. If (*THEN) appeared in a group that was called recursively or as a
416 subroutine, it did not work as intended. [But see next item.]
417
418 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
419 pattern fragments (but not containing any | characters). If A and B are
420 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
421 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
422 D. In other words, Perl considers parentheses that do not contain any |
423 characters to be part of a surrounding alternative, whereas PCRE was
424 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
425 differently. PCRE now behaves in the same way as Perl, except in the case
426 of subroutine/recursion calls such as (?1) which have in any case always
427 been different (but PCRE had them first :-).
428
429 16. Related to 15 above: Perl does not treat the | in a conditional group as
430 creating alternatives. Such a group is treated in the same way as an
431 ordinary group without any | characters when processing (*THEN). PCRE has
432 been changed to match Perl's behaviour.
433
434 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
435 RunGrepTest script failed.
436
437 18. Change 22 for version 13 caused atomic groups to use more stack. This is
438 inevitable for groups that contain captures, but it can lead to a lot of
439 stack use in large patterns. The old behaviour has been restored for atomic
440 groups that do not contain any capturing parentheses.
441
442 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
443 suppress the check for a minimum subject length at run time. (If it was
444 given to pcre_exec() or pcre_dfa_exec() it did work.)
445
446 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
447 fail to work when decoding hex characters in data strings in EBCDIC
448 environments.
449
450 21. It appears that in at least one Mac OS environment, the isxdigit() function
451 is implemented as a macro that evaluates to its argument more than once,
452 contravening the C 90 Standard (I haven't checked a later standard). There
453 was an instance in pcretest which caused it to go wrong when processing
454 \x{...} escapes in subject strings. The has been rewritten to avoid using
455 things like p++ in the argument of isxdigit().
456
457
458 Version 8.13 16-Aug-2011
459 ------------------------
460
461 1. The Unicode data tables have been updated to Unicode 6.0.0.
462
463 2. Two minor typos in pcre_internal.h have been fixed.
464
465 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
466 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
467 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
468
469 4. There were a number of related bugs in the code for matching backrefences
470 caselessly in UTF-8 mode when codes for the characters concerned were
471 different numbers of bytes. For example, U+023A and U+2C65 are an upper
472 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
473 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
474 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
475 2-byte code at the end of the subject (it thought there wasn't enough data
476 left).
477
478 5. Comprehensive information about what went wrong is now returned by
479 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
480 as the output vector has at least 2 elements. The offset of the start of
481 the failing character and a reason code are placed in the vector.
482
483 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
484 now returned is for the first byte of the failing character, instead of the
485 last byte inspected. This is an incompatible change, but I hope it is small
486 enough not to be a problem. It makes the returned offset consistent with
487 pcre_exec() and pcre_dfa_exec().
488
489 7. pcretest now gives a text phrase as well as the error number when
490 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
491 failure, the offset and reason code are output.
492
493 8. When \R was used with a maximizing quantifier it failed to skip backwards
494 over a \r\n pair if the subsequent match failed. Instead, it just skipped
495 back over a single character (\n). This seems wrong (because it treated the
496 two characters as a single entity when going forwards), conflicts with the
497 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
498 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
499 has been changed.
500
501 9. Some internal refactoring has changed the processing so that the handling
502 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
503 time (the PCRE_DOTALL option was changed this way some time ago: version
504 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
505 which was always a bit of a fudge. It also means that there is one less
506 argument for the match() function, which reduces its stack requirements
507 slightly. This change also fixes an incompatibility with Perl: the pattern
508 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
509
510 10. More internal refactoring has drastically reduced the number of recursive
511 calls to match() for possessively repeated groups such as (abc)++ when
512 using pcre_exec().
513
514 11. While implementing 10, a number of bugs in the handling of groups were
515 discovered and fixed:
516
517 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
518 (a|)*(?1) gave a compile-time internal error.
519 ((a|)+)+ did not notice that the outer group could match an empty string.
520 (^a|^)+ was not marked as anchored.
521 (.*a|.*)+ was not marked as matching at start or after a newline.
522
523 12. Yet more internal refactoring has removed another argument from the match()
524 function. Special calls to this function are now indicated by setting a
525 value in a variable in the "match data" data block.
526
527 13. Be more explicit in pcre_study() instead of relying on "default" for
528 opcodes that mean there is no starting character; this means that when new
529 ones are added and accidentally left out of pcre_study(), testing should
530 pick them up.
531
532 14. The -s option of pcretest has been documented for ages as being an old
533 synonym of -m (show memory usage). I have changed it to mean "force study
534 for every regex", that is, assume /S for every regex. This is similar to -i
535 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
536 using it. It makes it easier to run collections of tests with and without
537 study enabled, and thereby test pcre_study() more easily. All the standard
538 tests are now run with and without -s (but some patterns can be marked as
539 "never study" - see 20 below).
540
541 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
542 restoration of the capturing data to the outer values was not happening
543 correctly.
544
545 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
546 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
547 pattern had matched an empty string, and so incorrectly returned a no
548 match.
549
550 17. There was optimizing code for the last branch of non-capturing parentheses,
551 and also for the obeyed branch of a conditional subexpression, which used
552 tail recursion to cut down on stack usage. Unfortunately, now that there is
553 the possibility of (*THEN) occurring in these branches, tail recursion is
554 no longer possible because the return has to be checked for (*THEN). These
555 two optimizations have therefore been removed. [But see 8.20/11 above.]
556
557 18. If a pattern containing \R was studied, it was assumed that \R always
558 matched two bytes, thus causing the minimum subject length to be
559 incorrectly computed because \R can also match just one byte.
560
561 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
562 was incorrectly computed.
563
564 20. If /S is present twice on a test pattern in pcretest input, it now
565 *disables* studying, thereby overriding the use of -s on the command line
566 (see 14 above). This is necessary for one or two tests to keep the output
567 identical in both cases.
568
569 21. When (*ACCEPT) was used in an assertion that matched an empty string and
570 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
571
572 22. When an atomic group that contained a capturing parenthesis was
573 successfully matched, but the branch in which it appeared failed, the
574 capturing was not being forgotten if a higher numbered group was later
575 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
576 group 1 to "a", when in fact it should be unset. This applied to multi-
577 branched capturing and non-capturing groups, repeated or not, and also to
578 positive assertions (capturing in negative assertions does not happen
579 in PCRE) and also to nested atomic groups.
580
581 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
582 subject after a captured substring, to make it easier to tell which of a
583 number of identical substrings has been captured.
584
585 24. The way atomic groups are processed by pcre_exec() has been changed so that
586 if they are repeated, backtracking one repetition now resets captured
587 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
588 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
589 "aaa". Previously, it would have been "a". As part of this code
590 refactoring, the way recursive calls are handled has also been changed.
591
592 25. If an assertion condition captured any substrings, they were not passed
593 back unless some other capturing happened later. For example, if
594 (?(?=(a))a) was matched against "a", no capturing was returned.
595
596 26. When studying a pattern that contained subroutine calls or assertions,
597 the code for finding the minimum length of a possible match was handling
598 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
599 group 1 called group 2 while simultaneously a separate group 2 called group
600 1). A stack overflow occurred in this case. I have fixed this by limiting
601 the recursion depth to 10.
602
603 27. Updated RunTest.bat in the distribution to the version supplied by Tom
604 Fortmann. This supports explicit test numbers on the command line, and has
605 argument validation and error reporting.
606
607 28. An instance of \X with an unlimited repeat could fail if at any point the
608 first character it looked at was a mark character.
609
610 29. Some minor code refactoring concerning Unicode properties and scripts
611 should reduce the stack requirement of match() slightly.
612
613 30. Added the '=' option to pcretest to check the setting of unused capturing
614 slots at the end of the pattern, which are documented as being -1, but are
615 not included in the return count.
616
617 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
618 compiled something random. Now it gives a compile-time error (as does
619 Perl).
620
621 32. A *MARK encountered during the processing of a positive assertion is now
622 recorded and passed back (compatible with Perl).
623
624 33. If --only-matching or --colour was set on a pcregrep call whose pattern
625 had alternative anchored branches, the search for a second match in a line
626 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
627 matched the line "0102" twice. The same bug affected patterns that started
628 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
629 twice.
630
631 34. Previously, PCRE did not allow quantification of assertions. However, Perl
632 does, and because of capturing effects, quantifying parenthesized
633 assertions may at times be useful. Quantifiers are now allowed for
634 parenthesized assertions.
635
636 35. A minor code tidy in pcre_compile() when checking options for \R usage.
637
638 36. \g was being checked for fancy things in a character class, when it should
639 just be a literal "g".
640
641 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
642 appearance of a nested POSIX class supersedes an apparent external class.
643 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
644 unescaped square brackets may also appear as part of class names. For
645 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
646 more like Perl. (But see 8.20/1 above.)
647
648 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
649 was because it thought it was \N{name}, which is not supported).
650
651 39. Add minix to OS list not supporting the -S option in pcretest.
652
653 40. PCRE tries to detect cases of infinite recursion at compile time, but it
654 cannot analyze patterns in sufficient detail to catch mutual recursions
655 such as ((?1))((?2)). There is now a runtime test that gives an error if a
656 subgroup is called recursively as a subpattern for a second time at the
657 same position in the subject string. In previous releases this might have
658 been caught by the recursion limit, or it might have run out of stack.
659
660 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
661 happen only once. PCRE was, however incorrectly giving a compile time error
662 "recursive call could loop indefinitely" because it cannot analyze the
663 pattern in sufficient detail. The compile time test no longer happens when
664 PCRE is compiling a conditional subpattern, but actual runaway loops are
665 now caught at runtime (see 40 above).
666
667 42. It seems that Perl allows any characters other than a closing parenthesis
668 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
669 has been changed to be the same.
670
671 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
672 as not to get warnings when autogen.sh is called. Also changed
673 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
674
675 44. To help people who use pcregrep to scan files containing exceedingly long
676 lines, the following changes have been made:
677
678 (a) The default value of the buffer size parameter has been increased from
679 8K to 20K. (The actual buffer used is three times this size.)
680
681 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
682 PCRE is built.
683
684 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
685 to be set at run time.
686
687 (d) Numerical values in pcregrep options can be followed by K or M, for
688 example --buffer-size=50K.
689
690 (e) If a line being scanned overflows pcregrep's buffer, an error is now
691 given and the return code is set to 2.
692
693 45. Add a pointer to the latest mark to the callout data block.
694
695 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
696 partial match of an empty string instead of no match. This was specific to
697 the use of ".".
698
699 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
700 complete match instead of a partial match. This bug was dependent on both
701 the PCRE_UTF8 and PCRE_DOTALL options being set.
702
703 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
704 starting byte set, because \b was not being ignored.
705
706
707 Version 8.12 15-Jan-2011
708 ------------------------
709
710 1. Fixed some typos in the markup of the man pages, and wrote a script that
711 checks for such things as part of the documentation building process.
712
713 2. On a big-endian 64-bit system, pcregrep did not correctly process the
714 --match-limit and --recursion-limit options (added for 8.11). In
715 particular, this made one of the standard tests fail. (The integer value
716 went into the wrong half of a long int.)
717
718 3. If the --colour option was given to pcregrep with -v (invert match), it
719 did strange things, either producing crazy output, or crashing. It should,
720 of course, ignore a request for colour when reporting lines that do not
721 match.
722
723 4. Another pcregrep bug caused similar problems if --colour was specified with
724 -M (multiline) and the pattern match finished with a line ending.
725
726 5. In pcregrep, when a pattern that ended with a literal newline sequence was
727 matched in multiline mode, the following line was shown as part of the
728 match. This seems wrong, so I have changed it.
729
730 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
731 the check for further matches in the same line (so they could be coloured)
732 to overrun the end of the current line. If another match was found, it was
733 incorrectly shown (and then shown again when found in the next line).
734
735 7. If pcregrep was compiled under Windows, there was a reference to the
736 function pcregrep_exit() before it was defined. I am assuming this was
737 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
738 reported by a user. I've moved the definition above the reference.
739
740
741 Version 8.11 10-Dec-2010
742 ------------------------
743
744 1. (*THEN) was not working properly if there were untried alternatives prior
745 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
746 backtracked to try for "b" instead of moving to the next alternative branch
747 at the same level (in this case, to look for "c"). The Perl documentation
748 is clear that when (*THEN) is backtracked onto, it goes to the "next
749 alternative in the innermost enclosing group".
750
751 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
752 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
753 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
754 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
755 (*THEN).
756
757 3. If \s appeared in a character class, it removed the VT character from
758 the class, even if it had been included by some previous item, for example
759 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
760 of \s, but is part of the POSIX "space" class.)
761
762 4. A partial match never returns an empty string (because you can always
763 match an empty string at the end of the subject); however the checking for
764 an empty string was starting at the "start of match" point. This has been
765 changed to the "earliest inspected character" point, because the returned
766 data for a partial match starts at this character. This means that, for
767 example, /(?<=abc)def/ gives a partial match for the subject "abc"
768 (previously it gave "no match").
769
770 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
771 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
772 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
773 has an implication that the given string is incomplete (because a partial
774 match is preferred over a full match). For this reason, these items now
775 give a partial match in this situation. [Aside: previously, the one case
776 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
777 match rather than a full match, which was wrong by the old rules, but is
778 now correct.]
779
780 6. There was a bug in the handling of #-introduced comments, recognized when
781 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
782 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
783 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
784 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
785 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
786 places in pcre_compile().
787
788 7. Related to (6) above, when pcre_compile() was skipping #-introduced
789 comments when looking ahead for named forward references to subpatterns,
790 the only newline sequence it recognized was NL. It now handles newlines
791 according to the set newline convention.
792
793 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
794 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
795 cater for a lack of strerror(). These oversights have been fixed.
796
797 9. Added --match-limit and --recursion-limit to pcregrep.
798
799 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
800
801 11. When the -o option was used, pcregrep was setting a return code of 1, even
802 when matches were found, and --line-buffered was not being honoured.
803
804 12. Added an optional parentheses number to the -o and --only-matching options
805 of pcregrep.
806
807 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
808 can match an empty string. The code to do it in pcretest and pcredemo
809 needed fixing:
810
811 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
812 only one byte after an empty string match just before CRLF (this case
813 just got forgotten; "any" and "anycrlf" were OK).
814
815 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
816 mode when an empty string match preceded an ASCII character followed by
817 a non-ASCII character. (The code for advancing by one character rather
818 than one byte was nonsense.)
819
820 (c) The pcredemo.c sample program did not have any code at all to handle
821 the cases when CRLF is a valid newline sequence.
822
823 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
824 as a starting offset was within the subject string. There is now a new
825 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
826 negative or greater than the length of the string. In order to test this,
827 pcretest is extended to allow the setting of negative starting offsets.
828
829 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
830 starting offset points to the beginning of a UTF-8 character was
831 unnecessarily clumsy. I tidied it up.
832
833 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
834 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
835
836 17. Nobody had reported that the --include_dir option, which was added in
837 release 7.7 should have been called --include-dir (hyphen, not underscore)
838 for compatibility with GNU grep. I have changed it to --include-dir, but
839 left --include_dir as an undocumented synonym, and the same for
840 --exclude-dir, though that is not available in GNU grep, at least as of
841 release 2.5.4.
842
843 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
844 characters from a string of bytes) have been redefined so as not to use
845 loops, in order to improve performance in some environments. At the same
846 time, I abstracted some of the common code into auxiliary macros to save
847 repetition (this should not affect the compiled code).
848
849 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
850 compile-time error is now given if \c is not followed by an ASCII
851 character, that is, a byte less than 128. (In EBCDIC mode, the code is
852 different, and any byte value is allowed.)
853
854 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
855 START_OPTIMIZE option, which is now allowed at compile time - but just
856 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
857 to pcregrep and other applications that have no direct access to PCRE
858 options. The new /Y option in pcretest sets this option when calling
859 pcre_compile().
860
861 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
862 back references. Groups containing recursive back references were forced to
863 be atomic by that change, but in the case of named groups, the amount of
864 memory required was incorrectly computed, leading to "Failed: internal
865 error: code overflow". This has been fixed.
866
867 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
868 pcretest.c, to avoid build problems in some Borland environments.
869
870
871 Version 8.10 25-Jun-2010
872 ------------------------
873
874 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
875 THEN.
876
877 2. (*ACCEPT) was not working when inside an atomic group.
878
879 3. Inside a character class, \B is treated as a literal by default, but
880 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
881 causes the error). The code is unchanged, but I tidied the documentation.
882
883 4. Inside a character class, PCRE always treated \R and \X as literals,
884 whereas Perl faults them if its -w option is set. I have changed PCRE so
885 that it faults them when PCRE_EXTRA is set.
886
887 5. Added support for \N, which always matches any character other than
888 newline. (It is the same as "." when PCRE_DOTALL is not set.)
889
890 6. When compiling pcregrep with newer versions of gcc which may have
891 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
892 declared with attribute warn_unused_result" were given. Just casting the
893 result to (void) does not stop the warnings; a more elaborate fudge is
894 needed. I've used a macro to implement this.
895
896 7. Minor change to pcretest.c to avoid a compiler warning.
897
898 8. Added four artifical Unicode properties to help with an option to make
899 \s etc use properties (see next item). The new properties are: Xan
900 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
901
902 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
903 use Unicode properties. (*UCP) at the start of a pattern can be used to set
904 this option. Modified pcretest to add /W to test this facility. Added
905 REG_UCP to make it available via the POSIX interface.
906
907 10. Added --line-buffered to pcregrep.
908
909 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
910 studied, and the match started with a letter with a code point greater than
911 127 whose first byte was different to the first byte of the other case of
912 the letter, the other case of this starting letter was not recognized
913 (#976).
914
915 12. If a pattern that was studied started with a repeated Unicode property
916 test, for example, \p{Nd}+, there was the theoretical possibility of
917 setting up an incorrect bitmap of starting bytes, but fortunately it could
918 not have actually happened in practice until change 8 above was made (it
919 added property types that matched character-matching opcodes).
920
921 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
922 possible starting bytes for non-anchored patterns.
923
924 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
925 \R, and also a number of cases that involve Unicode properties, both
926 explicit and implicit when PCRE_UCP is set.
927
928 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
929 input, it could crash or give wrong results if characters with values
930 greater than 0xc0 were present in the subject string. (Detail: it assumed
931 UTF-8 input when processing these items.)
932
933 16. Added a lot of (int) casts to avoid compiler warnings in systems where
934 size_t is 64-bit (#991).
935
936 17. Added a check for running out of memory when PCRE is compiled with
937 --disable-stack-for-recursion (#990).
938
939 18. If the last data line in a file for pcretest does not have a newline on
940 the end, a newline was missing in the output.
941
942 19. The default pcre_chartables.c file recognizes only ASCII characters (values
943 less than 128) in its various bitmaps. However, there is a facility for
944 generating tables according to the current locale when PCRE is compiled. It
945 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
946 characters, are recognized by isspace() and therefore were getting set in
947 these tables, and indeed these tables seem to approximate to ISO 8859. This
948 caused a problem in UTF-8 mode when pcre_study() was used to create a list
949 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
950 which of course cannot start UTF-8 characters. I have changed the code so
951 that only real ASCII characters (less than 128) and the correct starting
952 bytes for UTF-8 encodings are set for characters greater than 127 when in
953 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
954 altogether.)
955
956 20. Added the /T option to pcretest so as to be able to run tests with non-
957 standard character tables, thus making it possible to include the tests
958 used for 19 above in the standard set of tests.
959
960 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
961 reference to a subpattern the other side of a comment that contains an
962 opening parenthesis caused either an internal compiling error, or a
963 reference to the wrong subpattern.
964
965
966 Version 8.02 19-Mar-2010
967 ------------------------
968
969 1. The Unicode data tables have been updated to Unicode 5.2.0.
970
971 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
972 configured.
973
974 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
975 original author of that file, following a query about its status.
976
977 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
978 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
979
980 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
981 quantifier applied to a forward-referencing subroutine call, could compile
982 incorrect code or give the error "internal error: previously-checked
983 referenced subpattern not found".
984
985 6. Both MS Visual Studio and Symbian OS have problems with initializing
986 variables to point to external functions. For these systems, therefore,
987 pcre_malloc etc. are now initialized to local functions that call the
988 relevant global functions.
989
990 7. There were two entries missing in the vectors called coptable and poptable
991 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
992 I've fixed the data, and added a kludgy way of testing at compile time that
993 the lengths are correct (equal to the number of opcodes).
994
995 8. Following on from 7, I added a similar kludge to check the length of the
996 eint vector in pcreposix.c.
997
998 9. Error texts for pcre_compile() are held as one long string to avoid too
999 much relocation at load time. To find a text, the string is searched,
1000 counting zeros. There was no check for running off the end of the string,
1001 which could happen if a new error number was added without updating the
1002 string.
1003
1004 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1005
1006 11. \K was not working if it appeared in an atomic group or in a group that
1007 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1008 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1009 the assertion is positive, but not if it is negative.
1010
1011 12. Change 11 fortuitously reduced the size of the stack frame used in the
1012 "match()" function of pcre_exec.c by one pointer. Forthcoming
1013 implementation of support for (*MARK) will need an extra pointer on the
1014 stack; I have reserved it now, so that the stack frame size does not
1015 decrease.
1016
1017 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1018 item in branch that calls a recursion is a subroutine call - as in the
1019 second branch in the above example - was incorrectly given the compile-
1020 time error "recursive call could loop indefinitely" because pcre_compile()
1021 was not correctly checking the subroutine for matching a non-empty string.
1022
1023 14. The checks for overrunning compiling workspace could trigger after an
1024 overrun had occurred. This is a "should never occur" error, but it can be
1025 triggered by pathological patterns such as hundreds of nested parentheses.
1026 The checks now trigger 100 bytes before the end of the workspace.
1027
1028 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1029
1030
1031 Version 8.01 19-Jan-2010
1032 ------------------------
1033
1034 1. If a pattern contained a conditional subpattern with only one branch (in
1035 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1036 computed the wrong minimum data length (which is of course zero for such
1037 subpatterns). This could cause incorrect "no match" results.
1038
1039 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1040 the pattern is reset in the first branch, pcre_compile() failed with
1041 "internal error: code overflow at offset...". This happened only when
1042 the reset was to the original external option setting. (An optimization
1043 abstracts leading options settings into an external setting, which was the
1044 cause of this.)
1045
1046 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1047 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1048 assertion pattern did not match (meaning that the assertion was true), it
1049 was incorrectly treated as false if the SKIP had been reached during the
1050 matching. This also applied to assertions used as conditions.
1051
1052 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1053 assertion subpattern, including such a pattern used as a condition,
1054 unpredictable results occurred, instead of the error return
1055 PCRE_ERROR_DFA_UITEM.
1056
1057 5. The C++ GlobalReplace function was not working like Perl for the special
1058 situation when an empty string is matched. It now does the fancy magic
1059 stuff that is necessary.
1060
1061 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1062 removed. (These were left over from very, very early versions of PCRE.)
1063
1064 7. Some cosmetic changes to the code to make life easier when compiling it
1065 as part of something else:
1066
1067 (a) Change DEBUG to PCRE_DEBUG.
1068
1069 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1070 called "current" as "current_branch", to prevent a collision with the
1071 Linux macro when compiled as a kernel module.
1072
1073 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1074 prevent a collision with the Linux macro when compiled as a kernel
1075 module.
1076
1077 8. In pcre_compile() there are some checks for integer overflows that used to
1078 cast potentially large values to (double). This has been changed to that
1079 when building, a check for int64_t is made, and if it is found, it is used
1080 instead, thus avoiding the use of floating point arithmetic. (There is no
1081 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1082 double.
1083
1084 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1085 2005 (difference between two addresses compared to an unsigned value).
1086
1087 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1088 custom one, because of the following reported problem in Windows:
1089
1090 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1091 under Win32.
1092 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1093 therefore missing the function definition.
1094 - The compiler thus generates a "C" signature for the test function.
1095 - The linker fails to find the "C" function.
1096 - PCRE fails to configure if asked to do so against libbz2.
1097
1098 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1099 messages were output:
1100
1101 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1102 rerunning libtoolize, to keep the correct libtool macros in-tree.
1103 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1104
1105 I have done both of these things.
1106
1107 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1108 most of the time, it *can* run out if it is given a pattern that contains a
1109 runaway infinite recursion. I updated the discussion in the pcrestack man
1110 page.
1111
1112 13. Now that we have gone to the x.xx style of version numbers, the minor
1113 version may start with zero. Using 08 or 09 is a bad idea because users
1114 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1115 interpreted as invalid octal numbers. I've updated the previous comment in
1116 configure.ac, and also added a check that gives an error if 08 or 09 are
1117 used.
1118
1119 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1120 causing partial matching to fail when the end of the subject matched \W
1121 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1122
1123 15. There were some discrepancies between the declarations in pcre_internal.h
1124 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1125 their definitions. The declarations used "const uschar *" and the
1126 definitions used USPTR. Even though USPTR is normally defined as "const
1127 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1128 reported that: "This difference in casting confuses some C++ compilers, for
1129 example, SunCC recognizes above declarations as different functions and
1130 generates broken code for hbpcre." I have changed the declarations to use
1131 USPTR.
1132
1133 16. GNU libtool is named differently on some systems. The autogen.sh script now
1134 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1135 (FreeBSD).
1136
1137 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1138 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1139 comment: "Figure out how to create a longlong from a string: strtoll and
1140 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1141 instance, but it only takes 2 args instead of 3!"
1142
1143 18. A subtle bug concerned with back references has been fixed by a change of
1144 specification, with a corresponding code fix. A pattern such as
1145 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1146 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1147 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1148 same bug. Such groups have to be quantified to be useful, or contained
1149 inside another quantified group. (If there's no repetition, the reference
1150 can never match.) The problem arises because, having left the group and
1151 moved on to the rest of the pattern, a later failure that backtracks into
1152 the group uses the captured value from the final iteration of the group
1153 rather than the correct earlier one. I have fixed this in PCRE by forcing
1154 any group that contains a reference to itself to be an atomic group; that
1155 is, there cannot be any backtracking into it once it has completed. This is
1156 similar to recursive and subroutine calls.
1157
1158
1159 Version 8.00 19-Oct-09
1160 ----------------------
1161
1162 1. The table for translating pcre_compile() error codes into POSIX error codes
1163 was out-of-date, and there was no check on the pcre_compile() error code
1164 being within the table. This could lead to an OK return being given in
1165 error.
1166
1167 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1168 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1169 in a Windows environment.
1170
1171 3. The pcregrep --count option prints the count for each file even when it is
1172 zero, as does GNU grep. However, pcregrep was also printing all files when
1173 --files-with-matches was added. Now, when both options are given, it prints
1174 counts only for those files that have at least one match. (GNU grep just
1175 prints the file name in this circumstance, but including the count seems
1176 more useful - otherwise, why use --count?) Also ensured that the
1177 combination -clh just lists non-zero counts, with no names.
1178
1179 4. The long form of the pcregrep -F option was incorrectly implemented as
1180 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1181 but it seems right to fix it, and I didn't think it was worth preserving
1182 the old behaviour.
1183
1184 5. The command line items --regex=pattern and --regexp=pattern were not
1185 recognized by pcregrep, which required --regex pattern or --regexp pattern
1186 (with a space rather than an '='). The man page documented the '=' forms,
1187 which are compatible with GNU grep; these now work.
1188
1189 6. No libpcreposix.pc file was created for pkg-config; there was just
1190 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1191
1192 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1193 when UCP support is not needed, by modifying the Python script that
1194 generates it from Unicode data files. This should not matter if the module
1195 is correctly used as a library, but I received one complaint about 50K of
1196 unwanted data. My guess is that the person linked everything into his
1197 program rather than using a library. Anyway, it does no harm.
1198
1199 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1200 was a minimum greater than 1 for a wide character in a possessive
1201 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1202 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1203 character. Chaos in the form of incorrect output or a compiling loop could
1204 result.
1205
1206 9. The restrictions on what a pattern can contain when partial matching is
1207 requested for pcre_exec() have been removed. All patterns can now be
1208 partially matched by this function. In addition, if there are at least two
1209 slots in the offset vector, the offset of the earliest inspected character
1210 for the match and the offset of the end of the subject are set in them when
1211 PCRE_ERROR_PARTIAL is returned.
1212
1213 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1214 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1215 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1216 and may be more useful for multi-segment matching.
1217
1218 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1219 used to be given if ever the end of the subject was reached; now it is
1220 given only if matching could not proceed because another character was
1221 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1222 string "Z", which now yields "no match" instead of "partial match". In the
1223 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1224 final character ended with (*FAIL).
1225
1226 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1227 if the pattern had a "must contain" character that was already found in the
1228 earlier partial match, unless partial matching was again requested. For
1229 example, with the pattern /dog.(body)?/, the "must contain" character is
1230 "g". If the first part-match was for the string "dog", restarting with
1231 "sbody" failed. This bug has been fixed.
1232
1233 13. The string returned by pcre_dfa_exec() after a partial match has been
1234 changed so that it starts at the first inspected character rather than the
1235 first character of the match. This makes a difference only if the pattern
1236 starts with a lookbehind assertion or \b or \B (\K is not supported by
1237 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1238 matching functions compatible, and I think it's the right thing to do.
1239
1240 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1241 so that the demonstration program is easily available in environments where
1242 PCRE has not been installed from source.
1243
1244 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1245 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1246 library.
1247
1248 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1249 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1250 is not the first non-POSIX option to be added. Clearly some people find
1251 these options useful.
1252
1253 17. If a caller to the POSIX matching function regexec() passes a non-zero
1254 value for nmatch with a NULL value for pmatch, the value of
1255 nmatch is forced to zero.
1256
1257 18. RunGrepTest did not have a test for the availability of the -u option of
1258 the diff command, as RunTest does. It now checks in the same way as
1259 RunTest, and also checks for the -b option.
1260
1261 19. If an odd number of negated classes containing just a single character
1262 interposed, within parentheses, between a forward reference to a named
1263 subpattern and the definition of the subpattern, compilation crashed with
1264 an internal error, complaining that it could not find the referenced
1265 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1266 [The bug was that it was starting one character too far in when skipping
1267 over the character class, thus treating the ] as data rather than
1268 terminating the class. This meant it could skip too much.]
1269
1270 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1271 /g option in pcretest when the pattern contains \K, which makes it possible
1272 to have an empty string match not at the start, even when the pattern is
1273 anchored. Updated pcretest and pcredemo to use this option.
1274
1275 21. If the maximum number of capturing subpatterns in a recursion was greater
1276 than the maximum at the outer level, the higher number was returned, but
1277 with unset values at the outer level. The correct (outer level) value is
1278 now given.
1279
1280 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1281 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1282 make it do so. The string so far is captured, making this feature
1283 compatible with Perl.
1284
1285 23. The tests have been re-organized, adding tests 11 and 12, to make it
1286 possible to check the Perl 5.10 features against Perl 5.10.
1287
1288 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1289 pattern matches a fixed length string. PCRE did not allow this; now it
1290 does. Neither allows recursion.
1291
1292 25. I finally figured out how to implement a request to provide the minimum
1293 length of subject string that was needed in order to match a given pattern.
1294 (It was back references and recursion that I had previously got hung up
1295 on.) This code has now been added to pcre_study(); it finds a lower bound
1296 to the length of subject needed. It is not necessarily the greatest lower
1297 bound, but using it to avoid searching strings that are too short does give
1298 some useful speed-ups. The value is available to calling programs via
1299 pcre_fullinfo().
1300
1301 26. While implementing 25, I discovered to my embarrassment that pcretest had
1302 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1303 study optimizations had never been tested with that matching function.
1304 Oops. What is worse, even when it was passed study data, there was a bug in
1305 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1306 were also very few tests of studied patterns with pcre_dfa_exec().
1307
1308 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1309 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1310 on the other side of the coin, they are no longer allowed to have different
1311 names, because these cannot be distinguished in PCRE, and this has caused
1312 confusion. (This is a difference from Perl.)
1313
1314 28. When duplicate subpattern names are present (necessarily with different
1315 numbers, as required by 27 above), and a test is made by name in a
1316 conditional pattern, either for a subpattern having been matched, or for
1317 recursion in such a pattern, all the associated numbered subpatterns are
1318 tested, and the overall condition is true if the condition is true for any
1319 one of them. This is the way Perl works, and is also more like the way
1320 testing by number works.
1321
1322
1323 Version 7.9 11-Apr-09
1324 ---------------------
1325
1326 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1327 (pcretest), all targets were linked against these libraries. This included
1328 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1329 libraries. This caused unwanted dependencies to be created. This problem
1330 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1331 pcretest is linked with readline.
1332
1333 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1334 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1335 moved outside it again, because FALSE and TRUE are already defined in AIX,
1336 but BOOL is not.
1337
1338 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1339 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1340
1341 4. The pcregrep documentation said spaces were inserted as well as colons (or
1342 hyphens) following file names and line numbers when outputting matching
1343 lines. This is not true; no spaces are inserted. I have also clarified the
1344 wording for the --colour (or --color) option.
1345
1346 5. In pcregrep, when --colour was used with -o, the list of matching strings
1347 was not coloured; this is different to GNU grep, so I have changed it to be
1348 the same.
1349
1350 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1351 each matching line was coloured. Now it goes on to look for further matches
1352 of any of the test patterns, which is the same behaviour as GNU grep.
1353
1354 7. A pattern that could match an empty string could cause pcregrep to loop; it
1355 doesn't make sense to accept an empty string match in pcregrep, so I have
1356 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1357 seems to be how GNU grep behaves.
1358
1359 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1360 start or after a newline", because the conditional assertion was not being
1361 correctly handled. The rule now is that both the assertion and what follows
1362 in the first alternative must satisfy the test.
1363
1364 9. If auto-callout was enabled in a pattern with a conditional group whose
1365 condition was an assertion, PCRE could crash during matching, both with
1366 pcre_exec() and pcre_dfa_exec().
1367
1368 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1369 used for matching.
1370
1371 11. Unicode property support in character classes was not working for
1372 characters (bytes) greater than 127 when not in UTF-8 mode.
1373
1374 12. Added the -M command line option to pcretest.
1375
1376 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1377
1378 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1379
1380 16. Added comments and documentation about mis-use of no_arg in the C++
1381 wrapper.
1382
1383 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1384 from Martin Jerabek that uses macro names for all relevant character and
1385 string constants.
1386
1387 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1388 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1389 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1390 these, but not everybody uses configure.
1391
1392 19. A conditional group that had only one branch was not being correctly
1393 recognized as an item that could match an empty string. This meant that an
1394 enclosing group might also not be so recognized, causing infinite looping
1395 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1396 with the subject "ab", where knowledge that the repeated group can match
1397 nothing is needed in order to break the loop.
1398
1399 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1400 exec(), but without supplying a callout function, matching went wrong.
1401
1402 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1403 leak if the size of the offset vector was greater than 30. When the vector
1404 is smaller, the saved offsets during recursion go onto a local stack
1405 vector, but for larger vectors malloc() is used. It was failing to free
1406 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1407 error, in fact).
1408
1409 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1410 heapframe that is used only when UTF-8 support is enabled. This caused no
1411 problem, but was untidy.
1412
1413 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1414 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1415 included within another project.
1416
1417 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1418 slightly modified by me:
1419
1420 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1421 not building pcregrep.
1422
1423 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1424 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1425
1426 25. Forward references, both numeric and by name, in patterns that made use of
1427 duplicate group numbers, could behave incorrectly or give incorrect errors,
1428 because when scanning forward to find the reference group, PCRE was not
1429 taking into account the duplicate group numbers. A pattern such as
1430 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1431
1432 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1433 the feature of a custom pointer more persuasive (as requested by a user).
1434
1435 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1436 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1437 pre-defined.
1438
1439 28. Added support for (*UTF8) at the start of a pattern.
1440
1441 29. Arrange for flags added by the "release type" setting in CMake to be shown
1442 in the configuration summary.
1443
1444
1445 Version 7.8 05-Sep-08
1446 ---------------------
1447
1448 1. Replaced UCP searching code with optimized version as implemented for Ad
1449 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1450 stage table and inline lookup instead of a function, giving speed ups of 2
1451 to 5 times on some simple patterns that I tested. Permission was given to
1452 distribute the MultiStage2.py script that generates the tables (it's not in
1453 the tarball, but is in the Subversion repository).
1454
1455 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1456 scripts.
1457
1458 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1459 a group with a zero qualifier. The result of the study could be incorrect,
1460 or the function might crash, depending on the pattern.
1461
1462 4. Caseless matching was not working for non-ASCII characters in back
1463 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1464 It now works when Unicode Property Support is available.
1465
1466 5. In pcretest, an escape such as \x{de} in the data was always generating
1467 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1468 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1469 truncation.
1470
1471 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1472
1473 7. Added two (int) casts to pcregrep when printing the difference of two
1474 pointers, in case they are 64-bit values.
1475
1476 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1477 test 2 if it fails.
1478
1479 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1480 and a #define of that name to empty if it is not externally set. This is to
1481 allow users of MSVC to set it if necessary.
1482
1483 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1484 the convenience functions in the pcre_get.c source file.
1485
1486 11. An option change at the start of a pattern that had top-level alternatives
1487 could cause overwriting and/or a crash. This command provoked a crash in
1488 some environments:
1489
1490 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1491
1492 This potential security problem was recorded as CVE-2008-2371.
1493
1494 12. For a pattern where the match had to start at the beginning or immediately
1495 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1496 pcre_dfa_exec() could read past the end of the passed subject if there was
1497 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1498 pcretest so that it places the subject at the end of its malloc-ed buffer.
1499
1500 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1501 exec() might read past the end of the data buffer in UTF-8 mode.
1502
1503 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1504 the data contained the byte 0x85 as part of a UTF-8 character within its
1505 first line. This applied both to normal and DFA matching.
1506
1507 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1508 /^[^d]*?$/8 failed to match "abc".
1509
1510 16. Added a missing copyright notice to pcrecpp_internal.h.
1511
1512 17. Make it more clear in the documentation that values returned from
1513 pcre_exec() in ovector are byte offsets, not character counts.
1514
1515 18. Tidied a few places to stop certain compilers from issuing warnings.
1516
1517 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1518 supplied by Stefan Weber. I made a further small update for 7.8 because
1519 there is a change of source arrangements: the pcre_searchfuncs.c module is
1520 replaced by pcre_ucd.c.
1521
1522
1523 Version 7.7 07-May-08
1524 ---------------------
1525
1526 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1527 a string to a long long, pretend we don't even have a long long." This is
1528 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1529
1530 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1531 pre-7.6 versions, which defined a global no_arg variable instead of putting
1532 it in the RE class. (See also #8 below.)
1533
1534 3. Remove a line of dead code, identified by coverity and reported by Nuno
1535 Lopes.
1536
1537 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1538
1539 (1) The include/exclude patterns were being applied to the whole pathnames
1540 of files, instead of just to the final components.
1541
1542 (2) If there was more than one level of directory, the subdirectories were
1543 skipped unless they satisfied the include/exclude conditions. This is
1544 inconsistent with GNU grep (and could even be seen as contrary to the
1545 pcregrep specification - which I improved to make it absolutely clear).
1546 The action now is always to scan all levels of directory, and just
1547 apply the include/exclude patterns to regular files.
1548
1549 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1550 --exclude_dir in the tests to avoid scanning .svn directories.
1551
1552 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1553 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1554 doesn't support NULs in patterns.
1555
1556 7. Added some missing "const"s to declarations of static tables in
1557 pcre_compile.c and pcre_dfa_exec.c.
1558
1559 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1560 caused by fix #2 above. (Subsequently also a second patch to fix the
1561 first patch. And a third patch - this was a messy problem.)
1562
1563 9. Applied Craig's patch to remove the use of push_back().
1564
1565 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1566 matching function regexec().
1567
1568 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1569 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1570 references. PCRE supports relative numbers with this syntax (I don't think
1571 Oniguruma does).
1572
1573 12. Previously, a group with a zero repeat such as (...){0} was completely
1574 omitted from the compiled regex. However, this means that if the group
1575 was called as a subroutine from elsewhere in the pattern, things went wrong
1576 (an internal error was given). Such groups are now left in the compiled
1577 pattern, with a new opcode that causes them to be skipped at execution
1578 time.
1579
1580 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1581 to the way PCRE behaves:
1582
1583 (a) A lone ] character is dis-allowed (Perl treats it as data).
1584
1585 (b) A back reference to an unmatched subpattern matches an empty string
1586 (Perl fails the current match path).
1587
1588 (c) A data ] in a character class must be notated as \] because if the
1589 first data character in a class is ], it defines an empty class. (In
1590 Perl it is not possible to have an empty class.) The empty class []
1591 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1592 The negative empty class [^] matches any one character, independently
1593 of the DOTALL setting.
1594
1595 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1596 non-existent subpattern following a character class starting with ']' and
1597 containing () gave an internal compiling error instead of "reference to
1598 non-existent subpattern". Fortunately, when the pattern did exist, the
1599 compiled code was correct. (When scanning forwards to check for the
1600 existencd of the subpattern, it was treating the data ']' as terminating
1601 the class, so got the count wrong. When actually compiling, the reference
1602 was subsequently set up correctly.)
1603
1604 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1605 it was being rejected as not supported by pcre_dfa_exec(), even though
1606 other assertions are supported. I have made pcre_dfa_exec() support
1607 (*FAIL).
1608
1609 16. The implementation of 13c above involved the invention of a new opcode,
1610 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1611 cannot be changed at match time, I realized I could make a small
1612 improvement to matching performance by compiling OP_ALLANY instead of
1613 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1614 on the OP_ANY path.
1615
1616 17. Compiling pcretest on Windows with readline support failed without the
1617 following two fixes: (1) Make the unistd.h include conditional on
1618 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1619
1620 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1621 ncurses library to be included for pcretest when ReadLine support is
1622 requested, but also to allow for it to be overridden. This patch came from
1623 Daniel Bergström.
1624
1625 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1626 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1627 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1628 spotting this.
1629
1630
1631 Version 7.6 28-Jan-08
1632 ---------------------
1633
1634 1. A character class containing a very large number of characters with
1635 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1636 overflow.
1637
1638 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1639 HAVE_LONG_LONG is not defined.
1640
1641 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1642 bring them up to date and include new features. This patch includes:
1643
1644 - Fixed PH's badly added libz and libbz2 support.
1645 - Fixed a problem with static linking.
1646 - Added pcredemo. [But later removed - see 7 below.]
1647 - Fixed dftables problem and added an option.
1648 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1649 HAVE_LONG_LONG.
1650 - Added readline support for pcretest.
1651 - Added an listing of the option settings after cmake has run.
1652
1653 4. A user submitted a patch to Makefile that makes it easy to create
1654 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1655 Makefile.am that cause it to include this special target, without
1656 affecting anything else. Note that the same mingw target plus all
1657 the other distribution libraries and programs are now supported
1658 when configuring with CMake (see 6 below) instead of with
1659 Configure/Make.
1660
1661 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1662 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1663 exported in the Windows port". It has not yet been confirmed that the patch
1664 solves the problem, but it does no harm.
1665
1666 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1667 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1668 with CMake, and also correct the comment about stack recursion.
1669
1670 7. Remove the automatic building of pcredemo from the ./configure system and
1671 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1672 of a program that users should build themselves after PCRE is installed, so
1673 building it automatically is not really right. What is more, it gave
1674 trouble in some build environments.
1675
1676 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1677
1678
1679 Version 7.5 10-Jan-08
1680 ---------------------
1681
1682 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1683 values in parens when parsing an RE using the C++ wrapper."
1684
1685 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1686 Characters greater than 255 were excluded from the class instead of being
1687 included.
1688
1689 3. The same bug as (2) above applied to negated POSIX classes such as
1690 [:^space:].
1691
1692 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1693 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1694 I have changed it.
1695
1696 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1697 first named subpattern) and a construct such as (?&a) would reference the
1698 first named subpattern whose name started with "a" (in other words, the
1699 length check was missing). Both these problems are fixed. "Subpattern name
1700 expected" is now given for (?&) (a zero-length name), and this patch also
1701 makes it give the same error for \k'' (previously it complained that that
1702 was a reference to a non-existent subpattern).
1703
1704 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1705 this is right because (?- can be followed by option settings as well as by
1706 digits. I have, however, made the messages clearer.
1707
1708 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1709 than the number used in the conditional) now cause a compile-time error.
1710 This is actually not compatible with Perl, which accepts such patterns, but
1711 treats the conditional as always being FALSE (as PCRE used to), but it
1712 seems to me that giving a diagnostic is better.
1713
1714 8. Change "alphameric" to the more common word "alphanumeric" in comments
1715 and messages.
1716
1717 9. Fix two occurrences of "backslash" in comments that should have been
1718 "backspace".
1719
1720 10. Remove two redundant lines of code that can never be obeyed (their function
1721 was moved elsewhere).
1722
1723 11. The program that makes PCRE's Unicode character property table had a bug
1724 which caused it to generate incorrect table entries for sequences of
1725 characters that have the same character type, but are in different scripts.
1726 It amalgamated them into a single range, with the script of the first of
1727 them. In other words, some characters were in the wrong script. There were
1728 thirteen such cases, affecting characters in the following ranges:
1729
1730 U+002b0 - U+002c1
1731 U+0060c - U+0060d
1732 U+0061e - U+00612
1733 U+0064b - U+0065e
1734 U+0074d - U+0076d
1735 U+01800 - U+01805
1736 U+01d00 - U+01d77
1737 U+01d9b - U+01dbf
1738 U+0200b - U+0200f
1739 U+030fc - U+030fe
1740 U+03260 - U+0327f
1741 U+0fb46 - U+0fbb1
1742 U+10450 - U+1049d
1743
1744 12. The -o option (show only the matching part of a line) for pcregrep was not
1745 compatible with GNU grep in that, if there was more than one match in a
1746 line, it showed only the first of them. It now behaves in the same way as
1747 GNU grep.
1748
1749 13. If the -o and -v options were combined for pcregrep, it printed a blank
1750 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1751 does the same. The return code can be used to tell if there were any
1752 non-matching lines.
1753
1754 14. Added --file-offsets and --line-offsets to pcregrep.
1755
1756 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1757 infinitely looping recursion. The bug was that positive lookaheads were not
1758 being skipped when checking for a possible empty match (negative lookaheads
1759 and both kinds of lookbehind were skipped).
1760
1761 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1762 inclusion of <windows.h> to before rather than after the definition of
1763 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1764
1765 17. Specifying a possessive quantifier with a specific limit for a Unicode
1766 character property caused pcre_compile() to compile bad code, which led at
1767 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1768 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1769 caused the error; without that there was no problem.
1770
1771 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1772
1773 19. Added --enable-pcretest-libreadline.
1774
1775 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1776 RE::GlobalReplace(). As a result, the number of replacements returned was
1777 double what it should be. I removed one of the increments, but Craig sent a
1778 later patch that removed the other one (the right fix) and added unit tests
1779 that check the return values (which was not done before).
1780
1781 21. Several CMake things:
1782
1783 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1784 the names libpcre and libpcreposix, not just pcre and pcreposix.
1785
1786 (2) The above change means that pcretest and pcregrep are now correctly
1787 linked with the newly-built libraries, not previously installed ones.
1788
1789 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1790
1791 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1792 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1793 UTF-8 newline character). The key issue is that the pattern starts .*;
1794 this means that the match must be either at the beginning, or after a
1795 newline. The bug was in the code for advancing after a failed match and
1796 checking that the new position followed a newline. It was not taking
1797 account of UTF-8 characters correctly.
1798
1799 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1800 character classes. PCRE was not treating the sequence [:...:] as a
1801 character class unless the ... were all letters. Perl, however, seems to
1802 allow any characters between [: and :], though of course it rejects as
1803 unknown any "names" that contain non-letters, because all the known class
1804 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1805 for example, whereas PCRE did not - it did not recognize a POSIX character
1806 class. This seemed a bit dangerous, so the code has been changed to be
1807 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1808 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1809 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1810 Perl does, and where it didn't before.
1811
1812 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1813 Windows environments %n is disabled by default.
1814
1815
1816 Version 7.4 21-Sep-07
1817 ---------------------
1818
1819 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1820 means that a class such as [\s] counted as "explicit reference to CR or
1821 LF". That isn't really right - the whole point of the change was to try to
1822 help when there was an actual mention of one of the two characters. So now
1823 the change happens only if \r or \n (or a literal CR or LF) character is
1824 encountered.
1825
1826 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1827 of both had grown to the point where there were only 3 bits left.
1828 Fortunately, there was spare space in the data structure, and so I have
1829 moved the internal flags into a new 16-bit field to free up more option
1830 bits.
1831
1832 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1833 but did not set the internal JCHANGED flag - either of these is enough to
1834 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1835 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1836 start sets both bits.
1837
1838 4. Added options (at build time, compile time, exec time) to change \R from
1839 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1840
1841 5. doc/pcresyntax.html was missing from the distribution.
1842
1843 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1844 compatibility, even though it is no longer used.
1845
1846 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1847 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1848 windows.h file is present (where different names are used). [This was
1849 reversed later after testing - see 16 below.]
1850
1851 8. Changed all #include <config.h> to #include "config.h". There were also
1852 some further <pcre.h> cases that I changed to "pcre.h".
1853
1854 9. When pcregrep was used with the --colour option, it missed the line ending
1855 sequence off the lines that it output.
1856
1857 10. It was pointed out to me that arrays of string pointers cause lots of
1858 relocations when a shared library is dynamically loaded. A technique of
1859 using a single long string with a table of offsets can drastically reduce
1860 these. I have refactored PCRE in four places to do this. The result is
1861 dramatic:
1862
1863 Originally: 290
1864 After changing UCP table: 187
1865 After changing error message table: 43
1866 After changing table of "verbs" 36
1867 After changing table of Posix names 22
1868
1869 Thanks to the folks working on Gregex for glib for this insight.
1870
1871 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1872 unicode-properties was also set.
1873
1874 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1875
1876 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1877 checked only for CRLF.
1878
1879 14. Added casts to pcretest.c to avoid compiler warnings.
1880
1881 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1882
1883 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1884 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1885 entirely. This removes changes made in 7 above.
1886
1887 17. The CMake files have been updated, and there is now more information about
1888 building with CMake in the NON-UNIX-USE document.
1889
1890
1891 Version 7.3 28-Aug-07
1892 ---------------------
1893
1894 1. In the rejigging of the build system that eventually resulted in 7.1, the
1895 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1896 brackets there is not right, since it causes compilers to look for an
1897 installed pcre.h, not the version that is in the source that is being
1898 compiled (which of course may be different). I have changed it back to:
1899
1900 #include "pcre.h"
1901
1902 I have a vague recollection that the change was concerned with compiling in
1903 different directories, but in the new build system, that is taken care of
1904 by the VPATH setting the Makefile.
1905
1906 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1907 when the subject happened to end in the byte 0x85 (e.g. if the last
1908 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1909 characters but of course it shouldn't be taken as a newline when it is part
1910 of another character. The bug was that, for an unlimited repeat of . in
1911 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1912 characters when looking for a newline.
1913
1914 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1915
1916 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1917 in debug output.
1918
1919 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1920 long printing in the pcrecpp unittest when running under MinGW.
1921
1922 6. ESC_K was left out of the EBCDIC table.
1923
1924 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1925 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1926 limit also applies to "virtual nesting" when a pattern is recursive, and in
1927 this case 1000 isn't so big. I have been able to remove this limit at the
1928 expense of backing off one optimization in certain circumstances. Normally,
1929 when pcre_exec() would call its internal match() function recursively and
1930 immediately return the result unconditionally, it uses a "tail recursion"
1931 feature to save stack. However, when a subpattern that can match an empty
1932 string has an unlimited repetition quantifier, it no longer makes this
1933 optimization. That gives it a stack frame in which to save the data for
1934 checking that an empty string has been matched. Previously this was taken
1935 from the 1000-entry workspace that had been reserved. So now there is no
1936 explicit limit, but more stack is used.
1937
1938 8. Applied Daniel's patches to solve problems with the import/export magic
1939 syntax that is required for Windows, and which was going wrong for the
1940 pcreposix and pcrecpp parts of the library. These were overlooked when this
1941 problem was solved for the main library.
1942
1943 9. There were some crude static tests to avoid integer overflow when computing
1944 the size of patterns that contain repeated groups with explicit upper
1945 limits. As the maximum quantifier is 65535, the maximum group length was
1946 set at 30,000 so that the product of these two numbers did not overflow a
1947 32-bit integer. However, it turns out that people want to use groups that
1948 are longer than 30,000 bytes (though not repeat them that many times).
1949 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1950 made it possible to implement the integer overflow checks in a much more
1951 dynamic way, which I have now done. The artificial limitation on group
1952 length has been removed - we now have only the limit on the total length of
1953 the compiled pattern, which depends on the LINK_SIZE setting.
1954
1955 10. Fixed a bug in the documentation for get/copy named substring when
1956 duplicate names are permitted. If none of the named substrings are set, the
1957 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1958 empty string.
1959
1960 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1961 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1962 because the ] is interpreted as the first data character and the
1963 terminating ] is not found. PCRE has been made compatible with Perl in this
1964 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1965 cause memory overwriting.
1966
1967 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1968 string has been matched (to stop an infinite loop). It was not recognizing
1969 a conditional subpattern that could match an empty string if that
1970 subpattern was within another subpattern. For example, it looped when
1971 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1972 condition was not nested. This bug has been fixed.
1973
1974 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1975 past the start of the subject in the presence of bytes with the top bit
1976 set, for example "\x8aBCD".
1977
1978 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1979 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1980
1981 14. Optimized (?!) to (*FAIL).
1982
1983 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1984 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1985 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1986 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1987 does: it's just the validity check that is more restrictive.
1988
1989 16. Inserted checks for integer overflows during escape sequence (backslash)
1990 processing, and also fixed erroneous offset values for syntax errors during
1991 backslash processing.
1992
1993 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1994 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1995
1996 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1997 caused an overrun.
1998
1999 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2000 something other than just ASCII characters) inside a group that had an
2001 unlimited repeat caused a loop at compile time (while checking to see
2002 whether the group could match an empty string).
2003
2004 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2005 [\P{Any}] did so. (Error in the code for printing property names.)
2006
2007 21. An orphan \E inside a character class could cause a crash.
2008
2009 22. A repeated capturing bracket such as (A)? could cause a wild memory
2010 reference during compilation.
2011
2012 23. There are several functions in pcre_compile() that scan along a compiled
2013 expression for various reasons (e.g. to see if it's fixed length for look
2014 behind). There were bugs in these functions when a repeated \p or \P was
2015 present in the pattern. These operators have additional parameters compared
2016 with \d, etc, and these were not being taken into account when moving along
2017 the compiled data. Specifically:
2018
2019 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2020 length.
2021
2022 (b) An item such as \pL+ within a repeated group could cause crashes or
2023 loops.
2024
2025 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2026 "reference to non-existent subpattern" error.
2027
2028 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2029
2030 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2031 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2032
2033 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2034
2035 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2036 character were causing crashes (broken optimization).
2037
2038 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2039 \p or \P) caused a compile-time loop.
2040
2041 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2042 break. For example, the unstudied pattern [\r\n]A does not match the string
2043 "\r\nA" because change 7.0/46 below moves the current point on by two
2044 characters after failing to match at the start. However, the pattern \nA
2045 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2046 the same is true. There doesn't seem any very clean way out of this, but
2047 what I have chosen to do makes the common cases work: PCRE now takes note
2048 of whether there can be an explicit match for \r or \n anywhere in the
2049 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2050 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2051 pattern has explicit CR or LF references.
2052
2053 29. Added (*CR) etc for changing newline setting at start of pattern.
2054
2055
2056 Version 7.2 19-Jun-07
2057 ---------------------
2058
2059 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2060 which is apparently normally available under Windows.
2061
2062 2. Re-jig the pcregrep tests with different newline settings in an attempt
2063 to make them independent of the local environment's newline setting.
2064
2065 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2066
2067 4. Some of the "internals" tests were previously cut out when the link size
2068 was not 2, because the output contained actual offsets. The recent new
2069 "Z" feature of pcretest means that these can be cut out, making the tests
2070 usable with all link sizes.
2071
2072 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2073 stack recursion. This gives a massive performance boost under BSD, but just
2074 a small improvement under Linux. However, it saves one field in the frame
2075 in all cases.
2076
2077 6. Added more features from the forthcoming Perl 5.10:
2078
2079 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2080 recursion call. It refers to the nth most recently opened parentheses.
2081
2082 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2083 to be opened parentheses.
2084
2085 (c) Conditions that refer to capturing parentheses can be specified
2086 relatively, for example, (?(-2)... or (?(+3)...
2087
2088 (d) \K resets the start of the current match so that everything before
2089 is not part of it.
2090
2091 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2092
2093 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2094 reference syntax.
2095
2096 (g) (?| introduces a group in which the numbering of parentheses in each
2097 alternative starts with the same number.
2098
2099 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2100
2101 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2102 PCRE_INFO_JCHANGED.
2103
2104 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2105 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2106 for detecting groups that can match an empty string.
2107
2108 9. A pattern with a very large number of alternatives (more than several
2109 hundred) was running out of internal workspace during the pre-compile
2110 phase, where pcre_compile() figures out how much memory will be needed. A
2111 bit of new cunning has reduced the workspace needed for groups with
2112 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2113 workspace instead of running out of the 4096 that are available.
2114
2115 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2116
2117 11. Applied patch from Google to remove an optimization that didn't quite work.
2118 The report of the bug said:
2119
2120 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2121 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2122 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2123
2124 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2125 it matched the wrong number of bytes.
2126
2127
2128 Version 7.1 24-Apr-07
2129 ---------------------
2130
2131 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2132 that is more "standard", making use of automake and other Autotools. There
2133 is some re-arrangement of the files and adjustment of comments consequent
2134 on this.
2135
2136 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2137 for recursive directory scanning broke on some systems because the files
2138 are not scanned in any specific order and on different systems the order
2139 was different. A call to "sort" has been inserted into RunGrepTest for the
2140 approprate test as a short-term fix. In the longer term there may be an
2141 alternative.
2142
2143 3. I had an email from Eric Raymond about problems translating some of PCRE's
2144 man pages to HTML (despite the fact that I distribute HTML pages, some
2145 people do their own conversions for various reasons). The problems
2146 concerned the use of low-level troff macros .br and .in. I have therefore
2147 removed all such uses from the man pages (some were redundant, some could
2148 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2149 HTML has been updated to handle .nf/.fi and to complain if it encounters
2150 .br or .in.
2151
2152 4. Updated comments in configure.ac that get placed in config.h.in and also
2153 arranged for config.h to be included in the distribution, with the name
2154 config.h.generic, for the benefit of those who have to compile without
2155 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2156
2157 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2158 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2159 makevp.bat for the current PCRE, using the additional files
2160 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2161
2162 6. A Windows user reported a minor discrepancy with test 2, which turned out
2163 to be caused by a trailing space on an input line that had got lost in his
2164 copy. The trailing space was an accident, so I've just removed it.
2165
2166 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2167 that is needed.
2168
2169 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2170 as "const" (a) because they are and (b) because it helps the PHP
2171 maintainers who have recently made a script to detect big data structures
2172 in the php code that should be moved to the .rodata section. I remembered
2173 to update Builducptable as well, so it won't revert if ucptable.h is ever
2174 re-created.
2175
2176 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2177 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2178 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2179 support is not required. This saves 1.5-2K of code, which is important in
2180 some applications.
2181
2182 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2183 so as not to refer to the tables, even though these functions will never be
2184 called when UTF-8 support is disabled. Otherwise there are problems with a
2185 shared library.
2186
2187 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2188
2189 (a) It was defining its arguments as char * instead of void *.
2190
2191 (b) It was assuming that all moves were upwards in memory; this was true
2192 a long time ago when I wrote it, but is no longer the case.
2193
2194 The emulated memove() is provided for those environments that have neither
2195 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2196 is clearly not the case, as these two bugs were recently reported.
2197
2198 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2199 and Detrail to create the HTML documentation, the .txt form of the man
2200 pages, and it removes trailing spaces from listed files. It also creates
2201 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2202 case, it wraps all the #defines with #ifndefs. This script should be run
2203 before "make dist".
2204
2205 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2206 with Unicode property support.
2207
2208 (a) For a maximizing quantifier, if the two different cases of the
2209 character were of different lengths in their UTF-8 codings (there are
2210 some cases like this - I found 11), and the matching function had to
2211 back up over a mixture of the two cases, it incorrectly assumed they
2212 were both the same length.
2213
2214 (b) When PCRE was configured to use the heap rather than the stack for
2215 recursion during matching, it was not correctly preserving the data for
2216 the other case of a UTF-8 character when checking ahead for a match
2217 while processing a minimizing repeat. If the check also involved
2218 matching a wide character, but failed, corruption could cause an
2219 erroneous result when trying to check for a repeat of the original
2220 character.
2221
2222 13. Some tidying changes to the testing mechanism:
2223
2224 (a) The RunTest script now detects the internal link size and whether there
2225 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2226 values substituted by "configure". (The RunGrepTest script already did
2227 this for UTF-8.) The configure.ac script no longer substitutes the
2228 relevant variables.
2229
2230 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2231 with length and offset values. This means that the output is different
2232 for different internal link sizes. Test 2 is skipped for link sizes
2233 other than 2 because of this, bypassing the problem. Unfortunately,
2234 there was also a test in test 3 (the locale tests) that used /B and
2235 failed for link sizes other than 2. Rather than cut the whole test out,
2236 I have added a new /Z option to pcretest that replaces the length and
2237 offset values with spaces. This is now used to make test 3 independent
2238 of link size. (Test 2 will be tidied up later.)
2239
2240 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2241 segmentation fault instead of returning the appropriate error message.
2242
2243 15. In multiline mode when the newline sequence was set to "any", the pattern
2244 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2245 This doesn't seem right; it now treats the CRLF combination as the line
2246 ending, and so does not match in that case. It's only a pattern such as ^$
2247 that would hit this one: something like ^ABC$ would have failed after \r
2248 and then tried again after \r\n.
2249
2250 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2251 in an attempt to make files that differ only in their line terminators
2252 compare equal. This works on Linux.
2253
2254 17. Under certain error circumstances pcregrep might try to free random memory
2255 as it exited. This is now fixed, thanks to valgrind.
2256
2257 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2258 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2259 was because its rules for how to advance for /g after matching an empty
2260 string at the end of a line did not allow for this case. They now check for
2261 it specially.
2262
2263 20. pcretest is supposed to handle patterns and data of any length, by
2264 extending its buffers when necessary. It was getting this wrong when the
2265 buffer for a data line had to be extended.
2266
2267 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2268 CRLF as a newline sequence.
2269
2270 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2271 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2272 I have nevertheless tidied it up.
2273
2274 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2275
2276 24. Added a man page for pcre-config.
2277
2278
2279 Version 7.0 19-Dec-06
2280 ---------------------
2281
2282 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2283 moving to gcc 4.1.1.
2284
2285 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2286 sys/time.h, which is documented as needed for this function. It doesn't
2287 seem to matter on Linux, but it showed up on some releases of OS X.
2288
2289 3. It seems that there are systems where bytes whose values are greater than
2290 127 match isprint() in the "C" locale. The "C" locale should be the
2291 default when a C program starts up. In most systems, only ASCII printing
2292 characters match isprint(). This difference caused the output from pcretest
2293 to vary, making some of the tests fail. I have changed pcretest so that:
2294
2295 (a) When it is outputting text in the compiled version of a pattern, bytes
2296 other than 32-126 are always shown as hex escapes.
2297
2298 (b) When it is outputting text that is a matched part of a subject string,
2299 it does the same, unless a different locale has been set for the match
2300 (using the /L modifier). In this case, it uses isprint() to decide.
2301
2302 4. Fixed a major bug that caused incorrect computation of the amount of memory
2303 required for a compiled pattern when options that changed within the
2304 pattern affected the logic of the preliminary scan that determines the
2305 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2306 that the computed length was too small. The symptoms of this bug were
2307 either the PCRE error "internal error: code overflow" from pcre_compile(),
2308 or a glibc crash with a message such as "pcretest: free(): invalid next
2309 size (fast)". Examples of patterns that provoked this bug (shown in
2310 pcretest format) are:
2311
2312 /(?-x: )/x
2313 /(?x)(?-x: \s*#\s*)/
2314 /((?i)[\x{c0}])/8
2315 /(?i:[\x{c0}])/8
2316
2317 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2318 is now done differently.
2319
2320 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2321 wrapper classes; (b) implement a new function in the C++ scanner that is
2322 more efficient than the old way of doing things because it avoids levels of
2323 recursion in the regex matching; (c) add a paragraph to the documentation
2324 for the FullMatch() function.
2325
2326 6. The escape sequence \n was being treated as whatever was defined as
2327 "newline". Not only was this contrary to the documentation, which states
2328 that \n is character 10 (hex 0A), but it also went horribly wrong when
2329 "newline" was defined as CRLF. This has been fixed.
2330
2331 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2332 was being set to -1 for the "end of line" case (supposedly a value that no
2333 character can have). Though this value is never used (the check for end of
2334 line is "zero bytes in current character"), it caused compiler complaints.
2335 I've changed it to 0xffffffff.
2336
2337 8. In pcre_version.c, the version string was being built by a sequence of
2338 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2339 string (as it is for production releases) called a macro with an empty
2340 argument. The C standard says the result of this is undefined. The gcc
2341 compiler treats it as an empty string (which was what was wanted) but it is
2342 reported that Visual C gives an error. The source has been hacked around to
2343 avoid this problem.
2344
2345 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2346 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2347 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2348 of them did).
2349
2350 10. Originally, pcretest opened its input and output without "b"; then I was
2351 told that "b" was needed in some environments, so it was added for release
2352 5.0 to both the input and output. (It makes no difference on Unix-like
2353 systems.) Later I was told that it is wrong for the input on Windows. I've
2354 now abstracted the modes into two macros, to make it easier to fiddle with
2355 them, and removed "b" from the input mode under Windows.
2356
2357 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2358
2359 12. Added -help and --help to pcretest as an official way of being reminded
2360 of the options.
2361
2362 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2363 and pcrecpp.cc because they annoy compilers at high warning levels.
2364
2365 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2366
2367 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2368 scripts are not C programs :-) and which was not noticed because it works
2369 on Linux.
2370
2371 16. pcretest is supposed to handle any length of pattern and data line (as one
2372 line or as a continued sequence of lines) by extending its input buffer if
2373 necessary. This feature was broken for very long pattern lines, leading to
2374 a string of junk being passed to pcre_compile() if the pattern was longer
2375 than about 50K.
2376
2377 17. I have done a major re-factoring of the way pcre_compile() computes the
2378 amount of memory needed for a compiled pattern. Previously, there was code
2379 that made a preliminary scan of the pattern in order to do this. That was
2380 OK when PCRE was new, but as the facilities have expanded, it has become
2381 harder and harder to keep it in step with the real compile phase, and there
2382 have been a number of bugs (see for example, 4 above). I have now found a
2383 cunning way of running the real compile function in a "fake" mode that
2384 enables it to compute how much memory it would need, while actually only
2385 ever using a few hundred bytes of working memory and without too many
2386 tests of the mode. This should make future maintenance and development
2387 easier. A side effect of this work is that the limit of 200 on the nesting
2388 depth of parentheses has been removed (though this was never a serious
2389 limitation, I suspect). However, there is a downside: pcre_compile() now
2390 runs more slowly than before (30% or more, depending on the pattern). I
2391 hope this isn't a big issue. There is no effect on runtime performance.
2392
2393 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2394 newline (only possible for the last line of a file) and it was a
2395 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2396
2397 19. Added additional timing features to pcretest. (1) The -tm option now times
2398 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2399 separate command line item, by a number that specifies the number of
2400 repeats to use when timing. The default is 50000; this gives better
2401 precision, but takes uncomfortably long for very large patterns.
2402
2403 20. Extended pcre_study() to be more clever in cases where a branch of a
2404 subpattern has no definite first character. For example, (a*|b*)[cd] would
2405 previously give no result from pcre_study(). Now it recognizes that the
2406 first character must be a, b, c, or d.
2407
2408 21. There was an incorrect error "recursive call could loop indefinitely" if
2409 a subpattern (or the entire pattern) that was being tested for matching an
2410 empty string contained only one non-empty item after a nested subpattern.
2411 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2412 incorrectly, because the \d was being skipped in the check.
2413
2414 22. The pcretest program now has a new pattern option /B and a command line
2415 option -b, which is equivalent to adding /B to every pattern. This causes
2416 it to show the compiled bytecode, without the additional information that
2417 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2418 is the same as /B/I).
2419
2420 23. A new optimization is now able automatically to treat some sequences such
2421 as a*b as a*+b. More specifically, if something simple (such as a character
2422 or a simple class like \d) has an unlimited quantifier, and is followed by
2423 something that cannot possibly match the quantified thing, the quantifier
2424 is automatically "possessified".
2425
2426 24. A recursive reference to a subpattern whose number was greater than 39
2427 went wrong under certain circumstances in UTF-8 mode. This bug could also
2428 have affected the operation of pcre_study().
2429
2430 25. Realized that a little bit of performance could be had by replacing
2431 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2432
2433 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2434
2435 27. Possessive quantifiers such as a++ were previously implemented by turning
2436 them into atomic groups such as ($>a+). Now they have their own opcodes,
2437 which improves performance. This includes the automatically created ones
2438 from 23 above.
2439
2440 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2441 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2442 the first matched character to be a colon. This applied both to named and
2443 numbered groups.
2444
2445 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2446
2447 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2448 building PCRE on Windows easier, so I have included it in the distribution.
2449
2450 31. There is now a check in pcretest against a ridiculously large number being
2451 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2452 loop, the loop is abandoned.
2453
2454 32. Forward references to subpatterns in conditions such as (?(2)...) where
2455 subpattern 2 is defined later cause pcre_compile() to search forwards in
2456 the pattern for the relevant set of parentheses. This search went wrong
2457 when there were unescaped parentheses in a character class, parentheses
2458 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2459
2460 33. "Subroutine" calls and backreferences were previously restricted to
2461 referencing subpatterns earlier in the regex. This restriction has now
2462 been removed.
2463
2464 34. Added a number of extra features that are going to be in Perl 5.10. On the
2465 whole, these are just syntactic alternatives for features that PCRE had
2466 previously implemented using the Python syntax or my own invention. The
2467 other formats are all retained for compatibility.
2468
2469 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2470 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2471 also .NET compatible.
2472
2473 (b) A recursion or subroutine call to a named group can now be defined as
2474 (?&name) as well as (?P>name).
2475
2476 (c) A backreference to a named group can now be defined as \k<name> or
2477 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2478 5.10, are also .NET compatible.
2479
2480 (d) A conditional reference to a named group can now use the syntax
2481 (?(<name>) or (?('name') as well as (?(name).
2482
2483 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2484 groups (named and numbered) that are never evaluated inline, but can be
2485 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2486 is always false. There may be only one alternative in such a group.
2487
2488 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2489 as the simple (?(R). The condition is true only if the most recent
2490 recursion is that of the given number or name. It does not search out
2491 through the entire recursion stack.
2492
2493 (g) The escape \gN or \g{N} has been added, where N is a positive or
2494 negative number, specifying an absolute or relative reference.
2495
2496 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2497 some "unreachable code" warnings.
2498
2499 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2500 things, this adds five new scripts.
2501
2502 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2503 There were also incompatibilities regarding the handling of \Q..\E inside
2504 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2505 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2506
2507 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2508 matches an empty string, and forcibly breaks the loop. There were bugs in
2509 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2510 against aaaa the result was just "a" rather than "aaaa", for example. Two
2511 separate and independent bugs (that affected different cases) have been
2512 fixed.
2513
2514 39. Refactored the code to abolish the use of different opcodes for small
2515 capturing bracket numbers. This is a tidy that I avoided doing when I
2516 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2517 The new approach is not only tidier, it makes it possible to reduce the
2518 memory needed to fix the previous bug (38).
2519
2520 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2521 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2522 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2523 mode.
2524
2525 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2526 report.
2527
2528 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2529 copy construction and assignment in the C++ wrapper.
2530
2531 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2532 couple of bugs that could have given wrong results in the "--newline=crlf"
2533 case.
2534
2535 44. Added a number of casts and did some reorganization of signed/unsigned int
2536 variables following suggestions from Dair Grant. Also renamed the variable
2537 "this" as "item" because it is a C++ keyword.
2538
2539 45. Arranged for dftables to add
2540
2541 #include "pcre_internal.h"
2542
2543 to pcre_chartables.c because without it, gcc 4.x may remove the array
2544 definition from the final binary if PCRE is built into a static library and
2545 dead code stripping is activated.
2546
2547 46. For an unanchored pattern, if a match attempt fails at the start of a
2548 newline sequence, and the newline setting is CRLF or ANY, and the next two
2549 characters are CRLF, advance by two characters instead of one.
2550
2551
2552 Version 6.7 04-Jul-06
2553 ---------------------
2554
2555 1. In order to handle tests when input lines are enormously long, pcretest has
2556 been re-factored so that it automatically extends its buffers when
2557 necessary. The code is crude, but this _is_ just a test program. The
2558 default size has been increased from 32K to 50K.
2559
2560 2. The code in pcre_study() was using the value of the re argument before
2561 testing it for NULL. (Of course, in any sensible call of the function, it
2562 won't be NULL.)
2563
2564 3. The memmove() emulation function in pcre_internal.h, which is used on
2565 systems that lack both memmove() and bcopy() - that is, hardly ever -
2566 was missing a "static" storage class specifier.
2567
2568 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2569 containing an extended class (one that cannot be represented by a bitmap
2570 because it contains high-valued characters or Unicode property items, e.g.
2571 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2572 pattern, but PCRE should not loop if you do not (it no longer does).
2573 [Detail: two cases were found: (a) a repeated subpattern containing an
2574 extended class; (b) a recursive reference to a subpattern that followed a
2575 previous extended class. It wasn't skipping over the extended class
2576 correctly when UTF-8 mode was not set.]
2577
2578 5. A negated single-character class was not being recognized as fixed-length
2579 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2580 compile error "lookbehind assertion is not fixed length".
2581
2582 6. The RunPerlTest auxiliary script was showing an unexpected difference
2583 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2584 write a Perl script that can interpret lines of an input file either as
2585 byte characters or as UTF-8, which is what "perltest" was being required to
2586 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2587 can't do is switch easily at run time between having the "use utf8;" pragma
2588 or not. In the end, I fudged it by using the RunPerlTest script to insert
2589 "use utf8;" explicitly for the UTF-8 tests.
2590
2591 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2592 the end of the subject string, contrary to the documentation and to what
2593 Perl does. This was true of both matching functions. Now it matches only at
2594 the start of the subject and immediately after *internal* newlines.
2595
2596 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2597 a pointer to an int instead of a pointer to an unsigned long int. This
2598 caused problems on 64-bit systems.
2599
2600 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2601 instance of the 'standard' template library not being so standard".
2602
2603 10. There was no check on the number of named subpatterns nor the maximum
2604 length of a subpattern name. The product of these values is used to compute
2605 the size of the memory block for a compiled pattern. By supplying a very
2606 long subpattern name and a large number of named subpatterns, the size
2607 computation could be caused to overflow. This is now prevented by limiting
2608 the length of names to 32 characters, and the number of named subpatterns
2609 to 10,000.
2610
2611 11. Subpatterns that are repeated with specific counts have to be replicated in
2612 the compiled pattern. The size of memory for this was computed from the
2613 length of the subpattern and the repeat count. The latter is limited to
2614 65535, but there was no limit on the former, meaning that integer overflow
2615 could in principle occur. The compiled length of a repeated subpattern is
2616 now limited to 30,000 bytes in order to prevent this.
2617
2618 12. Added the optional facility to have named substrings with the same name.
2619
2620 13. Added the ability to use a named substring as a condition, using the
2621 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2622 are numbers (not recommended). Forward references are permitted.
2623
2624 14. Added forward references in named backreferences (if you see what I mean).
2625
2626 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2627 pattern could run off the end of the subject. For example, the pattern
2628 "(?s)(.{1,5})"8 did this with the subject "ab".
2629
2630 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2631 PCRE_CASELESS was set when matching characters that were quantified with ?
2632 or *.
2633
2634 17. A character class other than a single negated character that had a minimum
2635 but no maximum quantifier - for example [ab]{6,} - was not handled
2636 correctly by pce_dfa_exec(). It would match only one character.
2637
2638 18. A valid (though odd) pattern that looked like a POSIX character
2639 class but used an invalid character after [ (for example [[,abc,]]) caused
2640 pcre_compile() to give the error "Failed: internal error: code overflow" or
2641 in some cases to crash with a glibc free() error. This could even happen if
2642 the pattern terminated after [[ but there just happened to be a sequence of
2643 letters, a binary zero, and a closing ] in the memory that followed.
2644
2645 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2646 over the years. Originally (before any Unicode support), just the bottom 8
2647 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2648 output from "man perlunicode" includes this:
2649
2650 The regular expression compiler produces polymorphic opcodes. That
2651 is, the pattern adapts to the data and automatically switches to
2652 the Unicode character scheme when presented with Unicode data--or
2653 instead uses a traditional byte scheme when presented with byte
2654 data.
2655
2656 Sadly, a wide octal escape does not cause a switch, and in a string with
2657 no other multibyte characters, these octal escapes are treated as before.
2658 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2659 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2660 Unicode string.
2661
2662 I have not perpetrated such confusion in PCRE. Up till now, it took just
2663 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2664 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2665 translate to the appropriate multibyte character.
2666
2667 29. Applied some refactoring to reduce the number of warnings from Microsoft
2668 and Borland compilers. This has included removing the fudge introduced
2669 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2670 a warning about an unused variable.
2671
2672 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2673 characters since release 4.0, because Perl (from release 5.004) does not.
2674 [Or at least, is documented not to: some releases seem to be in conflict
2675 with the documentation.] However, when a pattern was studied with
2676 pcre_study() and all its branches started with \s, PCRE still included VT
2677 as a possible starting character. Of course, this did no harm; it just
2678 caused an unnecessary match attempt.
2679
2680 22. Removed a now-redundant internal flag bit that recorded the fact that case
2681 dependency changed within the pattern. This was once needed for "required
2682 byte" processing, but is no longer used. This recovers a now-scarce options
2683 bit. Also moved the least significant internal flag bit to the most-
2684 significant bit of the word, which was not previously used (hangover from
2685 the days when it was an int rather than a uint) to free up another bit for
2686 the future.
2687
2688 23. Added support for CRLF line endings as well as CR and LF. As well as the
2689 default being selectable at build time, it can now be changed at runtime
2690 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2691 specify that it is scanning data with non-default line endings.
2692
2693 24. Changed the definition of CXXLINK to make it agree with the definition of
2694 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2695
2696 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2697 recursions. This makes a big different to stack usage for some patterns.
2698
2699 26. If a subpattern containing a named recursion or subroutine reference such
2700 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2701 the space required for the compiled pattern went wrong and gave too small a
2702 value. Depending on the environment, this could lead to "Failed: internal
2703 error: code overflow at offset 49" or "glibc detected double free or
2704 corruption" errors.
2705
2706 27. Applied patches from Google (a) to support the new newline modes and (b) to
2707 advance over multibyte UTF-8 characters in GlobalReplace.
2708
2709 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2710 difference for some implementation of PCRE in some Windows version.
2711
2712 29. Added some extra testing facilities to pcretest:
2713
2714 \q<number> in a data line sets the "match limit" value
2715 \Q<number> in a data line sets the "match recursion limt" value
2716 -S <number> sets the stack size, where <number> is in megabytes
2717
2718 The -S option isn't available for Windows.
2719
2720
2721 Version 6.6 06-Feb-06
2722 ---------------------
2723
2724 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2725 in pcreposix.h. I have copied the definition from pcre.h.
2726
2727 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2728 because pcre.h is no longer a built file.
2729
2730 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2731 not normally included in the compiled code.
2732
2733
2734 Version 6.5 01-Feb-06
2735 ---------------------
2736
2737 1. When using the partial match feature with pcre_dfa_exec(), it was not
2738 anchoring the second and subsequent partial matches at the new starting
2739 point. This could lead to incorrect results. For example, with the pattern
2740 /1234/, partially matching against "123" and then "a4" gave a match.
2741
2742 2. Changes to pcregrep:
2743
2744 (a) All non-match returns from pcre_exec() were being treated as failures
2745 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2746 error message is output. Some extra information is given for the
2747 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2748 probably the only errors that are likely to be caused by users (by
2749 specifying a regex that has nested indefinite repeats, for instance).
2750 If there are more than 20 of these errors, pcregrep is abandoned.
2751
2752 (b) A binary zero was treated as data while matching, but terminated the
2753 output line if it was written out. This has been fixed: binary zeroes
2754 are now no different to any other data bytes.
2755
2756 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2757 used to set a locale for matching. The --locale=xxxx long option has
2758 been added (no short equivalent) to specify a locale explicitly on the
2759 pcregrep command, overriding the environment variables.
2760
2761 (d) When -B was used with -n, some line numbers in the output were one less
2762 than they should have been.
2763
2764 (e) Added the -o (--only-matching) option.
2765
2766 (f) If -A or -C was used with -c (count only), some lines of context were
2767 accidentally printed for the final match.
2768
2769 (g) Added the -H (--with-filename) option.
2770
2771 (h) The combination of options -rh failed to suppress file names for files
2772 that were found from directory arguments.
2773
2774 (i) Added the -D (--devices) and -d (--directories) options.
2775
2776 (j) Added the -F (--fixed-strings) option.
2777
2778 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2779
2780 (l) Added the --colo(u)r option.
2781
2782 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2783 is not present by default.
2784
2785 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2786 items such as (?R) or (?1), when the recursion could match a number of
2787 alternatives. If it matched one of the alternatives, but subsequently,
2788 outside the recursion, there was a failure, the code tried to back up into
2789 the recursion. However, because of the way PCRE is implemented, this is not
2790 possible, and the result was an incorrect result from the match.
2791
2792 In order to prevent this happening, the specification of recursion has
2793 been changed so that all such subpatterns are automatically treated as
2794 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2795
2796 4. I had overlooked the fact that, in some locales, there are characters for
2797 which isalpha() is true but neither isupper() nor islower() are true. In
2798 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2799 and ordfeminine) are like this. This affected the treatment of \w and \W
2800 when they appeared in character classes, but not when they appeared outside
2801 a character class. The bit map for "word" characters is now created
2802 separately from the results of isalnum() instead of just taking it from the
2803 upper, lower, and digit maps. (Plus the underscore character, of course.)
2804
2805 5. The above bug also affected the handling of POSIX character classes such as
2806 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2807 permanent tables. Instead, the bit maps for such a class were previously
2808 created as the appropriate unions of the upper, lower, and digit bitmaps.
2809 Now they are created by subtraction from the [[:word:]] class, which has
2810 its own bitmap.
2811
2812 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2813 It is created by subtracting the vertical space characters (\x09, \x0a,
2814 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2815 subtraction was done in the overall bitmap for a character class, meaning
2816 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2817 be recognized. This bug has been fixed.
2818
2819 7. Patches from the folks at Google:
2820
2821 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2822 real life, but is still worth protecting against".
2823
2824 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2825 regular expressions".
2826
2827 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2828 have it.
2829
2830 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2831 "configure" and the latter not, in order to fix a problem somebody had
2832 with compiling the Arg class on HP-UX.
2833
2834 (e) Improve the error-handling of the C++ wrapper a little bit.
2835
2836 (f) New tests for checking recursion limiting.
2837
2838 8. The pcre_memmove() function, which is used only if the environment does not
2839 have a standard memmove() function (and is therefore rarely compiled),
2840 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2841 returning a result (though PCRE never actually uses the result).
2842
2843 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2844 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2845 returned instead of calling malloc() with an overflowing number that would
2846 most likely cause subsequent chaos.
2847
2848 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2849
2850 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2851 with this option is matched, the nmatch and pmatch options of regexec() are
2852 ignored.
2853
2854 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2855 provided in case anyone wants to the the POSIX interface with UTF-8
2856 strings.
2857
2858 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2859 C++ linking (needed for some HP-UX environments).
2860
2861 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2862 (unused parameter) and in the pcre_printint() function (omitted "default"
2863 switch label when the default is to do nothing).
2864
2865 15. Added some code to make it possible, when PCRE is compiled as a C++
2866 library, to replace subject pointers for pcre_exec() with a smart pointer
2867 class, thus making it possible to process discontinuous strings.
2868
2869 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2870 much the same function. They were added by different people who were trying
2871 to make PCRE easy to compile on non-Unix systems. It has been suggested
2872 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2873 for compiling on Windows systems. I have therefore replaced it with
2874 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2875 defaults to "extern" for C or "extern C" for C++, which works fine on
2876 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2877 SCOPE with something explicit in config.h. In addition:
2878
2879 (a) pcreposix.h still had just "extern" instead of either of these macros;
2880 I have replaced it with PCRE_DATA_SCOPE.
2881
2882 (b) Functions such as _pcre_xclass(), which are internal to the library,
2883 but external in the C sense, all had PCRE_EXPORT in their definitions.
2884 This is apparently wrong for the Windows case, so I have removed it.
2885 (It makes no difference on Unix-like systems.)
2886
2887 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2888 of recursive calls to match(). This is different to MATCH_LIMIT because
2889 that limits the total number of calls to match(), not all of which increase
2890 the depth of recursion. Limiting the recursion depth limits the amount of
2891 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2892 when PCRE is compiled, and changed at run time. A patch from Google adds
2893 this functionality to the C++ interface.
2894
2895 18. Changes to the handling of Unicode character properties:
2896
2897 (a) Updated the table to Unicode 4.1.0.
2898
2899 (b) Recognize characters that are not in the table as "Cn" (undefined).
2900
2901 (c) I revised the way the table is implemented to a much improved format
2902 which includes recognition of ranges. It now supports the ranges that
2903 are defined in UnicodeData.txt, and it also amalgamates other
2904 characters into ranges. This has reduced the number of entries in the
2905 table from around 16,000 to around 3,000, thus reducing its size
2906 considerably. I realized I did not need to use a tree structure after
2907 all - a binary chop search is just as efficient. Having reduced the
2908 number of entries, I extended their size from 6 bytes to 8 bytes to
2909 allow for more data.
2910
2911 (d) Added support for Unicode script names via properties such as \p{Han}.
2912
2913 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2914 matching that character.
2915
2916 20. When matching a repeated Unicode property with a minimum greater than zero,
2917 (for example \pL{2,}), PCRE could look past the end of the subject if it
2918 reached it while seeking the minimum number of characters. This could
2919 happen only if some of the characters were more than one byte long, because
2920 there is a check for at least the minimum number of bytes.
2921
2922 21. Refactored the implementation of \p and \P so as to be more general, to
2923 allow for more different types of property in future. This has changed the
2924 compiled form incompatibly. Anybody with saved compiled patterns that use
2925 \p or \P will have to recompile them.
2926
2927 22. Added "Any" and "L&" to the supported property types.
2928
2929 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2930 but give a compile time error if the value is greater than 0xff.
2931
2932 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2933 accidentally not being installed or uninstalled.
2934
2935 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2936 made were to insert the current release number. This seemed silly, because
2937 it made things harder for people building PCRE on systems that don't run
2938 "configure". I have turned pcre.h into a distributed file, no longer built
2939 by "configure", with the version identification directly included. There is
2940 no longer a pcre.h.in file.
2941
2942 However, this change necessitated a change to the pcre-config script as
2943 well. It is built from pcre-config.in, and one of the substitutions was the
2944 release number. I have updated configure.ac so that ./configure now finds
2945 the release number by grepping pcre.h.
2946
2947 26. Added the ability to run the tests under valgrind.
2948
2949
2950 Version 6.4 05-Sep-05
2951 ---------------------
2952
2953 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2954 "--" to be printed when multiple files were scanned, even when none of the
2955 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2956 consider it to be a bug, and have restored the previous behaviour.
2957
2958 2. A couple of code tidies to get rid of compiler warnings.
2959
2960 3. The pcretest program used to cheat by referring to symbols in the library
2961 whose names begin with _pcre_. These are internal symbols that are not
2962 really supposed to be visible externally, and in some environments it is
2963 possible to suppress them. The cheating is now confined to including
2964 certain files from the library's source, which is a bit cleaner.
2965
2966 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2967 file's purpose clearer.
2968
2969 5. Reorganized pcre_ucp_findchar().
2970
2971
2972 Version 6.3 15-Aug-05
2973 ---------------------
2974
2975 1. The file libpcre.pc.in did not have general read permission in the tarball.
2976
2977 2. There were some problems when building without C++ support:
2978
2979 (a) If C++ support was not built, "make install" and "make test" still
2980 tried to test it.
2981
2982 (b) There were problems when the value of CXX was explicitly set. Some
2983 changes have been made to try to fix these, and ...
2984
2985 (c) --disable-cpp can now be used to explicitly disable C++ support.
2986
2987 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2988 backslash in a target when C++ was disabled. This confuses some
2989 versions of "make", apparently. Using an intermediate variable solves
2990 this. (Same for CPP_LOBJ.)
2991
2992 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2993 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2994 necessary on certain architectures.
2995
2996 4. Added a setting of -export-symbols-regex to the link command to remove
2997 those symbols that are exported in the C sense, but actually are local
2998 within the library, and not documented. Their names all begin with
2999 "_pcre_". This is not a perfect job, because (a) we have to except some
3000 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3001 available (and never for static libraries). I have made a note to try to
3002 find a way round (a) in the future.
3003
3004
3005 Version 6.2 01-Aug-05
3006 ---------------------
3007
3008 1. There was no test for integer overflow of quantifier values. A construction
3009 such as {1111111111111111} would give undefined results. What is worse, if
3010 a minimum quantifier for a parenthesized subpattern overflowed and became
3011 negative, the calculation of the memory size went wrong. This could have
3012 led to memory overwriting.
3013
3014 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3015
3016 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3017 operating environments where this matters.
3018
3019 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3020 PCRE options from within the C++ wrapper.
3021
3022 5. Named capturing subpatterns were not being correctly counted when a pattern
3023 was compiled. This caused two problems: (a) If there were more than 100
3024 such subpatterns, the calculation of the memory needed for the whole
3025 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3026 back references of the form \12, where the number was greater than 9, were
3027 not recognized as back references, even though there were sufficient
3028 previous subpatterns.
3029
3030 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3031 versions of gcc, e.g. 2.95.4.
3032
3033
3034 Version 6.1 21-Jun-05
3035 ---------------------
3036
3037 1. There was one reference to the variable "posix" in pcretest.c that was not
3038 surrounded by "#if !defined NOPOSIX".
3039
3040 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3041 the cross-check on the old pcre_info() function, for the benefit of the
3042 cut-down version of PCRE that is currently imported into Exim.
3043
3044 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3045 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3046 patterns that start (?i) but I don't think that matters. The use of (?i) is
3047 just an example; this all applies to the other options as well.
3048
3049 4. Since libtool seems to echo the compile commands it is issuing, the output
3050 from "make" can be reduced a bit by putting "@" in front of each libtool
3051 compile command.
3052
3053 5. Patch from the folks at Google for configure.in to be a bit more thorough
3054 in checking for a suitable C++ installation before trying to compile the
3055 C++ stuff. This should fix a reported problem when a compiler was present,
3056 but no suitable headers.
3057
3058 6. The man pages all had just "PCRE" as their title. I have changed them to
3059 be the relevant file name. I have also arranged that these names are
3060 retained in the file doc/pcre.txt, which is a concatenation in text format
3061 of all the man pages except the little individual ones for each function.
3062
3063 7. The NON-UNIX-USE file had not been updated for the different set of source
3064 files that come with release 6. I also added a few comments about the C++
3065 wrapper.
3066
3067
3068 Version 6.0 07-Jun-05
3069 ---------------------
3070
3071 1. Some minor internal re-organization to help with my DFA experiments.
3072
3073 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3074 didn't matter for the library itself when fully configured, but did matter
3075 when compiling without UCP support, or within Exim, where the ucp files are
3076 not imported.
3077
3078 3. Refactoring of the library code to split up the various functions into
3079 different source modules. The addition of the new DFA matching code (see
3080 below) to a single monolithic source would have made it really too
3081 unwieldy, quite apart from causing all the code to be include in a
3082 statically linked application, when only some functions are used. This is
3083 relevant even without the DFA addition now that patterns can be compiled in
3084 one application and matched in another.
3085
3086 The downside of splitting up is that there have to be some external
3087 functions and data tables that are used internally in different modules of
3088 the library but which are not part of the API. These have all had their
3089 names changed to start with "_pcre_" so that they are unlikely to clash
3090 with other external names.
3091
3092 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3093 a different (DFA) algorithm. Although it is slower than the original
3094 function, it does have some advantages for certain types of matching
3095 problem.
3096
3097 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3098 including restarting after a partial match.
3099
3100 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3101 defined when compiling for Windows was sent to me. I have put it into the
3102 code, though I have no means of testing or verifying it.
3103
3104 7. Added the pcre_refcount() auxiliary function.
3105
3106 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3107 match before or at the first newline in the subject string. In pcretest,
3108 the /f option on a pattern can be used to set this.
3109
3110 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3111 would behave wrongly. This has been present in PCRE since release 4.0.
3112
3113 10. A number of changes to the pcregrep command:
3114
3115 (a) Refactored how -x works; insert ^(...)$ instead of setting
3116 PCRE_ANCHORED and checking the length, in preparation for adding
3117 something similar for -w.
3118
3119 (b) Added the -w (match as a word) option.
3120
3121 (c) Refactored the way lines are read and buffered so as to have more
3122 than one at a time available.
3123
3124 (d) Implemented a pcregrep test script.
3125
3126 (e) Added the -M (multiline match) option. This allows patterns to match
3127 over several lines of the subject. The buffering ensures that at least
3128 8K, or the rest of the document (whichever is the shorter) is available
3129 for matching (and similarly the previous 8K for lookbehind assertions).
3130
3131 (f) Changed the --help output so that it now says
3132
3133 -w, --word-regex(p)
3134
3135 instead of two lines, one with "regex" and the other with "regexp"
3136 because that confused at least one person since the short forms are the
3137 same. (This required a bit of code, as the output is generated
3138 automatically from a table. It wasn't just a text change.)
3139
3140 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3141 option but starts with a hyphen. Could be a pattern or a path name
3142 starting with a hyphen, for instance.
3143
3144 (h) "-" can be given as a file name to represent stdin.
3145
3146 (i) When file names are being printed, "(standard input)" is used for
3147 the standard input, for compatibility with GNU grep. Previously
3148 "<stdin>" was used.
3149
3150 (j) The option --label=xxx can be used to supply a name to be used for
3151 stdin when file names are being printed. There is no short form.
3152
3153 (k) Re-factored the options decoding logic because we are going to add
3154 two more options that take data. Such options can now be given in four
3155 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3156
3157 (l) Added the -A, -B, and -C options for requesting that lines of context
3158 around matches be printed.
3159
3160 (m) Added the -L option to print the names of files that do not contain
3161 any matching lines, that is, the complement of -l.
3162
3163 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3164 continue to scan other files.
3165
3166 (o) The -s option was incorrectly implemented. For compatibility with other
3167 greps, it now suppresses the error message for a non-existent or non-
3168 accessible file (but not the return code). There is a new option called
3169 -q that suppresses the output of matching lines, which was what -s was
3170 previously doing.
3171
3172 (p) Added --include and --exclude options to specify files for inclusion
3173 and exclusion when recursing.
3174
3175 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3176 Hopefully, it now does.
3177
3178 12. Missing cast in pcre_study().
3179
3180 13. Added an "uninstall" target to the makefile.
3181
3182 14. Replaced "extern" in the function prototypes in Makefile.in with
3183 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3184 world, but is set differently for Windows.
3185
3186 15. Added a second compiling function called pcre_compile2(). The only
3187 difference is that it has an extra argument, which is a pointer to an
3188 integer error code. When there is a compile-time failure, this is set
3189 non-zero, in addition to the error test pointer being set to point to an
3190 error message. The new argument may be NULL if no error number is required
3191 (but then you may as well call pcre_compile(), which is now just a
3192 wrapper). This facility is provided because some applications need a
3193 numeric error indication, but it has also enabled me to tidy up the way
3194 compile-time errors are handled in the POSIX wrapper.
3195
3196 16. Added VPATH=.libs to the makefile; this should help when building with one
3197 prefix path and installing with another. (Or so I'm told by someone who
3198 knows more about this stuff than I do.)
3199
3200 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3201 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3202 match everything, including newlines. This is not POSIX-compatible, but
3203 somebody wanted the feature. From pcretest it can be activated by using
3204 both the P and the s flags.
3205
3206 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3207
3208 19. libpcre.pc was being incorrectly installed as executable.
3209
3210 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3211 it now also looks for '\r' so that it will work unmodified on Windows.
3212
3213 21. Added Google's contributed C++ wrapper to the distribution.
3214
3215 22. Added some untidy missing memory free() calls in pcretest, to keep
3216 Electric Fence happy when testing.
3217
3218
3219
3220 Version 5.0 13-Sep-04
3221 ---------------------
3222
3223 1. Internal change: literal characters are no longer packed up into items
3224 containing multiple characters in a single byte-string. Each character
3225 is now matched using a separate opcode. However, there may be more than one
3226 byte in the character in UTF-8 mode.
3227
3228 2. The pcre_callout_block structure has two new fields: pattern_position and
3229 next_item_length. These contain the offset in the pattern to the next match
3230 item, and its length, respectively.
3231
3232 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3233 insertion of callouts before each pattern item. Added the /C option to
3234 pcretest to make use of this.
3235
3236 4. On the advice of a Windows user, the lines
3237
3238 #if defined(_WIN32) || defined(WIN32)
3239 _setmode( _fileno( stdout ), 0x8000 );
3240 #endif /* defined(_WIN32) || defined(WIN32) */
3241
3242 have been added to the source of pcretest. This apparently does useful
3243 magic in relation to line terminators.
3244
3245 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3246 for the benefit of those environments where the "b" makes a difference.
3247
3248 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3249 to know about it. I have put a hack into configure.in that adds in code
3250 to set GCC=yes if CC=icc. This seems to end up at a point in the
3251 generated configure script that is early enough to affect the setting of
3252 compiler options, which is what is needed, but I have no means of testing
3253 whether it really works. (The user who reported this had patched the
3254 generated configure script, which of course I cannot do.)
3255
3256 LATER: After change 22 below (new libtool files), the configure script
3257 seems to know about icc (and also ecc). Therefore, I have commented out
3258 this hack in configure.in.
3259
3260 7. Added support for pkg-config (2 patches were sent in).
3261
3262 8. Negated POSIX character classes that used a combination of internal tables
3263 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3264 [[:^ascii]]. Typically, they would match almost any characters. The other
3265 POSIX classes were not broken in this way.
3266
3267 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3268 to find the match, as PCRE was deluded into thinking that the match had to
3269 start at the start point or following a newline. The same bug applied to
3270 patterns with negative forward assertions or any backward assertions
3271 preceding ".*" at the start, unless the pattern required a fixed first
3272 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3273
3274 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3275 starting at the last subject character, bytes beyond the end of the subject
3276 string were read.
3277
3278 11. Renamed the variable "class" as "classbits" to make life easier for C++
3279 users. (Previously there was a macro definition, but it apparently wasn't
3280 enough.)
3281
3282 12. Added the new field "tables" to the extra data so that tables can be passed
3283 in at exec time, or the internal tables can be re-selected. This allows
3284 a compiled regex to be saved and re-used at a later time by a different
3285 program that might have everything at different addresses.
3286
3287 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3288 -R library as well as a -L library.
3289
3290 14. The debugging options of pcretest (-d on the command line or D on a
3291 pattern) showed incorrect output for anything following an extended class
3292 that contained multibyte characters and which was followed by a quantifier.
3293
3294 15. Added optional support for general category Unicode character properties
3295 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3296 support. It adds about 90K to the size of the library. The meanings of the
3297 inbuilt class escapes such as \d and \s have NOT been changed.
3298
3299 16. Updated pcredemo.c to include calls to free() to release the memory for the
3300 compiled pattern.
3301
3302 17. The generated file chartables.c was being created in the source directory
3303 instead of in the building directory. This caused the build to fail if the
3304 source directory was different from the building directory, and was
3305 read-only.
3306
3307 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3308 file. No doubt somebody will tell me if they don't make sense... Also added
3309 Dan Mooney's comments about building on OpenVMS.
3310
3311 19. Added support for partial matching via the PCRE_PARTIAL option for
3312 pcre_exec() and the \P data escape in pcretest.
3313
3314 20. Extended pcretest with 3 new pattern features:
3315
3316 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3317 write the compiled pattern to the file whose name is "rest-of-line".
3318 This is a straight binary dump of the data, with the saved pointer to
3319 the character tables forced to be NULL. The study data, if any, is
3320 written too. After writing, pcretest reads a new pattern.
3321
3322 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3323 compiled pattern from the given file. There must not be any
3324 occurrences of "<" in the file name (pretty unlikely); if there are,
3325 pcretest will instead treat the initial "<" as a pattern delimiter.
3326 After reading in the pattern, pcretest goes on to read data lines as
3327 usual.
3328
3329 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3330 and 16-bit fields in a compiled pattern, to simulate a pattern that
3331 was compiled on a host of opposite endianness.
3332
3333 21. The pcre-exec() function can now cope with patterns that were compiled on
3334 hosts of opposite endianness, with this restriction:
3335
3336 As for any compiled expression that is saved and used later, the tables
3337 pointer field cannot be preserved; the extra_data field in the arguments
3338 to pcre_exec() should be used to pass in a tables address if a value
3339 other than the default internal tables were used at compile time.
3340
3341 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3342 now diagnosed as an error. Previously, most of the time, a negative number
3343 would have been treated as zero, but if in addition "ovector" was passed as
3344 NULL, a crash could occur.
3345
3346 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3347 new versions from the libtool 1.5 distribution (the last one is a copy of
3348 a file called libtool.m4). This seems to have fixed the need to patch
3349 "configure" to support Darwin 1.3 (which I used to do). However, I still
3350 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3351 workstation).
3352
3353 24. Changed the PCRE licence to be the more standard "BSD" licence.
3354
3355
3356 Version 4.5 01-Dec-03
3357 ---------------------
3358
3359 1. There has been some re-arrangement of the code for the match() function so
3360 that it can be compiled in a version that does not call itself recursively.
3361 Instead, it keeps those local variables that need separate instances for
3362 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3363 needs to "recurse". Keeping track of where control must go is done by means
3364 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3365 hide most of the details from the main code, and operates only if
3366 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3367 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3368 operating.
3369
3370 To make it easier for callers to provide specially tailored get/free
3371 functions for this usage, two new functions, pcre_stack_malloc, and
3372 pcre_stack_free, are used. They are always called in strict stacking order,
3373 and the size of block requested is always the same.
3374
3375 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3376 PCRE has been compiled to use the stack or the heap for recursion. The
3377 -C option of pcretest uses this to show which version is compiled.
3378
3379 A new data escape \S, is added to pcretest; it causes the amounts of store
3380 obtained and freed by both kinds of malloc/free at match time to be added
3381 to the output.
3382
3383 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3384 what's available on my current Linux desktop machine.
3385
3386 3. When matching a UTF-8 string, the test for a valid string at the start has
3387 been extended. If start_offset is not zero, PCRE now checks that it points
3388 to a byte that is the start of a UTF-8 character. If not, it returns
3389 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3390 this is necessary because there may be backward assertions in the pattern.
3391 When matching the same subject several times, it may save resources to use
3392 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3393
3394 4. The code for checking the validity of UTF-8 strings has been tightened so
3395 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3396 containing "overlong sequences".
3397
3398 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3399 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3400 should have been "&", but it just so happened that all the cases this let
3401 through by mistake were picked up later in the function.
3402
3403 6. I had used a variable called "isblank" - this is a C99 function, causing
3404 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3405
3406 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3407 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3408 so the version is shown; (c) stop "make test" echoing "./RunTest".
3409
3410 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3411
3412 9. The prototype for memmove() for systems that don't have it was using
3413 size_t, but the inclusion of the header that defines size_t was later. I've
3414 moved the #includes for the C headers earlier to avoid this.
3415
3416 10. Added some adjustments to the code to make it easier to compiler on certain
3417 special systems:
3418
3419 (a) Some "const" qualifiers were missing.
3420 (b) Added the macro EXPORT before all exported functions; by default this
3421 is defined to be empty.
3422 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3423 that it reads its output file name as an argument instead of writing
3424 to the standard output and assuming this can be redirected.
3425
3426 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3427 class containing characters with values greater than 255, PCRE compilation
3428 went into a loop.
3429
3430 12. A recursive reference to a subpattern that was within another subpattern
3431 that had a minimum quantifier of zero caused PCRE to crash. For example,
3432 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3433 recursion. If the recursively-called subpattern itself had a zero repeat,
3434 that was OK.
3435
3436 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3437 buffer into which it was copied (for escape processing) was still set at
3438 1024, so long lines caused crashes.
3439
3440 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3441 "internal error: code overflow...". This applied to any character class
3442 that was followed by a possessive quantifier.
3443
3444 15. Modified the Makefile to add libpcre.la as a prerequisite for
3445 libpcreposix.la because I was told this is needed for a parallel build to
3446 work.
3447
3448 16. If a pattern that contained .* following optional items at the start was
3449 studied, the wrong optimizing data was generated, leading to matching
3450 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3451 matching string must start with a or b or c. The correct conclusion for
3452 this pattern is that a match can start with any character.
3453
3454
3455 Version 4.4 13-Aug-03
3456 ---------------------
3457
3458 1. In UTF-8 mode, a character class containing characters with values between
3459 127 and 255 was not handled correctly if the compiled pattern was studied.
3460 In fixing this, I have also improved the studying algorithm for such
3461 classes (slightly).
3462
3463 2. Three internal functions had redundant arguments passed to them. Removal
3464 might give a very teeny performance improvement.
3465
3466 3. Documentation bug: the value of the capture_top field in a callout is *one
3467 more than* the number of the hightest numbered captured substring.
3468
3469 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3470 in incorrectly linking with a previously installed version. They now link
3471 explicitly with libpcre.la.
3472
3473 5. configure.in no longer needs to recognize Cygwin specially.
3474
3475 6. A problem in pcre.in for Windows platforms is fixed.
3476
3477 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3478 pcretest, it used to include the size of the study block as part of its
3479 output. Unfortunately, the structure contains a field that has a different
3480 size on different hardware architectures. This meant that the tests that
3481 showed this size failed. As the block is currently always of a fixed size,
3482 this information isn't actually particularly useful in pcretest output, so
3483 I have just removed it.
3484
3485 8. Three pre-processor statements accidentally did not start in column 1.
3486 Sadly, there are *still* compilers around that complain, even though
3487 standard C has not required this for well over a decade. Sigh.
3488
3489 9. In pcretest, the code for checking callouts passed small integers in the
3490 callout_data field, which is a void * field. However, some picky compilers
3491 complained about the casts involved for this on 64-bit systems. Now
3492 pcretest passes the address of the small integer instead, which should get
3493 rid of the warnings.
3494
3495 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3496 both compile and run time, and gives an error if an invalid UTF-8 sequence
3497 is found. There is a option for disabling this check in cases where the
3498 string is known to be correct and/or the maximum performance is wanted.
3499
3500 11. In response to a bug report, I changed one line in Makefile.in from
3501
3502 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3503 to
3504 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3505
3506 to look similar to other lines, but I have no way of telling whether this
3507 is the right thing to do, as I do not use Windows. No doubt I'll get told
3508 if it's wrong...
3509
3510
3511 Version 4.3 21-May-03
3512 ---------------------
3513
3514 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3515 Makefile.
3516
3517 2. Some refactoring to improve the quality of the code:
3518
3519 (i) The utf8_table... variables are now declared "const".
3520
3521 (ii) The code for \cx, which used the "case flipping" table to upper case
3522 lower case letters, now just substracts 32. This is ASCII-specific,
3523 but the whole concept of \cx is ASCII-specific, so it seems
3524 reasonable.
3525
3526 (iii) PCRE was using its character types table to recognize decimal and
3527 hexadecimal digits in the pattern. This is silly, because it handles
3528 only 0-9, a-f, and A-F, but the character types table is locale-
3529 specific, which means strange things might happen. A private
3530 table is now used for this - though it costs 256 bytes, a table is
3531 much faster than multiple explicit tests. Of course, the standard
3532 character types table is still used for matching digits in subject
3533 strings against \d.
3534
3535 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3536 ending in _t are). So I've renamed it as ESC_tee.
3537
3538 3. The first argument for regexec() in the POSIX wrapper should have been
3539 defined as "const".
3540
3541 4. Changed pcretest to use malloc() for its buffers so that they can be
3542 Electric Fenced for debugging.
3543
3544 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3545 to read one or more bytes before the start of the subject string. Often this
3546 had no effect on PCRE's behaviour, but in some circumstances it could
3547 provoke a segmentation fault.
3548
3549 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3550 to try to read one or more bytes before the start of the subject string.
3551
3552 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3553 UTF-8 support could misbehave in various ways if the subject string
3554 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3555 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3556 back over UTF-8 characters.)
3557
3558
3559 Version 4.2 14-Apr-03
3560 ---------------------
3561
3562 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3563
3564 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3565 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3566 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3567 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3568 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3569 and BUILD_EXEEXT
3570 Note: automatic setting of the BUILD variables is not yet working
3571 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3572 compile-time but not at link-time
3573 [LINK]: use for linking executables only
3574 make different versions for Windows and non-Windows
3575 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3576 libraries
3577 [LINK_FOR_BUILD]: new variable
3578 [OBJEXT]: use throughout
3579 [EXEEXT]: use throughout
3580 <winshared>: new target
3581 <wininstall>: new target
3582 <dftables.o>: use native compiler
3583 <dftables>: use native linker
3584 <install>: handle Windows platform correctly
3585 <clean>: ditto
3586 <check>: ditto
3587 copy DLL to top builddir before testing
3588
3589 As part of these changes, -no-undefined was removed again. This was reported
3590 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3591 in any case.
3592
3593 3. Some tidies to get rid of compiler warnings:
3594
3595 . In the match_data structure, match_limit was an unsigned long int, whereas
3596 match_call_count was an int. I've made them both unsigned long ints.
3597
3598 . In pcretest the fact that a const uschar * doesn't automatically cast to
3599 a void * provoked a warning.
3600
3601 . Turning on some more compiler warnings threw up some "shadow" variables
3602 and a few more missing casts.
3603
3604 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3605 option, a class that contained a single character with a value between 128
3606 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3607
3608 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3609 option, a class that contained several characters, but with at least one
3610 whose value was between 128 and 255 caused PCRE to crash.
3611
3612
3613 Version 4.1 12-Mar-03
3614 ---------------------
3615
3616 1. Compiling with gcc -pedantic found a couple of places where casts were
3617 needed, and a string in dftables.c that was longer than standard compilers are
3618 required to support.
3619
3620 2. Compiling with Sun's compiler found a few more places where the code could
3621 be tidied up in order to avoid warnings.
3622
3623 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3624 first of these names is deprecated in the latest Autoconf in favour of the name
3625 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3626 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3627 analogy I have changed it to CFLAGS_FOR_BUILD.
3628
3629 4. Added -no-undefined to the linking command in the Makefile, because this is
3630 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3631 linking step for the pcreposix library.
3632
3633 5. PCRE was failing to diagnose the case of two named groups with the same
3634 name.
3635
3636 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3637 literal character that is needed in the subject for a match, and scans along to
3638 ensure that it is present before embarking on the full matching process. This
3639 saves time in cases of nested unlimited repeats that are never going to match.
3640 Problem: the scan can take a lot of time if the subject is very long (e.g.
3641 megabytes), thus penalizing straightforward matches. It is now done only if the
3642 amount of subject to be scanned is less than 1000 bytes.
3643
3644 7. A lesser problem with the same optimization is that it was recording the
3645 first character of an anchored pattern as "needed", thus provoking a search
3646 right along the subject, even when the first match of the pattern was going to
3647 fail. The "needed" character is now not set for anchored patterns, unless it
3648 follows something in the pattern that is of non-fixed length. Thus, it still
3649 fulfils its original purpose of finding quick non-matches in cases of nested
3650 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3651
3652
3653 Version 4.0 17-Feb-03
3654 ---------------------
3655
3656 1. If a comment in an extended regex that started immediately after a meta-item
3657 extended to the end of string, PCRE compiled incorrect data. This could lead to
3658 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3659
3660 2. Moved to autoconf 2.53 and libtool 1.4.2.
3661
3662 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3663 the special perltest8 script is no longer needed - all the tests can be run
3664 from a single perltest script.
3665
3666 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3667 by \s. It has now been removed in PCRE. This means it isn't recognized as
3668 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3669 class [:space:] *does* include VT, thereby creating a mess.
3670
3671 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3672 space and tab.
3673
3674 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3675 its new features into the main test script, reducing the number of scripts.
3676
3677 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3678 were backward compatible, and made the (?i) apply to the whole pattern, as if
3679 /i were given. Now it behaves more logically, and applies the option setting
3680 only to what follows. PCRE has been changed to follow suit. However, if it
3681 finds options settings right at the start of the pattern, it extracts them into
3682 the global options, as before. Thus, they show up in the info data.
3683
3684 8. Added support for the \Q...\E escape sequence. Characters in between are
3685 treated as literals. This is slightly different from Perl in that $ and @ are
3686 also handled as literals inside the quotes. In Perl, they will cause variable
3687 interpolation. Note the following examples:
3688
3689 Pattern PCRE matches Perl matches
3690
3691 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3692 \Qabc\$xyz\E abc\$xyz abc\$xyz
3693 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3694
3695 For compatibility with Perl, \Q...\E sequences are recognized inside character
3696 classes as well as outside them.
3697
3698 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3699 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3700 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3701 signed/unsigned warnings.
3702
3703 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3704 option for pcretest, so I've replaced it by a simple function that does just
3705 that job.
3706
3707 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3708 "pcregrep -".
3709
3710 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3711 Java package. This provides some syntactic sugar for simple cases of what my
3712 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3713 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3714 item, you can use this simplified notation. Note that only makes sense with
3715 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3716 greediness, whatever the setting of the PCRE_UNGREEDY option.
3717
3718 13. A change of greediness default within a pattern was not taking effect at
3719 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3720 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3721 was abstracted outside.
3722
3723 14. PCRE now supports the \G assertion. It is true when the current matching
3724 position is at the start point of the match. This differs from \A when the
3725 starting offset is non-zero. Used with the /g option of pcretest (or similar
3726 code), it works in the same way as it does for Perl's /g option. If all
3727 alternatives of a regex begin with \G, the expression is anchored to the start
3728 match position, and the "anchored" flag is set in the compiled expression.
3729
3730 15. Some bugs concerning the handling of certain option changes within patterns
3731 have been fixed. These applied to options other than (?ims). For example,
3732 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3733 been the other way round. Some of this was related to change 7 above.
3734
3735 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3736 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3737 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3738 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3739
3740 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3741 mode. Unlike ".", it always matches newline, whatever the setting of
3742 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3743 assertions. Perl allows it, but it doesn't (in general) work because it can't
3744 calculate the length of the lookbehind. At least, that's the case for Perl
3745 5.8.0 - I've been told they are going to document that it doesn't work in
3746 future.
3747
3748 18. Added an error diagnosis for escapes that PCRE does not support: these are
3749 \L, \l, \N, \P, \p, \U, \u, and \X.
3750
3751 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3752 reading past the end of the pattern in cases such as /[abcd/.
3753
3754 20. PCRE was getting more memory than necessary for patterns with classes that
3755 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3756
3757 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3758 compiling PCRE for use with Virtual Pascal.
3759
3760 22. Small fix to the Makefile to make it work properly if the build is done
3761 outside the source tree.
3762
3763 23. Added a new extension: a condition to go with recursion. If a conditional
3764 subpattern starts with (?(R) the "true" branch is used if recursion has
3765 happened, whereas the "false" branch is used only at the top level.
3766
3767 24. When there was a very long string of literal characters (over 255 bytes
3768 without UTF support, over 250 bytes with UTF support), the computation of how
3769 much memory was required could be incorrect, leading to segfaults or other
3770 strange effects.
3771
3772 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3773 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3774 there was a subsequent back reference to those brackets. This meant that, for
3775 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3776 possible to check for precisely this case. All we can do is abandon the
3777 optimization if .* occurs inside capturing brackets when there are any back
3778 references whatsoever. (See below for a better fix that came later.)
3779
3780 26. The handling of the optimization for finding the first character of a
3781 non-anchored pattern, and for finding a character that is required later in the
3782 match were failing in some cases. This didn't break the matching; it just
3783 failed to optimize when it could. The way this is done has been re-implemented.
3784
3785 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3786
3787 28. Added a new feature that provides some of the functionality that Perl
3788 provides with (?{...}). The facility is termed a "callout". The way it is done
3789 in PCRE is for the caller to provide an optional function, by setting
3790 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3791 global variable. By default it is unset, which disables all calling out. To get
3792 the function called, the regex must include (?C) at appropriate points. This
3793 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3794 This provides a means of identifying different callout points. When PCRE
3795 reaches such a point in the regex, if pcre_callout has been set, the external
3796 function is called. It is provided with data in a structure called
3797 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3798 matching continues; if it returns a non-zero value, the match at the current
3799 point fails. However, backtracking will occur if possible. [This was changed
3800 later and other features added - see item 49 below.]
3801
3802 29. pcretest is upgraded to test the callout functionality. It provides a
3803 callout function that displays information. By default, it shows the start of
3804 the match and the current position in the text. There are some new data escapes
3805 to vary what happens:
3806
3807 \C+ in addition, show current contents of captured substrings
3808 \C- do not supply a callout function
3809 \C!n return 1 when callout number n is reached
3810 \C!n!m return 1 when callout number n is reached for the mth time
3811
3812 30. If pcregrep was called with the -l option and just a single file name, it
3813 output "<stdin>" if a match was found, instead of the file name.
3814
3815 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3816 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3817 pcre_exec(). This saves a malloc/free per call. The default value of
3818 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3819 when configuring.
3820
3821 32. The default maximum size of a compiled pattern is 64K. There have been a
3822 few cases of people hitting this limit. The code now uses macros to handle the
3823 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3824 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3825 configuring. Tests 2 and 5 work only with 2-byte links because they output
3826 debugging information about compiled patterns.
3827
3828 33. Internal code re-arrangements:
3829
3830 (a) Moved the debugging function for printing out a compiled regex into
3831 its own source file (printint.c) and used #include to pull it into
3832 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3833 separate copies.
3834
3835 (b) Defined the list of op-code names for debugging as a macro in
3836 internal.h so that it is next to the definition of the opcodes.
3837
3838 (c) Defined a table of op-code lengths for simpler skipping along compiled
3839 code. This is again a macro in internal.h so that it is next to the
3840 definition of the opcodes.
3841
3842 34. Added support for recursive calls to individual subpatterns, along the
3843 lines of Robin Houston's patch (but implemented somewhat differently).
3844
3845 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3846 allow it to read and process whole directories in Win32. This code was
3847 contributed by Lionel Fourquaux; it has not been tested by me.
3848
3849 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3850 used to name a group. Names consist of alphanumerics and underscores, and must
3851 be unique. Back references use the syntax (?P=name) and recursive calls use
3852 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3853 numbers. The function pcre_fullinfo() can be used after compilation to extract
3854 a name/number map. There are three relevant calls:
3855
3856 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3857 PCRE_INFO_NAMECOUNT yields the number of entries
3858 PCRE_INFO_NAMETABLE yields a pointer to the map.
3859
3860 The map is a vector of fixed-size entries. The size of each entry depends on
3861 the length of the longest name used. The first two bytes of each entry are the
3862 group number, most significant byte first. There follows the corresponding
3863 name, zero terminated. The names are in alphabetical order.
3864
3865 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3866 case instead of 255. Making it the same both with and without UTF-8 support
3867 means that the same test output works with both.
3868
3869 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3870 calling malloc() with a zero argument.
3871
3872 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3873 optimization. I've improved things by keeping a bitmap of backreferences with
3874 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3875 fact referenced, the optimization can be applied. It is unlikely that a
3876 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3877 the match to follow \n) will appear inside brackets with a number greater than
3878 31, but if it does, any back reference > 31 suppresses the optimization.
3879
3880 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3881 of disabling numbered capturing parentheses. Any opening parenthesis that is
3882 not followed by ? behaves as if it were followed by ?: but named parentheses
3883 can still be used for capturing (and they will acquire numbers in the usual
3884 way).
3885
3886 41. Redesigned the return codes from the match() function into yes/no/error so
3887 that errors can be passed back from deep inside the nested calls. A malloc
3888 failure while inside a recursive subpattern call now causes the
3889 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3890
3891 42. It is now possible to set a limit on the number of times the match()
3892 function is called in a call to pcre_exec(). This facility makes it possible to
3893 limit the amount of recursion and backtracking, though not in a directly
3894 obvious way, because the match() function is used in a number of different
3895 circumstances. The count starts from zero for each position in the subject
3896 string (for non-anchored patterns). The default limit is, for compatibility, a
3897 large number, namely 10 000 000. You can change this in two ways:
3898
3899 (a) When configuring PCRE before making, you can use --with-match-limit=n
3900 to set a default value for the compiled library.
3901
3902 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3903 a different value is set. See 45 below.
3904
3905 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3906
3907 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3908 of things that can be changed at compile time. The first argument specifies
3909 what is wanted and the second points to where the information is to be placed.
3910 The current list of available information is:
3911
3912 PCRE_CONFIG_UTF8
3913
3914 The output is an integer that is set to one if UTF-8 support is available;
3915 otherwise it is set to zero.
3916
3917 PCRE_CONFIG_NEWLINE
3918
3919 The output is an integer that it set to the value of the code that is used for
3920 newline. It is either LF (10) or CR (13).
3921
3922 PCRE_CONFIG_LINK_SIZE
3923
3924 The output is an integer that contains the number of bytes used for internal
3925 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3926
3927 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3928
3929 The output is an integer that contains the threshold above which the POSIX
3930 interface uses malloc() for output vectors. See item 31 above.
3931
3932 PCRE_CONFIG_MATCH_LIMIT
3933
3934 The output is an unsigned integer that contains the default limit of the number
3935 of match() calls in a pcre_exec() execution. See 42 above.
3936
3937 44. pcretest has been upgraded by the addition of the -C option. This causes it
3938 to extract all the available output from the new pcre_config() function, and to
3939 output it. The program then exits immediately.
3940
3941 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3942 order to support additional features. One way would have been to define
3943 pcre_exec2() (for example) with extra arguments, but this would not have been
3944 extensible, and would also have required all calls to the original function to
3945 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3946 is used for passing in "extra" data from pcre_study().
3947
3948 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3949 contains the following fields:
3950
3951 flags a bitmap indicating which of the following fields are set
3952 study_data opaque data from pcre_study()
3953 match_limit a way of specifying a limit on match() calls for a specific
3954 call to pcre_exec()
3955 callout_data data for callouts (see 49 below)
3956
3957 The flag bits are also defined in pcre.h, and are
3958
3959 PCRE_EXTRA_STUDY_DATA
3960 PCRE_EXTRA_MATCH_LIMIT
3961 PCRE_EXTRA_CALLOUT_DATA
3962
3963 The pcre_study() function now returns one of these new pcre_extra blocks, with
3964 the actual study data pointed to by the study_data field, and the
3965 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3966 before. That is, this change is entirely upwards-compatible and requires no
3967 change to existing code.
3968
3969 If you want to pass in additional data to pcre_exec(), you can either place it
3970 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3971 block.
3972
3973 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3974 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3975 times with different match limits, until it finds the minimum value needed for
3976 pcre_exec() to complete. The value is then output. This can be instructive; for
3977 most simple matches the number is quite small, but for pathological cases it
3978 gets very large very quickly.
3979
3980 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3981 returns the size of the data block pointed to by the study_data field in a
3982 pcre_extra block, that is, the value that was passed as the argument to
3983 pcre_malloc() when PCRE was getting memory in which to place the information
3984 created by pcre_study(). The fourth argument should point to a size_t variable.
3985 pcretest has been extended so that this information is shown after a successful
3986 pcre_study() call when information about the compiled regex is being displayed.
3987
3988 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3989 because what follows is always an absolute path. (Later: it turns out that this
3990 is more than cosmetic for MinGW, because it doesn't like empty path
3991 components.)
3992
3993 49. Some changes have been made to the callout feature (see 28 above):
3994
3995 (i) A callout function now has three choices for what it returns:
3996
3997 0 => success, carry on matching
3998 > 0 => failure at this point, but backtrack if possible
3999 < 0 => serious error, return this value from pcre_exec()
4000
4001 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4002 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4003 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4004 use by callout functions. It will never be used by PCRE itself.
4005
4006 (ii) The pcre_extra structure (see 45 above) has a void * field called
4007 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4008 pcre_callout_block structure has a field of the same name. The contents of
4009 the field passed in the pcre_extra structure are passed to the callout
4010 function in the corresponding field in the callout block. This makes it
4011 easier to use the same callout-containing regex from multiple threads. For
4012 testing, the pcretest program has a new data escape
4013
4014 \C*n pass the number n (may be negative) as callout_data
4015
4016 If the callout function in pcretest receives a non-zero value as
4017 callout_data, it returns that value.
4018
4019 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4020 there were some redundant $(CFLAGS) in commands that are now specified as
4021 $(LINK), which already includes $(CFLAGS).
4022
4023 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4024 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4025 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4026 one-byte characters throughout. Note that case-insensitive matching applies
4027 only to characters whose values are less than 256. PCRE doesn't support the
4028 notion of cases for higher-valued characters.
4029
4030 (i) A character class whose characters are all within 0-255 is handled as
4031 a bit map, and the map is inverted for negative classes. Previously, a
4032 character > 255 always failed to match such a class; however it should
4033 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4034
4035 (ii) A negated character class with a single character < 255 is coded as
4036 "not this character" (OP_NOT). This wasn't working properly when the test
4037 character was multibyte, either singly or repeated.
4038
4039 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4040 mode, for example: \x{100}{2,3}.
4041
4042 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4043 singly or repeated) now correctly test multibyte characters. However,
4044 PCRE doesn't recognize any characters with values greater than 255 as
4045 digits, spaces, or word characters. Such characters always match \D, \S,
4046 and \W, and never match \d, \s, or \w.
4047
4048 (v) Classes may now contain characters and character ranges with values
4049 greater than 255. For example: [ab\x{100}-\x{400}].
4050
4051 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4052 PCRE in UTF-8 mode.
4053
4054 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4055 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4056 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4057 value.)
4058
4059 53. The single man page has become too large. I have therefore split it up into
4060 a number of separate man pages. These also give rise to individual HTML pages;
4061 these are now put in a separate directory, and there is an index.html page that
4062 lists them all. Some hyperlinking between the pages has been installed.
4063
4064 54. Added convenience functions for handling named capturing parentheses.
4065
4066 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4067 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4068 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4069 are faulted.
4070
4071 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4072 calling configure. These values are used when compiling the dftables.c program
4073 which is run to generate the source of the default character tables. They
4074 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4075 you will need to set these values.
4076
4077 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4078
4079
4080 Version 3.9 02-Jan-02
4081 ---------------------
4082
4083 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4084
4085 2. If --disable-static was given, the building process failed when trying to
4086 build pcretest and pcregrep. (For some reason it was using libtool to compile
4087 them, which is not right, as they aren't part of the library.)
4088
4089
4090 Version 3.8 18-Dec-01
4091 ---------------------
4092
4093 1. The experimental UTF-8 code was completely screwed up. It was packing the
4094 bytes in the wrong order. How dumb can you get?
4095
4096
4097 Version 3.7 29-Oct-01
4098 ---------------------
4099
4100 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4101 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4102 this didn't happen under Solaris 8, where I normally test things.
4103
4104 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4105 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4106 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4107 this fix an hour or so after the initial 3.7 release.)
4108
4109
4110 Version 3.6 23-Oct-01
4111 ---------------------
4112
4113 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4114 offsets passed as NULL with zero offset count.
4115
4116 2. The config.guess and config.sub files had not been updated when I moved to
4117 the latest autoconf.
4118
4119
4120 Version 3.5 15-Aug-01
4121 ---------------------
4122
4123 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4124 had been forgotten.
4125
4126 2. By using declared but undefined structures, we can avoid using "void"
4127 definitions in pcre.h while keeping the internal definitions of the structures
4128 private.
4129
4130 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4131 user point of view, this means that both static and shared libraries are built
4132 by default, but this can be individually controlled. More of the work of
4133 handling this static/shared cases is now inside libtool instead of PCRE's make
4134 file.
4135
4136 4. The pcretest utility is now installed along with pcregrep because it is
4137 useful for users (to test regexs) and by doing this, it automatically gets
4138 relinked by libtool. The documentation has been turned into a man page, so
4139 there are now .1, .txt, and .html versions in /doc.
4140
4141 5. Upgrades to pcregrep:
4142 (i) Added long-form option names like gnu grep.
4143 (ii) Added --help to list all options with an explanatory phrase.
4144 (iii) Added -r, --recursive to recurse into sub-directories.
4145 (iv) Added -f, --file to read patterns from a file.
4146
4147 6. pcre_exec() was referring to its "code" argument before testing that
4148 argument for NULL (and giving an error if it was NULL).
4149
4150 7. Upgraded Makefile.in to allow for compiling in a different directory from
4151 the source directory.
4152
4153 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4154 options bits, the pointer it was passed was to an int instead of to an unsigned
4155 long int. This mattered only on 64-bit systems.
4156
4157 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4158 generated) instead of pcre.in, which it its source. Also made the same change
4159 in several of the .c files.
4160
4161 10. A new release of gcc defines printf() as a macro, which broke pcretest
4162 because it had an ifdef in the middle of a string argument for printf(). Fixed
4163 by using separate calls to printf().
4164
4165 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4166 script, to force use of CR or LF instead of \n in the source. On non-Unix
4167 systems, the value can be set in config.h.
4168
4169 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4170 absolute limit. Changed the text of the error message to make this clear, and
4171 likewise updated the man page.
4172
4173 13. The limit of 99 on the number of capturing subpatterns has been removed.
4174 The new limit is 65535, which I hope will not be a "real" limit.
4175
4176
4177 Version 3.4 22-Aug-00
4178 ---------------------
4179
4180 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4181
4182 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4183
4184
4185 Version 3.3 01-Aug-00
4186 ---------------------
4187
4188 1. If an octal character was given, but the value was greater than \377, it
4189 was not getting masked to the least significant bits, as documented. This could
4190 lead to crashes in some systems.
4191
4192 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4193 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4194
4195 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4196 These just pass their arguments on to (pcre_free)(), but they are provided
4197 because some uses of PCRE bind it to non-C systems that can call its functions,
4198 but cannot call free() or pcre_free() directly.
4199
4200 4. Add "make test" as a synonym for "make check". Corrected some comments in
4201 the Makefile.
4202
4203 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4204 Makefile.
4205
4206 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4207 command called pgrep for grepping around the active processes.
4208
4209 7. Added the beginnings of support for UTF-8 character strings.
4210
4211 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4212 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4213 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4214 out for the ar command.)
4215
4216
4217 Version 3.2 12-May-00
4218 ---------------------
4219
4220 This is purely a bug fixing release.
4221
4222 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4223 of ZA. This was just one example of several cases that could provoke this bug,
4224 which was introduced by change 9 of version 2.00. The code for breaking
4225 infinite loops after an iteration that matches an empty string was't working
4226 correctly.
4227
4228 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4229 when matched against abbab (for example). After matching an empty string, it
4230 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4231 caused it to match further down the string than it should.
4232
4233 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4234 was there because it doesn't seem to be needed, and it causes trouble on some
4235 systems, as it is not a Standard C header. It has been removed.
4236
4237 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4238 were reported on the Macintosh. The changes were from
4239
4240 while ((c = *(++ptr)) != 0 && c != '\n');
4241 to
4242 while ((c = *(++ptr)) != 0 && c != '\n') ;
4243
4244 Totally extraordinary, but if that's what it takes...
4245
4246 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4247 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4248 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4249 assumes the way PCRE uses memmove() (always moving upwards).
4250
4251 6. PCRE is being used in one environment where strchr() is not available. There
4252 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4253 faster code anyway.
4254
4255
4256 Version 3.1 09-Feb-00
4257 ---------------------
4258
4259 The only change in this release is the fixing of some bugs in Makefile.in for
4260 the "install" target:
4261
4262 (1) It was failing to install pcreposix.h.
4263
4264 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4265
4266
4267 Version 3.0 01-Feb-00
4268 ---------------------
4269
4270 1. Add support for the /+ modifier to perltest (to output $` like it does in
4271 pcretest).
4272
4273 2. Add support for the /g modifier to perltest.
4274
4275 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4276 matches null strings.
4277
4278 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4279 pattern. Perl treats empty patterns specially - it reuses the most recent
4280 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4281 effect.
4282
4283 5. The POSIX interface was broken in that it was just handing over the POSIX
4284 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4285 required a bigger vector, with some working space on the end. This means that
4286 the POSIX wrapper now has to get and free some memory, and copy the results.
4287
4288 6. Added some simple autoconf support, placing the test data and the
4289 documentation in separate directories, re-organizing some of the
4290 information files, and making it build pcre-config (a GNU standard). Also added
4291 libtool support for building PCRE as a shared library, which is now the
4292 default.
4293
4294 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4295 09 are not valid octal constants. Single digits will be used for minor values
4296 less than 10.
4297
4298 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4299 existing programs that set these in the POSIX interface can use PCRE without
4300 modification.
4301
4302 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4303 return all that pcre_info() returns, plus additional data. The pcre_info()
4304 function is retained for compatibility, but is considered to be obsolete.
4305
4306 10. Added experimental recursion feature (?R) to handle one common case that
4307 Perl 5.6 will be able to do with (?p{...}).
4308
4309 11. Added support for POSIX character classes like [:alpha:], which Perl is
4310 adopting.
4311
4312
4313 Version 2.08 31-Aug-99
4314 ----------------------
4315
4316 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4317 trying to match at the startoffset position, but instead was moving forward to
4318 the next newline as if a previous match had failed.
4319
4320 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4321 and could get into a loop if a null string was matched other than at the start
4322 of the subject.
4323
4324 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4325 be distinguished at compile time, and for completeness also added PCRE_DATE.
4326
4327 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4328 in GnuWin32 environments.
4329
4330
4331 Version 2.07 29-Jul-99
4332 ----------------------
4333
4334 1. The documentation is now supplied in plain text form and HTML as well as in
4335 the form of man page sources.
4336
4337 2. C++ compilers don't like assigning (void *) values to other pointer types.
4338 In particular this affects malloc(). Although there is no problem in Standard
4339 C, I've put in casts to keep C++ compilers happy.
4340
4341 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4342 should be (const char *).
4343
4344 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4345 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4346 However, I haven't made this a standard facility. The documentation doesn't
4347 mention it, and the Makefile doesn't support it.
4348
4349 5. The Makefile now contains an "install" target, with editable destinations at
4350 the top of the file. The pcretest program is not installed.
4351
4352 6. pgrep -V now gives the PCRE version number and date.
4353
4354 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4355 causing the entire string to be ignored, instead of just the last character.
4356
4357 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4358 non-matching string, it can take a very, very long time, even for strings of
4359 quite modest length, because of the nested recursion. PCRE now does better in
4360 some of these cases. It does this by remembering the last required literal
4361 character in the pattern, and pre-searching the subject to ensure it is present
4362 before running the real match. In other words, it applies a heuristic to detect
4363 some types of certain failure quickly, and in the above example, if presented
4364 with a string that has no trailing " it gives "no match" very quickly.
4365
4366 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4367 other alternatives are tried instead.
4368
4369
4370 Version 2.06 09-Jun-99
4371 ----------------------
4372
4373 1. Change pcretest's output for amount of store used to show just the code
4374 space, because the remainder (the data block) varies in size between 32-bit and
4375 64-bit systems.
4376
4377 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4378 start matching at. This allows lookbehinds to work when searching for multiple
4379 occurrences in a string.
4380
4381 3. Added additional options to pcretest for testing multiple occurrences:
4382
4383 /+ outputs the rest of the string that follows a match
4384 /g loops for multiple occurrences, using the new startoffset argument
4385 /G loops for multiple occurrences by passing an incremented pointer
4386
4387 4. PCRE wasn't doing the "first character" optimization for patterns starting
4388 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4389 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4390 the letter 'x'. On long subject strings, this gives a significant speed-up.
4391
4392
4393 Version 2.05 21-Apr-99
4394 ----------------------
4395
4396 1. Changed the type of magic_number from int to long int so that it works
4397 properly on 16-bit systems.
4398
4399 2. Fixed a bug which caused patterns starting with .* not to work correctly
4400 when the subject string contained newline characters. PCRE was assuming
4401 anchoring for such patterns in all cases, which is not correct because .* will
4402 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4403 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4404 must be retried after every newline in the subject.
4405
4406
4407 Version 2.04 18-Feb-99
4408 ----------------------
4409
4410 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4411 computation of the store needed to hold the pattern was incorrect (too large).
4412 If such patterns were nested a few deep, this could multiply and become a real
4413 problem.
4414
4415 2. Added /M option to pcretest to show the memory requirement of a specific
4416 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4417
4418 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4419 compiled in such a way that the backtracking after subsequent failure was
4420 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4421 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4422
4423
4424 Version 2.03 02-Feb-99
4425 ----------------------
4426
4427 1. Fixed typo and small mistake in man page.
4428
4429 2. Added 4th condition (GPL supersedes if conflict) and created separate
4430 LICENCE file containing the conditions.
4431
4432 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4433 Perl, that is the internal \ allows the delimiter to be included in the
4434 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4435 the final delimiter, add \ to the end of the pattern (to test the error).
4436
4437 4. Added the convenience functions for extracting substrings after a successful
4438 match. Updated pcretest to make it able to test these functions.
4439
4440
4441 Version 2.02 14-Jan-99
4442 ----------------------
4443
4444 1. Initialized the working variables associated with each extraction so that
4445 their saving and restoring doesn't refer to uninitialized store.
4446
4447 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4448 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4449 fix the problem.
4450
4451 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4452 calls, and wasn't printing the correct value for compiling calls. Increased the
4453 default value of LOOPREPEAT, and the number of significant figures in the
4454 times.
4455
4456 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4457
4458 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4459 a building problem on Windows NT with a FAT file system.
4460
4461
4462 Version 2.01 21-Oct-98
4463 ----------------------
4464
4465 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4466 to character tables built by pcre_maketables() in the current locale. If NULL
4467 is passed, the default tables are used.
4468
4469
4470 Version 2.00 24-Sep-98
4471 ----------------------
4472
4473 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4474 it any more.
4475
4476 2. Allow quantification of (?>) groups, and make it work correctly.
4477
4478 3. The first character computation wasn't working for (?>) groups.
4479
4480 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4481 end of the subject) and add 5.005's \z, which really does match only at the
4482 very end of the subject.
4483
4484 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4485
4486 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4487 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4488 localized options. All options to pcre_study() were also removed.
4489
4490 7. Add other new features from 5.005:
4491
4492 $(?<= positive lookbehind
4493 $(?<! negative lookbehind
4494 (?imsx-imsx) added the unsetting capability
4495 such a setting is global if at outer level; local otherwise
4496 (?imsx-imsx:) non-capturing groups with option setting
4497 (?(cond)re|re) conditional pattern matching
4498
4499 A backreference to itself in a repeated group matches the previous
4500 captured string.
4501
4502 8. General tidying up of studying (both automatic and via "study")
4503 consequential on the addition of new assertions.
4504
4505 9. As in 5.005, unlimited repeated groups that could match an empty substring
4506 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4507 runtime if any iteration does actually match an empty substring.
4508
4509 10. Include the RunTest script in the distribution.
4510
4511 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4512 discrepancies, some of which were old and were also with respect to 5.004. They
4513 have now been fixed.
4514
4515
4516 Version 1.09 28-Apr-98
4517 ----------------------
4518
4519 1. A negated single character class followed by a quantifier with a minimum
4520 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4521 program crashes, or just wrong answers. This did not apply to negated classes
4522 containing more than one character, or to minima other than one.
4523
4524
4525 Version 1.08 27-Mar-98
4526 ----------------------
4527
4528 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4529
4530 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4531 latter must appear before anything that relies on it in the pattern.
4532
4533
4534 Version 1.07 16-Feb-98
4535 ----------------------
4536
4537 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4538 repeat of a potentially empty string).
4539
4540
4541 Version 1.06 23-Jan-98
4542 ----------------------
4543
4544 1. Added Markus Oberhumer's little patches for C++.
4545
4546 2. Literal strings longer than 255 characters were broken.
4547
4548
4549 Version 1.05 23-Dec-97
4550 ----------------------
4551
4552 1. Negated character classes containing more than one character were failing if
4553 PCRE_CASELESS was set at run time.
4554
4555
4556 Version 1.04 19-Dec-97
4557 ----------------------
4558
4559 1. Corrected the man page, where some "const" qualifiers had been omitted.
4560
4561 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4562 input syntax.
4563
4564 3. Fixed memory leak which occurred when a regex with back references was
4565 matched with an offsets vector that wasn't big enough. The temporary memory
4566 that is used in this case wasn't being freed if the match failed.
4567
4568 4. Tidied pcretest to ensure it frees memory that it gets.
4569
4570 5. Temporary memory was being obtained in the case where the passed offsets
4571 vector was exactly big enough.
4572
4573 6. Corrected definition of offsetof() from change 5 below.
4574
4575 7. I had screwed up change 6 below and broken the rules for the use of
4576 setjmp(). Now fixed.
4577
4578
4579 Version 1.03 18-Dec-97
4580 ----------------------
4581
4582 1. A erroneous regex with a missing opening parenthesis was correctly
4583 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4584 on some systems.
4585
4586 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4587 it was reported that one broken compiler failed on the former because "code" is
4588 also an independent variable.
4589
4590 3. The erroneous regex a[]b caused an array overrun reference.
4591
4592 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4593 fail on data ending with that character. (It was going on too far, and checking
4594 the next character, typically a binary zero.) This was specific to the
4595 optimized code for single-character negative classes.
4596
4597 5. Added a contributed patch from the TIN world which does the following:
4598
4599 + Add an undef for memmove, in case the the system defines a macro for it.
4600
4601 + Add a definition of offsetof(), in case there isn't one. (I don't know
4602 the reason behind this - offsetof() is part of the ANSI standard - but
4603 it does no harm).
4604
4605 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4606 most of the places where whitespace preceded '#'. I have given up and
4607 allowed the remaining 2 cases to be at the margin.
4608
4609 + Rename some variables in pcre to eliminate shadowing. This seems very
4610 pedantic, but does no harm, of course.
4611
4612 6. Moved the call to setjmp() into its own function, to get rid of warnings
4613 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4614
4615 7. Constructs such as \d{8,} were compiling into the equivalent of
4616 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4617 outcome, but in this particular case used more store than had been allocated,
4618 which caused the bug to be discovered because it threw up an internal error.
4619
4620 8. The debugging code in both pcre and pcretest for outputting the compiled
4621 form of a regex was going wrong in the case of back references followed by
4622 curly-bracketed repeats.
4623
4624
4625 Version 1.02 12-Dec-97
4626 ----------------------
4627
4628 1. Typos in pcre.3 and comments in the source fixed.
4629
4630 2. Applied a contributed patch to get rid of places where it used to remove
4631 'const' from variables, and fixed some signed/unsigned and uninitialized
4632 variable warnings.
4633
4634 3. Added the "runtest" target to Makefile.
4635
4636 4. Set default compiler flag to -O2 rather than just -O.
4637
4638
4639 Version 1.01 19-Nov-97
4640 ----------------------
4641
4642 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4643 like /([ab]*)*/, that is, for classes with more than one character in them.
4644
4645 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4646 as /((?>a*))*/ (a PCRE_EXTRA facility).
4647
4648
4649 Version 1.00 18-Nov-97
4650 ----------------------
4651
4652 1. Added compile-time macros to support systems such as SunOS4 which don't have
4653 memmove() or strerror() but have other things that can be used instead.
4654
4655 2. Arranged that "make clean" removes the executables.
4656
4657
4658 Version 0.99 27-Oct-97
4659 ----------------------
4660
4661 1. Fixed bug in code for optimizing classes with only one character. It was
4662 initializing a 32-byte map regardless, which could cause it to run off the end
4663 of the memory it had got.
4664
4665 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4666
4667
4668 Version 0.98 22-Oct-97
4669 ----------------------
4670
4671 1. Fixed bug in code for handling temporary memory usage when there are more
4672 back references than supplied space in the ovector. This could cause segfaults.
4673
4674
4675 Version 0.97 21-Oct-97
4676 ----------------------
4677
4678 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4679
4680 2. Optimized negated single characters not to use a bit map.
4681
4682 3. Brought error texts together as macro definitions; clarified some of them;
4683 fixed one that was wrong - it said "range out of order" when it meant "invalid
4684 escape sequence".
4685
4686 4. Changed some char * arguments to const char *.
4687
4688 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4689
4690 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4691 pcretest.
4692
4693
4694 Version 0.96 16-Oct-97
4695 ----------------------
4696
4697 1. Added a simple "pgrep" utility to the distribution.
4698
4699 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4700 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4701 where "ddd" means "one or more decimal digits".
4702
4703 3. Fixed serious bug. If a pattern had a back reference, but the call to
4704 pcre_exec() didn't supply a large enough ovector to record the related
4705 identifying subpattern, the match always failed. PCRE now remembers the number
4706 of the largest back reference, and gets some temporary memory in which to save
4707 the offsets during matching if necessary, in order to ensure that
4708 backreferences always work.
4709
4710 4. Increased the compatibility with Perl in a number of ways:
4711
4712 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4713 to request this handling. The option can be set at compile or exec time.
4714
4715 (b) $ matches before a terminating newline by default; an option
4716 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4717 mode). The option can be set at compile or exec time.
4718
4719 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4720 the same as Perl's. If the decimal number it represents is less than 10
4721 or there aren't that many previous left capturing parentheses, an octal
4722 escape is read. Inside a character class, it's always an octal escape,
4723 even if it is a single digit.
4724
4725 (d) An escaped but undefined alphabetic character is taken as a literal,
4726 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4727 escapes.
4728
4729 (e) {0} is now permitted. (The previous item is removed from the compiled
4730 pattern).
4731
4732 5. Changed all the names of code files so that the basic parts are no longer
4733 than 10 characters, and abolished the teeny "globals.c" file.
4734
4735 6. Changed the handling of character classes; they are now done with a 32-byte
4736 bit map always.
4737
4738 7. Added the -d and /D options to pcretest to make it possible to look at the
4739 internals of compilation without having to recompile pcre.
4740
4741
4742 Version 0.95 23-Sep-97
4743 ----------------------
4744
4745 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4746 \x20 at the start of a run of normal characters. These were being treated as
4747 real characters, instead of the source characters being re-checked.
4748
4749
4750 Version 0.94 18-Sep-97
4751 ----------------------
4752
4753 1. The functions are now thread-safe, with the caveat that the global variables
4754 containing pointers to malloc() and free() or alternative functions are the
4755 same for all threads.
4756
4757 2. Get pcre_study() to generate a bitmap of initial characters for non-
4758 anchored patterns when this is possible, and use it if passed to pcre_exec().
4759
4760
4761 Version 0.93 15-Sep-97
4762 ----------------------
4763
4764 1. /(b)|(:+)/ was computing an incorrect first character.
4765
4766 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4767 but not actually doing anything yet.
4768
4769 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4770 as Perl does (e.g. [-az] or [az-]).
4771
4772 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4773 all possible positions.
4774
4775 5. Split up into different modules to avoid including unneeded functions in a
4776 compiled binary. However, compile and exec are still in one module. The "study"
4777 function is split off.
4778
4779 6. The character tables are now in a separate module whose source is generated
4780 by an auxiliary program - but can then be edited by hand if required. There are
4781 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4782 toupper() in the code.
4783
4784 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4785 make them global. Abolish the function for setting them, as the caller can now
4786 set them directly.
4787
4788
4789 Version 0.92 11-Sep-97
4790 ----------------------
4791
4792 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4793 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4794
4795 2. Caseless matching was not working in character classes if the characters in
4796 the pattern were in upper case.
4797
4798 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4799
4800 4. Make PCRE_ANCHORED public and accept as a compile option.
4801
4802 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4803 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4804 pass them.
4805
4806 6. Give an error if bad option bits passed at compile or run time.
4807
4808 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4809 pcretest to cause it to pass that flag.
4810
4811 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4812 options, and the first character, if set.
4813
4814 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4815
4816
4817 Version 0.91 10-Sep-97
4818 ----------------------
4819
4820 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4821 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4822
4823 2. PCRE was looping on encountering an indefinitely repeated back reference to
4824 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4825 Perl does - treats the match as successful.
4826
4827 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12