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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 983 - (show annotations)
Fri Jul 6 08:55:10 2012 UTC (7 years, 5 months ago) by ph10
File size: 232754 byte(s)
Fix 2 incorrect #ifdefs in pcre_study.c.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.31 06-July-2012
5 -------------------------
6
7 1. Fixing a wrong JIT test case and some compiler warnings.
8
9 2. Removed a bashism from the RunTest script.
10
11 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12 to unsigned type, result still unsigned" that was given by an MS compiler
13 on encountering the code "-sizeof(xxx)".
14
15 4. Partial matching support is added to the JIT compiler.
16
17 5. Fixed several bugs concerned with partial matching of items that consist
18 of more than one character:
19
20 (a) /^(..)\1/ did not partially match "aba" because checking references was
21 done on an "all or nothing" basis. This also applied to repeated
22 references.
23
24 (b) \R did not give a hard partial match if \r was found at the end of the
25 subject.
26
27 (c) \X did not give a hard partial match after matching one or more
28 characters at the end of the subject.
29
30 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31 a partial match for the string "\r".
32
33 (e) When newline was set to CRLF, the metacharacter "." did not recognize
34 a partial match for a CR character at the end of the subject string.
35
36 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37 running pcretest, the text "(JIT)" added to the output whenever JIT is
38 actually used to run the match.
39
40 7. Individual JIT compile options can be set in pcretest by following -s+[+]
41 or /S+[+] with a digit between 1 and 7.
42
43 8. OP_NOT now supports any UTF character not just single-byte ones.
44
45 9. (*MARK) control verb is now supported by the JIT compiler.
46
47 10. The command "./RunTest list" lists the available tests without actually
48 running any of them. (Because I keep forgetting what they all are.)
49
50 11. Add PCRE_INFO_MAXLOOKBEHIND.
51
52 12. Applied a (slightly modified) user-supplied patch that improves performance
53 when the heap is used for recursion (compiled with --disable-stack-for-
54 recursion). Instead of malloc and free for each heap frame each time a
55 logical recursion happens, frames are retained on a chain and re-used where
56 possible. This sometimes gives as much as 30% improvement.
57
58 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59 call.
60
61 14. As documented, (*COMMIT) is now confined to within a positive assertion.
62
63 15. It is now possible to link pcretest with libedit as an alternative to
64 libreadline.
65
66 16. (*COMMIT) control verb is now supported by the JIT compiler.
67
68 17. The Unicode data tables have been updated to Unicode 6.1.0.
69
70 18. Added --file-list option to pcregrep.
71
72 19. Added binary file support to pcregrep, including the -a, --binary-files,
73 -I, and --text options.
74
75 20. The madvise function is renamed for posix_madvise for QNX compatibility
76 reasons. Fixed by Giuseppe D'Angelo.
77
78 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79
80 22. Changed the option for creating symbolic links for 16-bit man pages from
81 -s to -sf so that re-installing does not cause issues.
82
83 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
84
85 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86 matches in certain environments (the workspace was not being correctly
87 retained). Also added to pcre_dfa_exec() a simple plausibility check on
88 some of the workspace data at the beginning of a restart.
89
90 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
91 was not doing so when it should - probably a typo introduced by SVN 528
92 (change 8.10/14).
93
94 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
95 \w+ when the character tables indicated that \x{c4} was a word character.
96 There were several related cases, all because the tests for doing a table
97 lookup were testing for characters less than 127 instead of 255.
98
99 27. If a pattern contains capturing parentheses that are not used in a match,
100 their slots in the ovector are set to -1. For those that are higher than
101 any matched groups, this happens at the end of processing. In the case when
102 there were back references that the ovector was too small to contain
103 (causing temporary malloc'd memory to be used during matching), and the
104 highest capturing number was not used, memory off the end of the ovector
105 was incorrectly being set to -1. (It was using the size of the temporary
106 memory instead of the true size.)
107
108 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
109 ovector size, it uses memory at the end of the block that it has got.
110
111 29. Check for an overlong MARK name and give an error at compile time. The
112 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
113
114 30. JIT compiler update.
115
116 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
117 Rill for the patch.
118
119 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
120
121 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
122
123 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
124 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
125 was enabled.
126
127 35. Improve JIT code generation for greedy plus quantifier.
128
129 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
130 1 to "aa" instead of to an empty string. The bug affected repeated groups
131 that could potentially match an empty string.
132
133 37. Optimizing single character iterators in JIT.
134
135 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
136 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
137 codepoints that are too big for the mode are faulted, and in a UTF mode,
138 disallowed codepoints are also faulted.
139
140 39. If PCRE was compiled with UTF support, in three places in the DFA
141 matcher there was code that should only have been obeyed in UTF mode, but
142 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
143 processing when bytes with values greater than 127 were present. In 16-bit
144 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
145 both cases the values are those that cannot be the first data item in a UTF
146 character. The three items that might have provoked this were recursions,
147 possessively repeated groups, and atomic groups.
148
149 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
150 and pcregrep, because some OS require shared objects to be explicitly
151 passed to ld, causing the link step to fail if they are not.
152
153 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
154 mode, patterns that started with \h* or \R* might be incorrectly matched.
155
156
157 Version 8.30 04-February-2012
158 -----------------------------
159
160 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
161 name is defined in ctype.h.
162
163 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
164 only in quite long subpatterns.
165
166 3. Removed the function pcre_info(), which has been obsolete and deprecated
167 since it was replaced by pcre_fullinfo() in February 2000.
168
169 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
170 match a (*MARK), and the match failed at the start of the subject, a
171 reference to memory before the start of the subject could occur. This bug
172 was introduced by fix 17 of release 8.21.
173
174 5. A reference to an unset group with zero minimum repetition was giving
175 totally wrong answers (in non-JavaScript-compatibility mode). For example,
176 /(another)?(\1?)test/ matched against "hello world test". This bug was
177 introduced in release 8.13.
178
179 6. Add support for 16-bit character strings (a large amount of work involving
180 many changes and refactorings).
181
182 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
183 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
184 from a file.
185
186 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
187 rounding is not applied in this particular case).
188
189 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
190 if they appear, or are escaped, in patterns.
191
192 10. Get rid of a number of -Wunused-but-set-variable warnings.
193
194 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
195 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
196 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
197 also returns the mark "x". This bug applied to capturing parentheses,
198 non-capturing parentheses, and atomic parentheses. It also applied to some
199 assertions.
200
201 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
202 information out of configure.ac instead of relying on pcre.h.generic, which
203 is not stored in the repository.
204
205 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
206 -lreadline.
207
208 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
209
210 15. Applied Graycode's patch to put the top-level frame on the stack rather
211 than the heap when not using the stack for recursion. This gives a
212 performance improvement in many cases when recursion is not deep.
213
214 16. Experimental code added to "pcretest -C" to output the stack frame size.
215
216
217 Version 8.21 12-Dec-2011
218 ------------------------
219
220 1. Updating the JIT compiler.
221
222 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
223 are added as well.
224
225 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
226 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
227 calling _pcre_jit_exec. Some extra comments are added.
228
229 4. (*MARK) settings inside atomic groups that do not contain any capturing
230 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
231 was introduced by change 18 for 8.20.
232
233 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
234 ECMA-262 standard.
235
236 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
237 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
238 This bug was probably introduced by change 9 of 8.13.
239
240 7. While fixing 6 above, I noticed that a number of other items were being
241 incorrectly rejected as "not fixed length". This arose partly because newer
242 opcodes had not been added to the fixed-length checking code. I have (a)
243 corrected the bug and added tests for these items, and (b) arranged for an
244 error to occur if an unknown opcode is encountered while checking for fixed
245 length instead of just assuming "not fixed length". The items that were
246 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
247 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
248 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
249
250 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
251 being incorrectly compiled and would have given unpredicatble results.
252
253 9. A possessively repeated subpattern with minimum repeat count greater than
254 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
255 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
256 the first (A) could occur when it should not.
257
258 10. Add a cast and remove a redundant test from the code.
259
260 11. JIT should use pcre_malloc/pcre_free for allocation.
261
262 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
263 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
264 is anything other than /usr, -L is still shown).
265
266 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
267
268 14. Perl does not support \N without a following name in a [] class; PCRE now
269 also gives an error.
270
271 15. If a forward reference was repeated with an upper limit of around 2000,
272 it caused the error "internal error: overran compiling workspace". The
273 maximum number of forward references (including repeats) was limited by the
274 internal workspace, and dependent on the LINK_SIZE. The code has been
275 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
276 the default depends on LINK_SIZE. There is a new upper limit (for safety)
277 of around 200,000 forward references. While doing this, I also speeded up
278 the filling in of repeated forward references.
279
280 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
281 incorrectly expecting the subject to contain another "a" after the start.
282
283 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
284 in the match, the SKIP should be ignored. This was not happening; instead
285 the SKIP was being treated as NOMATCH. For patterns such as
286 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
287 tested.
288
289 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
290 now much more compatible with Perl, in particular in cases where the result
291 is a non-match for a non-anchored pattern. For example, if
292 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
293 "m", where previously it did not return a name. A side effect of this
294 change is that for partial matches, the last encountered mark name is
295 returned, as for non matches. A number of tests that were previously not
296 Perl-compatible have been moved into the Perl-compatible test files. The
297 refactoring has had the pleasing side effect of removing one argument from
298 the match() function, thus reducing its stack requirements.
299
300 19. If the /S+ option was used in pcretest to study a pattern using JIT,
301 subsequent uses of /S (without +) incorrectly behaved like /S+.
302
303 21. Retrieve executable code size support for the JIT compiler and fixing
304 some warnings.
305
306 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
307 not work when the shorter character appeared right at the end of the
308 subject string.
309
310 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
311 systems.
312
313 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
314 output it when the /M option is used in pcretest.
315
316 25. The CheckMan script was not being included in the distribution. Also, added
317 an explicit "perl" to run Perl scripts from the PrepareRelease script
318 because this is reportedly needed in Windows.
319
320 26. If study data was being save in a file and studying had not found a set of
321 "starts with" bytes for the pattern, the data written to the file (though
322 never used) was taken from uninitialized memory and so caused valgrind to
323 complain.
324
325 27. Updated RunTest.bat as provided by Sheri Pierce.
326
327 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
328
329 29. Computation of memory usage for the table of capturing group names was
330 giving an unnecessarily large value.
331
332
333 Version 8.20 21-Oct-2011
334 ------------------------
335
336 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
337 a POSIX class. After further experiments with Perl, which convinced me that
338 Perl has bugs and confusions, a closing square bracket is no longer allowed
339 in a POSIX name. This bug also affected patterns with classes that started
340 with full stops.
341
342 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
343 captured substring, but while checking the failing first alternative,
344 substring 1 is temporarily captured. If the output vector supplied to
345 pcre_exec() was not big enough for this capture, the yield of the function
346 was still zero ("insufficient space for captured substrings"). This cannot
347 be totally fixed without adding another stack variable, which seems a lot
348 of expense for a edge case. However, I have improved the situation in cases
349 such as /(a)(b)x|abc/ matched against "abc", where the return code
350 indicates that fewer than the maximum number of slots in the ovector have
351 been set.
352
353 3. Related to (2) above: when there are more back references in a pattern than
354 slots in the output vector, pcre_exec() uses temporary memory during
355 matching, and copies in the captures as far as possible afterwards. It was
356 using the entire output vector, but this conflicts with the specification
357 that only 2/3 is used for passing back captured substrings. Now it uses
358 only the first 2/3, for compatibility. This is, of course, another edge
359 case.
360
361 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
362 main code base, and can be used by building with --enable-jit. When this is
363 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
364 runtime --no-jit option is given.
365
366 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
367 ovector, the return from the function was zero, implying that there were
368 other matches that did not fit. The correct "exactly full" value is now
369 returned.
370
371 6. If a subpattern that was called recursively or as a subroutine contained
372 (*PRUNE) or any other control that caused it to give a non-standard return,
373 invalid errors such as "Error -26 (nested recursion at the same subject
374 position)" or even infinite loops could occur.
375
376 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
377 computing the minimum length on reaching *ACCEPT, and so ended up with the
378 wrong value of 1 rather than 0. Further investigation indicates that
379 computing a minimum subject length in the presence of *ACCEPT is difficult
380 (think back references, subroutine calls), and so I have changed the code
381 so that no minimum is registered for a pattern that contains *ACCEPT.
382
383 8. If (*THEN) was present in the first (true) branch of a conditional group,
384 it was not handled as intended. [But see 16 below.]
385
386 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
387 Sheri Pierce.
388
389 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
390 the first byte in a match must be "a".
391
392 11. Change 17 for 8.13 increased the recursion depth for patterns like
393 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
394 pattern contains any instances of (*THEN). If it does not, the old
395 optimizations are restored. It would be nice to do this on a per-group
396 basis, but at the moment that is not feasible.
397
398 12. In some environments, the output of pcretest -C is CRLF terminated. This
399 broke RunTest's code that checks for the link size. A single white space
400 character after the value is now allowed for.
401
402 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
403 For "fr", it uses the Windows-specific input and output files.
404
405 14. If (*THEN) appeared in a group that was called recursively or as a
406 subroutine, it did not work as intended. [But see next item.]
407
408 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
409 pattern fragments (but not containing any | characters). If A and B are
410 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
411 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
412 D. In other words, Perl considers parentheses that do not contain any |
413 characters to be part of a surrounding alternative, whereas PCRE was
414 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
415 differently. PCRE now behaves in the same way as Perl, except in the case
416 of subroutine/recursion calls such as (?1) which have in any case always
417 been different (but PCRE had them first :-).
418
419 16. Related to 15 above: Perl does not treat the | in a conditional group as
420 creating alternatives. Such a group is treated in the same way as an
421 ordinary group without any | characters when processing (*THEN). PCRE has
422 been changed to match Perl's behaviour.
423
424 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
425 RunGrepTest script failed.
426
427 18. Change 22 for version 13 caused atomic groups to use more stack. This is
428 inevitable for groups that contain captures, but it can lead to a lot of
429 stack use in large patterns. The old behaviour has been restored for atomic
430 groups that do not contain any capturing parentheses.
431
432 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
433 suppress the check for a minimum subject length at run time. (If it was
434 given to pcre_exec() or pcre_dfa_exec() it did work.)
435
436 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
437 fail to work when decoding hex characters in data strings in EBCDIC
438 environments.
439
440 21. It appears that in at least one Mac OS environment, the isxdigit() function
441 is implemented as a macro that evaluates to its argument more than once,
442 contravening the C 90 Standard (I haven't checked a later standard). There
443 was an instance in pcretest which caused it to go wrong when processing
444 \x{...} escapes in subject strings. The has been rewritten to avoid using
445 things like p++ in the argument of isxdigit().
446
447
448 Version 8.13 16-Aug-2011
449 ------------------------
450
451 1. The Unicode data tables have been updated to Unicode 6.0.0.
452
453 2. Two minor typos in pcre_internal.h have been fixed.
454
455 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
456 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
457 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
458
459 4. There were a number of related bugs in the code for matching backrefences
460 caselessly in UTF-8 mode when codes for the characters concerned were
461 different numbers of bytes. For example, U+023A and U+2C65 are an upper
462 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
463 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
464 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
465 2-byte code at the end of the subject (it thought there wasn't enough data
466 left).
467
468 5. Comprehensive information about what went wrong is now returned by
469 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
470 as the output vector has at least 2 elements. The offset of the start of
471 the failing character and a reason code are placed in the vector.
472
473 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
474 now returned is for the first byte of the failing character, instead of the
475 last byte inspected. This is an incompatible change, but I hope it is small
476 enough not to be a problem. It makes the returned offset consistent with
477 pcre_exec() and pcre_dfa_exec().
478
479 7. pcretest now gives a text phrase as well as the error number when
480 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
481 failure, the offset and reason code are output.
482
483 8. When \R was used with a maximizing quantifier it failed to skip backwards
484 over a \r\n pair if the subsequent match failed. Instead, it just skipped
485 back over a single character (\n). This seems wrong (because it treated the
486 two characters as a single entity when going forwards), conflicts with the
487 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
488 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
489 has been changed.
490
491 9. Some internal refactoring has changed the processing so that the handling
492 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
493 time (the PCRE_DOTALL option was changed this way some time ago: version
494 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
495 which was always a bit of a fudge. It also means that there is one less
496 argument for the match() function, which reduces its stack requirements
497 slightly. This change also fixes an incompatibility with Perl: the pattern
498 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
499
500 10. More internal refactoring has drastically reduced the number of recursive
501 calls to match() for possessively repeated groups such as (abc)++ when
502 using pcre_exec().
503
504 11. While implementing 10, a number of bugs in the handling of groups were
505 discovered and fixed:
506
507 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
508 (a|)*(?1) gave a compile-time internal error.
509 ((a|)+)+ did not notice that the outer group could match an empty string.
510 (^a|^)+ was not marked as anchored.
511 (.*a|.*)+ was not marked as matching at start or after a newline.
512
513 12. Yet more internal refactoring has removed another argument from the match()
514 function. Special calls to this function are now indicated by setting a
515 value in a variable in the "match data" data block.
516
517 13. Be more explicit in pcre_study() instead of relying on "default" for
518 opcodes that mean there is no starting character; this means that when new
519 ones are added and accidentally left out of pcre_study(), testing should
520 pick them up.
521
522 14. The -s option of pcretest has been documented for ages as being an old
523 synonym of -m (show memory usage). I have changed it to mean "force study
524 for every regex", that is, assume /S for every regex. This is similar to -i
525 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
526 using it. It makes it easier to run collections of tests with and without
527 study enabled, and thereby test pcre_study() more easily. All the standard
528 tests are now run with and without -s (but some patterns can be marked as
529 "never study" - see 20 below).
530
531 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
532 restoration of the capturing data to the outer values was not happening
533 correctly.
534
535 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
536 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
537 pattern had matched an empty string, and so incorrectly returned a no
538 match.
539
540 17. There was optimizing code for the last branch of non-capturing parentheses,
541 and also for the obeyed branch of a conditional subexpression, which used
542 tail recursion to cut down on stack usage. Unfortunately, now that there is
543 the possibility of (*THEN) occurring in these branches, tail recursion is
544 no longer possible because the return has to be checked for (*THEN). These
545 two optimizations have therefore been removed. [But see 8.20/11 above.]
546
547 18. If a pattern containing \R was studied, it was assumed that \R always
548 matched two bytes, thus causing the minimum subject length to be
549 incorrectly computed because \R can also match just one byte.
550
551 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
552 was incorrectly computed.
553
554 20. If /S is present twice on a test pattern in pcretest input, it now
555 *disables* studying, thereby overriding the use of -s on the command line
556 (see 14 above). This is necessary for one or two tests to keep the output
557 identical in both cases.
558
559 21. When (*ACCEPT) was used in an assertion that matched an empty string and
560 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
561
562 22. When an atomic group that contained a capturing parenthesis was
563 successfully matched, but the branch in which it appeared failed, the
564 capturing was not being forgotten if a higher numbered group was later
565 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
566 group 1 to "a", when in fact it should be unset. This applied to multi-
567 branched capturing and non-capturing groups, repeated or not, and also to
568 positive assertions (capturing in negative assertions does not happen
569 in PCRE) and also to nested atomic groups.
570
571 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
572 subject after a captured substring, to make it easier to tell which of a
573 number of identical substrings has been captured.
574
575 24. The way atomic groups are processed by pcre_exec() has been changed so that
576 if they are repeated, backtracking one repetition now resets captured
577 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
578 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
579 "aaa". Previously, it would have been "a". As part of this code
580 refactoring, the way recursive calls are handled has also been changed.
581
582 25. If an assertion condition captured any substrings, they were not passed
583 back unless some other capturing happened later. For example, if
584 (?(?=(a))a) was matched against "a", no capturing was returned.
585
586 26. When studying a pattern that contained subroutine calls or assertions,
587 the code for finding the minimum length of a possible match was handling
588 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
589 group 1 called group 2 while simultaneously a separate group 2 called group
590 1). A stack overflow occurred in this case. I have fixed this by limiting
591 the recursion depth to 10.
592
593 27. Updated RunTest.bat in the distribution to the version supplied by Tom
594 Fortmann. This supports explicit test numbers on the command line, and has
595 argument validation and error reporting.
596
597 28. An instance of \X with an unlimited repeat could fail if at any point the
598 first character it looked at was a mark character.
599
600 29. Some minor code refactoring concerning Unicode properties and scripts
601 should reduce the stack requirement of match() slightly.
602
603 30. Added the '=' option to pcretest to check the setting of unused capturing
604 slots at the end of the pattern, which are documented as being -1, but are
605 not included in the return count.
606
607 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
608 compiled something random. Now it gives a compile-time error (as does
609 Perl).
610
611 32. A *MARK encountered during the processing of a positive assertion is now
612 recorded and passed back (compatible with Perl).
613
614 33. If --only-matching or --colour was set on a pcregrep call whose pattern
615 had alternative anchored branches, the search for a second match in a line
616 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
617 matched the line "0102" twice. The same bug affected patterns that started
618 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
619 twice.
620
621 34. Previously, PCRE did not allow quantification of assertions. However, Perl
622 does, and because of capturing effects, quantifying parenthesized
623 assertions may at times be useful. Quantifiers are now allowed for
624 parenthesized assertions.
625
626 35. A minor code tidy in pcre_compile() when checking options for \R usage.
627
628 36. \g was being checked for fancy things in a character class, when it should
629 just be a literal "g".
630
631 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
632 appearance of a nested POSIX class supersedes an apparent external class.
633 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
634 unescaped square brackets may also appear as part of class names. For
635 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
636 more like Perl. (But see 8.20/1 above.)
637
638 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
639 was because it thought it was \N{name}, which is not supported).
640
641 39. Add minix to OS list not supporting the -S option in pcretest.
642
643 40. PCRE tries to detect cases of infinite recursion at compile time, but it
644 cannot analyze patterns in sufficient detail to catch mutual recursions
645 such as ((?1))((?2)). There is now a runtime test that gives an error if a
646 subgroup is called recursively as a subpattern for a second time at the
647 same position in the subject string. In previous releases this might have
648 been caught by the recursion limit, or it might have run out of stack.
649
650 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
651 happen only once. PCRE was, however incorrectly giving a compile time error
652 "recursive call could loop indefinitely" because it cannot analyze the
653 pattern in sufficient detail. The compile time test no longer happens when
654 PCRE is compiling a conditional subpattern, but actual runaway loops are
655 now caught at runtime (see 40 above).
656
657 42. It seems that Perl allows any characters other than a closing parenthesis
658 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
659 has been changed to be the same.
660
661 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
662 as not to get warnings when autogen.sh is called. Also changed
663 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
664
665 44. To help people who use pcregrep to scan files containing exceedingly long
666 lines, the following changes have been made:
667
668 (a) The default value of the buffer size parameter has been increased from
669 8K to 20K. (The actual buffer used is three times this size.)
670
671 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
672 PCRE is built.
673
674 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
675 to be set at run time.
676
677 (d) Numerical values in pcregrep options can be followed by K or M, for
678 example --buffer-size=50K.
679
680 (e) If a line being scanned overflows pcregrep's buffer, an error is now
681 given and the return code is set to 2.
682
683 45. Add a pointer to the latest mark to the callout data block.
684
685 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
686 partial match of an empty string instead of no match. This was specific to
687 the use of ".".
688
689 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
690 complete match instead of a partial match. This bug was dependent on both
691 the PCRE_UTF8 and PCRE_DOTALL options being set.
692
693 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
694 starting byte set, because \b was not being ignored.
695
696
697 Version 8.12 15-Jan-2011
698 ------------------------
699
700 1. Fixed some typos in the markup of the man pages, and wrote a script that
701 checks for such things as part of the documentation building process.
702
703 2. On a big-endian 64-bit system, pcregrep did not correctly process the
704 --match-limit and --recursion-limit options (added for 8.11). In
705 particular, this made one of the standard tests fail. (The integer value
706 went into the wrong half of a long int.)
707
708 3. If the --colour option was given to pcregrep with -v (invert match), it
709 did strange things, either producing crazy output, or crashing. It should,
710 of course, ignore a request for colour when reporting lines that do not
711 match.
712
713 4. Another pcregrep bug caused similar problems if --colour was specified with
714 -M (multiline) and the pattern match finished with a line ending.
715
716 5. In pcregrep, when a pattern that ended with a literal newline sequence was
717 matched in multiline mode, the following line was shown as part of the
718 match. This seems wrong, so I have changed it.
719
720 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
721 the check for further matches in the same line (so they could be coloured)
722 to overrun the end of the current line. If another match was found, it was
723 incorrectly shown (and then shown again when found in the next line).
724
725 7. If pcregrep was compiled under Windows, there was a reference to the
726 function pcregrep_exit() before it was defined. I am assuming this was
727 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
728 reported by a user. I've moved the definition above the reference.
729
730
731 Version 8.11 10-Dec-2010
732 ------------------------
733
734 1. (*THEN) was not working properly if there were untried alternatives prior
735 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
736 backtracked to try for "b" instead of moving to the next alternative branch
737 at the same level (in this case, to look for "c"). The Perl documentation
738 is clear that when (*THEN) is backtracked onto, it goes to the "next
739 alternative in the innermost enclosing group".
740
741 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
742 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
743 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
744 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
745 (*THEN).
746
747 3. If \s appeared in a character class, it removed the VT character from
748 the class, even if it had been included by some previous item, for example
749 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
750 of \s, but is part of the POSIX "space" class.)
751
752 4. A partial match never returns an empty string (because you can always
753 match an empty string at the end of the subject); however the checking for
754 an empty string was starting at the "start of match" point. This has been
755 changed to the "earliest inspected character" point, because the returned
756 data for a partial match starts at this character. This means that, for
757 example, /(?<=abc)def/ gives a partial match for the subject "abc"
758 (previously it gave "no match").
759
760 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
761 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
762 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
763 has an implication that the given string is incomplete (because a partial
764 match is preferred over a full match). For this reason, these items now
765 give a partial match in this situation. [Aside: previously, the one case
766 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
767 match rather than a full match, which was wrong by the old rules, but is
768 now correct.]
769
770 6. There was a bug in the handling of #-introduced comments, recognized when
771 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
772 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
773 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
774 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
775 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
776 places in pcre_compile().
777
778 7. Related to (6) above, when pcre_compile() was skipping #-introduced
779 comments when looking ahead for named forward references to subpatterns,
780 the only newline sequence it recognized was NL. It now handles newlines
781 according to the set newline convention.
782
783 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
784 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
785 cater for a lack of strerror(). These oversights have been fixed.
786
787 9. Added --match-limit and --recursion-limit to pcregrep.
788
789 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
790
791 11. When the -o option was used, pcregrep was setting a return code of 1, even
792 when matches were found, and --line-buffered was not being honoured.
793
794 12. Added an optional parentheses number to the -o and --only-matching options
795 of pcregrep.
796
797 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
798 can match an empty string. The code to do it in pcretest and pcredemo
799 needed fixing:
800
801 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
802 only one byte after an empty string match just before CRLF (this case
803 just got forgotten; "any" and "anycrlf" were OK).
804
805 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
806 mode when an empty string match preceded an ASCII character followed by
807 a non-ASCII character. (The code for advancing by one character rather
808 than one byte was nonsense.)
809
810 (c) The pcredemo.c sample program did not have any code at all to handle
811 the cases when CRLF is a valid newline sequence.
812
813 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
814 as a starting offset was within the subject string. There is now a new
815 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
816 negative or greater than the length of the string. In order to test this,
817 pcretest is extended to allow the setting of negative starting offsets.
818
819 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
820 starting offset points to the beginning of a UTF-8 character was
821 unnecessarily clumsy. I tidied it up.
822
823 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
824 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
825
826 17. Nobody had reported that the --include_dir option, which was added in
827 release 7.7 should have been called --include-dir (hyphen, not underscore)
828 for compatibility with GNU grep. I have changed it to --include-dir, but
829 left --include_dir as an undocumented synonym, and the same for
830 --exclude-dir, though that is not available in GNU grep, at least as of
831 release 2.5.4.
832
833 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
834 characters from a string of bytes) have been redefined so as not to use
835 loops, in order to improve performance in some environments. At the same
836 time, I abstracted some of the common code into auxiliary macros to save
837 repetition (this should not affect the compiled code).
838
839 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
840 compile-time error is now given if \c is not followed by an ASCII
841 character, that is, a byte less than 128. (In EBCDIC mode, the code is
842 different, and any byte value is allowed.)
843
844 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
845 START_OPTIMIZE option, which is now allowed at compile time - but just
846 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
847 to pcregrep and other applications that have no direct access to PCRE
848 options. The new /Y option in pcretest sets this option when calling
849 pcre_compile().
850
851 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
852 back references. Groups containing recursive back references were forced to
853 be atomic by that change, but in the case of named groups, the amount of
854 memory required was incorrectly computed, leading to "Failed: internal
855 error: code overflow". This has been fixed.
856
857 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
858 pcretest.c, to avoid build problems in some Borland environments.
859
860
861 Version 8.10 25-Jun-2010
862 ------------------------
863
864 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
865 THEN.
866
867 2. (*ACCEPT) was not working when inside an atomic group.
868
869 3. Inside a character class, \B is treated as a literal by default, but
870 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
871 causes the error). The code is unchanged, but I tidied the documentation.
872
873 4. Inside a character class, PCRE always treated \R and \X as literals,
874 whereas Perl faults them if its -w option is set. I have changed PCRE so
875 that it faults them when PCRE_EXTRA is set.
876
877 5. Added support for \N, which always matches any character other than
878 newline. (It is the same as "." when PCRE_DOTALL is not set.)
879
880 6. When compiling pcregrep with newer versions of gcc which may have
881 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
882 declared with attribute warn_unused_result" were given. Just casting the
883 result to (void) does not stop the warnings; a more elaborate fudge is
884 needed. I've used a macro to implement this.
885
886 7. Minor change to pcretest.c to avoid a compiler warning.
887
888 8. Added four artifical Unicode properties to help with an option to make
889 \s etc use properties (see next item). The new properties are: Xan
890 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
891
892 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
893 use Unicode properties. (*UCP) at the start of a pattern can be used to set
894 this option. Modified pcretest to add /W to test this facility. Added
895 REG_UCP to make it available via the POSIX interface.
896
897 10. Added --line-buffered to pcregrep.
898
899 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
900 studied, and the match started with a letter with a code point greater than
901 127 whose first byte was different to the first byte of the other case of
902 the letter, the other case of this starting letter was not recognized
903 (#976).
904
905 12. If a pattern that was studied started with a repeated Unicode property
906 test, for example, \p{Nd}+, there was the theoretical possibility of
907 setting up an incorrect bitmap of starting bytes, but fortunately it could
908 not have actually happened in practice until change 8 above was made (it
909 added property types that matched character-matching opcodes).
910
911 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
912 possible starting bytes for non-anchored patterns.
913
914 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
915 \R, and also a number of cases that involve Unicode properties, both
916 explicit and implicit when PCRE_UCP is set.
917
918 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
919 input, it could crash or give wrong results if characters with values
920 greater than 0xc0 were present in the subject string. (Detail: it assumed
921 UTF-8 input when processing these items.)
922
923 16. Added a lot of (int) casts to avoid compiler warnings in systems where
924 size_t is 64-bit (#991).
925
926 17. Added a check for running out of memory when PCRE is compiled with
927 --disable-stack-for-recursion (#990).
928
929 18. If the last data line in a file for pcretest does not have a newline on
930 the end, a newline was missing in the output.
931
932 19. The default pcre_chartables.c file recognizes only ASCII characters (values
933 less than 128) in its various bitmaps. However, there is a facility for
934 generating tables according to the current locale when PCRE is compiled. It
935 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
936 characters, are recognized by isspace() and therefore were getting set in
937 these tables, and indeed these tables seem to approximate to ISO 8859. This
938 caused a problem in UTF-8 mode when pcre_study() was used to create a list
939 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
940 which of course cannot start UTF-8 characters. I have changed the code so
941 that only real ASCII characters (less than 128) and the correct starting
942 bytes for UTF-8 encodings are set for characters greater than 127 when in
943 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
944 altogether.)
945
946 20. Added the /T option to pcretest so as to be able to run tests with non-
947 standard character tables, thus making it possible to include the tests
948 used for 19 above in the standard set of tests.
949
950 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
951 reference to a subpattern the other side of a comment that contains an
952 opening parenthesis caused either an internal compiling error, or a
953 reference to the wrong subpattern.
954
955
956 Version 8.02 19-Mar-2010
957 ------------------------
958
959 1. The Unicode data tables have been updated to Unicode 5.2.0.
960
961 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
962 configured.
963
964 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
965 original author of that file, following a query about its status.
966
967 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
968 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
969
970 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
971 quantifier applied to a forward-referencing subroutine call, could compile
972 incorrect code or give the error "internal error: previously-checked
973 referenced subpattern not found".
974
975 6. Both MS Visual Studio and Symbian OS have problems with initializing
976 variables to point to external functions. For these systems, therefore,
977 pcre_malloc etc. are now initialized to local functions that call the
978 relevant global functions.
979
980 7. There were two entries missing in the vectors called coptable and poptable
981 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
982 I've fixed the data, and added a kludgy way of testing at compile time that
983 the lengths are correct (equal to the number of opcodes).
984
985 8. Following on from 7, I added a similar kludge to check the length of the
986 eint vector in pcreposix.c.
987
988 9. Error texts for pcre_compile() are held as one long string to avoid too
989 much relocation at load time. To find a text, the string is searched,
990 counting zeros. There was no check for running off the end of the string,
991 which could happen if a new error number was added without updating the
992 string.
993
994 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
995
996 11. \K was not working if it appeared in an atomic group or in a group that
997 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
998 \K is "not well defined" if used in an assertion. PCRE now accepts it if
999 the assertion is positive, but not if it is negative.
1000
1001 12. Change 11 fortuitously reduced the size of the stack frame used in the
1002 "match()" function of pcre_exec.c by one pointer. Forthcoming
1003 implementation of support for (*MARK) will need an extra pointer on the
1004 stack; I have reserved it now, so that the stack frame size does not
1005 decrease.
1006
1007 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1008 item in branch that calls a recursion is a subroutine call - as in the
1009 second branch in the above example - was incorrectly given the compile-
1010 time error "recursive call could loop indefinitely" because pcre_compile()
1011 was not correctly checking the subroutine for matching a non-empty string.
1012
1013 14. The checks for overrunning compiling workspace could trigger after an
1014 overrun had occurred. This is a "should never occur" error, but it can be
1015 triggered by pathological patterns such as hundreds of nested parentheses.
1016 The checks now trigger 100 bytes before the end of the workspace.
1017
1018 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1019
1020
1021 Version 8.01 19-Jan-2010
1022 ------------------------
1023
1024 1. If a pattern contained a conditional subpattern with only one branch (in
1025 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1026 computed the wrong minimum data length (which is of course zero for such
1027 subpatterns). This could cause incorrect "no match" results.
1028
1029 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1030 the pattern is reset in the first branch, pcre_compile() failed with
1031 "internal error: code overflow at offset...". This happened only when
1032 the reset was to the original external option setting. (An optimization
1033 abstracts leading options settings into an external setting, which was the
1034 cause of this.)
1035
1036 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1037 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1038 assertion pattern did not match (meaning that the assertion was true), it
1039 was incorrectly treated as false if the SKIP had been reached during the
1040 matching. This also applied to assertions used as conditions.
1041
1042 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1043 assertion subpattern, including such a pattern used as a condition,
1044 unpredictable results occurred, instead of the error return
1045 PCRE_ERROR_DFA_UITEM.
1046
1047 5. The C++ GlobalReplace function was not working like Perl for the special
1048 situation when an empty string is matched. It now does the fancy magic
1049 stuff that is necessary.
1050
1051 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1052 removed. (These were left over from very, very early versions of PCRE.)
1053
1054 7. Some cosmetic changes to the code to make life easier when compiling it
1055 as part of something else:
1056
1057 (a) Change DEBUG to PCRE_DEBUG.
1058
1059 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1060 called "current" as "current_branch", to prevent a collision with the
1061 Linux macro when compiled as a kernel module.
1062
1063 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1064 prevent a collision with the Linux macro when compiled as a kernel
1065 module.
1066
1067 8. In pcre_compile() there are some checks for integer overflows that used to
1068 cast potentially large values to (double). This has been changed to that
1069 when building, a check for int64_t is made, and if it is found, it is used
1070 instead, thus avoiding the use of floating point arithmetic. (There is no
1071 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1072 double.
1073
1074 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1075 2005 (difference between two addresses compared to an unsigned value).
1076
1077 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1078 custom one, because of the following reported problem in Windows:
1079
1080 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1081 under Win32.
1082 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1083 therefore missing the function definition.
1084 - The compiler thus generates a "C" signature for the test function.
1085 - The linker fails to find the "C" function.
1086 - PCRE fails to configure if asked to do so against libbz2.
1087
1088 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1089 messages were output:
1090
1091 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1092 rerunning libtoolize, to keep the correct libtool macros in-tree.
1093 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1094
1095 I have done both of these things.
1096
1097 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1098 most of the time, it *can* run out if it is given a pattern that contains a
1099 runaway infinite recursion. I updated the discussion in the pcrestack man
1100 page.
1101
1102 13. Now that we have gone to the x.xx style of version numbers, the minor
1103 version may start with zero. Using 08 or 09 is a bad idea because users
1104 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1105 interpreted as invalid octal numbers. I've updated the previous comment in
1106 configure.ac, and also added a check that gives an error if 08 or 09 are
1107 used.
1108
1109 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1110 causing partial matching to fail when the end of the subject matched \W
1111 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1112
1113 15. There were some discrepancies between the declarations in pcre_internal.h
1114 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1115 their definitions. The declarations used "const uschar *" and the
1116 definitions used USPTR. Even though USPTR is normally defined as "const
1117 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1118 reported that: "This difference in casting confuses some C++ compilers, for
1119 example, SunCC recognizes above declarations as different functions and
1120 generates broken code for hbpcre." I have changed the declarations to use
1121 USPTR.
1122
1123 16. GNU libtool is named differently on some systems. The autogen.sh script now
1124 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1125 (FreeBSD).
1126
1127 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1128 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1129 comment: "Figure out how to create a longlong from a string: strtoll and
1130 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1131 instance, but it only takes 2 args instead of 3!"
1132
1133 18. A subtle bug concerned with back references has been fixed by a change of
1134 specification, with a corresponding code fix. A pattern such as
1135 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1136 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1137 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1138 same bug. Such groups have to be quantified to be useful, or contained
1139 inside another quantified group. (If there's no repetition, the reference
1140 can never match.) The problem arises because, having left the group and
1141 moved on to the rest of the pattern, a later failure that backtracks into
1142 the group uses the captured value from the final iteration of the group
1143 rather than the correct earlier one. I have fixed this in PCRE by forcing
1144 any group that contains a reference to itself to be an atomic group; that
1145 is, there cannot be any backtracking into it once it has completed. This is
1146 similar to recursive and subroutine calls.
1147
1148
1149 Version 8.00 19-Oct-09
1150 ----------------------
1151
1152 1. The table for translating pcre_compile() error codes into POSIX error codes
1153 was out-of-date, and there was no check on the pcre_compile() error code
1154 being within the table. This could lead to an OK return being given in
1155 error.
1156
1157 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1158 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1159 in a Windows environment.
1160
1161 3. The pcregrep --count option prints the count for each file even when it is
1162 zero, as does GNU grep. However, pcregrep was also printing all files when
1163 --files-with-matches was added. Now, when both options are given, it prints
1164 counts only for those files that have at least one match. (GNU grep just
1165 prints the file name in this circumstance, but including the count seems
1166 more useful - otherwise, why use --count?) Also ensured that the
1167 combination -clh just lists non-zero counts, with no names.
1168
1169 4. The long form of the pcregrep -F option was incorrectly implemented as
1170 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1171 but it seems right to fix it, and I didn't think it was worth preserving
1172 the old behaviour.
1173
1174 5. The command line items --regex=pattern and --regexp=pattern were not
1175 recognized by pcregrep, which required --regex pattern or --regexp pattern
1176 (with a space rather than an '='). The man page documented the '=' forms,
1177 which are compatible with GNU grep; these now work.
1178
1179 6. No libpcreposix.pc file was created for pkg-config; there was just
1180 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1181
1182 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1183 when UCP support is not needed, by modifying the Python script that
1184 generates it from Unicode data files. This should not matter if the module
1185 is correctly used as a library, but I received one complaint about 50K of
1186 unwanted data. My guess is that the person linked everything into his
1187 program rather than using a library. Anyway, it does no harm.
1188
1189 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1190 was a minimum greater than 1 for a wide character in a possessive
1191 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1192 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1193 character. Chaos in the form of incorrect output or a compiling loop could
1194 result.
1195
1196 9. The restrictions on what a pattern can contain when partial matching is
1197 requested for pcre_exec() have been removed. All patterns can now be
1198 partially matched by this function. In addition, if there are at least two
1199 slots in the offset vector, the offset of the earliest inspected character
1200 for the match and the offset of the end of the subject are set in them when
1201 PCRE_ERROR_PARTIAL is returned.
1202
1203 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1204 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1205 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1206 and may be more useful for multi-segment matching.
1207
1208 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1209 used to be given if ever the end of the subject was reached; now it is
1210 given only if matching could not proceed because another character was
1211 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1212 string "Z", which now yields "no match" instead of "partial match". In the
1213 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1214 final character ended with (*FAIL).
1215
1216 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1217 if the pattern had a "must contain" character that was already found in the
1218 earlier partial match, unless partial matching was again requested. For
1219 example, with the pattern /dog.(body)?/, the "must contain" character is
1220 "g". If the first part-match was for the string "dog", restarting with
1221 "sbody" failed. This bug has been fixed.
1222
1223 13. The string returned by pcre_dfa_exec() after a partial match has been
1224 changed so that it starts at the first inspected character rather than the
1225 first character of the match. This makes a difference only if the pattern
1226 starts with a lookbehind assertion or \b or \B (\K is not supported by
1227 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1228 matching functions compatible, and I think it's the right thing to do.
1229
1230 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1231 so that the demonstration program is easily available in environments where
1232 PCRE has not been installed from source.
1233
1234 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1235 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1236 library.
1237
1238 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1239 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1240 is not the first non-POSIX option to be added. Clearly some people find
1241 these options useful.
1242
1243 17. If a caller to the POSIX matching function regexec() passes a non-zero
1244 value for nmatch with a NULL value for pmatch, the value of
1245 nmatch is forced to zero.
1246
1247 18. RunGrepTest did not have a test for the availability of the -u option of
1248 the diff command, as RunTest does. It now checks in the same way as
1249 RunTest, and also checks for the -b option.
1250
1251 19. If an odd number of negated classes containing just a single character
1252 interposed, within parentheses, between a forward reference to a named
1253 subpattern and the definition of the subpattern, compilation crashed with
1254 an internal error, complaining that it could not find the referenced
1255 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1256 [The bug was that it was starting one character too far in when skipping
1257 over the character class, thus treating the ] as data rather than
1258 terminating the class. This meant it could skip too much.]
1259
1260 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1261 /g option in pcretest when the pattern contains \K, which makes it possible
1262 to have an empty string match not at the start, even when the pattern is
1263 anchored. Updated pcretest and pcredemo to use this option.
1264
1265 21. If the maximum number of capturing subpatterns in a recursion was greater
1266 than the maximum at the outer level, the higher number was returned, but
1267 with unset values at the outer level. The correct (outer level) value is
1268 now given.
1269
1270 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1271 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1272 make it do so. The string so far is captured, making this feature
1273 compatible with Perl.
1274
1275 23. The tests have been re-organized, adding tests 11 and 12, to make it
1276 possible to check the Perl 5.10 features against Perl 5.10.
1277
1278 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1279 pattern matches a fixed length string. PCRE did not allow this; now it
1280 does. Neither allows recursion.
1281
1282 25. I finally figured out how to implement a request to provide the minimum
1283 length of subject string that was needed in order to match a given pattern.
1284 (It was back references and recursion that I had previously got hung up
1285 on.) This code has now been added to pcre_study(); it finds a lower bound
1286 to the length of subject needed. It is not necessarily the greatest lower
1287 bound, but using it to avoid searching strings that are too short does give
1288 some useful speed-ups. The value is available to calling programs via
1289 pcre_fullinfo().
1290
1291 26. While implementing 25, I discovered to my embarrassment that pcretest had
1292 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1293 study optimizations had never been tested with that matching function.
1294 Oops. What is worse, even when it was passed study data, there was a bug in
1295 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1296 were also very few tests of studied patterns with pcre_dfa_exec().
1297
1298 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1299 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1300 on the other side of the coin, they are no longer allowed to have different
1301 names, because these cannot be distinguished in PCRE, and this has caused
1302 confusion. (This is a difference from Perl.)
1303
1304 28. When duplicate subpattern names are present (necessarily with different
1305 numbers, as required by 27 above), and a test is made by name in a
1306 conditional pattern, either for a subpattern having been matched, or for
1307 recursion in such a pattern, all the associated numbered subpatterns are
1308 tested, and the overall condition is true if the condition is true for any
1309 one of them. This is the way Perl works, and is also more like the way
1310 testing by number works.
1311
1312
1313 Version 7.9 11-Apr-09
1314 ---------------------
1315
1316 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1317 (pcretest), all targets were linked against these libraries. This included
1318 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1319 libraries. This caused unwanted dependencies to be created. This problem
1320 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1321 pcretest is linked with readline.
1322
1323 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1324 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1325 moved outside it again, because FALSE and TRUE are already defined in AIX,
1326 but BOOL is not.
1327
1328 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1329 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1330
1331 4. The pcregrep documentation said spaces were inserted as well as colons (or
1332 hyphens) following file names and line numbers when outputting matching
1333 lines. This is not true; no spaces are inserted. I have also clarified the
1334 wording for the --colour (or --color) option.
1335
1336 5. In pcregrep, when --colour was used with -o, the list of matching strings
1337 was not coloured; this is different to GNU grep, so I have changed it to be
1338 the same.
1339
1340 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1341 each matching line was coloured. Now it goes on to look for further matches
1342 of any of the test patterns, which is the same behaviour as GNU grep.
1343
1344 7. A pattern that could match an empty string could cause pcregrep to loop; it
1345 doesn't make sense to accept an empty string match in pcregrep, so I have
1346 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1347 seems to be how GNU grep behaves.
1348
1349 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1350 start or after a newline", because the conditional assertion was not being
1351 correctly handled. The rule now is that both the assertion and what follows
1352 in the first alternative must satisfy the test.
1353
1354 9. If auto-callout was enabled in a pattern with a conditional group whose
1355 condition was an assertion, PCRE could crash during matching, both with
1356 pcre_exec() and pcre_dfa_exec().
1357
1358 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1359 used for matching.
1360
1361 11. Unicode property support in character classes was not working for
1362 characters (bytes) greater than 127 when not in UTF-8 mode.
1363
1364 12. Added the -M command line option to pcretest.
1365
1366 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1367
1368 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1369
1370 16. Added comments and documentation about mis-use of no_arg in the C++
1371 wrapper.
1372
1373 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1374 from Martin Jerabek that uses macro names for all relevant character and
1375 string constants.
1376
1377 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1378 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1379 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1380 these, but not everybody uses configure.
1381
1382 19. A conditional group that had only one branch was not being correctly
1383 recognized as an item that could match an empty string. This meant that an
1384 enclosing group might also not be so recognized, causing infinite looping
1385 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1386 with the subject "ab", where knowledge that the repeated group can match
1387 nothing is needed in order to break the loop.
1388
1389 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1390 exec(), but without supplying a callout function, matching went wrong.
1391
1392 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1393 leak if the size of the offset vector was greater than 30. When the vector
1394 is smaller, the saved offsets during recursion go onto a local stack
1395 vector, but for larger vectors malloc() is used. It was failing to free
1396 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1397 error, in fact).
1398
1399 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1400 heapframe that is used only when UTF-8 support is enabled. This caused no
1401 problem, but was untidy.
1402
1403 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1404 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1405 included within another project.
1406
1407 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1408 slightly modified by me:
1409
1410 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1411 not building pcregrep.
1412
1413 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1414 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1415
1416 25. Forward references, both numeric and by name, in patterns that made use of
1417 duplicate group numbers, could behave incorrectly or give incorrect errors,
1418 because when scanning forward to find the reference group, PCRE was not
1419 taking into account the duplicate group numbers. A pattern such as
1420 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1421
1422 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1423 the feature of a custom pointer more persuasive (as requested by a user).
1424
1425 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1426 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1427 pre-defined.
1428
1429 28. Added support for (*UTF8) at the start of a pattern.
1430
1431 29. Arrange for flags added by the "release type" setting in CMake to be shown
1432 in the configuration summary.
1433
1434
1435 Version 7.8 05-Sep-08
1436 ---------------------
1437
1438 1. Replaced UCP searching code with optimized version as implemented for Ad
1439 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1440 stage table and inline lookup instead of a function, giving speed ups of 2
1441 to 5 times on some simple patterns that I tested. Permission was given to
1442 distribute the MultiStage2.py script that generates the tables (it's not in
1443 the tarball, but is in the Subversion repository).
1444
1445 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1446 scripts.
1447
1448 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1449 a group with a zero qualifier. The result of the study could be incorrect,
1450 or the function might crash, depending on the pattern.
1451
1452 4. Caseless matching was not working for non-ASCII characters in back
1453 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1454 It now works when Unicode Property Support is available.
1455
1456 5. In pcretest, an escape such as \x{de} in the data was always generating
1457 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1458 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1459 truncation.
1460
1461 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1462
1463 7. Added two (int) casts to pcregrep when printing the difference of two
1464 pointers, in case they are 64-bit values.
1465
1466 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1467 test 2 if it fails.
1468
1469 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1470 and a #define of that name to empty if it is not externally set. This is to
1471 allow users of MSVC to set it if necessary.
1472
1473 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1474 the convenience functions in the pcre_get.c source file.
1475
1476 11. An option change at the start of a pattern that had top-level alternatives
1477 could cause overwriting and/or a crash. This command provoked a crash in
1478 some environments:
1479
1480 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1481
1482 This potential security problem was recorded as CVE-2008-2371.
1483
1484 12. For a pattern where the match had to start at the beginning or immediately
1485 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1486 pcre_dfa_exec() could read past the end of the passed subject if there was
1487 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1488 pcretest so that it places the subject at the end of its malloc-ed buffer.
1489
1490 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1491 exec() might read past the end of the data buffer in UTF-8 mode.
1492
1493 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1494 the data contained the byte 0x85 as part of a UTF-8 character within its
1495 first line. This applied both to normal and DFA matching.
1496
1497 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1498 /^[^d]*?$/8 failed to match "abc".
1499
1500 16. Added a missing copyright notice to pcrecpp_internal.h.
1501
1502 17. Make it more clear in the documentation that values returned from
1503 pcre_exec() in ovector are byte offsets, not character counts.
1504
1505 18. Tidied a few places to stop certain compilers from issuing warnings.
1506
1507 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1508 supplied by Stefan Weber. I made a further small update for 7.8 because
1509 there is a change of source arrangements: the pcre_searchfuncs.c module is
1510 replaced by pcre_ucd.c.
1511
1512
1513 Version 7.7 07-May-08
1514 ---------------------
1515
1516 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1517 a string to a long long, pretend we don't even have a long long." This is
1518 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1519
1520 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1521 pre-7.6 versions, which defined a global no_arg variable instead of putting
1522 it in the RE class. (See also #8 below.)
1523
1524 3. Remove a line of dead code, identified by coverity and reported by Nuno
1525 Lopes.
1526
1527 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1528
1529 (1) The include/exclude patterns were being applied to the whole pathnames
1530 of files, instead of just to the final components.
1531
1532 (2) If there was more than one level of directory, the subdirectories were
1533 skipped unless they satisfied the include/exclude conditions. This is
1534 inconsistent with GNU grep (and could even be seen as contrary to the
1535 pcregrep specification - which I improved to make it absolutely clear).
1536 The action now is always to scan all levels of directory, and just
1537 apply the include/exclude patterns to regular files.
1538
1539 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1540 --exclude_dir in the tests to avoid scanning .svn directories.
1541
1542 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1543 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1544 doesn't support NULs in patterns.
1545
1546 7. Added some missing "const"s to declarations of static tables in
1547 pcre_compile.c and pcre_dfa_exec.c.
1548
1549 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1550 caused by fix #2 above. (Subsequently also a second patch to fix the
1551 first patch. And a third patch - this was a messy problem.)
1552
1553 9. Applied Craig's patch to remove the use of push_back().
1554
1555 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1556 matching function regexec().
1557
1558 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1559 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1560 references. PCRE supports relative numbers with this syntax (I don't think
1561 Oniguruma does).
1562
1563 12. Previously, a group with a zero repeat such as (...){0} was completely
1564 omitted from the compiled regex. However, this means that if the group
1565 was called as a subroutine from elsewhere in the pattern, things went wrong
1566 (an internal error was given). Such groups are now left in the compiled
1567 pattern, with a new opcode that causes them to be skipped at execution
1568 time.
1569
1570 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1571 to the way PCRE behaves:
1572
1573 (a) A lone ] character is dis-allowed (Perl treats it as data).
1574
1575 (b) A back reference to an unmatched subpattern matches an empty string
1576 (Perl fails the current match path).
1577
1578 (c) A data ] in a character class must be notated as \] because if the
1579 first data character in a class is ], it defines an empty class. (In
1580 Perl it is not possible to have an empty class.) The empty class []
1581 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1582 The negative empty class [^] matches any one character, independently
1583 of the DOTALL setting.
1584
1585 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1586 non-existent subpattern following a character class starting with ']' and
1587 containing () gave an internal compiling error instead of "reference to
1588 non-existent subpattern". Fortunately, when the pattern did exist, the
1589 compiled code was correct. (When scanning forwards to check for the
1590 existencd of the subpattern, it was treating the data ']' as terminating
1591 the class, so got the count wrong. When actually compiling, the reference
1592 was subsequently set up correctly.)
1593
1594 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1595 it was being rejected as not supported by pcre_dfa_exec(), even though
1596 other assertions are supported. I have made pcre_dfa_exec() support
1597 (*FAIL).
1598
1599 16. The implementation of 13c above involved the invention of a new opcode,
1600 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1601 cannot be changed at match time, I realized I could make a small
1602 improvement to matching performance by compiling OP_ALLANY instead of
1603 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1604 on the OP_ANY path.
1605
1606 17. Compiling pcretest on Windows with readline support failed without the
1607 following two fixes: (1) Make the unistd.h include conditional on
1608 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1609
1610 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1611 ncurses library to be included for pcretest when ReadLine support is
1612 requested, but also to allow for it to be overridden. This patch came from
1613 Daniel Bergström.
1614
1615 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1616 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1617 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1618 spotting this.
1619
1620
1621 Version 7.6 28-Jan-08
1622 ---------------------
1623
1624 1. A character class containing a very large number of characters with
1625 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1626 overflow.
1627
1628 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1629 HAVE_LONG_LONG is not defined.
1630
1631 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1632 bring them up to date and include new features. This patch includes:
1633
1634 - Fixed PH's badly added libz and libbz2 support.
1635 - Fixed a problem with static linking.
1636 - Added pcredemo. [But later removed - see 7 below.]
1637 - Fixed dftables problem and added an option.
1638 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1639 HAVE_LONG_LONG.
1640 - Added readline support for pcretest.
1641 - Added an listing of the option settings after cmake has run.
1642
1643 4. A user submitted a patch to Makefile that makes it easy to create
1644 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1645 Makefile.am that cause it to include this special target, without
1646 affecting anything else. Note that the same mingw target plus all
1647 the other distribution libraries and programs are now supported
1648 when configuring with CMake (see 6 below) instead of with
1649 Configure/Make.
1650
1651 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1652 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1653 exported in the Windows port". It has not yet been confirmed that the patch
1654 solves the problem, but it does no harm.
1655
1656 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1657 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1658 with CMake, and also correct the comment about stack recursion.
1659
1660 7. Remove the automatic building of pcredemo from the ./configure system and
1661 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1662 of a program that users should build themselves after PCRE is installed, so
1663 building it automatically is not really right. What is more, it gave
1664 trouble in some build environments.
1665
1666 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1667
1668
1669 Version 7.5 10-Jan-08
1670 ---------------------
1671
1672 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1673 values in parens when parsing an RE using the C++ wrapper."
1674
1675 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1676 Characters greater than 255 were excluded from the class instead of being
1677 included.
1678
1679 3. The same bug as (2) above applied to negated POSIX classes such as
1680 [:^space:].
1681
1682 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1683 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1684 I have changed it.
1685
1686 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1687 first named subpattern) and a construct such as (?&a) would reference the
1688 first named subpattern whose name started with "a" (in other words, the
1689 length check was missing). Both these problems are fixed. "Subpattern name
1690 expected" is now given for (?&) (a zero-length name), and this patch also
1691 makes it give the same error for \k'' (previously it complained that that
1692 was a reference to a non-existent subpattern).
1693
1694 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1695 this is right because (?- can be followed by option settings as well as by
1696 digits. I have, however, made the messages clearer.
1697
1698 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1699 than the number used in the conditional) now cause a compile-time error.
1700 This is actually not compatible with Perl, which accepts such patterns, but
1701 treats the conditional as always being FALSE (as PCRE used to), but it
1702 seems to me that giving a diagnostic is better.
1703
1704 8. Change "alphameric" to the more common word "alphanumeric" in comments
1705 and messages.
1706
1707 9. Fix two occurrences of "backslash" in comments that should have been
1708 "backspace".
1709
1710 10. Remove two redundant lines of code that can never be obeyed (their function
1711 was moved elsewhere).
1712
1713 11. The program that makes PCRE's Unicode character property table had a bug
1714 which caused it to generate incorrect table entries for sequences of
1715 characters that have the same character type, but are in different scripts.
1716 It amalgamated them into a single range, with the script of the first of
1717 them. In other words, some characters were in the wrong script. There were
1718 thirteen such cases, affecting characters in the following ranges:
1719
1720 U+002b0 - U+002c1
1721 U+0060c - U+0060d
1722 U+0061e - U+00612
1723 U+0064b - U+0065e
1724 U+0074d - U+0076d
1725 U+01800 - U+01805
1726 U+01d00 - U+01d77
1727 U+01d9b - U+01dbf
1728 U+0200b - U+0200f
1729 U+030fc - U+030fe
1730 U+03260 - U+0327f
1731 U+0fb46 - U+0fbb1
1732 U+10450 - U+1049d
1733
1734 12. The -o option (show only the matching part of a line) for pcregrep was not
1735 compatible with GNU grep in that, if there was more than one match in a
1736 line, it showed only the first of them. It now behaves in the same way as
1737 GNU grep.
1738
1739 13. If the -o and -v options were combined for pcregrep, it printed a blank
1740 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1741 does the same. The return code can be used to tell if there were any
1742 non-matching lines.
1743
1744 14. Added --file-offsets and --line-offsets to pcregrep.
1745
1746 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1747 infinitely looping recursion. The bug was that positive lookaheads were not
1748 being skipped when checking for a possible empty match (negative lookaheads
1749 and both kinds of lookbehind were skipped).
1750
1751 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1752 inclusion of <windows.h> to before rather than after the definition of
1753 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1754
1755 17. Specifying a possessive quantifier with a specific limit for a Unicode
1756 character property caused pcre_compile() to compile bad code, which led at
1757 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1758 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1759 caused the error; without that there was no problem.
1760
1761 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1762
1763 19. Added --enable-pcretest-libreadline.
1764
1765 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1766 RE::GlobalReplace(). As a result, the number of replacements returned was
1767 double what it should be. I removed one of the increments, but Craig sent a
1768 later patch that removed the other one (the right fix) and added unit tests
1769 that check the return values (which was not done before).
1770
1771 21. Several CMake things:
1772
1773 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1774 the names libpcre and libpcreposix, not just pcre and pcreposix.
1775
1776 (2) The above change means that pcretest and pcregrep are now correctly
1777 linked with the newly-built libraries, not previously installed ones.
1778
1779 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1780
1781 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1782 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1783 UTF-8 newline character). The key issue is that the pattern starts .*;
1784 this means that the match must be either at the beginning, or after a
1785 newline. The bug was in the code for advancing after a failed match and
1786 checking that the new position followed a newline. It was not taking
1787 account of UTF-8 characters correctly.
1788
1789 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1790 character classes. PCRE was not treating the sequence [:...:] as a
1791 character class unless the ... were all letters. Perl, however, seems to
1792 allow any characters between [: and :], though of course it rejects as
1793 unknown any "names" that contain non-letters, because all the known class
1794 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1795 for example, whereas PCRE did not - it did not recognize a POSIX character
1796 class. This seemed a bit dangerous, so the code has been changed to be
1797 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1798 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1799 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1800 Perl does, and where it didn't before.
1801
1802 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1803 Windows environments %n is disabled by default.
1804
1805
1806 Version 7.4 21-Sep-07
1807 ---------------------
1808
1809 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1810 means that a class such as [\s] counted as "explicit reference to CR or
1811 LF". That isn't really right - the whole point of the change was to try to
1812 help when there was an actual mention of one of the two characters. So now
1813 the change happens only if \r or \n (or a literal CR or LF) character is
1814 encountered.
1815
1816 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1817 of both had grown to the point where there were only 3 bits left.
1818 Fortunately, there was spare space in the data structure, and so I have
1819 moved the internal flags into a new 16-bit field to free up more option
1820 bits.
1821
1822 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1823 but did not set the internal JCHANGED flag - either of these is enough to
1824 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1825 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1826 start sets both bits.
1827
1828 4. Added options (at build time, compile time, exec time) to change \R from
1829 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1830
1831 5. doc/pcresyntax.html was missing from the distribution.
1832
1833 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1834 compatibility, even though it is no longer used.
1835
1836 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1837 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1838 windows.h file is present (where different names are used). [This was
1839 reversed later after testing - see 16 below.]
1840
1841 8. Changed all #include <config.h> to #include "config.h". There were also
1842 some further <pcre.h> cases that I changed to "pcre.h".
1843
1844 9. When pcregrep was used with the --colour option, it missed the line ending
1845 sequence off the lines that it output.
1846
1847 10. It was pointed out to me that arrays of string pointers cause lots of
1848 relocations when a shared library is dynamically loaded. A technique of
1849 using a single long string with a table of offsets can drastically reduce
1850 these. I have refactored PCRE in four places to do this. The result is
1851 dramatic:
1852
1853 Originally: 290
1854 After changing UCP table: 187
1855 After changing error message table: 43
1856 After changing table of "verbs" 36
1857 After changing table of Posix names 22
1858
1859 Thanks to the folks working on Gregex for glib for this insight.
1860
1861 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1862 unicode-properties was also set.
1863
1864 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1865
1866 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1867 checked only for CRLF.
1868
1869 14. Added casts to pcretest.c to avoid compiler warnings.
1870
1871 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1872
1873 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1874 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1875 entirely. This removes changes made in 7 above.
1876
1877 17. The CMake files have been updated, and there is now more information about
1878 building with CMake in the NON-UNIX-USE document.
1879
1880
1881 Version 7.3 28-Aug-07
1882 ---------------------
1883
1884 1. In the rejigging of the build system that eventually resulted in 7.1, the
1885 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1886 brackets there is not right, since it causes compilers to look for an
1887 installed pcre.h, not the version that is in the source that is being
1888 compiled (which of course may be different). I have changed it back to:
1889
1890 #include "pcre.h"
1891
1892 I have a vague recollection that the change was concerned with compiling in
1893 different directories, but in the new build system, that is taken care of
1894 by the VPATH setting the Makefile.
1895
1896 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1897 when the subject happened to end in the byte 0x85 (e.g. if the last
1898 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1899 characters but of course it shouldn't be taken as a newline when it is part
1900 of another character. The bug was that, for an unlimited repeat of . in
1901 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1902 characters when looking for a newline.
1903
1904 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1905
1906 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1907 in debug output.
1908
1909 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1910 long printing in the pcrecpp unittest when running under MinGW.
1911
1912 6. ESC_K was left out of the EBCDIC table.
1913
1914 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1915 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1916 limit also applies to "virtual nesting" when a pattern is recursive, and in
1917 this case 1000 isn't so big. I have been able to remove this limit at the
1918 expense of backing off one optimization in certain circumstances. Normally,
1919 when pcre_exec() would call its internal match() function recursively and
1920 immediately return the result unconditionally, it uses a "tail recursion"
1921 feature to save stack. However, when a subpattern that can match an empty
1922 string has an unlimited repetition quantifier, it no longer makes this
1923 optimization. That gives it a stack frame in which to save the data for
1924 checking that an empty string has been matched. Previously this was taken
1925 from the 1000-entry workspace that had been reserved. So now there is no
1926 explicit limit, but more stack is used.
1927
1928 8. Applied Daniel's patches to solve problems with the import/export magic
1929 syntax that is required for Windows, and which was going wrong for the
1930 pcreposix and pcrecpp parts of the library. These were overlooked when this
1931 problem was solved for the main library.
1932
1933 9. There were some crude static tests to avoid integer overflow when computing
1934 the size of patterns that contain repeated groups with explicit upper
1935 limits. As the maximum quantifier is 65535, the maximum group length was
1936 set at 30,000 so that the product of these two numbers did not overflow a
1937 32-bit integer. However, it turns out that people want to use groups that
1938 are longer than 30,000 bytes (though not repeat them that many times).
1939 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1940 made it possible to implement the integer overflow checks in a much more
1941 dynamic way, which I have now done. The artificial limitation on group
1942 length has been removed - we now have only the limit on the total length of
1943 the compiled pattern, which depends on the LINK_SIZE setting.
1944
1945 10. Fixed a bug in the documentation for get/copy named substring when
1946 duplicate names are permitted. If none of the named substrings are set, the
1947 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1948 empty string.
1949
1950 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1951 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1952 because the ] is interpreted as the first data character and the
1953 terminating ] is not found. PCRE has been made compatible with Perl in this
1954 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1955 cause memory overwriting.
1956
1957 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1958 string has been matched (to stop an infinite loop). It was not recognizing
1959 a conditional subpattern that could match an empty string if that
1960 subpattern was within another subpattern. For example, it looped when
1961 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1962 condition was not nested. This bug has been fixed.
1963
1964 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1965 past the start of the subject in the presence of bytes with the top bit
1966 set, for example "\x8aBCD".
1967
1968 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1969 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1970
1971 14. Optimized (?!) to (*FAIL).
1972
1973 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1974 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1975 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1976 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1977 does: it's just the validity check that is more restrictive.
1978
1979 16. Inserted checks for integer overflows during escape sequence (backslash)
1980 processing, and also fixed erroneous offset values for syntax errors during
1981 backslash processing.
1982
1983 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1984 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1985
1986 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1987 caused an overrun.
1988
1989 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1990 something other than just ASCII characters) inside a group that had an
1991 unlimited repeat caused a loop at compile time (while checking to see
1992 whether the group could match an empty string).
1993
1994 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1995 [\P{Any}] did so. (Error in the code for printing property names.)
1996
1997 21. An orphan \E inside a character class could cause a crash.
1998
1999 22. A repeated capturing bracket such as (A)? could cause a wild memory
2000 reference during compilation.
2001
2002 23. There are several functions in pcre_compile() that scan along a compiled
2003 expression for various reasons (e.g. to see if it's fixed length for look
2004 behind). There were bugs in these functions when a repeated \p or \P was
2005 present in the pattern. These operators have additional parameters compared
2006 with \d, etc, and these were not being taken into account when moving along
2007 the compiled data. Specifically:
2008
2009 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2010 length.
2011
2012 (b) An item such as \pL+ within a repeated group could cause crashes or
2013 loops.
2014
2015 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2016 "reference to non-existent subpattern" error.
2017
2018 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2019
2020 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2021 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2022
2023 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2024
2025 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2026 character were causing crashes (broken optimization).
2027
2028 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2029 \p or \P) caused a compile-time loop.
2030
2031 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2032 break. For example, the unstudied pattern [\r\n]A does not match the string
2033 "\r\nA" because change 7.0/46 below moves the current point on by two
2034 characters after failing to match at the start. However, the pattern \nA
2035 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2036 the same is true. There doesn't seem any very clean way out of this, but
2037 what I have chosen to do makes the common cases work: PCRE now takes note
2038 of whether there can be an explicit match for \r or \n anywhere in the
2039 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2040 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2041 pattern has explicit CR or LF references.
2042
2043 29. Added (*CR) etc for changing newline setting at start of pattern.
2044
2045
2046 Version 7.2 19-Jun-07
2047 ---------------------
2048
2049 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2050 which is apparently normally available under Windows.
2051
2052 2. Re-jig the pcregrep tests with different newline settings in an attempt
2053 to make them independent of the local environment's newline setting.
2054
2055 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2056
2057 4. Some of the "internals" tests were previously cut out when the link size
2058 was not 2, because the output contained actual offsets. The recent new
2059 "Z" feature of pcretest means that these can be cut out, making the tests
2060 usable with all link sizes.
2061
2062 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2063 stack recursion. This gives a massive performance boost under BSD, but just
2064 a small improvement under Linux. However, it saves one field in the frame
2065 in all cases.
2066
2067 6. Added more features from the forthcoming Perl 5.10:
2068
2069 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2070 recursion call. It refers to the nth most recently opened parentheses.
2071
2072 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2073 to be opened parentheses.
2074
2075 (c) Conditions that refer to capturing parentheses can be specified
2076 relatively, for example, (?(-2)... or (?(+3)...
2077
2078 (d) \K resets the start of the current match so that everything before
2079 is not part of it.
2080
2081 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2082
2083 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2084 reference syntax.
2085
2086 (g) (?| introduces a group in which the numbering of parentheses in each
2087 alternative starts with the same number.
2088
2089 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2090
2091 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2092 PCRE_INFO_JCHANGED.
2093
2094 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2095 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2096 for detecting groups that can match an empty string.
2097
2098 9. A pattern with a very large number of alternatives (more than several
2099 hundred) was running out of internal workspace during the pre-compile
2100 phase, where pcre_compile() figures out how much memory will be needed. A
2101 bit of new cunning has reduced the workspace needed for groups with
2102 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2103 workspace instead of running out of the 4096 that are available.
2104
2105 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2106
2107 11. Applied patch from Google to remove an optimization that didn't quite work.
2108 The report of the bug said:
2109
2110 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2111 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2112 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2113
2114 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2115 it matched the wrong number of bytes.
2116
2117
2118 Version 7.1 24-Apr-07
2119 ---------------------
2120
2121 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2122 that is more "standard", making use of automake and other Autotools. There
2123 is some re-arrangement of the files and adjustment of comments consequent
2124 on this.
2125
2126 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2127 for recursive directory scanning broke on some systems because the files
2128 are not scanned in any specific order and on different systems the order
2129 was different. A call to "sort" has been inserted into RunGrepTest for the
2130 approprate test as a short-term fix. In the longer term there may be an
2131 alternative.
2132
2133 3. I had an email from Eric Raymond about problems translating some of PCRE's
2134 man pages to HTML (despite the fact that I distribute HTML pages, some
2135 people do their own conversions for various reasons). The problems
2136 concerned the use of low-level troff macros .br and .in. I have therefore
2137 removed all such uses from the man pages (some were redundant, some could
2138 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2139 HTML has been updated to handle .nf/.fi and to complain if it encounters
2140 .br or .in.
2141
2142 4. Updated comments in configure.ac that get placed in config.h.in and also
2143 arranged for config.h to be included in the distribution, with the name
2144 config.h.generic, for the benefit of those who have to compile without
2145 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2146
2147 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2148 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2149 makevp.bat for the current PCRE, using the additional files
2150 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2151
2152 6. A Windows user reported a minor discrepancy with test 2, which turned out
2153 to be caused by a trailing space on an input line that had got lost in his
2154 copy. The trailing space was an accident, so I've just removed it.
2155
2156 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2157 that is needed.
2158
2159 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2160 as "const" (a) because they are and (b) because it helps the PHP
2161 maintainers who have recently made a script to detect big data structures
2162 in the php code that should be moved to the .rodata section. I remembered
2163 to update Builducptable as well, so it won't revert if ucptable.h is ever
2164 re-created.
2165
2166 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2167 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2168 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2169 support is not required. This saves 1.5-2K of code, which is important in
2170 some applications.
2171
2172 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2173 so as not to refer to the tables, even though these functions will never be
2174 called when UTF-8 support is disabled. Otherwise there are problems with a
2175 shared library.
2176
2177 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2178
2179 (a) It was defining its arguments as char * instead of void *.
2180
2181 (b) It was assuming that all moves were upwards in memory; this was true
2182 a long time ago when I wrote it, but is no longer the case.
2183
2184 The emulated memove() is provided for those environments that have neither
2185 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2186 is clearly not the case, as these two bugs were recently reported.
2187
2188 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2189 and Detrail to create the HTML documentation, the .txt form of the man
2190 pages, and it removes trailing spaces from listed files. It also creates
2191 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2192 case, it wraps all the #defines with #ifndefs. This script should be run
2193 before "make dist".
2194
2195 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2196 with Unicode property support.
2197
2198 (a) For a maximizing quantifier, if the two different cases of the
2199 character were of different lengths in their UTF-8 codings (there are
2200 some cases like this - I found 11), and the matching function had to
2201 back up over a mixture of the two cases, it incorrectly assumed they
2202 were both the same length.
2203
2204 (b) When PCRE was configured to use the heap rather than the stack for
2205 recursion during matching, it was not correctly preserving the data for
2206 the other case of a UTF-8 character when checking ahead for a match
2207 while processing a minimizing repeat. If the check also involved
2208 matching a wide character, but failed, corruption could cause an
2209 erroneous result when trying to check for a repeat of the original
2210 character.
2211
2212 13. Some tidying changes to the testing mechanism:
2213
2214 (a) The RunTest script now detects the internal link size and whether there
2215 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2216 values substituted by "configure". (The RunGrepTest script already did
2217 this for UTF-8.) The configure.ac script no longer substitutes the
2218 relevant variables.
2219
2220 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2221 with length and offset values. This means that the output is different
2222 for different internal link sizes. Test 2 is skipped for link sizes
2223 other than 2 because of this, bypassing the problem. Unfortunately,
2224 there was also a test in test 3 (the locale tests) that used /B and
2225 failed for link sizes other than 2. Rather than cut the whole test out,
2226 I have added a new /Z option to pcretest that replaces the length and
2227 offset values with spaces. This is now used to make test 3 independent
2228 of link size. (Test 2 will be tidied up later.)
2229
2230 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2231 segmentation fault instead of returning the appropriate error message.
2232
2233 15. In multiline mode when the newline sequence was set to "any", the pattern
2234 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2235 This doesn't seem right; it now treats the CRLF combination as the line
2236 ending, and so does not match in that case. It's only a pattern such as ^$
2237 that would hit this one: something like ^ABC$ would have failed after \r
2238 and then tried again after \r\n.
2239
2240 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2241 in an attempt to make files that differ only in their line terminators
2242 compare equal. This works on Linux.
2243
2244 17. Under certain error circumstances pcregrep might try to free random memory
2245 as it exited. This is now fixed, thanks to valgrind.
2246
2247 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2248 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2249 was because its rules for how to advance for /g after matching an empty
2250 string at the end of a line did not allow for this case. They now check for
2251 it specially.
2252
2253 20. pcretest is supposed to handle patterns and data of any length, by
2254 extending its buffers when necessary. It was getting this wrong when the
2255 buffer for a data line had to be extended.
2256
2257 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2258 CRLF as a newline sequence.
2259
2260 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2261 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2262 I have nevertheless tidied it up.
2263
2264 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2265
2266 24. Added a man page for pcre-config.
2267
2268
2269 Version 7.0 19-Dec-06
2270 ---------------------
2271
2272 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2273 moving to gcc 4.1.1.
2274
2275 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2276 sys/time.h, which is documented as needed for this function. It doesn't
2277 seem to matter on Linux, but it showed up on some releases of OS X.
2278
2279 3. It seems that there are systems where bytes whose values are greater than
2280 127 match isprint() in the "C" locale. The "C" locale should be the
2281 default when a C program starts up. In most systems, only ASCII printing
2282 characters match isprint(). This difference caused the output from pcretest
2283 to vary, making some of the tests fail. I have changed pcretest so that:
2284
2285 (a) When it is outputting text in the compiled version of a pattern, bytes
2286 other than 32-126 are always shown as hex escapes.
2287
2288 (b) When it is outputting text that is a matched part of a subject string,
2289 it does the same, unless a different locale has been set for the match
2290 (using the /L modifier). In this case, it uses isprint() to decide.
2291
2292 4. Fixed a major bug that caused incorrect computation of the amount of memory
2293 required for a compiled pattern when options that changed within the
2294 pattern affected the logic of the preliminary scan that determines the
2295 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2296 that the computed length was too small. The symptoms of this bug were
2297 either the PCRE error "internal error: code overflow" from pcre_compile(),
2298 or a glibc crash with a message such as "pcretest: free(): invalid next
2299 size (fast)". Examples of patterns that provoked this bug (shown in
2300 pcretest format) are:
2301
2302 /(?-x: )/x
2303 /(?x)(?-x: \s*#\s*)/
2304 /((?i)[\x{c0}])/8
2305 /(?i:[\x{c0}])/8
2306
2307 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2308 is now done differently.
2309
2310 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2311 wrapper classes; (b) implement a new function in the C++ scanner that is
2312 more efficient than the old way of doing things because it avoids levels of
2313 recursion in the regex matching; (c) add a paragraph to the documentation
2314 for the FullMatch() function.
2315
2316 6. The escape sequence \n was being treated as whatever was defined as
2317 "newline". Not only was this contrary to the documentation, which states
2318 that \n is character 10 (hex 0A), but it also went horribly wrong when
2319 "newline" was defined as CRLF. This has been fixed.
2320
2321 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2322 was being set to -1 for the "end of line" case (supposedly a value that no
2323 character can have). Though this value is never used (the check for end of
2324 line is "zero bytes in current character"), it caused compiler complaints.
2325 I've changed it to 0xffffffff.
2326
2327 8. In pcre_version.c, the version string was being built by a sequence of
2328 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2329 string (as it is for production releases) called a macro with an empty
2330 argument. The C standard says the result of this is undefined. The gcc
2331 compiler treats it as an empty string (which was what was wanted) but it is
2332 reported that Visual C gives an error. The source has been hacked around to
2333 avoid this problem.
2334
2335 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2336 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2337 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2338 of them did).
2339
2340 10. Originally, pcretest opened its input and output without "b"; then I was
2341 told that "b" was needed in some environments, so it was added for release
2342 5.0 to both the input and output. (It makes no difference on Unix-like
2343 systems.) Later I was told that it is wrong for the input on Windows. I've
2344 now abstracted the modes into two macros, to make it easier to fiddle with
2345 them, and removed "b" from the input mode under Windows.
2346
2347 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2348
2349 12. Added -help and --help to pcretest as an official way of being reminded
2350 of the options.
2351
2352 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2353 and pcrecpp.cc because they annoy compilers at high warning levels.
2354
2355 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2356
2357 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2358 scripts are not C programs :-) and which was not noticed because it works
2359 on Linux.
2360
2361 16. pcretest is supposed to handle any length of pattern and data line (as one
2362 line or as a continued sequence of lines) by extending its input buffer if
2363 necessary. This feature was broken for very long pattern lines, leading to
2364 a string of junk being passed to pcre_compile() if the pattern was longer
2365 than about 50K.
2366
2367 17. I have done a major re-factoring of the way pcre_compile() computes the
2368 amount of memory needed for a compiled pattern. Previously, there was code
2369 that made a preliminary scan of the pattern in order to do this. That was
2370 OK when PCRE was new, but as the facilities have expanded, it has become
2371 harder and harder to keep it in step with the real compile phase, and there
2372 have been a number of bugs (see for example, 4 above). I have now found a
2373 cunning way of running the real compile function in a "fake" mode that
2374 enables it to compute how much memory it would need, while actually only
2375 ever using a few hundred bytes of working memory and without too many
2376 tests of the mode. This should make future maintenance and development
2377 easier. A side effect of this work is that the limit of 200 on the nesting
2378 depth of parentheses has been removed (though this was never a serious
2379 limitation, I suspect). However, there is a downside: pcre_compile() now
2380 runs more slowly than before (30% or more, depending on the pattern). I
2381 hope this isn't a big issue. There is no effect on runtime performance.
2382
2383 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2384 newline (only possible for the last line of a file) and it was a
2385 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2386
2387 19. Added additional timing features to pcretest. (1) The -tm option now times
2388 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2389 separate command line item, by a number that specifies the number of
2390 repeats to use when timing. The default is 50000; this gives better
2391 precision, but takes uncomfortably long for very large patterns.
2392
2393 20. Extended pcre_study() to be more clever in cases where a branch of a
2394 subpattern has no definite first character. For example, (a*|b*)[cd] would
2395 previously give no result from pcre_study(). Now it recognizes that the
2396 first character must be a, b, c, or d.
2397
2398 21. There was an incorrect error "recursive call could loop indefinitely" if
2399 a subpattern (or the entire pattern) that was being tested for matching an
2400 empty string contained only one non-empty item after a nested subpattern.
2401 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2402 incorrectly, because the \d was being skipped in the check.
2403
2404 22. The pcretest program now has a new pattern option /B and a command line
2405 option -b, which is equivalent to adding /B to every pattern. This causes
2406 it to show the compiled bytecode, without the additional information that
2407 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2408 is the same as /B/I).
2409
2410 23. A new optimization is now able automatically to treat some sequences such
2411 as a*b as a*+b. More specifically, if something simple (such as a character
2412 or a simple class like \d) has an unlimited quantifier, and is followed by
2413 something that cannot possibly match the quantified thing, the quantifier
2414 is automatically "possessified".
2415
2416 24. A recursive reference to a subpattern whose number was greater than 39
2417 went wrong under certain circumstances in UTF-8 mode. This bug could also
2418 have affected the operation of pcre_study().
2419
2420 25. Realized that a little bit of performance could be had by replacing
2421 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2422
2423 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2424
2425 27. Possessive quantifiers such as a++ were previously implemented by turning
2426 them into atomic groups such as ($>a+). Now they have their own opcodes,
2427 which improves performance. This includes the automatically created ones
2428 from 23 above.
2429
2430 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2431 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2432 the first matched character to be a colon. This applied both to named and
2433 numbered groups.
2434
2435 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2436
2437 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2438 building PCRE on Windows easier, so I have included it in the distribution.
2439
2440 31. There is now a check in pcretest against a ridiculously large number being
2441 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2442 loop, the loop is abandoned.
2443
2444 32. Forward references to subpatterns in conditions such as (?(2)...) where
2445 subpattern 2 is defined later cause pcre_compile() to search forwards in
2446 the pattern for the relevant set of parentheses. This search went wrong
2447 when there were unescaped parentheses in a character class, parentheses
2448 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2449
2450 33. "Subroutine" calls and backreferences were previously restricted to
2451 referencing subpatterns earlier in the regex. This restriction has now
2452 been removed.
2453
2454 34. Added a number of extra features that are going to be in Perl 5.10. On the
2455 whole, these are just syntactic alternatives for features that PCRE had
2456 previously implemented using the Python syntax or my own invention. The
2457 other formats are all retained for compatibility.
2458
2459 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2460 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2461 also .NET compatible.
2462
2463 (b) A recursion or subroutine call to a named group can now be defined as
2464 (?&name) as well as (?P>name).
2465
2466 (c) A backreference to a named group can now be defined as \k<name> or
2467 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2468 5.10, are also .NET compatible.
2469
2470 (d) A conditional reference to a named group can now use the syntax
2471 (?(<name>) or (?('name') as well as (?(name).
2472
2473 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2474 groups (named and numbered) that are never evaluated inline, but can be
2475 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2476 is always false. There may be only one alternative in such a group.
2477
2478 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2479 as the simple (?(R). The condition is true only if the most recent
2480 recursion is that of the given number or name. It does not search out
2481 through the entire recursion stack.
2482
2483 (g) The escape \gN or \g{N} has been added, where N is a positive or
2484 negative number, specifying an absolute or relative reference.
2485
2486 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2487 some "unreachable code" warnings.
2488
2489 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2490 things, this adds five new scripts.
2491
2492 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2493 There were also incompatibilities regarding the handling of \Q..\E inside
2494 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2495 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2496
2497 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2498 matches an empty string, and forcibly breaks the loop. There were bugs in
2499 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2500 against aaaa the result was just "a" rather than "aaaa", for example. Two
2501 separate and independent bugs (that affected different cases) have been
2502 fixed.
2503
2504 39. Refactored the code to abolish the use of different opcodes for small
2505 capturing bracket numbers. This is a tidy that I avoided doing when I
2506 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2507 The new approach is not only tidier, it makes it possible to reduce the
2508 memory needed to fix the previous bug (38).
2509
2510 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2511 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2512 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2513 mode.
2514
2515 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2516 report.
2517
2518 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2519 copy construction and assignment in the C++ wrapper.
2520
2521 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2522 couple of bugs that could have given wrong results in the "--newline=crlf"
2523 case.
2524
2525 44. Added a number of casts and did some reorganization of signed/unsigned int
2526 variables following suggestions from Dair Grant. Also renamed the variable
2527 "this" as "item" because it is a C++ keyword.
2528
2529 45. Arranged for dftables to add
2530
2531 #include "pcre_internal.h"
2532
2533 to pcre_chartables.c because without it, gcc 4.x may remove the array
2534 definition from the final binary if PCRE is built into a static library and
2535 dead code stripping is activated.
2536
2537 46. For an unanchored pattern, if a match attempt fails at the start of a
2538 newline sequence, and the newline setting is CRLF or ANY, and the next two
2539 characters are CRLF, advance by two characters instead of one.
2540
2541
2542 Version 6.7 04-Jul-06
2543 ---------------------
2544
2545 1. In order to handle tests when input lines are enormously long, pcretest has
2546 been re-factored so that it automatically extends its buffers when
2547 necessary. The code is crude, but this _is_ just a test program. The
2548 default size has been increased from 32K to 50K.
2549
2550 2. The code in pcre_study() was using the value of the re argument before
2551 testing it for NULL. (Of course, in any sensible call of the function, it
2552 won't be NULL.)
2553
2554 3. The memmove() emulation function in pcre_internal.h, which is used on
2555 systems that lack both memmove() and bcopy() - that is, hardly ever -
2556 was missing a "static" storage class specifier.
2557
2558 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2559 containing an extended class (one that cannot be represented by a bitmap
2560 because it contains high-valued characters or Unicode property items, e.g.
2561 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2562 pattern, but PCRE should not loop if you do not (it no longer does).
2563 [Detail: two cases were found: (a) a repeated subpattern containing an
2564 extended class; (b) a recursive reference to a subpattern that followed a
2565 previous extended class. It wasn't skipping over the extended class
2566 correctly when UTF-8 mode was not set.]
2567
2568 5. A negated single-character class was not being recognized as fixed-length
2569 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2570 compile error "lookbehind assertion is not fixed length".
2571
2572 6. The RunPerlTest auxiliary script was showing an unexpected difference
2573 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2574 write a Perl script that can interpret lines of an input file either as
2575 byte characters or as UTF-8, which is what "perltest" was being required to
2576 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2577 can't do is switch easily at run time between having the "use utf8;" pragma
2578 or not. In the end, I fudged it by using the RunPerlTest script to insert
2579 "use utf8;" explicitly for the UTF-8 tests.
2580
2581 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2582 the end of the subject string, contrary to the documentation and to what
2583 Perl does. This was true of both matching functions. Now it matches only at
2584 the start of the subject and immediately after *internal* newlines.
2585
2586 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2587 a pointer to an int instead of a pointer to an unsigned long int. This
2588 caused problems on 64-bit systems.
2589
2590 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2591 instance of the 'standard' template library not being so standard".
2592
2593 10. There was no check on the number of named subpatterns nor the maximum
2594 length of a subpattern name. The product of these values is used to compute
2595 the size of the memory block for a compiled pattern. By supplying a very
2596 long subpattern name and a large number of named subpatterns, the size
2597 computation could be caused to overflow. This is now prevented by limiting
2598 the length of names to 32 characters, and the number of named subpatterns
2599 to 10,000.
2600
2601 11. Subpatterns that are repeated with specific counts have to be replicated in
2602 the compiled pattern. The size of memory for this was computed from the
2603 length of the subpattern and the repeat count. The latter is limited to
2604 65535, but there was no limit on the former, meaning that integer overflow
2605 could in principle occur. The compiled length of a repeated subpattern is
2606 now limited to 30,000 bytes in order to prevent this.
2607
2608 12. Added the optional facility to have named substrings with the same name.
2609
2610 13. Added the ability to use a named substring as a condition, using the
2611 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2612 are numbers (not recommended). Forward references are permitted.
2613
2614 14. Added forward references in named backreferences (if you see what I mean).
2615
2616 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2617 pattern could run off the end of the subject. For example, the pattern
2618 "(?s)(.{1,5})"8 did this with the subject "ab".
2619
2620 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2621 PCRE_CASELESS was set when matching characters that were quantified with ?
2622 or *.
2623
2624 17. A character class other than a single negated character that had a minimum
2625 but no maximum quantifier - for example [ab]{6,} - was not handled
2626 correctly by pce_dfa_exec(). It would match only one character.
2627
2628 18. A valid (though odd) pattern that looked like a POSIX character
2629 class but used an invalid character after [ (for example [[,abc,]]) caused
2630 pcre_compile() to give the error "Failed: internal error: code overflow" or
2631 in some cases to crash with a glibc free() error. This could even happen if
2632 the pattern terminated after [[ but there just happened to be a sequence of
2633 letters, a binary zero, and a closing ] in the memory that followed.
2634
2635 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2636 over the years. Originally (before any Unicode support), just the bottom 8
2637 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2638 output from "man perlunicode" includes this:
2639
2640 The regular expression compiler produces polymorphic opcodes. That
2641 is, the pattern adapts to the data and automatically switches to
2642 the Unicode character scheme when presented with Unicode data--or
2643 instead uses a traditional byte scheme when presented with byte
2644 data.
2645
2646 Sadly, a wide octal escape does not cause a switch, and in a string with
2647 no other multibyte characters, these octal escapes are treated as before.
2648 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2649 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2650 Unicode string.
2651
2652 I have not perpetrated such confusion in PCRE. Up till now, it took just
2653 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2654 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2655 translate to the appropriate multibyte character.
2656
2657 29. Applied some refactoring to reduce the number of warnings from Microsoft
2658 and Borland compilers. This has included removing the fudge introduced
2659 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2660 a warning about an unused variable.
2661
2662 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2663 characters since release 4.0, because Perl (from release 5.004) does not.
2664 [Or at least, is documented not to: some releases seem to be in conflict
2665 with the documentation.] However, when a pattern was studied with
2666 pcre_study() and all its branches started with \s, PCRE still included VT
2667 as a possible starting character. Of course, this did no harm; it just
2668 caused an unnecessary match attempt.
2669
2670 22. Removed a now-redundant internal flag bit that recorded the fact that case
2671 dependency changed within the pattern. This was once needed for "required
2672 byte" processing, but is no longer used. This recovers a now-scarce options
2673 bit. Also moved the least significant internal flag bit to the most-
2674 significant bit of the word, which was not previously used (hangover from
2675 the days when it was an int rather than a uint) to free up another bit for
2676 the future.
2677
2678 23. Added support for CRLF line endings as well as CR and LF. As well as the
2679 default being selectable at build time, it can now be changed at runtime
2680 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2681 specify that it is scanning data with non-default line endings.
2682
2683 24. Changed the definition of CXXLINK to make it agree with the definition of
2684 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2685
2686 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2687 recursions. This makes a big different to stack usage for some patterns.
2688
2689 26. If a subpattern containing a named recursion or subroutine reference such
2690 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2691 the space required for the compiled pattern went wrong and gave too small a
2692 value. Depending on the environment, this could lead to "Failed: internal
2693 error: code overflow at offset 49" or "glibc detected double free or
2694 corruption" errors.
2695
2696 27. Applied patches from Google (a) to support the new newline modes and (b) to
2697 advance over multibyte UTF-8 characters in GlobalReplace.
2698
2699 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2700 difference for some implementation of PCRE in some Windows version.
2701
2702 29. Added some extra testing facilities to pcretest:
2703
2704 \q<number> in a data line sets the "match limit" value
2705 \Q<number> in a data line sets the "match recursion limt" value
2706 -S <number> sets the stack size, where <number> is in megabytes
2707
2708 The -S option isn't available for Windows.
2709
2710
2711 Version 6.6 06-Feb-06
2712 ---------------------
2713
2714 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2715 in pcreposix.h. I have copied the definition from pcre.h.
2716
2717 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2718 because pcre.h is no longer a built file.
2719
2720 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2721 not normally included in the compiled code.
2722
2723
2724 Version 6.5 01-Feb-06
2725 ---------------------
2726
2727 1. When using the partial match feature with pcre_dfa_exec(), it was not
2728 anchoring the second and subsequent partial matches at the new starting
2729 point. This could lead to incorrect results. For example, with the pattern
2730 /1234/, partially matching against "123" and then "a4" gave a match.
2731
2732 2. Changes to pcregrep:
2733
2734 (a) All non-match returns from pcre_exec() were being treated as failures
2735 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2736 error message is output. Some extra information is given for the
2737 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2738 probably the only errors that are likely to be caused by users (by
2739 specifying a regex that has nested indefinite repeats, for instance).
2740 If there are more than 20 of these errors, pcregrep is abandoned.
2741
2742 (b) A binary zero was treated as data while matching, but terminated the
2743 output line if it was written out. This has been fixed: binary zeroes
2744 are now no different to any other data bytes.
2745
2746 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2747 used to set a locale for matching. The --locale=xxxx long option has
2748 been added (no short equivalent) to specify a locale explicitly on the
2749 pcregrep command, overriding the environment variables.
2750
2751 (d) When -B was used with -n, some line numbers in the output were one less
2752 than they should have been.
2753
2754 (e) Added the -o (--only-matching) option.
2755
2756 (f) If -A or -C was used with -c (count only), some lines of context were
2757 accidentally printed for the final match.
2758
2759 (g) Added the -H (--with-filename) option.
2760
2761 (h) The combination of options -rh failed to suppress file names for files
2762 that were found from directory arguments.
2763
2764 (i) Added the -D (--devices) and -d (--directories) options.
2765
2766 (j) Added the -F (--fixed-strings) option.
2767
2768 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2769
2770 (l) Added the --colo(u)r option.
2771
2772 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2773 is not present by default.
2774
2775 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2776 items such as (?R) or (?1), when the recursion could match a number of
2777 alternatives. If it matched one of the alternatives, but subsequently,
2778 outside the recursion, there was a failure, the code tried to back up into
2779 the recursion. However, because of the way PCRE is implemented, this is not
2780 possible, and the result was an incorrect result from the match.
2781
2782 In order to prevent this happening, the specification of recursion has
2783 been changed so that all such subpatterns are automatically treated as
2784 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2785
2786 4. I had overlooked the fact that, in some locales, there are characters for
2787 which isalpha() is true but neither isupper() nor islower() are true. In
2788 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2789 and ordfeminine) are like this. This affected the treatment of \w and \W
2790 when they appeared in character classes, but not when they appeared outside
2791 a character class. The bit map for "word" characters is now created
2792 separately from the results of isalnum() instead of just taking it from the
2793 upper, lower, and digit maps. (Plus the underscore character, of course.)
2794
2795 5. The above bug also affected the handling of POSIX character classes such as
2796 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2797 permanent tables. Instead, the bit maps for such a class were previously
2798 created as the appropriate unions of the upper, lower, and digit bitmaps.
2799 Now they are created by subtraction from the [[:word:]] class, which has
2800 its own bitmap.
2801
2802 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2803 It is created by subtracting the vertical space characters (\x09, \x0a,
2804 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2805 subtraction was done in the overall bitmap for a character class, meaning
2806 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2807 be recognized. This bug has been fixed.
2808
2809 7. Patches from the folks at Google:
2810
2811 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2812 real life, but is still worth protecting against".
2813
2814 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2815 regular expressions".
2816
2817 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2818 have it.
2819
2820 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2821 "configure" and the latter not, in order to fix a problem somebody had
2822 with compiling the Arg class on HP-UX.
2823
2824 (e) Improve the error-handling of the C++ wrapper a little bit.
2825
2826 (f) New tests for checking recursion limiting.
2827
2828 8. The pcre_memmove() function, which is used only if the environment does not
2829 have a standard memmove() function (and is therefore rarely compiled),
2830 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2831 returning a result (though PCRE never actually uses the result).
2832
2833 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2834 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2835 returned instead of calling malloc() with an overflowing number that would
2836 most likely cause subsequent chaos.
2837
2838 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2839
2840 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2841 with this option is matched, the nmatch and pmatch options of regexec() are
2842 ignored.
2843
2844 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2845 provided in case anyone wants to the the POSIX interface with UTF-8
2846 strings.
2847
2848 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2849 C++ linking (needed for some HP-UX environments).
2850
2851 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2852 (unused parameter) and in the pcre_printint() function (omitted "default"
2853 switch label when the default is to do nothing).
2854
2855 15. Added some code to make it possible, when PCRE is compiled as a C++
2856 library, to replace subject pointers for pcre_exec() with a smart pointer
2857 class, thus making it possible to process discontinuous strings.
2858
2859 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2860 much the same function. They were added by different people who were trying
2861 to make PCRE easy to compile on non-Unix systems. It has been suggested
2862 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2863 for compiling on Windows systems. I have therefore replaced it with
2864 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2865 defaults to "extern" for C or "extern C" for C++, which works fine on
2866 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2867 SCOPE with something explicit in config.h. In addition:
2868
2869 (a) pcreposix.h still had just "extern" instead of either of these macros;
2870 I have replaced it with PCRE_DATA_SCOPE.
2871
2872 (b) Functions such as _pcre_xclass(), which are internal to the library,
2873 but external in the C sense, all had PCRE_EXPORT in their definitions.
2874 This is apparently wrong for the Windows case, so I have removed it.
2875 (It makes no difference on Unix-like systems.)
2876
2877 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2878 of recursive calls to match(). This is different to MATCH_LIMIT because
2879 that limits the total number of calls to match(), not all of which increase
2880 the depth of recursion. Limiting the recursion depth limits the amount of
2881 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2882 when PCRE is compiled, and changed at run time. A patch from Google adds
2883 this functionality to the C++ interface.
2884
2885 18. Changes to the handling of Unicode character properties:
2886
2887 (a) Updated the table to Unicode 4.1.0.
2888
2889 (b) Recognize characters that are not in the table as "Cn" (undefined).
2890
2891 (c) I revised the way the table is implemented to a much improved format
2892 which includes recognition of ranges. It now supports the ranges that
2893 are defined in UnicodeData.txt, and it also amalgamates other
2894 characters into ranges. This has reduced the number of entries in the
2895 table from around 16,000 to around 3,000, thus reducing its size
2896 considerably. I realized I did not need to use a tree structure after
2897 all - a binary chop search is just as efficient. Having reduced the
2898 number of entries, I extended their size from 6 bytes to 8 bytes to
2899 allow for more data.
2900
2901 (d) Added support for Unicode script names via properties such as \p{Han}.
2902
2903 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2904 matching that character.
2905
2906 20. When matching a repeated Unicode property with a minimum greater than zero,
2907 (for example \pL{2,}), PCRE could look past the end of the subject if it
2908 reached it while seeking the minimum number of characters. This could
2909 happen only if some of the characters were more than one byte long, because
2910 there is a check for at least the minimum number of bytes.
2911
2912 21. Refactored the implementation of \p and \P so as to be more general, to
2913 allow for more different types of property in future. This has changed the
2914 compiled form incompatibly. Anybody with saved compiled patterns that use
2915 \p or \P will have to recompile them.
2916
2917 22. Added "Any" and "L&" to the supported property types.
2918
2919 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2920 but give a compile time error if the value is greater than 0xff.
2921
2922 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2923 accidentally not being installed or uninstalled.
2924
2925 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2926 made were to insert the current release number. This seemed silly, because
2927 it made things harder for people building PCRE on systems that don't run
2928 "configure". I have turned pcre.h into a distributed file, no longer built
2929 by "configure", with the version identification directly included. There is
2930 no longer a pcre.h.in file.
2931
2932 However, this change necessitated a change to the pcre-config script as
2933 well. It is built from pcre-config.in, and one of the substitutions was the
2934 release number. I have updated configure.ac so that ./configure now finds
2935 the release number by grepping pcre.h.
2936
2937 26. Added the ability to run the tests under valgrind.
2938
2939
2940 Version 6.4 05-Sep-05
2941 ---------------------
2942
2943 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2944 "--" to be printed when multiple files were scanned, even when none of the
2945 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2946 consider it to be a bug, and have restored the previous behaviour.
2947
2948 2. A couple of code tidies to get rid of compiler warnings.
2949
2950 3. The pcretest program used to cheat by referring to symbols in the library
2951 whose names begin with _pcre_. These are internal symbols that are not
2952 really supposed to be visible externally, and in some environments it is
2953 possible to suppress them. The cheating is now confined to including
2954 certain files from the library's source, which is a bit cleaner.
2955
2956 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2957 file's purpose clearer.
2958
2959 5. Reorganized pcre_ucp_findchar().
2960
2961
2962 Version 6.3 15-Aug-05
2963 ---------------------
2964
2965 1. The file libpcre.pc.in did not have general read permission in the tarball.
2966
2967 2. There were some problems when building without C++ support:
2968
2969 (a) If C++ support was not built, "make install" and "make test" still
2970 tried to test it.
2971
2972 (b) There were problems when the value of CXX was explicitly set. Some
2973 changes have been made to try to fix these, and ...
2974
2975 (c) --disable-cpp can now be used to explicitly disable C++ support.
2976
2977 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2978 backslash in a target when C++ was disabled. This confuses some
2979 versions of "make", apparently. Using an intermediate variable solves
2980 this. (Same for CPP_LOBJ.)
2981
2982 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2983 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2984 necessary on certain architectures.
2985
2986 4. Added a setting of -export-symbols-regex to the link command to remove
2987 those symbols that are exported in the C sense, but actually are local
2988 within the library, and not documented. Their names all begin with
2989 "_pcre_". This is not a perfect job, because (a) we have to except some
2990 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2991 available (and never for static libraries). I have made a note to try to
2992 find a way round (a) in the future.
2993
2994
2995 Version 6.2 01-Aug-05
2996 ---------------------
2997
2998 1. There was no test for integer overflow of quantifier values. A construction
2999 such as {1111111111111111} would give undefined results. What is worse, if
3000 a minimum quantifier for a parenthesized subpattern overflowed and became
3001 negative, the calculation of the memory size went wrong. This could have
3002 led to memory overwriting.
3003
3004 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3005
3006 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3007 operating environments where this matters.
3008
3009 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3010 PCRE options from within the C++ wrapper.
3011
3012 5. Named capturing subpatterns were not being correctly counted when a pattern
3013 was compiled. This caused two problems: (a) If there were more than 100
3014 such subpatterns, the calculation of the memory needed for the whole
3015 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3016 back references of the form \12, where the number was greater than 9, were
3017 not recognized as back references, even though there were sufficient
3018 previous subpatterns.
3019
3020 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3021 versions of gcc, e.g. 2.95.4.
3022
3023
3024 Version 6.1 21-Jun-05
3025 ---------------------
3026
3027 1. There was one reference to the variable "posix" in pcretest.c that was not
3028 surrounded by "#if !defined NOPOSIX".
3029
3030 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3031 the cross-check on the old pcre_info() function, for the benefit of the
3032 cut-down version of PCRE that is currently imported into Exim.
3033
3034 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3035 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3036 patterns that start (?i) but I don't think that matters. The use of (?i) is
3037 just an example; this all applies to the other options as well.
3038
3039 4. Since libtool seems to echo the compile commands it is issuing, the output
3040 from "make" can be reduced a bit by putting "@" in front of each libtool
3041 compile command.
3042
3043 5. Patch from the folks at Google for configure.in to be a bit more thorough
3044 in checking for a suitable C++ installation before trying to compile the
3045 C++ stuff. This should fix a reported problem when a compiler was present,
3046 but no suitable headers.
3047
3048 6. The man pages all had just "PCRE" as their title. I have changed them to
3049 be the relevant file name. I have also arranged that these names are
3050 retained in the file doc/pcre.txt, which is a concatenation in text format
3051 of all the man pages except the little individual ones for each function.
3052
3053 7. The NON-UNIX-USE file had not been updated for the different set of source
3054 files that come with release 6. I also added a few comments about the C++
3055 wrapper.
3056
3057
3058 Version 6.0 07-Jun-05
3059 ---------------------
3060
3061 1. Some minor internal re-organization to help with my DFA experiments.
3062
3063 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3064 didn't matter for the library itself when fully configured, but did matter
3065 when compiling without UCP support, or within Exim, where the ucp files are
3066 not imported.
3067
3068 3. Refactoring of the library code to split up the various functions into
3069 different source modules. The addition of the new DFA matching code (see
3070 below) to a single monolithic source would have made it really too
3071 unwieldy, quite apart from causing all the code to be include in a
3072 statically linked application, when only some functions are used. This is
3073 relevant even without the DFA addition now that patterns can be compiled in
3074 one application and matched in another.
3075
3076 The downside of splitting up is that there have to be some external
3077 functions and data tables that are used internally in different modules of
3078 the library but which are not part of the API. These have all had their
3079 names changed to start with "_pcre_" so that they are unlikely to clash
3080 with other external names.
3081
3082 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3083 a different (DFA) algorithm. Although it is slower than the original
3084 function, it does have some advantages for certain types of matching
3085 problem.
3086
3087 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3088 including restarting after a partial match.
3089
3090 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3091 defined when compiling for Windows was sent to me. I have put it into the
3092 code, though I have no means of testing or verifying it.
3093
3094 7. Added the pcre_refcount() auxiliary function.
3095
3096 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3097 match before or at the first newline in the subject string. In pcretest,
3098 the /f option on a pattern can be used to set this.
3099
3100 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3101 would behave wrongly. This has been present in PCRE since release 4.0.
3102
3103 10. A number of changes to the pcregrep command:
3104
3105 (a) Refactored how -x works; insert ^(...)$ instead of setting
3106 PCRE_ANCHORED and checking the length, in preparation for adding
3107 something similar for -w.
3108
3109 (b) Added the -w (match as a word) option.
3110
3111 (c) Refactored the way lines are read and buffered so as to have more
3112 than one at a time available.
3113
3114 (d) Implemented a pcregrep test script.
3115
3116 (e) Added the -M (multiline match) option. This allows patterns to match
3117 over several lines of the subject. The buffering ensures that at least
3118 8K, or the rest of the document (whichever is the shorter) is available
3119 for matching (and similarly the previous 8K for lookbehind assertions).
3120
3121 (f) Changed the --help output so that it now says
3122
3123 -w, --word-regex(p)
3124
3125 instead of two lines, one with "regex" and the other with "regexp"
3126 because that confused at least one person since the short forms are the
3127 same. (This required a bit of code, as the output is generated
3128 automatically from a table. It wasn't just a text change.)
3129
3130 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3131 option but starts with a hyphen. Could be a pattern or a path name
3132 starting with a hyphen, for instance.
3133
3134 (h) "-" can be given as a file name to represent stdin.
3135
3136 (i) When file names are being printed, "(standard input)" is used for
3137 the standard input, for compatibility with GNU grep. Previously
3138 "<stdin>" was used.
3139
3140 (j) The option --label=xxx can be used to supply a name to be used for
3141 stdin when file names are being printed. There is no short form.
3142
3143 (k) Re-factored the options decoding logic because we are going to add
3144 two more options that take data. Such options can now be given in four
3145 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3146
3147 (l) Added the -A, -B, and -C options for requesting that lines of context
3148 around matches be printed.
3149
3150 (m) Added the -L option to print the names of files that do not contain
3151 any matching lines, that is, the complement of -l.
3152
3153 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3154 continue to scan other files.
3155
3156 (o) The -s option was incorrectly implemented. For compatibility with other
3157 greps, it now suppresses the error message for a non-existent or non-
3158 accessible file (but not the return code). There is a new option called
3159 -q that suppresses the output of matching lines, which was what -s was
3160 previously doing.
3161
3162 (p) Added --include and --exclude options to specify files for inclusion
3163 and exclusion when recursing.
3164
3165 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3166 Hopefully, it now does.
3167
3168 12. Missing cast in pcre_study().
3169
3170 13. Added an "uninstall" target to the makefile.
3171
3172 14. Replaced "extern" in the function prototypes in Makefile.in with
3173 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3174 world, but is set differently for Windows.
3175
3176 15. Added a second compiling function called pcre_compile2(). The only
3177 difference is that it has an extra argument, which is a pointer to an
3178 integer error code. When there is a compile-time failure, this is set
3179 non-zero, in addition to the error test pointer being set to point to an
3180 error message. The new argument may be NULL if no error number is required
3181 (but then you may as well call pcre_compile(), which is now just a
3182 wrapper). This facility is provided because some applications need a
3183 numeric error indication, but it has also enabled me to tidy up the way
3184 compile-time errors are handled in the POSIX wrapper.
3185
3186 16. Added VPATH=.libs to the makefile; this should help when building with one
3187 prefix path and installing with another. (Or so I'm told by someone who
3188 knows more about this stuff than I do.)
3189
3190 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3191 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3192 match everything, including newlines. This is not POSIX-compatible, but
3193 somebody wanted the feature. From pcretest it can be activated by using
3194 both the P and the s flags.
3195
3196 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3197
3198 19. libpcre.pc was being incorrectly installed as executable.
3199
3200 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3201 it now also looks for '\r' so that it will work unmodified on Windows.
3202
3203 21. Added Google's contributed C++ wrapper to the distribution.
3204
3205 22. Added some untidy missing memory free() calls in pcretest, to keep
3206 Electric Fence happy when testing.
3207
3208
3209
3210 Version 5.0 13-Sep-04
3211 ---------------------
3212
3213 1. Internal change: literal characters are no longer packed up into items
3214 containing multiple characters in a single byte-string. Each character
3215 is now matched using a separate opcode. However, there may be more than one
3216 byte in the character in UTF-8 mode.
3217
3218 2. The pcre_callout_block structure has two new fields: pattern_position and
3219 next_item_length. These contain the offset in the pattern to the next match
3220 item, and its length, respectively.
3221
3222 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3223 insertion of callouts before each pattern item. Added the /C option to
3224 pcretest to make use of this.
3225
3226 4. On the advice of a Windows user, the lines
3227
3228 #if defined(_WIN32) || defined(WIN32)
3229 _setmode( _fileno( stdout ), 0x8000 );
3230 #endif /* defined(_WIN32) || defined(WIN32) */
3231
3232 have been added to the source of pcretest. This apparently does useful
3233 magic in relation to line terminators.
3234
3235 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3236 for the benefit of those environments where the "b" makes a difference.
3237
3238 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3239 to know about it. I have put a hack into configure.in that adds in code
3240 to set GCC=yes if CC=icc. This seems to end up at a point in the
3241 generated configure script that is early enough to affect the setting of
3242 compiler options, which is what is needed, but I have no means of testing
3243 whether it really works. (The user who reported this had patched the
3244 generated configure script, which of course I cannot do.)
3245
3246 LATER: After change 22 below (new libtool files), the configure script
3247 seems to know about icc (and also ecc). Therefore, I have commented out
3248 this hack in configure.in.
3249
3250 7. Added support for pkg-config (2 patches were sent in).
3251
3252 8. Negated POSIX character classes that used a combination of internal tables
3253 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3254 [[:^ascii]]. Typically, they would match almost any characters. The other
3255 POSIX classes were not broken in this way.
3256
3257 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3258 to find the match, as PCRE was deluded into thinking that the match had to
3259 start at the start point or following a newline. The same bug applied to
3260 patterns with negative forward assertions or any backward assertions
3261 preceding ".*" at the start, unless the pattern required a fixed first
3262 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3263
3264 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3265 starting at the last subject character, bytes beyond the end of the subject
3266 string were read.
3267
3268 11. Renamed the variable "class" as "classbits" to make life easier for C++
3269 users. (Previously there was a macro definition, but it apparently wasn't
3270 enough.)
3271
3272 12. Added the new field "tables" to the extra data so that tables can be passed
3273 in at exec time, or the internal tables can be re-selected. This allows
3274 a compiled regex to be saved and re-used at a later time by a different
3275 program that might have everything at different addresses.
3276
3277 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3278 -R library as well as a -L library.
3279
3280 14. The debugging options of pcretest (-d on the command line or D on a
3281 pattern) showed incorrect output for anything following an extended class
3282 that contained multibyte characters and which was followed by a quantifier.
3283
3284 15. Added optional support for general category Unicode character properties
3285 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3286 support. It adds about 90K to the size of the library. The meanings of the
3287 inbuilt class escapes such as \d and \s have NOT been changed.
3288
3289 16. Updated pcredemo.c to include calls to free() to release the memory for the
3290 compiled pattern.
3291
3292 17. The generated file chartables.c was being created in the source directory
3293 instead of in the building directory. This caused the build to fail if the
3294 source directory was different from the building directory, and was
3295 read-only.
3296
3297 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3298 file. No doubt somebody will tell me if they don't make sense... Also added
3299 Dan Mooney's comments about building on OpenVMS.
3300
3301 19. Added support for partial matching via the PCRE_PARTIAL option for
3302 pcre_exec() and the \P data escape in pcretest.
3303
3304 20. Extended pcretest with 3 new pattern features:
3305
3306 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3307 write the compiled pattern to the file whose name is "rest-of-line".
3308 This is a straight binary dump of the data, with the saved pointer to
3309 the character tables forced to be NULL. The study data, if any, is
3310 written too. After writing, pcretest reads a new pattern.
3311
3312 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3313 compiled pattern from the given file. There must not be any
3314 occurrences of "<" in the file name (pretty unlikely); if there are,
3315 pcretest will instead treat the initial "<" as a pattern delimiter.
3316 After reading in the pattern, pcretest goes on to read data lines as
3317 usual.
3318
3319 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3320 and 16-bit fields in a compiled pattern, to simulate a pattern that
3321 was compiled on a host of opposite endianness.
3322
3323 21. The pcre-exec() function can now cope with patterns that were compiled on
3324 hosts of opposite endianness, with this restriction:
3325
3326 As for any compiled expression that is saved and used later, the tables
3327 pointer field cannot be preserved; the extra_data field in the arguments
3328 to pcre_exec() should be used to pass in a tables address if a value
3329 other than the default internal tables were used at compile time.
3330
3331 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3332 now diagnosed as an error. Previously, most of the time, a negative number
3333 would have been treated as zero, but if in addition "ovector" was passed as
3334 NULL, a crash could occur.
3335
3336 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3337 new versions from the libtool 1.5 distribution (the last one is a copy of
3338 a file called libtool.m4). This seems to have fixed the need to patch
3339 "configure" to support Darwin 1.3 (which I used to do). However, I still
3340 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3341 workstation).
3342
3343 24. Changed the PCRE licence to be the more standard "BSD" licence.
3344
3345
3346 Version 4.5 01-Dec-03
3347 ---------------------
3348
3349 1. There has been some re-arrangement of the code for the match() function so
3350 that it can be compiled in a version that does not call itself recursively.
3351 Instead, it keeps those local variables that need separate instances for
3352 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3353 needs to "recurse". Keeping track of where control must go is done by means
3354 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3355 hide most of the details from the main code, and operates only if
3356 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3357 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3358 operating.
3359
3360 To make it easier for callers to provide specially tailored get/free
3361 functions for this usage, two new functions, pcre_stack_malloc, and
3362 pcre_stack_free, are used. They are always called in strict stacking order,
3363 and the size of block requested is always the same.
3364
3365 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3366 PCRE has been compiled to use the stack or the heap for recursion. The
3367 -C option of pcretest uses this to show which version is compiled.
3368
3369 A new data escape \S, is added to pcretest; it causes the amounts of store
3370 obtained and freed by both kinds of malloc/free at match time to be added
3371 to the output.
3372
3373 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3374 what's available on my current Linux desktop machine.
3375
3376 3. When matching a UTF-8 string, the test for a valid string at the start has
3377 been extended. If start_offset is not zero, PCRE now checks that it points
3378 to a byte that is the start of a UTF-8 character. If not, it returns
3379 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3380 this is necessary because there may be backward assertions in the pattern.
3381 When matching the same subject several times, it may save resources to use
3382 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3383
3384 4. The code for checking the validity of UTF-8 strings has been tightened so
3385 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3386 containing "overlong sequences".
3387
3388 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3389 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3390 should have been "&", but it just so happened that all the cases this let
3391 through by mistake were picked up later in the function.
3392
3393 6. I had used a variable called "isblank" - this is a C99 function, causing
3394 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3395
3396 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3397 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3398 so the version is shown; (c) stop "make test" echoing "./RunTest".
3399
3400 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3401
3402 9. The prototype for memmove() for systems that don't have it was using
3403 size_t, but the inclusion of the header that defines size_t was later. I've
3404 moved the #includes for the C headers earlier to avoid this.
3405
3406 10. Added some adjustments to the code to make it easier to compiler on certain
3407 special systems:
3408
3409 (a) Some "const" qualifiers were missing.
3410 (b) Added the macro EXPORT before all exported functions; by default this
3411 is defined to be empty.
3412 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3413 that it reads its output file name as an argument instead of writing
3414 to the standard output and assuming this can be redirected.
3415
3416 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3417 class containing characters with values greater than 255, PCRE compilation
3418 went into a loop.
3419
3420 12. A recursive reference to a subpattern that was within another subpattern
3421 that had a minimum quantifier of zero caused PCRE to crash. For example,
3422 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3423 recursion. If the recursively-called subpattern itself had a zero repeat,
3424 that was OK.
3425
3426 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3427 buffer into which it was copied (for escape processing) was still set at
3428 1024, so long lines caused crashes.
3429
3430 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3431 "internal error: code overflow...". This applied to any character class
3432 that was followed by a possessive quantifier.
3433
3434 15. Modified the Makefile to add libpcre.la as a prerequisite for
3435 libpcreposix.la because I was told this is needed for a parallel build to
3436 work.
3437
3438 16. If a pattern that contained .* following optional items at the start was
3439 studied, the wrong optimizing data was generated, leading to matching
3440 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3441 matching string must start with a or b or c. The correct conclusion for
3442 this pattern is that a match can start with any character.
3443
3444
3445 Version 4.4 13-Aug-03
3446 ---------------------
3447
3448 1. In UTF-8 mode, a character class containing characters with values between
3449 127 and 255 was not handled correctly if the compiled pattern was studied.
3450 In fixing this, I have also improved the studying algorithm for such
3451 classes (slightly).
3452
3453 2. Three internal functions had redundant arguments passed to them. Removal
3454 might give a very teeny performance improvement.
3455
3456 3. Documentation bug: the value of the capture_top field in a callout is *one
3457 more than* the number of the hightest numbered captured substring.
3458
3459 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3460 in incorrectly linking with a previously installed version. They now link
3461 explicitly with libpcre.la.
3462
3463 5. configure.in no longer needs to recognize Cygwin specially.
3464
3465 6. A problem in pcre.in for Windows platforms is fixed.
3466
3467 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3468 pcretest, it used to include the size of the study block as part of its
3469 output. Unfortunately, the structure contains a field that has a different
3470 size on different hardware architectures. This meant that the tests that
3471 showed this size failed. As the block is currently always of a fixed size,
3472 this information isn't actually particularly useful in pcretest output, so
3473 I have just removed it.
3474
3475 8. Three pre-processor statements accidentally did not start in column 1.
3476 Sadly, there are *still* compilers around that complain, even though
3477 standard C has not required this for well over a decade. Sigh.
3478
3479 9. In pcretest, the code for checking callouts passed small integers in the
3480 callout_data field, which is a void * field. However, some picky compilers
3481 complained about the casts involved for this on 64-bit systems. Now
3482 pcretest passes the address of the small integer instead, which should get
3483 rid of the warnings.
3484
3485 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3486 both compile and run time, and gives an error if an invalid UTF-8 sequence
3487 is found. There is a option for disabling this check in cases where the
3488 string is known to be correct and/or the maximum performance is wanted.
3489
3490 11. In response to a bug report, I changed one line in Makefile.in from
3491
3492 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3493 to
3494 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3495
3496 to look similar to other lines, but I have no way of telling whether this
3497 is the right thing to do, as I do not use Windows. No doubt I'll get told
3498 if it's wrong...
3499
3500
3501 Version 4.3 21-May-03
3502 ---------------------
3503
3504 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3505 Makefile.
3506
3507 2. Some refactoring to improve the quality of the code:
3508
3509 (i) The utf8_table... variables are now declared "const".
3510
3511 (ii) The code for \cx, which used the "case flipping" table to upper case
3512 lower case letters, now just substracts 32. This is ASCII-specific,
3513 but the whole concept of \cx is ASCII-specific, so it seems
3514 reasonable.
3515
3516 (iii) PCRE was using its character types table to recognize decimal and
3517 hexadecimal digits in the pattern. This is silly, because it handles
3518 only 0-9, a-f, and A-F, but the character types table is locale-
3519 specific, which means strange things might happen. A private
3520 table is now used for this - though it costs 256 bytes, a table is
3521 much faster than multiple explicit tests. Of course, the standard
3522 character types table is still used for matching digits in subject
3523 strings against \d.
3524
3525 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3526 ending in _t are). So I've renamed it as ESC_tee.
3527
3528 3. The first argument for regexec() in the POSIX wrapper should have been
3529 defined as "const".
3530
3531 4. Changed pcretest to use malloc() for its buffers so that they can be
3532 Electric Fenced for debugging.
3533
3534 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3535 to read one or more bytes before the start of the subject string. Often this
3536 had no effect on PCRE's behaviour, but in some circumstances it could
3537 provoke a segmentation fault.
3538
3539 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3540 to try to read one or more bytes before the start of the subject string.
3541
3542 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3543 UTF-8 support could misbehave in various ways if the subject string
3544 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3545 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3546 back over UTF-8 characters.)
3547
3548
3549 Version 4.2 14-Apr-03
3550 ---------------------
3551
3552 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3553
3554 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3555 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3556 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3557 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3558 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3559 and BUILD_EXEEXT
3560 Note: automatic setting of the BUILD variables is not yet working
3561 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3562 compile-time but not at link-time
3563 [LINK]: use for linking executables only
3564 make different versions for Windows and non-Windows
3565 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3566 libraries
3567 [LINK_FOR_BUILD]: new variable
3568 [OBJEXT]: use throughout
3569 [EXEEXT]: use throughout
3570 <winshared>: new target
3571 <wininstall>: new target
3572 <dftables.o>: use native compiler
3573 <dftables>: use native linker
3574 <install>: handle Windows platform correctly
3575 <clean>: ditto
3576 <check>: ditto
3577 copy DLL to top builddir before testing
3578
3579 As part of these changes, -no-undefined was removed again. This was reported
3580 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3581 in any case.
3582
3583 3. Some tidies to get rid of compiler warnings:
3584
3585 . In the match_data structure, match_limit was an unsigned long int, whereas
3586 match_call_count was an int. I've made them both unsigned long ints.
3587
3588 . In pcretest the fact that a const uschar * doesn't automatically cast to
3589 a void * provoked a warning.
3590
3591 . Turning on some more compiler warnings threw up some "shadow" variables
3592 and a few more missing casts.
3593
3594 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3595 option, a class that contained a single character with a value between 128
3596 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3597
3598 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3599 option, a class that contained several characters, but with at least one
3600 whose value was between 128 and 255 caused PCRE to crash.
3601
3602
3603 Version 4.1 12-Mar-03
3604 ---------------------
3605
3606 1. Compiling with gcc -pedantic found a couple of places where casts were
3607 needed, and a string in dftables.c that was longer than standard compilers are
3608 required to support.
3609
3610 2. Compiling with Sun's compiler found a few more places where the code could
3611 be tidied up in order to avoid warnings.
3612
3613 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3614 first of these names is deprecated in the latest Autoconf in favour of the name
3615 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3616 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3617 analogy I have changed it to CFLAGS_FOR_BUILD.
3618
3619 4. Added -no-undefined to the linking command in the Makefile, because this is
3620 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3621 linking step for the pcreposix library.
3622
3623 5. PCRE was failing to diagnose the case of two named groups with the same
3624 name.
3625
3626 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3627 literal character that is needed in the subject for a match, and scans along to
3628 ensure that it is present before embarking on the full matching process. This
3629 saves time in cases of nested unlimited repeats that are never going to match.
3630 Problem: the scan can take a lot of time if the subject is very long (e.g.
3631 megabytes), thus penalizing straightforward matches. It is now done only if the
3632 amount of subject to be scanned is less than 1000 bytes.
3633
3634 7. A lesser problem with the same optimization is that it was recording the
3635 first character of an anchored pattern as "needed", thus provoking a search
3636 right along the subject, even when the first match of the pattern was going to
3637 fail. The "needed" character is now not set for anchored patterns, unless it
3638 follows something in the pattern that is of non-fixed length. Thus, it still
3639 fulfils its original purpose of finding quick non-matches in cases of nested
3640 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3641
3642
3643 Version 4.0 17-Feb-03
3644 ---------------------
3645
3646 1. If a comment in an extended regex that started immediately after a meta-item
3647 extended to the end of string, PCRE compiled incorrect data. This could lead to
3648 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3649
3650 2. Moved to autoconf 2.53 and libtool 1.4.2.
3651
3652 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3653 the special perltest8 script is no longer needed - all the tests can be run
3654 from a single perltest script.
3655
3656 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3657 by \s. It has now been removed in PCRE. This means it isn't recognized as
3658 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3659 class [:space:] *does* include VT, thereby creating a mess.
3660
3661 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3662 space and tab.
3663
3664 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3665 its new features into the main test script, reducing the number of scripts.
3666
3667 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3668 were backward compatible, and made the (?i) apply to the whole pattern, as if
3669 /i were given. Now it behaves more logically, and applies the option setting
3670 only to what follows. PCRE has been changed to follow suit. However, if it
3671 finds options settings right at the start of the pattern, it extracts them into
3672 the global options, as before. Thus, they show up in the info data.
3673
3674 8. Added support for the \Q...\E escape sequence. Characters in between are
3675 treated as literals. This is slightly different from Perl in that $ and @ are
3676 also handled as literals inside the quotes. In Perl, they will cause variable
3677 interpolation. Note the following examples:
3678
3679 Pattern PCRE matches Perl matches
3680
3681 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3682 \Qabc\$xyz\E abc\$xyz abc\$xyz
3683 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3684
3685 For compatibility with Perl, \Q...\E sequences are recognized inside character
3686 classes as well as outside them.
3687
3688 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3689 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3690 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3691 signed/unsigned warnings.
3692
3693 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3694 option for pcretest, so I've replaced it by a simple function that does just
3695 that job.
3696
3697 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3698 "pcregrep -".
3699
3700 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3701 Java package. This provides some syntactic sugar for simple cases of what my
3702 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3703 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3704 item, you can use this simplified notation. Note that only makes sense with
3705 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3706 greediness, whatever the setting of the PCRE_UNGREEDY option.
3707
3708 13. A change of greediness default within a pattern was not taking effect at
3709 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3710 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3711 was abstracted outside.
3712
3713 14. PCRE now supports the \G assertion. It is true when the current matching
3714 position is at the start point of the match. This differs from \A when the
3715 starting offset is non-zero. Used with the /g option of pcretest (or similar
3716 code), it works in the same way as it does for Perl's /g option. If all
3717 alternatives of a regex begin with \G, the expression is anchored to the start
3718 match position, and the "anchored" flag is set in the compiled expression.
3719
3720 15. Some bugs concerning the handling of certain option changes within patterns
3721 have been fixed. These applied to options other than (?ims). For example,
3722 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3723 been the other way round. Some of this was related to change 7 above.
3724
3725 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3726 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3727 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3728 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3729
3730 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3731 mode. Unlike ".", it always matches newline, whatever the setting of
3732 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3733 assertions. Perl allows it, but it doesn't (in general) work because it can't
3734 calculate the length of the lookbehind. At least, that's the case for Perl
3735 5.8.0 - I've been told they are going to document that it doesn't work in
3736 future.
3737
3738 18. Added an error diagnosis for escapes that PCRE does not support: these are
3739 \L, \l, \N, \P, \p, \U, \u, and \X.
3740
3741 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3742 reading past the end of the pattern in cases such as /[abcd/.
3743
3744 20. PCRE was getting more memory than necessary for patterns with classes that
3745 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3746
3747 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3748 compiling PCRE for use with Virtual Pascal.
3749
3750 22. Small fix to the Makefile to make it work properly if the build is done
3751 outside the source tree.
3752
3753 23. Added a new extension: a condition to go with recursion. If a conditional
3754 subpattern starts with (?(R) the "true" branch is used if recursion has
3755 happened, whereas the "false" branch is used only at the top level.
3756
3757 24. When there was a very long string of literal characters (over 255 bytes
3758 without UTF support, over 250 bytes with UTF support), the computation of how
3759 much memory was required could be incorrect, leading to segfaults or other
3760 strange effects.
3761
3762 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3763 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3764 there was a subsequent back reference to those brackets. This meant that, for
3765 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3766 possible to check for precisely this case. All we can do is abandon the
3767 optimization if .* occurs inside capturing brackets when there are any back
3768 references whatsoever. (See below for a better fix that came later.)
3769
3770 26. The handling of the optimization for finding the first character of a
3771 non-anchored pattern, and for finding a character that is required later in the
3772 match were failing in some cases. This didn't break the matching; it just
3773 failed to optimize when it could. The way this is done has been re-implemented.
3774
3775 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3776
3777 28. Added a new feature that provides some of the functionality that Perl
3778 provides with (?{...}). The facility is termed a "callout". The way it is done
3779 in PCRE is for the caller to provide an optional function, by setting
3780 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3781 global variable. By default it is unset, which disables all calling out. To get
3782 the function called, the regex must include (?C) at appropriate points. This
3783 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3784 This provides a means of identifying different callout points. When PCRE
3785 reaches such a point in the regex, if pcre_callout has been set, the external
3786 function is called. It is provided with data in a structure called
3787 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3788 matching continues; if it returns a non-zero value, the match at the current
3789 point fails. However, backtracking will occur if possible. [This was changed
3790 later and other features added - see item 49 below.]
3791
3792 29. pcretest is upgraded to test the callout functionality. It provides a
3793 callout function that displays information. By default, it shows the start of
3794 the match and the current position in the text. There are some new data escapes
3795 to vary what happens:
3796
3797 \C+ in addition, show current contents of captured substrings
3798 \C- do not supply a callout function
3799 \C!n return 1 when callout number n is reached
3800 \C!n!m return 1 when callout number n is reached for the mth time
3801
3802 30. If pcregrep was called with the -l option and just a single file name, it
3803 output "<stdin>" if a match was found, instead of the file name.
3804
3805 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3806 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3807 pcre_exec(). This saves a malloc/free per call. The default value of
3808 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3809 when configuring.
3810
3811 32. The default maximum size of a compiled pattern is 64K. There have been a
3812 few cases of people hitting this limit. The code now uses macros to handle the
3813 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3814 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3815 configuring. Tests 2 and 5 work only with 2-byte links because they output
3816 debugging information about compiled patterns.
3817
3818 33. Internal code re-arrangements:
3819
3820 (a) Moved the debugging function for printing out a compiled regex into
3821 its own source file (printint.c) and used #include to pull it into
3822 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3823 separate copies.
3824
3825 (b) Defined the list of op-code names for debugging as a macro in
3826 internal.h so that it is next to the definition of the opcodes.
3827
3828 (c) Defined a table of op-code lengths for simpler skipping along compiled
3829 code. This is again a macro in internal.h so that it is next to the
3830 definition of the opcodes.
3831
3832 34. Added support for recursive calls to individual subpatterns, along the
3833 lines of Robin Houston's patch (but implemented somewhat differently).
3834
3835 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3836 allow it to read and process whole directories in Win32. This code was
3837 contributed by Lionel Fourquaux; it has not been tested by me.
3838
3839 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3840 used to name a group. Names consist of alphanumerics and underscores, and must
3841 be unique. Back references use the syntax (?P=name) and recursive calls use
3842 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3843 numbers. The function pcre_fullinfo() can be used after compilation to extract
3844 a name/number map. There are three relevant calls:
3845
3846 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3847 PCRE_INFO_NAMECOUNT yields the number of entries
3848 PCRE_INFO_NAMETABLE yields a pointer to the map.
3849
3850 The map is a vector of fixed-size entries. The size of each entry depends on
3851 the length of the longest name used. The first two bytes of each entry are the
3852 group number, most significant byte first. There follows the corresponding
3853 name, zero terminated. The names are in alphabetical order.
3854
3855 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3856 case instead of 255. Making it the same both with and without UTF-8 support
3857 means that the same test output works with both.
3858
3859 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3860 calling malloc() with a zero argument.
3861
3862 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3863 optimization. I've improved things by keeping a bitmap of backreferences with
3864 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3865 fact referenced, the optimization can be applied. It is unlikely that a
3866 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3867 the match to follow \n) will appear inside brackets with a number greater than
3868 31, but if it does, any back reference > 31 suppresses the optimization.
3869
3870 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3871 of disabling numbered capturing parentheses. Any opening parenthesis that is
3872 not followed by ? behaves as if it were followed by ?: but named parentheses
3873 can still be used for capturing (and they will acquire numbers in the usual
3874 way).
3875
3876 41. Redesigned the return codes from the match() function into yes/no/error so
3877 that errors can be passed back from deep inside the nested calls. A malloc
3878 failure while inside a recursive subpattern call now causes the
3879 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3880
3881 42. It is now possible to set a limit on the number of times the match()
3882 function is called in a call to pcre_exec(). This facility makes it possible to
3883 limit the amount of recursion and backtracking, though not in a directly
3884 obvious way, because the match() function is used in a number of different
3885 circumstances. The count starts from zero for each position in the subject
3886 string (for non-anchored patterns). The default limit is, for compatibility, a
3887 large number, namely 10 000 000. You can change this in two ways:
3888
3889 (a) When configuring PCRE before making, you can use --with-match-limit=n
3890 to set a default value for the compiled library.
3891
3892 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3893 a different value is set. See 45 below.
3894
3895 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3896
3897 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3898 of things that can be changed at compile time. The first argument specifies
3899 what is wanted and the second points to where the information is to be placed.
3900 The current list of available information is:
3901
3902 PCRE_CONFIG_UTF8
3903
3904 The output is an integer that is set to one if UTF-8 support is available;
3905 otherwise it is set to zero.
3906
3907 PCRE_CONFIG_NEWLINE
3908
3909 The output is an integer that it set to the value of the code that is used for
3910 newline. It is either LF (10) or CR (13).
3911
3912 PCRE_CONFIG_LINK_SIZE
3913
3914 The output is an integer that contains the number of bytes used for internal
3915 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3916
3917 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3918
3919 The output is an integer that contains the threshold above which the POSIX
3920 interface uses malloc() for output vectors. See item 31 above.
3921
3922 PCRE_CONFIG_MATCH_LIMIT
3923
3924 The output is an unsigned integer that contains the default limit of the number
3925 of match() calls in a pcre_exec() execution. See 42 above.
3926
3927 44. pcretest has been upgraded by the addition of the -C option. This causes it
3928 to extract all the available output from the new pcre_config() function, and to
3929 output it. The program then exits immediately.
3930
3931 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3932 order to support additional features. One way would have been to define
3933 pcre_exec2() (for example) with extra arguments, but this would not have been
3934 extensible, and would also have required all calls to the original function to
3935 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3936 is used for passing in "extra" data from pcre_study().
3937
3938 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3939 contains the following fields:
3940
3941 flags a bitmap indicating which of the following fields are set
3942 study_data opaque data from pcre_study()
3943 match_limit a way of specifying a limit on match() calls for a specific
3944 call to pcre_exec()
3945 callout_data data for callouts (see 49 below)
3946
3947 The flag bits are also defined in pcre.h, and are
3948
3949 PCRE_EXTRA_STUDY_DATA
3950 PCRE_EXTRA_MATCH_LIMIT
3951 PCRE_EXTRA_CALLOUT_DATA
3952
3953 The pcre_study() function now returns one of these new pcre_extra blocks, with
3954 the actual study data pointed to by the study_data field, and the
3955 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3956 before. That is, this change is entirely upwards-compatible and requires no
3957 change to existing code.
3958
3959 If you want to pass in additional data to pcre_exec(), you can either place it
3960 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3961 block.
3962
3963 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3964 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3965 times with different match limits, until it finds the minimum value needed for
3966 pcre_exec() to complete. The value is then output. This can be instructive; for
3967 most simple matches the number is quite small, but for pathological cases it
3968 gets very large very quickly.
3969
3970 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3971 returns the size of the data block pointed to by the study_data field in a
3972 pcre_extra block, that is, the value that was passed as the argument to
3973 pcre_malloc() when PCRE was getting memory in which to place the information
3974 created by pcre_study(). The fourth argument should point to a size_t variable.
3975 pcretest has been extended so that this information is shown after a successful
3976 pcre_study() call when information about the compiled regex is being displayed.
3977
3978 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3979 because what follows is always an absolute path. (Later: it turns out that this
3980 is more than cosmetic for MinGW, because it doesn't like empty path
3981 components.)
3982
3983 49. Some changes have been made to the callout feature (see 28 above):
3984
3985 (i) A callout function now has three choices for what it returns:
3986
3987 0 => success, carry on matching
3988 > 0 => failure at this point, but backtrack if possible
3989 < 0 => serious error, return this value from pcre_exec()
3990
3991 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3992 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3993 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3994 use by callout functions. It will never be used by PCRE itself.
3995
3996 (ii) The pcre_extra structure (see 45 above) has a void * field called
3997 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3998 pcre_callout_block structure has a field of the same name. The contents of
3999 the field passed in the pcre_extra structure are passed to the callout
4000 function in the corresponding field in the callout block. This makes it
4001 easier to use the same callout-containing regex from multiple threads. For
4002 testing, the pcretest program has a new data escape
4003
4004 \C*n pass the number n (may be negative) as callout_data
4005
4006 If the callout function in pcretest receives a non-zero value as
4007 callout_data, it returns that value.
4008
4009 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4010 there were some redundant $(CFLAGS) in commands that are now specified as
4011 $(LINK), which already includes $(CFLAGS).
4012
4013 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4014 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4015 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4016 one-byte characters throughout. Note that case-insensitive matching applies
4017 only to characters whose values are less than 256. PCRE doesn't support the
4018 notion of cases for higher-valued characters.
4019
4020 (i) A character class whose characters are all within 0-255 is handled as
4021 a bit map, and the map is inverted for negative classes. Previously, a
4022 character > 255 always failed to match such a class; however it should
4023 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4024
4025 (ii) A negated character class with a single character < 255 is coded as
4026 "not this character" (OP_NOT). This wasn't working properly when the test
4027 character was multibyte, either singly or repeated.
4028
4029 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4030 mode, for example: \x{100}{2,3}.
4031
4032 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4033 singly or repeated) now correctly test multibyte characters. However,
4034 PCRE doesn't recognize any characters with values greater than 255 as
4035 digits, spaces, or word characters. Such characters always match \D, \S,
4036 and \W, and never match \d, \s, or \w.
4037
4038 (v) Classes may now contain characters and character ranges with values
4039 greater than 255. For example: [ab\x{100}-\x{400}].
4040
4041 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4042 PCRE in UTF-8 mode.
4043
4044 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4045 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4046 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4047 value.)
4048
4049 53. The single man page has become too large. I have therefore split it up into
4050 a number of separate man pages. These also give rise to individual HTML pages;
4051 these are now put in a separate directory, and there is an index.html page that
4052 lists them all. Some hyperlinking between the pages has been installed.
4053
4054 54. Added convenience functions for handling named capturing parentheses.
4055
4056 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4057 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4058 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4059 are faulted.
4060
4061 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4062 calling configure. These values are used when compiling the dftables.c program
4063 which is run to generate the source of the default character tables. They
4064 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4065 you will need to set these values.
4066
4067 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4068
4069
4070 Version 3.9 02-Jan-02
4071 ---------------------
4072
4073 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4074
4075 2. If --disable-static was given, the building process failed when trying to
4076 build pcretest and pcregrep. (For some reason it was using libtool to compile
4077 them, which is not right, as they aren't part of the library.)
4078
4079
4080 Version 3.8 18-Dec-01
4081 ---------------------
4082
4083 1. The experimental UTF-8 code was completely screwed up. It was packing the
4084 bytes in the wrong order. How dumb can you get?
4085
4086
4087 Version 3.7 29-Oct-01
4088 ---------------------
4089
4090 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4091 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4092 this didn't happen under Solaris 8, where I normally test things.
4093
4094 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4095 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4096 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4097 this fix an hour or so after the initial 3.7 release.)
4098
4099
4100 Version 3.6 23-Oct-01
4101 ---------------------
4102
4103 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4104 offsets passed as NULL with zero offset count.
4105
4106 2. The config.guess and config.sub files had not been updated when I moved to
4107 the latest autoconf.
4108
4109
4110 Version 3.5 15-Aug-01
4111 ---------------------
4112
4113 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4114 had been forgotten.
4115
4116 2. By using declared but undefined structures, we can avoid using "void"
4117 definitions in pcre.h while keeping the internal definitions of the structures
4118 private.
4119
4120 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4121 user point of view, this means that both static and shared libraries are built
4122 by default, but this can be individually controlled. More of the work of
4123 handling this static/shared cases is now inside libtool instead of PCRE's make
4124 file.
4125
4126 4. The pcretest utility is now installed along with pcregrep because it is
4127 useful for users (to test regexs) and by doing this, it automatically gets
4128 relinked by libtool. The documentation has been turned into a man page, so
4129 there are now .1, .txt, and .html versions in /doc.
4130
4131 5. Upgrades to pcregrep:
4132 (i) Added long-form option names like gnu grep.
4133 (ii) Added --help to list all options with an explanatory phrase.
4134 (iii) Added -r, --recursive to recurse into sub-directories.
4135 (iv) Added -f, --file to read patterns from a file.
4136
4137 6. pcre_exec() was referring to its "code" argument before testing that
4138 argument for NULL (and giving an error if it was NULL).
4139
4140 7. Upgraded Makefile.in to allow for compiling in a different directory from
4141 the source directory.
4142
4143 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4144 options bits, the pointer it was passed was to an int instead of to an unsigned
4145 long int. This mattered only on 64-bit systems.
4146
4147 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4148 generated) instead of pcre.in, which it its source. Also made the same change
4149 in several of the .c files.
4150
4151 10. A new release of gcc defines printf() as a macro, which broke pcretest
4152 because it had an ifdef in the middle of a string argument for printf(). Fixed
4153 by using separate calls to printf().
4154
4155 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4156 script, to force use of CR or LF instead of \n in the source. On non-Unix
4157 systems, the value can be set in config.h.
4158
4159 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4160 absolute limit. Changed the text of the error message to make this clear, and
4161 likewise updated the man page.
4162
4163 13. The limit of 99 on the number of capturing subpatterns has been removed.
4164 The new limit is 65535, which I hope will not be a "real" limit.
4165
4166
4167 Version 3.4 22-Aug-00
4168 ---------------------
4169
4170 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4171
4172 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4173
4174
4175 Version 3.3 01-Aug-00
4176 ---------------------
4177
4178 1. If an octal character was given, but the value was greater than \377, it
4179 was not getting masked to the least significant bits, as documented. This could
4180 lead to crashes in some systems.
4181
4182 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4183 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4184
4185 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4186 These just pass their arguments on to (pcre_free)(), but they are provided
4187 because some uses of PCRE bind it to non-C systems that can call its functions,
4188 but cannot call free() or pcre_free() directly.
4189
4190 4. Add "make test" as a synonym for "make check". Corrected some comments in
4191 the Makefile.
4192
4193 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4194 Makefile.
4195
4196 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4197 command called pgrep for grepping around the active processes.
4198
4199 7. Added the beginnings of support for UTF-8 character strings.
4200
4201 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4202 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4203 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4204 out for the ar command.)
4205
4206
4207 Version 3.2 12-May-00
4208 ---------------------
4209
4210 This is purely a bug fixing release.
4211
4212 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4213 of ZA. This was just one example of several cases that could provoke this bug,
4214 which was introduced by change 9 of version 2.00. The code for breaking
4215 infinite loops after an iteration that matches an empty string was't working
4216 correctly.
4217
4218 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4219 when matched against abbab (for example). After matching an empty string, it
4220 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4221 caused it to match further down the string than it should.
4222
4223 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4224 was there because it doesn't seem to be needed, and it causes trouble on some
4225 systems, as it is not a Standard C header. It has been removed.
4226
4227 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4228 were reported on the Macintosh. The changes were from
4229
4230 while ((c = *(++ptr)) != 0 && c != '\n');
4231 to
4232 while ((c = *(++ptr)) != 0 && c != '\n') ;
4233
4234 Totally extraordinary, but if that's what it takes...
4235
4236 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4237 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4238 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4239 assumes the way PCRE uses memmove() (always moving upwards).
4240
4241 6. PCRE is being used in one environment where strchr() is not available. There
4242 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4243 faster code anyway.
4244
4245
4246 Version 3.1 09-Feb-00
4247 ---------------------
4248
4249 The only change in this release is the fixing of some bugs in Makefile.in for
4250 the "install" target:
4251
4252 (1) It was failing to install pcreposix.h.
4253
4254 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4255
4256
4257 Version 3.0 01-Feb-00
4258 ---------------------
4259
4260 1. Add support for the /+ modifier to perltest (to output $` like it does in
4261 pcretest).
4262
4263 2. Add support for the /g modifier to perltest.
4264
4265 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4266 matches null strings.
4267
4268 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4269 pattern. Perl treats empty patterns specially - it reuses the most recent
4270 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4271 effect.
4272
4273 5. The POSIX interface was broken in that it was just handing over the POSIX
4274 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4275 required a bigger vector, with some working space on the end. This means that
4276 the POSIX wrapper now has to get and free some memory, and copy the results.
4277
4278 6. Added some simple autoconf support, placing the test data and the
4279 documentation in separate directories, re-organizing some of the
4280 information files, and making it build pcre-config (a GNU standard). Also added
4281 libtool support for building PCRE as a shared library, which is now the
4282 default.
4283
4284 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4285 09 are not valid octal constants. Single digits will be used for minor values
4286 less than 10.
4287
4288 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4289 existing programs that set these in the POSIX interface can use PCRE without
4290 modification.
4291
4292 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4293 return all that pcre_info() returns, plus additional data. The pcre_info()
4294 function is retained for compatibility, but is considered to be obsolete.
4295
4296 10. Added experimental recursion feature (?R) to handle one common case that
4297 Perl 5.6 will be able to do with (?p{...}).
4298
4299 11. Added support for POSIX character classes like [:alpha:], which Perl is
4300 adopting.
4301
4302
4303 Version 2.08 31-Aug-99
4304 ----------------------
4305
4306 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4307 trying to match at the startoffset position, but instead was moving forward to
4308 the next newline as if a previous match had failed.
4309
4310 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4311 and could get into a loop if a null string was matched other than at the start
4312 of the subject.
4313
4314 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4315 be distinguished at compile time, and for completeness also added PCRE_DATE.
4316
4317 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4318 in GnuWin32 environments.
4319
4320
4321 Version 2.07 29-Jul-99
4322 ----------------------
4323
4324 1. The documentation is now supplied in plain text form and HTML as well as in
4325 the form of man page sources.
4326
4327 2. C++ compilers don't like assigning (void *) values to other pointer types.
4328 In particular this affects malloc(). Although there is no problem in Standard
4329 C, I've put in casts to keep C++ compilers happy.
4330
4331 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4332 should be (const char *).
4333
4334 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4335 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4336 However, I haven't made this a standard facility. The documentation doesn't
4337 mention it, and the Makefile doesn't support it.
4338
4339 5. The Makefile now contains an "install" target, with editable destinations at
4340 the top of the file. The pcretest program is not installed.
4341
4342 6. pgrep -V now gives the PCRE version number and date.
4343
4344 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4345 causing the entire string to be ignored, instead of just the last character.
4346
4347 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4348 non-matching string, it can take a very, very long time, even for strings of
4349 quite modest length, because of the nested recursion. PCRE now does better in
4350 some of these cases. It does this by remembering the last required literal
4351 character in the pattern, and pre-searching the subject to ensure it is present
4352 before running the real match. In other words, it applies a heuristic to detect
4353 some types of certain failure quickly, and in the above example, if presented
4354 with a string that has no trailing " it gives "no match" very quickly.
4355
4356 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4357 other alternatives are tried instead.
4358
4359
4360 Version 2.06 09-Jun-99
4361 ----------------------
4362
4363 1. Change pcretest's output for amount of store used to show just the code
4364 space, because the remainder (the data block) varies in size between 32-bit and
4365 64-bit systems.
4366
4367 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4368 start matching at. This allows lookbehinds to work when searching for multiple
4369 occurrences in a string.
4370
4371 3. Added additional options to pcretest for testing multiple occurrences:
4372
4373 /+ outputs the rest of the string that follows a match
4374 /g loops for multiple occurrences, using the new startoffset argument
4375 /G loops for multiple occurrences by passing an incremented pointer
4376
4377 4. PCRE wasn't doing the "first character" optimization for patterns starting
4378 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4379 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4380 the letter 'x'. On long subject strings, this gives a significant speed-up.
4381
4382
4383 Version 2.05 21-Apr-99
4384 ----------------------
4385
4386 1. Changed the type of magic_number from int to long int so that it works
4387 properly on 16-bit systems.
4388
4389 2. Fixed a bug which caused patterns starting with .* not to work correctly
4390 when the subject string contained newline characters. PCRE was assuming
4391 anchoring for such patterns in all cases, which is not correct because .* will
4392 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4393 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4394 must be retried after every newline in the subject.
4395
4396
4397 Version 2.04 18-Feb-99
4398 ----------------------
4399
4400 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4401 computation of the store needed to hold the pattern was incorrect (too large).
4402 If such patterns were nested a few deep, this could multiply and become a real
4403 problem.
4404
4405 2. Added /M option to pcretest to show the memory requirement of a specific
4406 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4407
4408 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4409 compiled in such a way that the backtracking after subsequent failure was
4410 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4411 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4412
4413
4414 Version 2.03 02-Feb-99
4415 ----------------------
4416
4417 1. Fixed typo and small mistake in man page.
4418
4419 2. Added 4th condition (GPL supersedes if conflict) and created separate
4420 LICENCE file containing the conditions.
4421
4422 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4423 Perl, that is the internal \ allows the delimiter to be included in the
4424 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4425 the final delimiter, add \ to the end of the pattern (to test the error).
4426
4427 4. Added the convenience functions for extracting substrings after a successful
4428 match. Updated pcretest to make it able to test these functions.
4429
4430
4431 Version 2.02 14-Jan-99
4432 ----------------------
4433
4434 1. Initialized the working variables associated with each extraction so that
4435 their saving and restoring doesn't refer to uninitialized store.
4436
4437 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4438 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4439 fix the problem.
4440
4441 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4442 calls, and wasn't printing the correct value for compiling calls. Increased the
4443 default value of LOOPREPEAT, and the number of significant figures in the
4444 times.
4445
4446 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4447
4448 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4449 a building problem on Windows NT with a FAT file system.
4450
4451
4452 Version 2.01 21-Oct-98
4453 ----------------------
4454
4455 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4456 to character tables built by pcre_maketables() in the current locale. If NULL
4457 is passed, the default tables are used.
4458
4459
4460 Version 2.00 24-Sep-98
4461 ----------------------
4462
4463 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4464 it any more.
4465
4466 2. Allow quantification of (?>) groups, and make it work correctly.
4467
4468 3. The first character computation wasn't working for (?>) groups.
4469
4470 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4471 end of the subject) and add 5.005's \z, which really does match only at the
4472 very end of the subject.
4473
4474 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4475
4476 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4477 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4478 localized options. All options to pcre_study() were also removed.
4479
4480 7. Add other new features from 5.005:
4481
4482 $(?<= positive lookbehind
4483 $(?<! negative lookbehind
4484 (?imsx-imsx) added the unsetting capability
4485 such a setting is global if at outer level; local otherwise
4486 (?imsx-imsx:) non-capturing groups with option setting
4487 (?(cond)re|re) conditional pattern matching
4488
4489 A backreference to itself in a repeated group matches the previous
4490 captured string.
4491
4492 8. General tidying up of studying (both automatic and via "study")
4493 consequential on the addition of new assertions.
4494
4495 9. As in 5.005, unlimited repeated groups that could match an empty substring
4496 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4497 runtime if any iteration does actually match an empty substring.
4498
4499 10. Include the RunTest script in the distribution.
4500
4501 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4502 discrepancies, some of which were old and were also with respect to 5.004. They
4503 have now been fixed.
4504
4505
4506 Version 1.09 28-Apr-98
4507 ----------------------
4508
4509 1. A negated single character class followed by a quantifier with a minimum
4510 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4511 program crashes, or just wrong answers. This did not apply to negated classes
4512 containing more than one character, or to minima other than one.
4513
4514
4515 Version 1.08 27-Mar-98
4516 ----------------------
4517
4518 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4519
4520 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4521 latter must appear before anything that relies on it in the pattern.
4522
4523
4524 Version 1.07 16-Feb-98
4525 ----------------------
4526
4527 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4528 repeat of a potentially empty string).
4529
4530
4531 Version 1.06 23-Jan-98
4532 ----------------------
4533
4534 1. Added Markus Oberhumer's little patches for C++.
4535
4536 2. Literal strings longer than 255 characters were broken.
4537
4538
4539 Version 1.05 23-Dec-97
4540 ----------------------
4541
4542 1. Negated character classes containing more than one character were failing if
4543 PCRE_CASELESS was set at run time.
4544
4545
4546 Version 1.04 19-Dec-97
4547 ----------------------
4548
4549 1. Corrected the man page, where some "const" qualifiers had been omitted.
4550
4551 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4552 input syntax.
4553
4554 3. Fixed memory leak which occurred when a regex with back references was
4555 matched with an offsets vector that wasn't big enough. The temporary memory
4556 that is used in this case wasn't being freed if the match failed.
4557
4558 4. Tidied pcretest to ensure it frees memory that it gets.
4559
4560 5. Temporary memory was being obtained in the case where the passed offsets
4561 vector was exactly big enough.
4562
4563 6. Corrected definition of offsetof() from change 5 below.
4564
4565 7. I had screwed up change 6 below and broken the rules for the use of
4566 setjmp(). Now fixed.
4567
4568
4569 Version 1.03 18-Dec-97
4570 ----------------------
4571
4572 1. A erroneous regex with a missing opening parenthesis was correctly
4573 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4574 on some systems.
4575
4576 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4577 it was reported that one broken compiler failed on the former because "code" is
4578 also an independent variable.
4579
4580 3. The erroneous regex a[]b caused an array overrun reference.
4581
4582 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4583 fail on data ending with that character. (It was going on too far, and checking
4584 the next character, typically a binary zero.) This was specific to the
4585 optimized code for single-character negative classes.
4586
4587 5. Added a contributed patch from the TIN world which does the following:
4588
4589 + Add an undef for memmove, in case the the system defines a macro for it.
4590
4591 + Add a definition of offsetof(), in case there isn't one. (I don't know
4592 the reason behind this - offsetof() is part of the ANSI standard - but
4593 it does no harm).
4594
4595 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4596 most of the places where whitespace preceded '#'. I have given up and
4597 allowed the remaining 2 cases to be at the margin.
4598
4599 + Rename some variables in pcre to eliminate shadowing. This seems very
4600 pedantic, but does no harm, of course.
4601
4602 6. Moved the call to setjmp() into its own function, to get rid of warnings
4603 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4604
4605 7. Constructs such as \d{8,} were compiling into the equivalent of
4606 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4607 outcome, but in this particular case used more store than had been allocated,
4608 which caused the bug to be discovered because it threw up an internal error.
4609
4610 8. The debugging code in both pcre and pcretest for outputting the compiled
4611 form of a regex was going wrong in the case of back references followed by
4612 curly-bracketed repeats.
4613
4614
4615 Version 1.02 12-Dec-97
4616 ----------------------
4617
4618 1. Typos in pcre.3 and comments in the source fixed.
4619
4620 2. Applied a contributed patch to get rid of places where it used to remove
4621 'const' from variables, and fixed some signed/unsigned and uninitialized
4622 variable warnings.
4623
4624 3. Added the "runtest" target to Makefile.
4625
4626 4. Set default compiler flag to -O2 rather than just -O.
4627
4628
4629 Version 1.01 19-Nov-97
4630 ----------------------
4631
4632 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4633 like /([ab]*)*/, that is, for classes with more than one character in them.
4634
4635 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4636 as /((?>a*))*/ (a PCRE_EXTRA facility).
4637
4638
4639 Version 1.00 18-Nov-97
4640 ----------------------
4641
4642 1. Added compile-time macros to support systems such as SunOS4 which don't have
4643 memmove() or strerror() but have other things that can be used instead.
4644
4645 2. Arranged that "make clean" removes the executables.
4646
4647
4648 Version 0.99 27-Oct-97
4649 ----------------------
4650
4651 1. Fixed bug in code for optimizing classes with only one character. It was
4652 initializing a 32-byte map regardless, which could cause it to run off the end
4653 of the memory it had got.
4654
4655 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4656
4657
4658 Version 0.98 22-Oct-97
4659 ----------------------
4660
4661 1. Fixed bug in code for handling temporary memory usage when there are more
4662 back references than supplied space in the ovector. This could cause segfaults.
4663
4664
4665 Version 0.97 21-Oct-97
4666 ----------------------
4667
4668 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4669
4670 2. Optimized negated single characters not to use a bit map.
4671
4672 3. Brought error texts together as macro definitions; clarified some of them;
4673 fixed one that was wrong - it said "range out of order" when it meant "invalid
4674 escape sequence".
4675
4676 4. Changed some char * arguments to const char *.
4677
4678 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4679
4680 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4681 pcretest.
4682
4683
4684 Version 0.96 16-Oct-97
4685 ----------------------
4686
4687 1. Added a simple "pgrep" utility to the distribution.
4688
4689 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4690 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4691 where "ddd" means "one or more decimal digits".
4692
4693 3. Fixed serious bug. If a pattern had a back reference, but the call to
4694 pcre_exec() didn't supply a large enough ovector to record the related
4695 identifying subpattern, the match always failed. PCRE now remembers the number
4696 of the largest back reference, and gets some temporary memory in which to save
4697 the offsets during matching if necessary, in order to ensure that
4698 backreferences always work.
4699
4700 4. Increased the compatibility with Perl in a number of ways:
4701
4702 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4703 to request this handling. The option can be set at compile or exec time.
4704
4705 (b) $ matches before a terminating newline by default; an option
4706 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4707 mode). The option can be set at compile or exec time.
4708
4709 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4710 the same as Perl's. If the decimal number it represents is less than 10
4711 or there aren't that many previous left capturing parentheses, an octal
4712 escape is read. Inside a character class, it's always an octal escape,
4713 even if it is a single digit.
4714
4715 (d) An escaped but undefined alphabetic character is taken as a literal,
4716 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4717 escapes.
4718
4719 (e) {0} is now permitted. (The previous item is removed from the compiled
4720 pattern).
4721
4722 5. Changed all the names of code files so that the basic parts are no longer
4723 than 10 characters, and abolished the teeny "globals.c" file.
4724
4725 6. Changed the handling of character classes; they are now done with a 32-byte
4726 bit map always.
4727
4728 7. Added the -d and /D options to pcretest to make it possible to look at the
4729 internals of compilation without having to recompile pcre.
4730
4731
4732 Version 0.95 23-Sep-97
4733 ----------------------
4734
4735 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4736 \x20 at the start of a run of normal characters. These were being treated as
4737 real characters, instead of the source characters being re-checked.
4738
4739
4740 Version 0.94 18-Sep-97
4741 ----------------------
4742
4743 1. The functions are now thread-safe, with the caveat that the global variables
4744 containing pointers to malloc() and free() or alternative functions are the
4745 same for all threads.
4746
4747 2. Get pcre_study() to generate a bitmap of initial characters for non-
4748 anchored patterns when this is possible, and use it if passed to pcre_exec().
4749
4750
4751 Version 0.93 15-Sep-97
4752 ----------------------
4753
4754 1. /(b)|(:+)/ was computing an incorrect first character.
4755
4756 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4757 but not actually doing anything yet.
4758
4759 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4760 as Perl does (e.g. [-az] or [az-]).
4761
4762 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4763 all possible positions.
4764
4765 5. Split up into different modules to avoid including unneeded functions in a
4766 compiled binary. However, compile and exec are still in one module. The "study"
4767 function is split off.
4768
4769 6. The character tables are now in a separate module whose source is generated
4770 by an auxiliary program - but can then be edited by hand if required. There are
4771 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4772 toupper() in the code.
4773
4774 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4775 make them global. Abolish the function for setting them, as the caller can now
4776 set them directly.
4777
4778
4779 Version 0.92 11-Sep-97
4780 ----------------------
4781
4782 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4783 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4784
4785 2. Caseless matching was not working in character classes if the characters in
4786 the pattern were in upper case.
4787
4788 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4789
4790 4. Make PCRE_ANCHORED public and accept as a compile option.
4791
4792 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4793 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4794 pass them.
4795
4796 6. Give an error if bad option bits passed at compile or run time.
4797
4798 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4799 pcretest to cause it to pass that flag.
4800
4801 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4802 options, and the first character, if set.
4803
4804 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4805
4806
4807 Version 0.91 10-Sep-97
4808 ----------------------
4809
4810 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4811 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4812
4813 2. PCRE was looping on encountering an indefinitely repeated back reference to
4814 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4815 Perl does - treats the match as successful.
4816
4817 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5