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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 994 - (show annotations)
Tue Jul 10 14:29:26 2012 UTC (7 years, 3 months ago) by ph10
File size: 233503 byte(s)
Error occurred while calculating annotation data.
Fix patterns that incorrectly set "anchored" or "start of line" for .* inside 
atomic parentheses or when *PRUNE or *SKIP is present.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25
26 Version 8.31 06-July-2012
27 -------------------------
28
29 1. Fixing a wrong JIT test case and some compiler warnings.
30
31 2. Removed a bashism from the RunTest script.
32
33 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
34 to unsigned type, result still unsigned" that was given by an MS compiler
35 on encountering the code "-sizeof(xxx)".
36
37 4. Partial matching support is added to the JIT compiler.
38
39 5. Fixed several bugs concerned with partial matching of items that consist
40 of more than one character:
41
42 (a) /^(..)\1/ did not partially match "aba" because checking references was
43 done on an "all or nothing" basis. This also applied to repeated
44 references.
45
46 (b) \R did not give a hard partial match if \r was found at the end of the
47 subject.
48
49 (c) \X did not give a hard partial match after matching one or more
50 characters at the end of the subject.
51
52 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
53 a partial match for the string "\r".
54
55 (e) When newline was set to CRLF, the metacharacter "." did not recognize
56 a partial match for a CR character at the end of the subject string.
57
58 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
59 running pcretest, the text "(JIT)" added to the output whenever JIT is
60 actually used to run the match.
61
62 7. Individual JIT compile options can be set in pcretest by following -s+[+]
63 or /S+[+] with a digit between 1 and 7.
64
65 8. OP_NOT now supports any UTF character not just single-byte ones.
66
67 9. (*MARK) control verb is now supported by the JIT compiler.
68
69 10. The command "./RunTest list" lists the available tests without actually
70 running any of them. (Because I keep forgetting what they all are.)
71
72 11. Add PCRE_INFO_MAXLOOKBEHIND.
73
74 12. Applied a (slightly modified) user-supplied patch that improves performance
75 when the heap is used for recursion (compiled with --disable-stack-for-
76 recursion). Instead of malloc and free for each heap frame each time a
77 logical recursion happens, frames are retained on a chain and re-used where
78 possible. This sometimes gives as much as 30% improvement.
79
80 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
81 call.
82
83 14. As documented, (*COMMIT) is now confined to within a positive assertion.
84
85 15. It is now possible to link pcretest with libedit as an alternative to
86 libreadline.
87
88 16. (*COMMIT) control verb is now supported by the JIT compiler.
89
90 17. The Unicode data tables have been updated to Unicode 6.1.0.
91
92 18. Added --file-list option to pcregrep.
93
94 19. Added binary file support to pcregrep, including the -a, --binary-files,
95 -I, and --text options.
96
97 20. The madvise function is renamed for posix_madvise for QNX compatibility
98 reasons. Fixed by Giuseppe D'Angelo.
99
100 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
101
102 22. Changed the option for creating symbolic links for 16-bit man pages from
103 -s to -sf so that re-installing does not cause issues.
104
105 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
106
107 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
108 matches in certain environments (the workspace was not being correctly
109 retained). Also added to pcre_dfa_exec() a simple plausibility check on
110 some of the workspace data at the beginning of a restart.
111
112 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
113 was not doing so when it should - probably a typo introduced by SVN 528
114 (change 8.10/14).
115
116 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
117 \w+ when the character tables indicated that \x{c4} was a word character.
118 There were several related cases, all because the tests for doing a table
119 lookup were testing for characters less than 127 instead of 255.
120
121 27. If a pattern contains capturing parentheses that are not used in a match,
122 their slots in the ovector are set to -1. For those that are higher than
123 any matched groups, this happens at the end of processing. In the case when
124 there were back references that the ovector was too small to contain
125 (causing temporary malloc'd memory to be used during matching), and the
126 highest capturing number was not used, memory off the end of the ovector
127 was incorrectly being set to -1. (It was using the size of the temporary
128 memory instead of the true size.)
129
130 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
131 ovector size, it uses memory at the end of the block that it has got.
132
133 29. Check for an overlong MARK name and give an error at compile time. The
134 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
135
136 30. JIT compiler update.
137
138 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
139 Rill for the patch.
140
141 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
142
143 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
144
145 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
146 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
147 was enabled.
148
149 35. Improve JIT code generation for greedy plus quantifier.
150
151 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
152 1 to "aa" instead of to an empty string. The bug affected repeated groups
153 that could potentially match an empty string.
154
155 37. Optimizing single character iterators in JIT.
156
157 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
158 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
159 codepoints that are too big for the mode are faulted, and in a UTF mode,
160 disallowed codepoints are also faulted.
161
162 39. If PCRE was compiled with UTF support, in three places in the DFA
163 matcher there was code that should only have been obeyed in UTF mode, but
164 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
165 processing when bytes with values greater than 127 were present. In 16-bit
166 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
167 both cases the values are those that cannot be the first data item in a UTF
168 character. The three items that might have provoked this were recursions,
169 possessively repeated groups, and atomic groups.
170
171 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
172 and pcregrep, because some OS require shared objects to be explicitly
173 passed to ld, causing the link step to fail if they are not.
174
175 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
176 mode, patterns that started with \h* or \R* might be incorrectly matched.
177
178
179 Version 8.30 04-February-2012
180 -----------------------------
181
182 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
183 name is defined in ctype.h.
184
185 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
186 only in quite long subpatterns.
187
188 3. Removed the function pcre_info(), which has been obsolete and deprecated
189 since it was replaced by pcre_fullinfo() in February 2000.
190
191 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
192 match a (*MARK), and the match failed at the start of the subject, a
193 reference to memory before the start of the subject could occur. This bug
194 was introduced by fix 17 of release 8.21.
195
196 5. A reference to an unset group with zero minimum repetition was giving
197 totally wrong answers (in non-JavaScript-compatibility mode). For example,
198 /(another)?(\1?)test/ matched against "hello world test". This bug was
199 introduced in release 8.13.
200
201 6. Add support for 16-bit character strings (a large amount of work involving
202 many changes and refactorings).
203
204 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
205 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
206 from a file.
207
208 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
209 rounding is not applied in this particular case).
210
211 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
212 if they appear, or are escaped, in patterns.
213
214 10. Get rid of a number of -Wunused-but-set-variable warnings.
215
216 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
217 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
218 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
219 also returns the mark "x". This bug applied to capturing parentheses,
220 non-capturing parentheses, and atomic parentheses. It also applied to some
221 assertions.
222
223 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
224 information out of configure.ac instead of relying on pcre.h.generic, which
225 is not stored in the repository.
226
227 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
228 -lreadline.
229
230 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
231
232 15. Applied Graycode's patch to put the top-level frame on the stack rather
233 than the heap when not using the stack for recursion. This gives a
234 performance improvement in many cases when recursion is not deep.
235
236 16. Experimental code added to "pcretest -C" to output the stack frame size.
237
238
239 Version 8.21 12-Dec-2011
240 ------------------------
241
242 1. Updating the JIT compiler.
243
244 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
245 are added as well.
246
247 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
248 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
249 calling _pcre_jit_exec. Some extra comments are added.
250
251 4. (*MARK) settings inside atomic groups that do not contain any capturing
252 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
253 was introduced by change 18 for 8.20.
254
255 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
256 ECMA-262 standard.
257
258 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
259 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
260 This bug was probably introduced by change 9 of 8.13.
261
262 7. While fixing 6 above, I noticed that a number of other items were being
263 incorrectly rejected as "not fixed length". This arose partly because newer
264 opcodes had not been added to the fixed-length checking code. I have (a)
265 corrected the bug and added tests for these items, and (b) arranged for an
266 error to occur if an unknown opcode is encountered while checking for fixed
267 length instead of just assuming "not fixed length". The items that were
268 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
269 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
270 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
271
272 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
273 being incorrectly compiled and would have given unpredicatble results.
274
275 9. A possessively repeated subpattern with minimum repeat count greater than
276 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
277 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
278 the first (A) could occur when it should not.
279
280 10. Add a cast and remove a redundant test from the code.
281
282 11. JIT should use pcre_malloc/pcre_free for allocation.
283
284 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
285 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
286 is anything other than /usr, -L is still shown).
287
288 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
289
290 14. Perl does not support \N without a following name in a [] class; PCRE now
291 also gives an error.
292
293 15. If a forward reference was repeated with an upper limit of around 2000,
294 it caused the error "internal error: overran compiling workspace". The
295 maximum number of forward references (including repeats) was limited by the
296 internal workspace, and dependent on the LINK_SIZE. The code has been
297 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
298 the default depends on LINK_SIZE. There is a new upper limit (for safety)
299 of around 200,000 forward references. While doing this, I also speeded up
300 the filling in of repeated forward references.
301
302 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
303 incorrectly expecting the subject to contain another "a" after the start.
304
305 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
306 in the match, the SKIP should be ignored. This was not happening; instead
307 the SKIP was being treated as NOMATCH. For patterns such as
308 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
309 tested.
310
311 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
312 now much more compatible with Perl, in particular in cases where the result
313 is a non-match for a non-anchored pattern. For example, if
314 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
315 "m", where previously it did not return a name. A side effect of this
316 change is that for partial matches, the last encountered mark name is
317 returned, as for non matches. A number of tests that were previously not
318 Perl-compatible have been moved into the Perl-compatible test files. The
319 refactoring has had the pleasing side effect of removing one argument from
320 the match() function, thus reducing its stack requirements.
321
322 19. If the /S+ option was used in pcretest to study a pattern using JIT,
323 subsequent uses of /S (without +) incorrectly behaved like /S+.
324
325 21. Retrieve executable code size support for the JIT compiler and fixing
326 some warnings.
327
328 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
329 not work when the shorter character appeared right at the end of the
330 subject string.
331
332 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
333 systems.
334
335 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
336 output it when the /M option is used in pcretest.
337
338 25. The CheckMan script was not being included in the distribution. Also, added
339 an explicit "perl" to run Perl scripts from the PrepareRelease script
340 because this is reportedly needed in Windows.
341
342 26. If study data was being save in a file and studying had not found a set of
343 "starts with" bytes for the pattern, the data written to the file (though
344 never used) was taken from uninitialized memory and so caused valgrind to
345 complain.
346
347 27. Updated RunTest.bat as provided by Sheri Pierce.
348
349 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
350
351 29. Computation of memory usage for the table of capturing group names was
352 giving an unnecessarily large value.
353
354
355 Version 8.20 21-Oct-2011
356 ------------------------
357
358 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
359 a POSIX class. After further experiments with Perl, which convinced me that
360 Perl has bugs and confusions, a closing square bracket is no longer allowed
361 in a POSIX name. This bug also affected patterns with classes that started
362 with full stops.
363
364 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
365 captured substring, but while checking the failing first alternative,
366 substring 1 is temporarily captured. If the output vector supplied to
367 pcre_exec() was not big enough for this capture, the yield of the function
368 was still zero ("insufficient space for captured substrings"). This cannot
369 be totally fixed without adding another stack variable, which seems a lot
370 of expense for a edge case. However, I have improved the situation in cases
371 such as /(a)(b)x|abc/ matched against "abc", where the return code
372 indicates that fewer than the maximum number of slots in the ovector have
373 been set.
374
375 3. Related to (2) above: when there are more back references in a pattern than
376 slots in the output vector, pcre_exec() uses temporary memory during
377 matching, and copies in the captures as far as possible afterwards. It was
378 using the entire output vector, but this conflicts with the specification
379 that only 2/3 is used for passing back captured substrings. Now it uses
380 only the first 2/3, for compatibility. This is, of course, another edge
381 case.
382
383 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
384 main code base, and can be used by building with --enable-jit. When this is
385 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
386 runtime --no-jit option is given.
387
388 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
389 ovector, the return from the function was zero, implying that there were
390 other matches that did not fit. The correct "exactly full" value is now
391 returned.
392
393 6. If a subpattern that was called recursively or as a subroutine contained
394 (*PRUNE) or any other control that caused it to give a non-standard return,
395 invalid errors such as "Error -26 (nested recursion at the same subject
396 position)" or even infinite loops could occur.
397
398 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
399 computing the minimum length on reaching *ACCEPT, and so ended up with the
400 wrong value of 1 rather than 0. Further investigation indicates that
401 computing a minimum subject length in the presence of *ACCEPT is difficult
402 (think back references, subroutine calls), and so I have changed the code
403 so that no minimum is registered for a pattern that contains *ACCEPT.
404
405 8. If (*THEN) was present in the first (true) branch of a conditional group,
406 it was not handled as intended. [But see 16 below.]
407
408 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
409 Sheri Pierce.
410
411 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
412 the first byte in a match must be "a".
413
414 11. Change 17 for 8.13 increased the recursion depth for patterns like
415 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
416 pattern contains any instances of (*THEN). If it does not, the old
417 optimizations are restored. It would be nice to do this on a per-group
418 basis, but at the moment that is not feasible.
419
420 12. In some environments, the output of pcretest -C is CRLF terminated. This
421 broke RunTest's code that checks for the link size. A single white space
422 character after the value is now allowed for.
423
424 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
425 For "fr", it uses the Windows-specific input and output files.
426
427 14. If (*THEN) appeared in a group that was called recursively or as a
428 subroutine, it did not work as intended. [But see next item.]
429
430 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
431 pattern fragments (but not containing any | characters). If A and B are
432 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
433 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
434 D. In other words, Perl considers parentheses that do not contain any |
435 characters to be part of a surrounding alternative, whereas PCRE was
436 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
437 differently. PCRE now behaves in the same way as Perl, except in the case
438 of subroutine/recursion calls such as (?1) which have in any case always
439 been different (but PCRE had them first :-).
440
441 16. Related to 15 above: Perl does not treat the | in a conditional group as
442 creating alternatives. Such a group is treated in the same way as an
443 ordinary group without any | characters when processing (*THEN). PCRE has
444 been changed to match Perl's behaviour.
445
446 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
447 RunGrepTest script failed.
448
449 18. Change 22 for version 13 caused atomic groups to use more stack. This is
450 inevitable for groups that contain captures, but it can lead to a lot of
451 stack use in large patterns. The old behaviour has been restored for atomic
452 groups that do not contain any capturing parentheses.
453
454 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
455 suppress the check for a minimum subject length at run time. (If it was
456 given to pcre_exec() or pcre_dfa_exec() it did work.)
457
458 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
459 fail to work when decoding hex characters in data strings in EBCDIC
460 environments.
461
462 21. It appears that in at least one Mac OS environment, the isxdigit() function
463 is implemented as a macro that evaluates to its argument more than once,
464 contravening the C 90 Standard (I haven't checked a later standard). There
465 was an instance in pcretest which caused it to go wrong when processing
466 \x{...} escapes in subject strings. The has been rewritten to avoid using
467 things like p++ in the argument of isxdigit().
468
469
470 Version 8.13 16-Aug-2011
471 ------------------------
472
473 1. The Unicode data tables have been updated to Unicode 6.0.0.
474
475 2. Two minor typos in pcre_internal.h have been fixed.
476
477 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
478 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
479 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
480
481 4. There were a number of related bugs in the code for matching backrefences
482 caselessly in UTF-8 mode when codes for the characters concerned were
483 different numbers of bytes. For example, U+023A and U+2C65 are an upper
484 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
485 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
486 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
487 2-byte code at the end of the subject (it thought there wasn't enough data
488 left).
489
490 5. Comprehensive information about what went wrong is now returned by
491 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
492 as the output vector has at least 2 elements. The offset of the start of
493 the failing character and a reason code are placed in the vector.
494
495 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
496 now returned is for the first byte of the failing character, instead of the
497 last byte inspected. This is an incompatible change, but I hope it is small
498 enough not to be a problem. It makes the returned offset consistent with
499 pcre_exec() and pcre_dfa_exec().
500
501 7. pcretest now gives a text phrase as well as the error number when
502 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
503 failure, the offset and reason code are output.
504
505 8. When \R was used with a maximizing quantifier it failed to skip backwards
506 over a \r\n pair if the subsequent match failed. Instead, it just skipped
507 back over a single character (\n). This seems wrong (because it treated the
508 two characters as a single entity when going forwards), conflicts with the
509 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
510 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
511 has been changed.
512
513 9. Some internal refactoring has changed the processing so that the handling
514 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
515 time (the PCRE_DOTALL option was changed this way some time ago: version
516 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
517 which was always a bit of a fudge. It also means that there is one less
518 argument for the match() function, which reduces its stack requirements
519 slightly. This change also fixes an incompatibility with Perl: the pattern
520 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
521
522 10. More internal refactoring has drastically reduced the number of recursive
523 calls to match() for possessively repeated groups such as (abc)++ when
524 using pcre_exec().
525
526 11. While implementing 10, a number of bugs in the handling of groups were
527 discovered and fixed:
528
529 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
530 (a|)*(?1) gave a compile-time internal error.
531 ((a|)+)+ did not notice that the outer group could match an empty string.
532 (^a|^)+ was not marked as anchored.
533 (.*a|.*)+ was not marked as matching at start or after a newline.
534
535 12. Yet more internal refactoring has removed another argument from the match()
536 function. Special calls to this function are now indicated by setting a
537 value in a variable in the "match data" data block.
538
539 13. Be more explicit in pcre_study() instead of relying on "default" for
540 opcodes that mean there is no starting character; this means that when new
541 ones are added and accidentally left out of pcre_study(), testing should
542 pick them up.
543
544 14. The -s option of pcretest has been documented for ages as being an old
545 synonym of -m (show memory usage). I have changed it to mean "force study
546 for every regex", that is, assume /S for every regex. This is similar to -i
547 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
548 using it. It makes it easier to run collections of tests with and without
549 study enabled, and thereby test pcre_study() more easily. All the standard
550 tests are now run with and without -s (but some patterns can be marked as
551 "never study" - see 20 below).
552
553 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
554 restoration of the capturing data to the outer values was not happening
555 correctly.
556
557 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
558 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
559 pattern had matched an empty string, and so incorrectly returned a no
560 match.
561
562 17. There was optimizing code for the last branch of non-capturing parentheses,
563 and also for the obeyed branch of a conditional subexpression, which used
564 tail recursion to cut down on stack usage. Unfortunately, now that there is
565 the possibility of (*THEN) occurring in these branches, tail recursion is
566 no longer possible because the return has to be checked for (*THEN). These
567 two optimizations have therefore been removed. [But see 8.20/11 above.]
568
569 18. If a pattern containing \R was studied, it was assumed that \R always
570 matched two bytes, thus causing the minimum subject length to be
571 incorrectly computed because \R can also match just one byte.
572
573 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
574 was incorrectly computed.
575
576 20. If /S is present twice on a test pattern in pcretest input, it now
577 *disables* studying, thereby overriding the use of -s on the command line
578 (see 14 above). This is necessary for one or two tests to keep the output
579 identical in both cases.
580
581 21. When (*ACCEPT) was used in an assertion that matched an empty string and
582 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
583
584 22. When an atomic group that contained a capturing parenthesis was
585 successfully matched, but the branch in which it appeared failed, the
586 capturing was not being forgotten if a higher numbered group was later
587 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
588 group 1 to "a", when in fact it should be unset. This applied to multi-
589 branched capturing and non-capturing groups, repeated or not, and also to
590 positive assertions (capturing in negative assertions does not happen
591 in PCRE) and also to nested atomic groups.
592
593 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
594 subject after a captured substring, to make it easier to tell which of a
595 number of identical substrings has been captured.
596
597 24. The way atomic groups are processed by pcre_exec() has been changed so that
598 if they are repeated, backtracking one repetition now resets captured
599 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
600 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
601 "aaa". Previously, it would have been "a". As part of this code
602 refactoring, the way recursive calls are handled has also been changed.
603
604 25. If an assertion condition captured any substrings, they were not passed
605 back unless some other capturing happened later. For example, if
606 (?(?=(a))a) was matched against "a", no capturing was returned.
607
608 26. When studying a pattern that contained subroutine calls or assertions,
609 the code for finding the minimum length of a possible match was handling
610 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
611 group 1 called group 2 while simultaneously a separate group 2 called group
612 1). A stack overflow occurred in this case. I have fixed this by limiting
613 the recursion depth to 10.
614
615 27. Updated RunTest.bat in the distribution to the version supplied by Tom
616 Fortmann. This supports explicit test numbers on the command line, and has
617 argument validation and error reporting.
618
619 28. An instance of \X with an unlimited repeat could fail if at any point the
620 first character it looked at was a mark character.
621
622 29. Some minor code refactoring concerning Unicode properties and scripts
623 should reduce the stack requirement of match() slightly.
624
625 30. Added the '=' option to pcretest to check the setting of unused capturing
626 slots at the end of the pattern, which are documented as being -1, but are
627 not included in the return count.
628
629 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
630 compiled something random. Now it gives a compile-time error (as does
631 Perl).
632
633 32. A *MARK encountered during the processing of a positive assertion is now
634 recorded and passed back (compatible with Perl).
635
636 33. If --only-matching or --colour was set on a pcregrep call whose pattern
637 had alternative anchored branches, the search for a second match in a line
638 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
639 matched the line "0102" twice. The same bug affected patterns that started
640 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
641 twice.
642
643 34. Previously, PCRE did not allow quantification of assertions. However, Perl
644 does, and because of capturing effects, quantifying parenthesized
645 assertions may at times be useful. Quantifiers are now allowed for
646 parenthesized assertions.
647
648 35. A minor code tidy in pcre_compile() when checking options for \R usage.
649
650 36. \g was being checked for fancy things in a character class, when it should
651 just be a literal "g".
652
653 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
654 appearance of a nested POSIX class supersedes an apparent external class.
655 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
656 unescaped square brackets may also appear as part of class names. For
657 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
658 more like Perl. (But see 8.20/1 above.)
659
660 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
661 was because it thought it was \N{name}, which is not supported).
662
663 39. Add minix to OS list not supporting the -S option in pcretest.
664
665 40. PCRE tries to detect cases of infinite recursion at compile time, but it
666 cannot analyze patterns in sufficient detail to catch mutual recursions
667 such as ((?1))((?2)). There is now a runtime test that gives an error if a
668 subgroup is called recursively as a subpattern for a second time at the
669 same position in the subject string. In previous releases this might have
670 been caught by the recursion limit, or it might have run out of stack.
671
672 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
673 happen only once. PCRE was, however incorrectly giving a compile time error
674 "recursive call could loop indefinitely" because it cannot analyze the
675 pattern in sufficient detail. The compile time test no longer happens when
676 PCRE is compiling a conditional subpattern, but actual runaway loops are
677 now caught at runtime (see 40 above).
678
679 42. It seems that Perl allows any characters other than a closing parenthesis
680 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
681 has been changed to be the same.
682
683 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
684 as not to get warnings when autogen.sh is called. Also changed
685 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
686
687 44. To help people who use pcregrep to scan files containing exceedingly long
688 lines, the following changes have been made:
689
690 (a) The default value of the buffer size parameter has been increased from
691 8K to 20K. (The actual buffer used is three times this size.)
692
693 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
694 PCRE is built.
695
696 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
697 to be set at run time.
698
699 (d) Numerical values in pcregrep options can be followed by K or M, for
700 example --buffer-size=50K.
701
702 (e) If a line being scanned overflows pcregrep's buffer, an error is now
703 given and the return code is set to 2.
704
705 45. Add a pointer to the latest mark to the callout data block.
706
707 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
708 partial match of an empty string instead of no match. This was specific to
709 the use of ".".
710
711 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
712 complete match instead of a partial match. This bug was dependent on both
713 the PCRE_UTF8 and PCRE_DOTALL options being set.
714
715 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
716 starting byte set, because \b was not being ignored.
717
718
719 Version 8.12 15-Jan-2011
720 ------------------------
721
722 1. Fixed some typos in the markup of the man pages, and wrote a script that
723 checks for such things as part of the documentation building process.
724
725 2. On a big-endian 64-bit system, pcregrep did not correctly process the
726 --match-limit and --recursion-limit options (added for 8.11). In
727 particular, this made one of the standard tests fail. (The integer value
728 went into the wrong half of a long int.)
729
730 3. If the --colour option was given to pcregrep with -v (invert match), it
731 did strange things, either producing crazy output, or crashing. It should,
732 of course, ignore a request for colour when reporting lines that do not
733 match.
734
735 4. Another pcregrep bug caused similar problems if --colour was specified with
736 -M (multiline) and the pattern match finished with a line ending.
737
738 5. In pcregrep, when a pattern that ended with a literal newline sequence was
739 matched in multiline mode, the following line was shown as part of the
740 match. This seems wrong, so I have changed it.
741
742 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
743 the check for further matches in the same line (so they could be coloured)
744 to overrun the end of the current line. If another match was found, it was
745 incorrectly shown (and then shown again when found in the next line).
746
747 7. If pcregrep was compiled under Windows, there was a reference to the
748 function pcregrep_exit() before it was defined. I am assuming this was
749 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
750 reported by a user. I've moved the definition above the reference.
751
752
753 Version 8.11 10-Dec-2010
754 ------------------------
755
756 1. (*THEN) was not working properly if there were untried alternatives prior
757 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
758 backtracked to try for "b" instead of moving to the next alternative branch
759 at the same level (in this case, to look for "c"). The Perl documentation
760 is clear that when (*THEN) is backtracked onto, it goes to the "next
761 alternative in the innermost enclosing group".
762
763 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
764 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
765 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
766 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
767 (*THEN).
768
769 3. If \s appeared in a character class, it removed the VT character from
770 the class, even if it had been included by some previous item, for example
771 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
772 of \s, but is part of the POSIX "space" class.)
773
774 4. A partial match never returns an empty string (because you can always
775 match an empty string at the end of the subject); however the checking for
776 an empty string was starting at the "start of match" point. This has been
777 changed to the "earliest inspected character" point, because the returned
778 data for a partial match starts at this character. This means that, for
779 example, /(?<=abc)def/ gives a partial match for the subject "abc"
780 (previously it gave "no match").
781
782 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
783 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
784 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
785 has an implication that the given string is incomplete (because a partial
786 match is preferred over a full match). For this reason, these items now
787 give a partial match in this situation. [Aside: previously, the one case
788 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
789 match rather than a full match, which was wrong by the old rules, but is
790 now correct.]
791
792 6. There was a bug in the handling of #-introduced comments, recognized when
793 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
794 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
795 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
796 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
797 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
798 places in pcre_compile().
799
800 7. Related to (6) above, when pcre_compile() was skipping #-introduced
801 comments when looking ahead for named forward references to subpatterns,
802 the only newline sequence it recognized was NL. It now handles newlines
803 according to the set newline convention.
804
805 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
806 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
807 cater for a lack of strerror(). These oversights have been fixed.
808
809 9. Added --match-limit and --recursion-limit to pcregrep.
810
811 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
812
813 11. When the -o option was used, pcregrep was setting a return code of 1, even
814 when matches were found, and --line-buffered was not being honoured.
815
816 12. Added an optional parentheses number to the -o and --only-matching options
817 of pcregrep.
818
819 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
820 can match an empty string. The code to do it in pcretest and pcredemo
821 needed fixing:
822
823 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
824 only one byte after an empty string match just before CRLF (this case
825 just got forgotten; "any" and "anycrlf" were OK).
826
827 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
828 mode when an empty string match preceded an ASCII character followed by
829 a non-ASCII character. (The code for advancing by one character rather
830 than one byte was nonsense.)
831
832 (c) The pcredemo.c sample program did not have any code at all to handle
833 the cases when CRLF is a valid newline sequence.
834
835 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
836 as a starting offset was within the subject string. There is now a new
837 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
838 negative or greater than the length of the string. In order to test this,
839 pcretest is extended to allow the setting of negative starting offsets.
840
841 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
842 starting offset points to the beginning of a UTF-8 character was
843 unnecessarily clumsy. I tidied it up.
844
845 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
846 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
847
848 17. Nobody had reported that the --include_dir option, which was added in
849 release 7.7 should have been called --include-dir (hyphen, not underscore)
850 for compatibility with GNU grep. I have changed it to --include-dir, but
851 left --include_dir as an undocumented synonym, and the same for
852 --exclude-dir, though that is not available in GNU grep, at least as of
853 release 2.5.4.
854
855 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
856 characters from a string of bytes) have been redefined so as not to use
857 loops, in order to improve performance in some environments. At the same
858 time, I abstracted some of the common code into auxiliary macros to save
859 repetition (this should not affect the compiled code).
860
861 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
862 compile-time error is now given if \c is not followed by an ASCII
863 character, that is, a byte less than 128. (In EBCDIC mode, the code is
864 different, and any byte value is allowed.)
865
866 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
867 START_OPTIMIZE option, which is now allowed at compile time - but just
868 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
869 to pcregrep and other applications that have no direct access to PCRE
870 options. The new /Y option in pcretest sets this option when calling
871 pcre_compile().
872
873 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
874 back references. Groups containing recursive back references were forced to
875 be atomic by that change, but in the case of named groups, the amount of
876 memory required was incorrectly computed, leading to "Failed: internal
877 error: code overflow". This has been fixed.
878
879 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
880 pcretest.c, to avoid build problems in some Borland environments.
881
882
883 Version 8.10 25-Jun-2010
884 ------------------------
885
886 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
887 THEN.
888
889 2. (*ACCEPT) was not working when inside an atomic group.
890
891 3. Inside a character class, \B is treated as a literal by default, but
892 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
893 causes the error). The code is unchanged, but I tidied the documentation.
894
895 4. Inside a character class, PCRE always treated \R and \X as literals,
896 whereas Perl faults them if its -w option is set. I have changed PCRE so
897 that it faults them when PCRE_EXTRA is set.
898
899 5. Added support for \N, which always matches any character other than
900 newline. (It is the same as "." when PCRE_DOTALL is not set.)
901
902 6. When compiling pcregrep with newer versions of gcc which may have
903 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
904 declared with attribute warn_unused_result" were given. Just casting the
905 result to (void) does not stop the warnings; a more elaborate fudge is
906 needed. I've used a macro to implement this.
907
908 7. Minor change to pcretest.c to avoid a compiler warning.
909
910 8. Added four artifical Unicode properties to help with an option to make
911 \s etc use properties (see next item). The new properties are: Xan
912 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
913
914 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
915 use Unicode properties. (*UCP) at the start of a pattern can be used to set
916 this option. Modified pcretest to add /W to test this facility. Added
917 REG_UCP to make it available via the POSIX interface.
918
919 10. Added --line-buffered to pcregrep.
920
921 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
922 studied, and the match started with a letter with a code point greater than
923 127 whose first byte was different to the first byte of the other case of
924 the letter, the other case of this starting letter was not recognized
925 (#976).
926
927 12. If a pattern that was studied started with a repeated Unicode property
928 test, for example, \p{Nd}+, there was the theoretical possibility of
929 setting up an incorrect bitmap of starting bytes, but fortunately it could
930 not have actually happened in practice until change 8 above was made (it
931 added property types that matched character-matching opcodes).
932
933 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
934 possible starting bytes for non-anchored patterns.
935
936 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
937 \R, and also a number of cases that involve Unicode properties, both
938 explicit and implicit when PCRE_UCP is set.
939
940 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
941 input, it could crash or give wrong results if characters with values
942 greater than 0xc0 were present in the subject string. (Detail: it assumed
943 UTF-8 input when processing these items.)
944
945 16. Added a lot of (int) casts to avoid compiler warnings in systems where
946 size_t is 64-bit (#991).
947
948 17. Added a check for running out of memory when PCRE is compiled with
949 --disable-stack-for-recursion (#990).
950
951 18. If the last data line in a file for pcretest does not have a newline on
952 the end, a newline was missing in the output.
953
954 19. The default pcre_chartables.c file recognizes only ASCII characters (values
955 less than 128) in its various bitmaps. However, there is a facility for
956 generating tables according to the current locale when PCRE is compiled. It
957 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
958 characters, are recognized by isspace() and therefore were getting set in
959 these tables, and indeed these tables seem to approximate to ISO 8859. This
960 caused a problem in UTF-8 mode when pcre_study() was used to create a list
961 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
962 which of course cannot start UTF-8 characters. I have changed the code so
963 that only real ASCII characters (less than 128) and the correct starting
964 bytes for UTF-8 encodings are set for characters greater than 127 when in
965 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
966 altogether.)
967
968 20. Added the /T option to pcretest so as to be able to run tests with non-
969 standard character tables, thus making it possible to include the tests
970 used for 19 above in the standard set of tests.
971
972 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
973 reference to a subpattern the other side of a comment that contains an
974 opening parenthesis caused either an internal compiling error, or a
975 reference to the wrong subpattern.
976
977
978 Version 8.02 19-Mar-2010
979 ------------------------
980
981 1. The Unicode data tables have been updated to Unicode 5.2.0.
982
983 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
984 configured.
985
986 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
987 original author of that file, following a query about its status.
988
989 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
990 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
991
992 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
993 quantifier applied to a forward-referencing subroutine call, could compile
994 incorrect code or give the error "internal error: previously-checked
995 referenced subpattern not found".
996
997 6. Both MS Visual Studio and Symbian OS have problems with initializing
998 variables to point to external functions. For these systems, therefore,
999 pcre_malloc etc. are now initialized to local functions that call the
1000 relevant global functions.
1001
1002 7. There were two entries missing in the vectors called coptable and poptable
1003 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1004 I've fixed the data, and added a kludgy way of testing at compile time that
1005 the lengths are correct (equal to the number of opcodes).
1006
1007 8. Following on from 7, I added a similar kludge to check the length of the
1008 eint vector in pcreposix.c.
1009
1010 9. Error texts for pcre_compile() are held as one long string to avoid too
1011 much relocation at load time. To find a text, the string is searched,
1012 counting zeros. There was no check for running off the end of the string,
1013 which could happen if a new error number was added without updating the
1014 string.
1015
1016 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1017
1018 11. \K was not working if it appeared in an atomic group or in a group that
1019 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1020 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1021 the assertion is positive, but not if it is negative.
1022
1023 12. Change 11 fortuitously reduced the size of the stack frame used in the
1024 "match()" function of pcre_exec.c by one pointer. Forthcoming
1025 implementation of support for (*MARK) will need an extra pointer on the
1026 stack; I have reserved it now, so that the stack frame size does not
1027 decrease.
1028
1029 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1030 item in branch that calls a recursion is a subroutine call - as in the
1031 second branch in the above example - was incorrectly given the compile-
1032 time error "recursive call could loop indefinitely" because pcre_compile()
1033 was not correctly checking the subroutine for matching a non-empty string.
1034
1035 14. The checks for overrunning compiling workspace could trigger after an
1036 overrun had occurred. This is a "should never occur" error, but it can be
1037 triggered by pathological patterns such as hundreds of nested parentheses.
1038 The checks now trigger 100 bytes before the end of the workspace.
1039
1040 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1041
1042
1043 Version 8.01 19-Jan-2010
1044 ------------------------
1045
1046 1. If a pattern contained a conditional subpattern with only one branch (in
1047 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1048 computed the wrong minimum data length (which is of course zero for such
1049 subpatterns). This could cause incorrect "no match" results.
1050
1051 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1052 the pattern is reset in the first branch, pcre_compile() failed with
1053 "internal error: code overflow at offset...". This happened only when
1054 the reset was to the original external option setting. (An optimization
1055 abstracts leading options settings into an external setting, which was the
1056 cause of this.)
1057
1058 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1059 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1060 assertion pattern did not match (meaning that the assertion was true), it
1061 was incorrectly treated as false if the SKIP had been reached during the
1062 matching. This also applied to assertions used as conditions.
1063
1064 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1065 assertion subpattern, including such a pattern used as a condition,
1066 unpredictable results occurred, instead of the error return
1067 PCRE_ERROR_DFA_UITEM.
1068
1069 5. The C++ GlobalReplace function was not working like Perl for the special
1070 situation when an empty string is matched. It now does the fancy magic
1071 stuff that is necessary.
1072
1073 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1074 removed. (These were left over from very, very early versions of PCRE.)
1075
1076 7. Some cosmetic changes to the code to make life easier when compiling it
1077 as part of something else:
1078
1079 (a) Change DEBUG to PCRE_DEBUG.
1080
1081 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1082 called "current" as "current_branch", to prevent a collision with the
1083 Linux macro when compiled as a kernel module.
1084
1085 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1086 prevent a collision with the Linux macro when compiled as a kernel
1087 module.
1088
1089 8. In pcre_compile() there are some checks for integer overflows that used to
1090 cast potentially large values to (double). This has been changed to that
1091 when building, a check for int64_t is made, and if it is found, it is used
1092 instead, thus avoiding the use of floating point arithmetic. (There is no
1093 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1094 double.
1095
1096 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1097 2005 (difference between two addresses compared to an unsigned value).
1098
1099 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1100 custom one, because of the following reported problem in Windows:
1101
1102 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1103 under Win32.
1104 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1105 therefore missing the function definition.
1106 - The compiler thus generates a "C" signature for the test function.
1107 - The linker fails to find the "C" function.
1108 - PCRE fails to configure if asked to do so against libbz2.
1109
1110 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1111 messages were output:
1112
1113 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1114 rerunning libtoolize, to keep the correct libtool macros in-tree.
1115 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1116
1117 I have done both of these things.
1118
1119 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1120 most of the time, it *can* run out if it is given a pattern that contains a
1121 runaway infinite recursion. I updated the discussion in the pcrestack man
1122 page.
1123
1124 13. Now that we have gone to the x.xx style of version numbers, the minor
1125 version may start with zero. Using 08 or 09 is a bad idea because users
1126 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1127 interpreted as invalid octal numbers. I've updated the previous comment in
1128 configure.ac, and also added a check that gives an error if 08 or 09 are
1129 used.
1130
1131 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1132 causing partial matching to fail when the end of the subject matched \W
1133 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1134
1135 15. There were some discrepancies between the declarations in pcre_internal.h
1136 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1137 their definitions. The declarations used "const uschar *" and the
1138 definitions used USPTR. Even though USPTR is normally defined as "const
1139 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1140 reported that: "This difference in casting confuses some C++ compilers, for
1141 example, SunCC recognizes above declarations as different functions and
1142 generates broken code for hbpcre." I have changed the declarations to use
1143 USPTR.
1144
1145 16. GNU libtool is named differently on some systems. The autogen.sh script now
1146 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1147 (FreeBSD).
1148
1149 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1150 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1151 comment: "Figure out how to create a longlong from a string: strtoll and
1152 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1153 instance, but it only takes 2 args instead of 3!"
1154
1155 18. A subtle bug concerned with back references has been fixed by a change of
1156 specification, with a corresponding code fix. A pattern such as
1157 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1158 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1159 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1160 same bug. Such groups have to be quantified to be useful, or contained
1161 inside another quantified group. (If there's no repetition, the reference
1162 can never match.) The problem arises because, having left the group and
1163 moved on to the rest of the pattern, a later failure that backtracks into
1164 the group uses the captured value from the final iteration of the group
1165 rather than the correct earlier one. I have fixed this in PCRE by forcing
1166 any group that contains a reference to itself to be an atomic group; that
1167 is, there cannot be any backtracking into it once it has completed. This is
1168 similar to recursive and subroutine calls.
1169
1170
1171 Version 8.00 19-Oct-09
1172 ----------------------
1173
1174 1. The table for translating pcre_compile() error codes into POSIX error codes
1175 was out-of-date, and there was no check on the pcre_compile() error code
1176 being within the table. This could lead to an OK return being given in
1177 error.
1178
1179 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1180 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1181 in a Windows environment.
1182
1183 3. The pcregrep --count option prints the count for each file even when it is
1184 zero, as does GNU grep. However, pcregrep was also printing all files when
1185 --files-with-matches was added. Now, when both options are given, it prints
1186 counts only for those files that have at least one match. (GNU grep just
1187 prints the file name in this circumstance, but including the count seems
1188 more useful - otherwise, why use --count?) Also ensured that the
1189 combination -clh just lists non-zero counts, with no names.
1190
1191 4. The long form of the pcregrep -F option was incorrectly implemented as
1192 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1193 but it seems right to fix it, and I didn't think it was worth preserving
1194 the old behaviour.
1195
1196 5. The command line items --regex=pattern and --regexp=pattern were not
1197 recognized by pcregrep, which required --regex pattern or --regexp pattern
1198 (with a space rather than an '='). The man page documented the '=' forms,
1199 which are compatible with GNU grep; these now work.
1200
1201 6. No libpcreposix.pc file was created for pkg-config; there was just
1202 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1203
1204 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1205 when UCP support is not needed, by modifying the Python script that
1206 generates it from Unicode data files. This should not matter if the module
1207 is correctly used as a library, but I received one complaint about 50K of
1208 unwanted data. My guess is that the person linked everything into his
1209 program rather than using a library. Anyway, it does no harm.
1210
1211 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1212 was a minimum greater than 1 for a wide character in a possessive
1213 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1214 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1215 character. Chaos in the form of incorrect output or a compiling loop could
1216 result.
1217
1218 9. The restrictions on what a pattern can contain when partial matching is
1219 requested for pcre_exec() have been removed. All patterns can now be
1220 partially matched by this function. In addition, if there are at least two
1221 slots in the offset vector, the offset of the earliest inspected character
1222 for the match and the offset of the end of the subject are set in them when
1223 PCRE_ERROR_PARTIAL is returned.
1224
1225 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1226 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1227 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1228 and may be more useful for multi-segment matching.
1229
1230 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1231 used to be given if ever the end of the subject was reached; now it is
1232 given only if matching could not proceed because another character was
1233 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1234 string "Z", which now yields "no match" instead of "partial match". In the
1235 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1236 final character ended with (*FAIL).
1237
1238 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1239 if the pattern had a "must contain" character that was already found in the
1240 earlier partial match, unless partial matching was again requested. For
1241 example, with the pattern /dog.(body)?/, the "must contain" character is
1242 "g". If the first part-match was for the string "dog", restarting with
1243 "sbody" failed. This bug has been fixed.
1244
1245 13. The string returned by pcre_dfa_exec() after a partial match has been
1246 changed so that it starts at the first inspected character rather than the
1247 first character of the match. This makes a difference only if the pattern
1248 starts with a lookbehind assertion or \b or \B (\K is not supported by
1249 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1250 matching functions compatible, and I think it's the right thing to do.
1251
1252 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1253 so that the demonstration program is easily available in environments where
1254 PCRE has not been installed from source.
1255
1256 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1257 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1258 library.
1259
1260 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1261 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1262 is not the first non-POSIX option to be added. Clearly some people find
1263 these options useful.
1264
1265 17. If a caller to the POSIX matching function regexec() passes a non-zero
1266 value for nmatch with a NULL value for pmatch, the value of
1267 nmatch is forced to zero.
1268
1269 18. RunGrepTest did not have a test for the availability of the -u option of
1270 the diff command, as RunTest does. It now checks in the same way as
1271 RunTest, and also checks for the -b option.
1272
1273 19. If an odd number of negated classes containing just a single character
1274 interposed, within parentheses, between a forward reference to a named
1275 subpattern and the definition of the subpattern, compilation crashed with
1276 an internal error, complaining that it could not find the referenced
1277 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1278 [The bug was that it was starting one character too far in when skipping
1279 over the character class, thus treating the ] as data rather than
1280 terminating the class. This meant it could skip too much.]
1281
1282 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1283 /g option in pcretest when the pattern contains \K, which makes it possible
1284 to have an empty string match not at the start, even when the pattern is
1285 anchored. Updated pcretest and pcredemo to use this option.
1286
1287 21. If the maximum number of capturing subpatterns in a recursion was greater
1288 than the maximum at the outer level, the higher number was returned, but
1289 with unset values at the outer level. The correct (outer level) value is
1290 now given.
1291
1292 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1293 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1294 make it do so. The string so far is captured, making this feature
1295 compatible with Perl.
1296
1297 23. The tests have been re-organized, adding tests 11 and 12, to make it
1298 possible to check the Perl 5.10 features against Perl 5.10.
1299
1300 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1301 pattern matches a fixed length string. PCRE did not allow this; now it
1302 does. Neither allows recursion.
1303
1304 25. I finally figured out how to implement a request to provide the minimum
1305 length of subject string that was needed in order to match a given pattern.
1306 (It was back references and recursion that I had previously got hung up
1307 on.) This code has now been added to pcre_study(); it finds a lower bound
1308 to the length of subject needed. It is not necessarily the greatest lower
1309 bound, but using it to avoid searching strings that are too short does give
1310 some useful speed-ups. The value is available to calling programs via
1311 pcre_fullinfo().
1312
1313 26. While implementing 25, I discovered to my embarrassment that pcretest had
1314 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1315 study optimizations had never been tested with that matching function.
1316 Oops. What is worse, even when it was passed study data, there was a bug in
1317 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1318 were also very few tests of studied patterns with pcre_dfa_exec().
1319
1320 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1321 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1322 on the other side of the coin, they are no longer allowed to have different
1323 names, because these cannot be distinguished in PCRE, and this has caused
1324 confusion. (This is a difference from Perl.)
1325
1326 28. When duplicate subpattern names are present (necessarily with different
1327 numbers, as required by 27 above), and a test is made by name in a
1328 conditional pattern, either for a subpattern having been matched, or for
1329 recursion in such a pattern, all the associated numbered subpatterns are
1330 tested, and the overall condition is true if the condition is true for any
1331 one of them. This is the way Perl works, and is also more like the way
1332 testing by number works.
1333
1334
1335 Version 7.9 11-Apr-09
1336 ---------------------
1337
1338 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1339 (pcretest), all targets were linked against these libraries. This included
1340 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1341 libraries. This caused unwanted dependencies to be created. This problem
1342 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1343 pcretest is linked with readline.
1344
1345 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1346 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1347 moved outside it again, because FALSE and TRUE are already defined in AIX,
1348 but BOOL is not.
1349
1350 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1351 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1352
1353 4. The pcregrep documentation said spaces were inserted as well as colons (or
1354 hyphens) following file names and line numbers when outputting matching
1355 lines. This is not true; no spaces are inserted. I have also clarified the
1356 wording for the --colour (or --color) option.
1357
1358 5. In pcregrep, when --colour was used with -o, the list of matching strings
1359 was not coloured; this is different to GNU grep, so I have changed it to be
1360 the same.
1361
1362 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1363 each matching line was coloured. Now it goes on to look for further matches
1364 of any of the test patterns, which is the same behaviour as GNU grep.
1365
1366 7. A pattern that could match an empty string could cause pcregrep to loop; it
1367 doesn't make sense to accept an empty string match in pcregrep, so I have
1368 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1369 seems to be how GNU grep behaves.
1370
1371 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1372 start or after a newline", because the conditional assertion was not being
1373 correctly handled. The rule now is that both the assertion and what follows
1374 in the first alternative must satisfy the test.
1375
1376 9. If auto-callout was enabled in a pattern with a conditional group whose
1377 condition was an assertion, PCRE could crash during matching, both with
1378 pcre_exec() and pcre_dfa_exec().
1379
1380 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1381 used for matching.
1382
1383 11. Unicode property support in character classes was not working for
1384 characters (bytes) greater than 127 when not in UTF-8 mode.
1385
1386 12. Added the -M command line option to pcretest.
1387
1388 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1389
1390 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1391
1392 16. Added comments and documentation about mis-use of no_arg in the C++
1393 wrapper.
1394
1395 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1396 from Martin Jerabek that uses macro names for all relevant character and
1397 string constants.
1398
1399 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1400 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1401 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1402 these, but not everybody uses configure.
1403
1404 19. A conditional group that had only one branch was not being correctly
1405 recognized as an item that could match an empty string. This meant that an
1406 enclosing group might also not be so recognized, causing infinite looping
1407 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1408 with the subject "ab", where knowledge that the repeated group can match
1409 nothing is needed in order to break the loop.
1410
1411 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1412 exec(), but without supplying a callout function, matching went wrong.
1413
1414 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1415 leak if the size of the offset vector was greater than 30. When the vector
1416 is smaller, the saved offsets during recursion go onto a local stack
1417 vector, but for larger vectors malloc() is used. It was failing to free
1418 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1419 error, in fact).
1420
1421 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1422 heapframe that is used only when UTF-8 support is enabled. This caused no
1423 problem, but was untidy.
1424
1425 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1426 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1427 included within another project.
1428
1429 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1430 slightly modified by me:
1431
1432 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1433 not building pcregrep.
1434
1435 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1436 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1437
1438 25. Forward references, both numeric and by name, in patterns that made use of
1439 duplicate group numbers, could behave incorrectly or give incorrect errors,
1440 because when scanning forward to find the reference group, PCRE was not
1441 taking into account the duplicate group numbers. A pattern such as
1442 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1443
1444 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1445 the feature of a custom pointer more persuasive (as requested by a user).
1446
1447 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1448 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1449 pre-defined.
1450
1451 28. Added support for (*UTF8) at the start of a pattern.
1452
1453 29. Arrange for flags added by the "release type" setting in CMake to be shown
1454 in the configuration summary.
1455
1456
1457 Version 7.8 05-Sep-08
1458 ---------------------
1459
1460 1. Replaced UCP searching code with optimized version as implemented for Ad
1461 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1462 stage table and inline lookup instead of a function, giving speed ups of 2
1463 to 5 times on some simple patterns that I tested. Permission was given to
1464 distribute the MultiStage2.py script that generates the tables (it's not in
1465 the tarball, but is in the Subversion repository).
1466
1467 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1468 scripts.
1469
1470 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1471 a group with a zero qualifier. The result of the study could be incorrect,
1472 or the function might crash, depending on the pattern.
1473
1474 4. Caseless matching was not working for non-ASCII characters in back
1475 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1476 It now works when Unicode Property Support is available.
1477
1478 5. In pcretest, an escape such as \x{de} in the data was always generating
1479 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1480 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1481 truncation.
1482
1483 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1484
1485 7. Added two (int) casts to pcregrep when printing the difference of two
1486 pointers, in case they are 64-bit values.
1487
1488 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1489 test 2 if it fails.
1490
1491 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1492 and a #define of that name to empty if it is not externally set. This is to
1493 allow users of MSVC to set it if necessary.
1494
1495 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1496 the convenience functions in the pcre_get.c source file.
1497
1498 11. An option change at the start of a pattern that had top-level alternatives
1499 could cause overwriting and/or a crash. This command provoked a crash in
1500 some environments:
1501
1502 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1503
1504 This potential security problem was recorded as CVE-2008-2371.
1505
1506 12. For a pattern where the match had to start at the beginning or immediately
1507 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1508 pcre_dfa_exec() could read past the end of the passed subject if there was
1509 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1510 pcretest so that it places the subject at the end of its malloc-ed buffer.
1511
1512 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1513 exec() might read past the end of the data buffer in UTF-8 mode.
1514
1515 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1516 the data contained the byte 0x85 as part of a UTF-8 character within its
1517 first line. This applied both to normal and DFA matching.
1518
1519 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1520 /^[^d]*?$/8 failed to match "abc".
1521
1522 16. Added a missing copyright notice to pcrecpp_internal.h.
1523
1524 17. Make it more clear in the documentation that values returned from
1525 pcre_exec() in ovector are byte offsets, not character counts.
1526
1527 18. Tidied a few places to stop certain compilers from issuing warnings.
1528
1529 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1530 supplied by Stefan Weber. I made a further small update for 7.8 because
1531 there is a change of source arrangements: the pcre_searchfuncs.c module is
1532 replaced by pcre_ucd.c.
1533
1534
1535 Version 7.7 07-May-08
1536 ---------------------
1537
1538 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1539 a string to a long long, pretend we don't even have a long long." This is
1540 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1541
1542 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1543 pre-7.6 versions, which defined a global no_arg variable instead of putting
1544 it in the RE class. (See also #8 below.)
1545
1546 3. Remove a line of dead code, identified by coverity and reported by Nuno
1547 Lopes.
1548
1549 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1550
1551 (1) The include/exclude patterns were being applied to the whole pathnames
1552 of files, instead of just to the final components.
1553
1554 (2) If there was more than one level of directory, the subdirectories were
1555 skipped unless they satisfied the include/exclude conditions. This is
1556 inconsistent with GNU grep (and could even be seen as contrary to the
1557 pcregrep specification - which I improved to make it absolutely clear).
1558 The action now is always to scan all levels of directory, and just
1559 apply the include/exclude patterns to regular files.
1560
1561 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1562 --exclude_dir in the tests to avoid scanning .svn directories.
1563
1564 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1565 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1566 doesn't support NULs in patterns.
1567
1568 7. Added some missing "const"s to declarations of static tables in
1569 pcre_compile.c and pcre_dfa_exec.c.
1570
1571 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1572 caused by fix #2 above. (Subsequently also a second patch to fix the
1573 first patch. And a third patch - this was a messy problem.)
1574
1575 9. Applied Craig's patch to remove the use of push_back().
1576
1577 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1578 matching function regexec().
1579
1580 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1581 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1582 references. PCRE supports relative numbers with this syntax (I don't think
1583 Oniguruma does).
1584
1585 12. Previously, a group with a zero repeat such as (...){0} was completely
1586 omitted from the compiled regex. However, this means that if the group
1587 was called as a subroutine from elsewhere in the pattern, things went wrong
1588 (an internal error was given). Such groups are now left in the compiled
1589 pattern, with a new opcode that causes them to be skipped at execution
1590 time.
1591
1592 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1593 to the way PCRE behaves:
1594
1595 (a) A lone ] character is dis-allowed (Perl treats it as data).
1596
1597 (b) A back reference to an unmatched subpattern matches an empty string
1598 (Perl fails the current match path).
1599
1600 (c) A data ] in a character class must be notated as \] because if the
1601 first data character in a class is ], it defines an empty class. (In
1602 Perl it is not possible to have an empty class.) The empty class []
1603 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1604 The negative empty class [^] matches any one character, independently
1605 of the DOTALL setting.
1606
1607 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1608 non-existent subpattern following a character class starting with ']' and
1609 containing () gave an internal compiling error instead of "reference to
1610 non-existent subpattern". Fortunately, when the pattern did exist, the
1611 compiled code was correct. (When scanning forwards to check for the
1612 existencd of the subpattern, it was treating the data ']' as terminating
1613 the class, so got the count wrong. When actually compiling, the reference
1614 was subsequently set up correctly.)
1615
1616 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1617 it was being rejected as not supported by pcre_dfa_exec(), even though
1618 other assertions are supported. I have made pcre_dfa_exec() support
1619 (*FAIL).
1620
1621 16. The implementation of 13c above involved the invention of a new opcode,
1622 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1623 cannot be changed at match time, I realized I could make a small
1624 improvement to matching performance by compiling OP_ALLANY instead of
1625 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1626 on the OP_ANY path.
1627
1628 17. Compiling pcretest on Windows with readline support failed without the
1629 following two fixes: (1) Make the unistd.h include conditional on
1630 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1631
1632 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1633 ncurses library to be included for pcretest when ReadLine support is
1634 requested, but also to allow for it to be overridden. This patch came from
1635 Daniel Bergström.
1636
1637 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1638 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1639 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1640 spotting this.
1641
1642
1643 Version 7.6 28-Jan-08
1644 ---------------------
1645
1646 1. A character class containing a very large number of characters with
1647 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1648 overflow.
1649
1650 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1651 HAVE_LONG_LONG is not defined.
1652
1653 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1654 bring them up to date and include new features. This patch includes:
1655
1656 - Fixed PH's badly added libz and libbz2 support.
1657 - Fixed a problem with static linking.
1658 - Added pcredemo. [But later removed - see 7 below.]
1659 - Fixed dftables problem and added an option.
1660 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1661 HAVE_LONG_LONG.
1662 - Added readline support for pcretest.
1663 - Added an listing of the option settings after cmake has run.
1664
1665 4. A user submitted a patch to Makefile that makes it easy to create
1666 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1667 Makefile.am that cause it to include this special target, without
1668 affecting anything else. Note that the same mingw target plus all
1669 the other distribution libraries and programs are now supported
1670 when configuring with CMake (see 6 below) instead of with
1671 Configure/Make.
1672
1673 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1674 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1675 exported in the Windows port". It has not yet been confirmed that the patch
1676 solves the problem, but it does no harm.
1677
1678 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1679 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1680 with CMake, and also correct the comment about stack recursion.
1681
1682 7. Remove the automatic building of pcredemo from the ./configure system and
1683 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1684 of a program that users should build themselves after PCRE is installed, so
1685 building it automatically is not really right. What is more, it gave
1686 trouble in some build environments.
1687
1688 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1689
1690
1691 Version 7.5 10-Jan-08
1692 ---------------------
1693
1694 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1695 values in parens when parsing an RE using the C++ wrapper."
1696
1697 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1698 Characters greater than 255 were excluded from the class instead of being
1699 included.
1700
1701 3. The same bug as (2) above applied to negated POSIX classes such as
1702 [:^space:].
1703
1704 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1705 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1706 I have changed it.
1707
1708 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1709 first named subpattern) and a construct such as (?&a) would reference the
1710 first named subpattern whose name started with "a" (in other words, the
1711 length check was missing). Both these problems are fixed. "Subpattern name
1712 expected" is now given for (?&) (a zero-length name), and this patch also
1713 makes it give the same error for \k'' (previously it complained that that
1714 was a reference to a non-existent subpattern).
1715
1716 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1717 this is right because (?- can be followed by option settings as well as by
1718 digits. I have, however, made the messages clearer.
1719
1720 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1721 than the number used in the conditional) now cause a compile-time error.
1722 This is actually not compatible with Perl, which accepts such patterns, but
1723 treats the conditional as always being FALSE (as PCRE used to), but it
1724 seems to me that giving a diagnostic is better.
1725
1726 8. Change "alphameric" to the more common word "alphanumeric" in comments
1727 and messages.
1728
1729 9. Fix two occurrences of "backslash" in comments that should have been
1730 "backspace".
1731
1732 10. Remove two redundant lines of code that can never be obeyed (their function
1733 was moved elsewhere).
1734
1735 11. The program that makes PCRE's Unicode character property table had a bug
1736 which caused it to generate incorrect table entries for sequences of
1737 characters that have the same character type, but are in different scripts.
1738 It amalgamated them into a single range, with the script of the first of
1739 them. In other words, some characters were in the wrong script. There were
1740 thirteen such cases, affecting characters in the following ranges:
1741
1742 U+002b0 - U+002c1
1743 U+0060c - U+0060d
1744 U+0061e - U+00612
1745 U+0064b - U+0065e
1746 U+0074d - U+0076d
1747 U+01800 - U+01805
1748 U+01d00 - U+01d77
1749 U+01d9b - U+01dbf
1750 U+0200b - U+0200f
1751 U+030fc - U+030fe
1752 U+03260 - U+0327f
1753 U+0fb46 - U+0fbb1
1754 U+10450 - U+1049d
1755
1756 12. The -o option (show only the matching part of a line) for pcregrep was not
1757 compatible with GNU grep in that, if there was more than one match in a
1758 line, it showed only the first of them. It now behaves in the same way as
1759 GNU grep.
1760
1761 13. If the -o and -v options were combined for pcregrep, it printed a blank
1762 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1763 does the same. The return code can be used to tell if there were any
1764 non-matching lines.
1765
1766 14. Added --file-offsets and --line-offsets to pcregrep.
1767
1768 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1769 infinitely looping recursion. The bug was that positive lookaheads were not
1770 being skipped when checking for a possible empty match (negative lookaheads
1771 and both kinds of lookbehind were skipped).
1772
1773 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1774 inclusion of <windows.h> to before rather than after the definition of
1775 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1776
1777 17. Specifying a possessive quantifier with a specific limit for a Unicode
1778 character property caused pcre_compile() to compile bad code, which led at
1779 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1780 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1781 caused the error; without that there was no problem.
1782
1783 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1784
1785 19. Added --enable-pcretest-libreadline.
1786
1787 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1788 RE::GlobalReplace(). As a result, the number of replacements returned was
1789 double what it should be. I removed one of the increments, but Craig sent a
1790 later patch that removed the other one (the right fix) and added unit tests
1791 that check the return values (which was not done before).
1792
1793 21. Several CMake things:
1794
1795 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1796 the names libpcre and libpcreposix, not just pcre and pcreposix.
1797
1798 (2) The above change means that pcretest and pcregrep are now correctly
1799 linked with the newly-built libraries, not previously installed ones.
1800
1801 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1802
1803 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1804 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1805 UTF-8 newline character). The key issue is that the pattern starts .*;
1806 this means that the match must be either at the beginning, or after a
1807 newline. The bug was in the code for advancing after a failed match and
1808 checking that the new position followed a newline. It was not taking
1809 account of UTF-8 characters correctly.
1810
1811 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1812 character classes. PCRE was not treating the sequence [:...:] as a
1813 character class unless the ... were all letters. Perl, however, seems to
1814 allow any characters between [: and :], though of course it rejects as
1815 unknown any "names" that contain non-letters, because all the known class
1816 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1817 for example, whereas PCRE did not - it did not recognize a POSIX character
1818 class. This seemed a bit dangerous, so the code has been changed to be
1819 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1820 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1821 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1822 Perl does, and where it didn't before.
1823
1824 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1825 Windows environments %n is disabled by default.
1826
1827
1828 Version 7.4 21-Sep-07
1829 ---------------------
1830
1831 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1832 means that a class such as [\s] counted as "explicit reference to CR or
1833 LF". That isn't really right - the whole point of the change was to try to
1834 help when there was an actual mention of one of the two characters. So now
1835 the change happens only if \r or \n (or a literal CR or LF) character is
1836 encountered.
1837
1838 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1839 of both had grown to the point where there were only 3 bits left.
1840 Fortunately, there was spare space in the data structure, and so I have
1841 moved the internal flags into a new 16-bit field to free up more option
1842 bits.
1843
1844 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1845 but did not set the internal JCHANGED flag - either of these is enough to
1846 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1847 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1848 start sets both bits.
1849
1850 4. Added options (at build time, compile time, exec time) to change \R from
1851 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1852
1853 5. doc/pcresyntax.html was missing from the distribution.
1854
1855 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1856 compatibility, even though it is no longer used.
1857
1858 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1859 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1860 windows.h file is present (where different names are used). [This was
1861 reversed later after testing - see 16 below.]
1862
1863 8. Changed all #include <config.h> to #include "config.h". There were also
1864 some further <pcre.h> cases that I changed to "pcre.h".
1865
1866 9. When pcregrep was used with the --colour option, it missed the line ending
1867 sequence off the lines that it output.
1868
1869 10. It was pointed out to me that arrays of string pointers cause lots of
1870 relocations when a shared library is dynamically loaded. A technique of
1871 using a single long string with a table of offsets can drastically reduce
1872 these. I have refactored PCRE in four places to do this. The result is
1873 dramatic:
1874
1875 Originally: 290
1876 After changing UCP table: 187
1877 After changing error message table: 43
1878 After changing table of "verbs" 36
1879 After changing table of Posix names 22
1880
1881 Thanks to the folks working on Gregex for glib for this insight.
1882
1883 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1884 unicode-properties was also set.
1885
1886 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1887
1888 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1889 checked only for CRLF.
1890
1891 14. Added casts to pcretest.c to avoid compiler warnings.
1892
1893 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1894
1895 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1896 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1897 entirely. This removes changes made in 7 above.
1898
1899 17. The CMake files have been updated, and there is now more information about
1900 building with CMake in the NON-UNIX-USE document.
1901
1902
1903 Version 7.3 28-Aug-07
1904 ---------------------
1905
1906 1. In the rejigging of the build system that eventually resulted in 7.1, the
1907 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1908 brackets there is not right, since it causes compilers to look for an
1909 installed pcre.h, not the version that is in the source that is being
1910 compiled (which of course may be different). I have changed it back to:
1911
1912 #include "pcre.h"
1913
1914 I have a vague recollection that the change was concerned with compiling in
1915 different directories, but in the new build system, that is taken care of
1916 by the VPATH setting the Makefile.
1917
1918 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1919 when the subject happened to end in the byte 0x85 (e.g. if the last
1920 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1921 characters but of course it shouldn't be taken as a newline when it is part
1922 of another character. The bug was that, for an unlimited repeat of . in
1923 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1924 characters when looking for a newline.
1925
1926 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1927
1928 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1929 in debug output.
1930
1931 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1932 long printing in the pcrecpp unittest when running under MinGW.
1933
1934 6. ESC_K was left out of the EBCDIC table.
1935
1936 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1937 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1938 limit also applies to "virtual nesting" when a pattern is recursive, and in
1939 this case 1000 isn't so big. I have been able to remove this limit at the
1940 expense of backing off one optimization in certain circumstances. Normally,
1941 when pcre_exec() would call its internal match() function recursively and
1942 immediately return the result unconditionally, it uses a "tail recursion"
1943 feature to save stack. However, when a subpattern that can match an empty
1944 string has an unlimited repetition quantifier, it no longer makes this
1945 optimization. That gives it a stack frame in which to save the data for
1946 checking that an empty string has been matched. Previously this was taken
1947 from the 1000-entry workspace that had been reserved. So now there is no
1948 explicit limit, but more stack is used.
1949
1950 8. Applied Daniel's patches to solve problems with the import/export magic
1951 syntax that is required for Windows, and which was going wrong for the
1952 pcreposix and pcrecpp parts of the library. These were overlooked when this
1953 problem was solved for the main library.
1954
1955 9. There were some crude static tests to avoid integer overflow when computing
1956 the size of patterns that contain repeated groups with explicit upper
1957 limits. As the maximum quantifier is 65535, the maximum group length was
1958 set at 30,000 so that the product of these two numbers did not overflow a
1959 32-bit integer. However, it turns out that people want to use groups that
1960 are longer than 30,000 bytes (though not repeat them that many times).
1961 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1962 made it possible to implement the integer overflow checks in a much more
1963 dynamic way, which I have now done. The artificial limitation on group
1964 length has been removed - we now have only the limit on the total length of
1965 the compiled pattern, which depends on the LINK_SIZE setting.
1966
1967 10. Fixed a bug in the documentation for get/copy named substring when
1968 duplicate names are permitted. If none of the named substrings are set, the
1969 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1970 empty string.
1971
1972 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1973 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1974 because the ] is interpreted as the first data character and the
1975 terminating ] is not found. PCRE has been made compatible with Perl in this
1976 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1977 cause memory overwriting.
1978
1979 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1980 string has been matched (to stop an infinite loop). It was not recognizing
1981 a conditional subpattern that could match an empty string if that
1982 subpattern was within another subpattern. For example, it looped when
1983 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1984 condition was not nested. This bug has been fixed.
1985
1986 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1987 past the start of the subject in the presence of bytes with the top bit
1988 set, for example "\x8aBCD".
1989
1990 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1991 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1992
1993 14. Optimized (?!) to (*FAIL).
1994
1995 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1996 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1997 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1998 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1999 does: it's just the validity check that is more restrictive.
2000
2001 16. Inserted checks for integer overflows during escape sequence (backslash)
2002 processing, and also fixed erroneous offset values for syntax errors during
2003 backslash processing.
2004
2005 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2006 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2007
2008 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2009 caused an overrun.
2010
2011 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2012 something other than just ASCII characters) inside a group that had an
2013 unlimited repeat caused a loop at compile time (while checking to see
2014 whether the group could match an empty string).
2015
2016 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2017 [\P{Any}] did so. (Error in the code for printing property names.)
2018
2019 21. An orphan \E inside a character class could cause a crash.
2020
2021 22. A repeated capturing bracket such as (A)? could cause a wild memory
2022 reference during compilation.
2023
2024 23. There are several functions in pcre_compile() that scan along a compiled
2025 expression for various reasons (e.g. to see if it's fixed length for look
2026 behind). There were bugs in these functions when a repeated \p or \P was
2027 present in the pattern. These operators have additional parameters compared
2028 with \d, etc, and these were not being taken into account when moving along
2029 the compiled data. Specifically:
2030
2031 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2032 length.
2033
2034 (b) An item such as \pL+ within a repeated group could cause crashes or
2035 loops.
2036
2037 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2038 "reference to non-existent subpattern" error.
2039
2040 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2041
2042 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2043 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2044
2045 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2046
2047 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2048 character were causing crashes (broken optimization).
2049
2050 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2051 \p or \P) caused a compile-time loop.
2052
2053 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2054 break. For example, the unstudied pattern [\r\n]A does not match the string
2055 "\r\nA" because change 7.0/46 below moves the current point on by two
2056 characters after failing to match at the start. However, the pattern \nA
2057 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2058 the same is true. There doesn't seem any very clean way out of this, but
2059 what I have chosen to do makes the common cases work: PCRE now takes note
2060 of whether there can be an explicit match for \r or \n anywhere in the
2061 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2062 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2063 pattern has explicit CR or LF references.
2064
2065 29. Added (*CR) etc for changing newline setting at start of pattern.
2066
2067
2068 Version 7.2 19-Jun-07
2069 ---------------------
2070
2071 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2072 which is apparently normally available under Windows.
2073
2074 2. Re-jig the pcregrep tests with different newline settings in an attempt
2075 to make them independent of the local environment's newline setting.
2076
2077 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2078
2079 4. Some of the "internals" tests were previously cut out when the link size
2080 was not 2, because the output contained actual offsets. The recent new
2081 "Z" feature of pcretest means that these can be cut out, making the tests
2082 usable with all link sizes.
2083
2084 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2085 stack recursion. This gives a massive performance boost under BSD, but just
2086 a small improvement under Linux. However, it saves one field in the frame
2087 in all cases.
2088
2089 6. Added more features from the forthcoming Perl 5.10:
2090
2091 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2092 recursion call. It refers to the nth most recently opened parentheses.
2093
2094 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2095 to be opened parentheses.
2096
2097 (c) Conditions that refer to capturing parentheses can be specified
2098 relatively, for example, (?(-2)... or (?(+3)...
2099
2100 (d) \K resets the start of the current match so that everything before
2101 is not part of it.
2102
2103 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2104
2105 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2106 reference syntax.
2107
2108 (g) (?| introduces a group in which the numbering of parentheses in each
2109 alternative starts with the same number.
2110
2111 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2112
2113 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2114 PCRE_INFO_JCHANGED.
2115
2116 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2117 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2118 for detecting groups that can match an empty string.
2119
2120 9. A pattern with a very large number of alternatives (more than several
2121 hundred) was running out of internal workspace during the pre-compile
2122 phase, where pcre_compile() figures out how much memory will be needed. A
2123 bit of new cunning has reduced the workspace needed for groups with
2124 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2125 workspace instead of running out of the 4096 that are available.
2126
2127 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2128
2129 11. Applied patch from Google to remove an optimization that didn't quite work.
2130 The report of the bug said:
2131
2132 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2133 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2134 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2135
2136 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2137 it matched the wrong number of bytes.
2138
2139
2140 Version 7.1 24-Apr-07
2141 ---------------------
2142
2143 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2144 that is more "standard", making use of automake and other Autotools. There
2145 is some re-arrangement of the files and adjustment of comments consequent
2146 on this.
2147
2148 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2149 for recursive directory scanning broke on some systems because the files
2150 are not scanned in any specific order and on different systems the order
2151 was different. A call to "sort" has been inserted into RunGrepTest for the
2152 approprate test as a short-term fix. In the longer term there may be an
2153 alternative.
2154
2155 3. I had an email from Eric Raymond about problems translating some of PCRE's
2156 man pages to HTML (despite the fact that I distribute HTML pages, some
2157 people do their own conversions for various reasons). The problems
2158 concerned the use of low-level troff macros .br and .in. I have therefore
2159 removed all such uses from the man pages (some were redundant, some could
2160 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2161 HTML has been updated to handle .nf/.fi and to complain if it encounters
2162 .br or .in.
2163
2164 4. Updated comments in configure.ac that get placed in config.h.in and also
2165 arranged for config.h to be included in the distribution, with the name
2166 config.h.generic, for the benefit of those who have to compile without
2167 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2168
2169 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2170 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2171 makevp.bat for the current PCRE, using the additional files
2172 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2173
2174 6. A Windows user reported a minor discrepancy with test 2, which turned out
2175 to be caused by a trailing space on an input line that had got lost in his
2176 copy. The trailing space was an accident, so I've just removed it.
2177
2178 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2179 that is needed.
2180
2181 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2182 as "const" (a) because they are and (b) because it helps the PHP
2183 maintainers who have recently made a script to detect big data structures
2184 in the php code that should be moved to the .rodata section. I remembered
2185 to update Builducptable as well, so it won't revert if ucptable.h is ever
2186 re-created.
2187
2188 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2189 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2190 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2191 support is not required. This saves 1.5-2K of code, which is important in
2192 some applications.
2193
2194 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2195 so as not to refer to the tables, even though these functions will never be
2196 called when UTF-8 support is disabled. Otherwise there are problems with a
2197 shared library.
2198
2199 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2200
2201 (a) It was defining its arguments as char * instead of void *.
2202
2203 (b) It was assuming that all moves were upwards in memory; this was true
2204 a long time ago when I wrote it, but is no longer the case.
2205
2206 The emulated memove() is provided for those environments that have neither
2207 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2208 is clearly not the case, as these two bugs were recently reported.
2209
2210 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2211 and Detrail to create the HTML documentation, the .txt form of the man
2212 pages, and it removes trailing spaces from listed files. It also creates
2213 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2214 case, it wraps all the #defines with #ifndefs. This script should be run
2215 before "make dist".
2216
2217 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2218 with Unicode property support.
2219
2220 (a) For a maximizing quantifier, if the two different cases of the
2221 character were of different lengths in their UTF-8 codings (there are
2222 some cases like this - I found 11), and the matching function had to
2223 back up over a mixture of the two cases, it incorrectly assumed they
2224 were both the same length.
2225
2226 (b) When PCRE was configured to use the heap rather than the stack for
2227 recursion during matching, it was not correctly preserving the data for
2228 the other case of a UTF-8 character when checking ahead for a match
2229 while processing a minimizing repeat. If the check also involved
2230 matching a wide character, but failed, corruption could cause an
2231 erroneous result when trying to check for a repeat of the original
2232 character.
2233
2234 13. Some tidying changes to the testing mechanism:
2235
2236 (a) The RunTest script now detects the internal link size and whether there
2237 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2238 values substituted by "configure". (The RunGrepTest script already did
2239 this for UTF-8.) The configure.ac script no longer substitutes the
2240 relevant variables.
2241
2242 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2243 with length and offset values. This means that the output is different
2244 for different internal link sizes. Test 2 is skipped for link sizes
2245 other than 2 because of this, bypassing the problem. Unfortunately,
2246 there was also a test in test 3 (the locale tests) that used /B and
2247 failed for link sizes other than 2. Rather than cut the whole test out,
2248 I have added a new /Z option to pcretest that replaces the length and
2249 offset values with spaces. This is now used to make test 3 independent
2250 of link size. (Test 2 will be tidied up later.)
2251
2252 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2253 segmentation fault instead of returning the appropriate error message.
2254
2255 15. In multiline mode when the newline sequence was set to "any", the pattern
2256 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2257 This doesn't seem right; it now treats the CRLF combination as the line
2258 ending, and so does not match in that case. It's only a pattern such as ^$
2259 that would hit this one: something like ^ABC$ would have failed after \r
2260 and then tried again after \r\n.
2261
2262 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2263 in an attempt to make files that differ only in their line terminators
2264 compare equal. This works on Linux.
2265
2266 17. Under certain error circumstances pcregrep might try to free random memory
2267 as it exited. This is now fixed, thanks to valgrind.
2268
2269 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2270 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2271 was because its rules for how to advance for /g after matching an empty
2272 string at the end of a line did not allow for this case. They now check for
2273 it specially.
2274
2275 20. pcretest is supposed to handle patterns and data of any length, by
2276 extending its buffers when necessary. It was getting this wrong when the
2277 buffer for a data line had to be extended.
2278
2279 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2280 CRLF as a newline sequence.
2281
2282 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2283 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2284 I have nevertheless tidied it up.
2285
2286 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2287
2288 24. Added a man page for pcre-config.
2289
2290
2291 Version 7.0 19-Dec-06
2292 ---------------------
2293
2294 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2295 moving to gcc 4.1.1.
2296
2297 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2298 sys/time.h, which is documented as needed for this function. It doesn't
2299 seem to matter on Linux, but it showed up on some releases of OS X.
2300
2301 3. It seems that there are systems where bytes whose values are greater than
2302 127 match isprint() in the "C" locale. The "C" locale should be the
2303 default when a C program starts up. In most systems, only ASCII printing
2304 characters match isprint(). This difference caused the output from pcretest
2305 to vary, making some of the tests fail. I have changed pcretest so that:
2306
2307 (a) When it is outputting text in the compiled version of a pattern, bytes
2308 other than 32-126 are always shown as hex escapes.
2309
2310 (b) When it is outputting text that is a matched part of a subject string,
2311 it does the same, unless a different locale has been set for the match
2312 (using the /L modifier). In this case, it uses isprint() to decide.
2313
2314 4. Fixed a major bug that caused incorrect computation of the amount of memory
2315 required for a compiled pattern when options that changed within the
2316 pattern affected the logic of the preliminary scan that determines the
2317 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2318 that the computed length was too small. The symptoms of this bug were
2319 either the PCRE error "internal error: code overflow" from pcre_compile(),
2320 or a glibc crash with a message such as "pcretest: free(): invalid next
2321 size (fast)". Examples of patterns that provoked this bug (shown in
2322 pcretest format) are:
2323
2324 /(?-x: )/x
2325 /(?x)(?-x: \s*#\s*)/
2326 /((?i)[\x{c0}])/8
2327 /(?i:[\x{c0}])/8
2328
2329 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2330 is now done differently.
2331
2332 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2333 wrapper classes; (b) implement a new function in the C++ scanner that is
2334 more efficient than the old way of doing things because it avoids levels of
2335 recursion in the regex matching; (c) add a paragraph to the documentation
2336 for the FullMatch() function.
2337
2338 6. The escape sequence \n was being treated as whatever was defined as
2339 "newline". Not only was this contrary to the documentation, which states
2340 that \n is character 10 (hex 0A), but it also went horribly wrong when
2341 "newline" was defined as CRLF. This has been fixed.
2342
2343 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2344 was being set to -1 for the "end of line" case (supposedly a value that no
2345 character can have). Though this value is never used (the check for end of
2346 line is "zero bytes in current character"), it caused compiler complaints.
2347 I've changed it to 0xffffffff.
2348
2349 8. In pcre_version.c, the version string was being built by a sequence of
2350 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2351 string (as it is for production releases) called a macro with an empty
2352 argument. The C standard says the result of this is undefined. The gcc
2353 compiler treats it as an empty string (which was what was wanted) but it is
2354 reported that Visual C gives an error. The source has been hacked around to
2355 avoid this problem.
2356
2357 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2358 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2359 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2360 of them did).
2361
2362 10. Originally, pcretest opened its input and output without "b"; then I was
2363 told that "b" was needed in some environments, so it was added for release
2364 5.0 to both the input and output. (It makes no difference on Unix-like
2365 systems.) Later I was told that it is wrong for the input on Windows. I've
2366 now abstracted the modes into two macros, to make it easier to fiddle with
2367 them, and removed "b" from the input mode under Windows.
2368
2369 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2370
2371 12. Added -help and --help to pcretest as an official way of being reminded
2372 of the options.
2373
2374 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2375 and pcrecpp.cc because they annoy compilers at high warning levels.
2376
2377 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2378
2379 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2380 scripts are not C programs :-) and which was not noticed because it works
2381 on Linux.
2382
2383 16. pcretest is supposed to handle any length of pattern and data line (as one
2384 line or as a continued sequence of lines) by extending its input buffer if
2385 necessary. This feature was broken for very long pattern lines, leading to
2386 a string of junk being passed to pcre_compile() if the pattern was longer
2387 than about 50K.
2388
2389 17. I have done a major re-factoring of the way pcre_compile() computes the
2390 amount of memory needed for a compiled pattern. Previously, there was code
2391 that made a preliminary scan of the pattern in order to do this. That was
2392 OK when PCRE was new, but as the facilities have expanded, it has become
2393 harder and harder to keep it in step with the real compile phase, and there
2394 have been a number of bugs (see for example, 4 above). I have now found a
2395 cunning way of running the real compile function in a "fake" mode that
2396 enables it to compute how much memory it would need, while actually only
2397 ever using a few hundred bytes of working memory and without too many
2398 tests of the mode. This should make future maintenance and development
2399 easier. A side effect of this work is that the limit of 200 on the nesting
2400 depth of parentheses has been removed (though this was never a serious
2401 limitation, I suspect). However, there is a downside: pcre_compile() now
2402 runs more slowly than before (30% or more, depending on the pattern). I
2403 hope this isn't a big issue. There is no effect on runtime performance.
2404
2405 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2406 newline (only possible for the last line of a file) and it was a
2407 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2408
2409 19. Added additional timing features to pcretest. (1) The -tm option now times
2410 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2411 separate command line item, by a number that specifies the number of
2412 repeats to use when timing. The default is 50000; this gives better
2413 precision, but takes uncomfortably long for very large patterns.
2414
2415 20. Extended pcre_study() to be more clever in cases where a branch of a
2416 subpattern has no definite first character. For example, (a*|b*)[cd] would
2417 previously give no result from pcre_study(). Now it recognizes that the
2418 first character must be a, b, c, or d.
2419
2420 21. There was an incorrect error "recursive call could loop indefinitely" if
2421 a subpattern (or the entire pattern) that was being tested for matching an
2422 empty string contained only one non-empty item after a nested subpattern.
2423 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2424 incorrectly, because the \d was being skipped in the check.
2425
2426 22. The pcretest program now has a new pattern option /B and a command line
2427 option -b, which is equivalent to adding /B to every pattern. This causes
2428 it to show the compiled bytecode, without the additional information that
2429 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2430 is the same as /B/I).
2431
2432 23. A new optimization is now able automatically to treat some sequences such
2433 as a*b as a*+b. More specifically, if something simple (such as a character
2434 or a simple class like \d) has an unlimited quantifier, and is followed by
2435 something that cannot possibly match the quantified thing, the quantifier
2436 is automatically "possessified".
2437
2438 24. A recursive reference to a subpattern whose number was greater than 39
2439 went wrong under certain circumstances in UTF-8 mode. This bug could also
2440 have affected the operation of pcre_study().
2441
2442 25. Realized that a little bit of performance could be had by replacing
2443 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2444
2445 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2446
2447 27. Possessive quantifiers such as a++ were previously implemented by turning
2448 them into atomic groups such as ($>a+). Now they have their own opcodes,
2449 which improves performance. This includes the automatically created ones
2450 from 23 above.
2451
2452 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2453 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2454 the first matched character to be a colon. This applied both to named and
2455 numbered groups.
2456
2457 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2458
2459 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2460 building PCRE on Windows easier, so I have included it in the distribution.
2461
2462 31. There is now a check in pcretest against a ridiculously large number being
2463 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2464 loop, the loop is abandoned.
2465
2466 32. Forward references to subpatterns in conditions such as (?(2)...) where
2467 subpattern 2 is defined later cause pcre_compile() to search forwards in
2468 the pattern for the relevant set of parentheses. This search went wrong
2469 when there were unescaped parentheses in a character class, parentheses
2470 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2471
2472 33. "Subroutine" calls and backreferences were previously restricted to
2473 referencing subpatterns earlier in the regex. This restriction has now
2474 been removed.
2475
2476 34. Added a number of extra features that are going to be in Perl 5.10. On the
2477 whole, these are just syntactic alternatives for features that PCRE had
2478 previously implemented using the Python syntax or my own invention. The
2479 other formats are all retained for compatibility.
2480
2481 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2482 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2483 also .NET compatible.
2484
2485 (b) A recursion or subroutine call to a named group can now be defined as
2486 (?&name) as well as (?P>name).
2487
2488 (c) A backreference to a named group can now be defined as \k<name> or
2489 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2490 5.10, are also .NET compatible.
2491
2492 (d) A conditional reference to a named group can now use the syntax
2493 (?(<name>) or (?('name') as well as (?(name).
2494
2495 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2496 groups (named and numbered) that are never evaluated inline, but can be
2497 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2498 is always false. There may be only one alternative in such a group.
2499
2500 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2501 as the simple (?(R). The condition is true only if the most recent
2502 recursion is that of the given number or name. It does not search out
2503 through the entire recursion stack.
2504
2505 (g) The escape \gN or \g{N} has been added, where N is a positive or
2506 negative number, specifying an absolute or relative reference.
2507
2508 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2509 some "unreachable code" warnings.
2510
2511 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2512 things, this adds five new scripts.
2513
2514 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2515 There were also incompatibilities regarding the handling of \Q..\E inside
2516 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2517 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2518
2519 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2520 matches an empty string, and forcibly breaks the loop. There were bugs in
2521 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2522 against aaaa the result was just "a" rather than "aaaa", for example. Two
2523 separate and independent bugs (that affected different cases) have been
2524 fixed.
2525
2526 39. Refactored the code to abolish the use of different opcodes for small
2527 capturing bracket numbers. This is a tidy that I avoided doing when I
2528 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2529 The new approach is not only tidier, it makes it possible to reduce the
2530 memory needed to fix the previous bug (38).
2531
2532 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2533 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2534 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2535 mode.
2536
2537 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2538 report.
2539
2540 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2541 copy construction and assignment in the C++ wrapper.
2542
2543 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2544 couple of bugs that could have given wrong results in the "--newline=crlf"
2545 case.
2546
2547 44. Added a number of casts and did some reorganization of signed/unsigned int
2548 variables following suggestions from Dair Grant. Also renamed the variable
2549 "this" as "item" because it is a C++ keyword.
2550
2551 45. Arranged for dftables to add
2552
2553 #include "pcre_internal.h"
2554
2555 to pcre_chartables.c because without it, gcc 4.x may remove the array
2556 definition from the final binary if PCRE is built into a static library and
2557 dead code stripping is activated.
2558
2559 46. For an unanchored pattern, if a match attempt fails at the start of a
2560 newline sequence, and the newline setting is CRLF or ANY, and the next two
2561 characters are CRLF, advance by two characters instead of one.
2562
2563
2564 Version 6.7 04-Jul-06
2565 ---------------------
2566
2567 1. In order to handle tests when input lines are enormously long, pcretest has
2568 been re-factored so that it automatically extends its buffers when
2569 necessary. The code is crude, but this _is_ just a test program. The
2570 default size has been increased from 32K to 50K.
2571
2572 2. The code in pcre_study() was using the value of the re argument before
2573 testing it for NULL. (Of course, in any sensible call of the function, it
2574 won't be NULL.)
2575
2576 3. The memmove() emulation function in pcre_internal.h, which is used on
2577 systems that lack both memmove() and bcopy() - that is, hardly ever -
2578 was missing a "static" storage class specifier.
2579
2580 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2581 containing an extended class (one that cannot be represented by a bitmap
2582 because it contains high-valued characters or Unicode property items, e.g.
2583 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2584 pattern, but PCRE should not loop if you do not (it no longer does).
2585 [Detail: two cases were found: (a) a repeated subpattern containing an
2586 extended class; (b) a recursive reference to a subpattern that followed a
2587 previous extended class. It wasn't skipping over the extended class
2588 correctly when UTF-8 mode was not set.]
2589
2590 5. A negated single-character class was not being recognized as fixed-length
2591 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2592 compile error "lookbehind assertion is not fixed length".
2593
2594 6. The RunPerlTest auxiliary script was showing an unexpected difference
2595 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2596 write a Perl script that can interpret lines of an input file either as
2597 byte characters or as UTF-8, which is what "perltest" was being required to
2598 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2599 can't do is switch easily at run time between having the "use utf8;" pragma
2600 or not. In the end, I fudged it by using the RunPerlTest script to insert
2601 "use utf8;" explicitly for the UTF-8 tests.
2602
2603 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2604 the end of the subject string, contrary to the documentation and to what
2605 Perl does. This was true of both matching functions. Now it matches only at
2606 the start of the subject and immediately after *internal* newlines.
2607
2608 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2609 a pointer to an int instead of a pointer to an unsigned long int. This
2610 caused problems on 64-bit systems.
2611
2612 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2613 instance of the 'standard' template library not being so standard".
2614
2615 10. There was no check on the number of named subpatterns nor the maximum
2616 length of a subpattern name. The product of these values is used to compute
2617 the size of the memory block for a compiled pattern. By supplying a very
2618 long subpattern name and a large number of named subpatterns, the size
2619 computation could be caused to overflow. This is now prevented by limiting
2620 the length of names to 32 characters, and the number of named subpatterns
2621 to 10,000.
2622
2623 11. Subpatterns that are repeated with specific counts have to be replicated in
2624 the compiled pattern. The size of memory for this was computed from the
2625 length of the subpattern and the repeat count. The latter is limited to
2626 65535, but there was no limit on the former, meaning that integer overflow
2627 could in principle occur. The compiled length of a repeated subpattern is
2628 now limited to 30,000 bytes in order to prevent this.
2629
2630 12. Added the optional facility to have named substrings with the same name.
2631
2632 13. Added the ability to use a named substring as a condition, using the
2633 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2634 are numbers (not recommended). Forward references are permitted.
2635
2636 14. Added forward references in named backreferences (if you see what I mean).
2637
2638 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2639 pattern could run off the end of the subject. For example, the pattern
2640 "(?s)(.{1,5})"8 did this with the subject "ab".
2641
2642 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2643 PCRE_CASELESS was set when matching characters that were quantified with ?
2644 or *.
2645
2646 17. A character class other than a single negated character that had a minimum
2647 but no maximum quantifier - for example [ab]{6,} - was not handled
2648 correctly by pce_dfa_exec(). It would match only one character.
2649
2650 18. A valid (though odd) pattern that looked like a POSIX character
2651 class but used an invalid character after [ (for example [[,abc,]]) caused
2652 pcre_compile() to give the error "Failed: internal error: code overflow" or
2653 in some cases to crash with a glibc free() error. This could even happen if
2654 the pattern terminated after [[ but there just happened to be a sequence of
2655 letters, a binary zero, and a closing ] in the memory that followed.
2656
2657 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2658 over the years. Originally (before any Unicode support), just the bottom 8
2659 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2660 output from "man perlunicode" includes this:
2661
2662 The regular expression compiler produces polymorphic opcodes. That
2663 is, the pattern adapts to the data and automatically switches to
2664 the Unicode character scheme when presented with Unicode data--or
2665 instead uses a traditional byte scheme when presented with byte
2666 data.
2667
2668 Sadly, a wide octal escape does not cause a switch, and in a string with
2669 no other multibyte characters, these octal escapes are treated as before.
2670 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2671 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2672 Unicode string.
2673
2674 I have not perpetrated such confusion in PCRE. Up till now, it took just
2675 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2676 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2677 translate to the appropriate multibyte character.
2678
2679 29. Applied some refactoring to reduce the number of warnings from Microsoft
2680 and Borland compilers. This has included removing the fudge introduced
2681 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2682 a warning about an unused variable.
2683
2684 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2685 characters since release 4.0, because Perl (from release 5.004) does not.
2686 [Or at least, is documented not to: some releases seem to be in conflict
2687 with the documentation.] However, when a pattern was studied with
2688 pcre_study() and all its branches started with \s, PCRE still included VT
2689 as a possible starting character. Of course, this did no harm; it just
2690 caused an unnecessary match attempt.
2691
2692 22. Removed a now-redundant internal flag bit that recorded the fact that case
2693 dependency changed within the pattern. This was once needed for "required
2694 byte" processing, but is no longer used. This recovers a now-scarce options
2695 bit. Also moved the least significant internal flag bit to the most-
2696 significant bit of the word, which was not previously used (hangover from
2697 the days when it was an int rather than a uint) to free up another bit for
2698 the future.
2699
2700 23. Added support for CRLF line endings as well as CR and LF. As well as the
2701 default being selectable at build time, it can now be changed at runtime
2702 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2703 specify that it is scanning data with non-default line endings.
2704
2705 24. Changed the definition of CXXLINK to make it agree with the definition of
2706 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2707
2708 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2709 recursions. This makes a big different to stack usage for some patterns.
2710
2711 26. If a subpattern containing a named recursion or subroutine reference such
2712 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2713 the space required for the compiled pattern went wrong and gave too small a
2714 value. Depending on the environment, this could lead to "Failed: internal
2715 error: code overflow at offset 49" or "glibc detected double free or
2716 corruption" errors.
2717
2718 27. Applied patches from Google (a) to support the new newline modes and (b) to
2719 advance over multibyte UTF-8 characters in GlobalReplace.
2720
2721 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2722 difference for some implementation of PCRE in some Windows version.
2723
2724 29. Added some extra testing facilities to pcretest:
2725
2726 \q<number> in a data line sets the "match limit" value
2727 \Q<number> in a data line sets the "match recursion limt" value
2728 -S <number> sets the stack size, where <number> is in megabytes
2729
2730 The -S option isn't available for Windows.
2731
2732
2733 Version 6.6 06-Feb-06
2734 ---------------------
2735
2736 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2737 in pcreposix.h. I have copied the definition from pcre.h.
2738
2739 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2740 because pcre.h is no longer a built file.
2741
2742 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2743 not normally included in the compiled code.
2744
2745
2746 Version 6.5 01-Feb-06
2747 ---------------------
2748
2749 1. When using the partial match feature with pcre_dfa_exec(), it was not
2750 anchoring the second and subsequent partial matches at the new starting
2751 point. This could lead to incorrect results. For example, with the pattern
2752 /1234/, partially matching against "123" and then "a4" gave a match.
2753
2754 2. Changes to pcregrep:
2755
2756 (a) All non-match returns from pcre_exec() were being treated as failures
2757 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2758 error message is output. Some extra information is given for the
2759 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2760 probably the only errors that are likely to be caused by users (by
2761 specifying a regex that has nested indefinite repeats, for instance).
2762 If there are more than 20 of these errors, pcregrep is abandoned.
2763
2764 (b) A binary zero was treated as data while matching, but terminated the
2765 output line if it was written out. This has been fixed: binary zeroes
2766 are now no different to any other data bytes.
2767
2768 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2769 used to set a locale for matching. The --locale=xxxx long option has
2770 been added (no short equivalent) to specify a locale explicitly on the
2771 pcregrep command, overriding the environment variables.
2772
2773 (d) When -B was used with -n, some line numbers in the output were one less
2774 than they should have been.
2775
2776 (e) Added the -o (--only-matching) option.
2777
2778 (f) If -A or -C was used with -c (count only), some lines of context were
2779 accidentally printed for the final match.
2780
2781 (g) Added the -H (--with-filename) option.
2782
2783 (h) The combination of options -rh failed to suppress file names for files
2784 that were found from directory arguments.
2785
2786 (i) Added the -D (--devices) and -d (--directories) options.
2787
2788 (j) Added the -F (--fixed-strings) option.
2789
2790 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2791
2792 (l) Added the --colo(u)r option.
2793
2794 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2795 is not present by default.
2796
2797 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2798 items such as (?R) or (?1), when the recursion could match a number of
2799 alternatives. If it matched one of the alternatives, but subsequently,
2800 outside the recursion, there was a failure, the code tried to back up into
2801 the recursion. However, because of the way PCRE is implemented, this is not
2802 possible, and the result was an incorrect result from the match.
2803
2804 In order to prevent this happening, the specification of recursion has
2805 been changed so that all such subpatterns are automatically treated as
2806 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2807
2808 4. I had overlooked the fact that, in some locales, there are characters for
2809 which isalpha() is true but neither isupper() nor islower() are true. In
2810 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2811 and ordfeminine) are like this. This affected the treatment of \w and \W
2812 when they appeared in character classes, but not when they appeared outside
2813 a character class. The bit map for "word" characters is now created
2814 separately from the results of isalnum() instead of just taking it from the
2815 upper, lower, and digit maps. (Plus the underscore character, of course.)
2816
2817 5. The above bug also affected the handling of POSIX character classes such as
2818 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2819 permanent tables. Instead, the bit maps for such a class were previously
2820 created as the appropriate unions of the upper, lower, and digit bitmaps.
2821 Now they are created by subtraction from the [[:word:]] class, which has
2822 its own bitmap.
2823
2824 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2825 It is created by subtracting the vertical space characters (\x09, \x0a,
2826 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2827 subtraction was done in the overall bitmap for a character class, meaning
2828 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2829 be recognized. This bug has been fixed.
2830
2831 7. Patches from the folks at Google:
2832
2833 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2834 real life, but is still worth protecting against".
2835
2836 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2837 regular expressions".
2838
2839 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2840 have it.
2841
2842 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2843 "configure" and the latter not, in order to fix a problem somebody had
2844 with compiling the Arg class on HP-UX.
2845
2846 (e) Improve the error-handling of the C++ wrapper a little bit.
2847
2848 (f) New tests for checking recursion limiting.
2849
2850 8. The pcre_memmove() function, which is used only if the environment does not
2851 have a standard memmove() function (and is therefore rarely compiled),
2852 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2853 returning a result (though PCRE never actually uses the result).
2854
2855 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2856 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2857 returned instead of calling malloc() with an overflowing number that would
2858 most likely cause subsequent chaos.
2859
2860 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2861
2862 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2863 with this option is matched, the nmatch and pmatch options of regexec() are
2864 ignored.
2865
2866 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2867 provided in case anyone wants to the the POSIX interface with UTF-8
2868 strings.
2869
2870 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2871 C++ linking (needed for some HP-UX environments).
2872
2873 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2874 (unused parameter) and in the pcre_printint() function (omitted "default"
2875 switch label when the default is to do nothing).
2876
2877 15. Added some code to make it possible, when PCRE is compiled as a C++
2878 library, to replace subject pointers for pcre_exec() with a smart pointer
2879 class, thus making it possible to process discontinuous strings.
2880
2881 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2882 much the same function. They were added by different people who were trying
2883 to make PCRE easy to compile on non-Unix systems. It has been suggested
2884 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2885 for compiling on Windows systems. I have therefore replaced it with
2886 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2887 defaults to "extern" for C or "extern C" for C++, which works fine on
2888 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2889 SCOPE with something explicit in config.h. In addition:
2890
2891 (a) pcreposix.h still had just "extern" instead of either of these macros;
2892 I have replaced it with PCRE_DATA_SCOPE.
2893
2894 (b) Functions such as _pcre_xclass(), which are internal to the library,
2895 but external in the C sense, all had PCRE_EXPORT in their definitions.
2896 This is apparently wrong for the Windows case, so I have removed it.
2897 (It makes no difference on Unix-like systems.)
2898
2899 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2900 of recursive calls to match(). This is different to MATCH_LIMIT because
2901 that limits the total number of calls to match(), not all of which increase
2902 the depth of recursion. Limiting the recursion depth limits the amount of
2903 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2904 when PCRE is compiled, and changed at run time. A patch from Google adds
2905 this functionality to the C++ interface.
2906
2907 18. Changes to the handling of Unicode character properties:
2908
2909 (a) Updated the table to Unicode 4.1.0.
2910
2911 (b) Recognize characters that are not in the table as "Cn" (undefined).
2912
2913 (c) I revised the way the table is implemented to a much improved format
2914 which includes recognition of ranges. It now supports the ranges that
2915 are defined in UnicodeData.txt, and it also amalgamates other
2916 characters into ranges. This has reduced the number of entries in the
2917 table from around 16,000 to around 3,000, thus reducing its size
2918 considerably. I realized I did not need to use a tree structure after
2919 all - a binary chop search is just as efficient. Having reduced the
2920 number of entries, I extended their size from 6 bytes to 8 bytes to
2921 allow for more data.
2922
2923 (d) Added support for Unicode script names via properties such as \p{Han}.
2924
2925 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2926 matching that character.
2927
2928 20. When matching a repeated Unicode property with a minimum greater than zero,
2929 (for example \pL{2,}), PCRE could look past the end of the subject if it
2930 reached it while seeking the minimum number of characters. This could
2931 happen only if some of the characters were more than one byte long, because
2932 there is a check for at least the minimum number of bytes.
2933
2934 21. Refactored the implementation of \p and \P so as to be more general, to
2935 allow for more different types of property in future. This has changed the
2936 compiled form incompatibly. Anybody with saved compiled patterns that use
2937 \p or \P will have to recompile them.
2938
2939 22. Added "Any" and "L&" to the supported property types.
2940
2941 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2942 but give a compile time error if the value is greater than 0xff.
2943
2944 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2945 accidentally not being installed or uninstalled.
2946
2947 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2948 made were to insert the current release number. This seemed silly, because
2949 it made things harder for people building PCRE on systems that don't run
2950 "configure". I have turned pcre.h into a distributed file, no longer built
2951 by "configure", with the version identification directly included. There is
2952 no longer a pcre.h.in file.
2953
2954 However, this change necessitated a change to the pcre-config script as
2955 well. It is built from pcre-config.in, and one of the substitutions was the
2956 release number. I have updated configure.ac so that ./configure now finds
2957 the release number by grepping pcre.h.
2958
2959 26. Added the ability to run the tests under valgrind.
2960
2961
2962 Version 6.4 05-Sep-05
2963 ---------------------
2964
2965 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2966 "--" to be printed when multiple files were scanned, even when none of the
2967 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2968 consider it to be a bug, and have restored the previous behaviour.
2969
2970 2. A couple of code tidies to get rid of compiler warnings.
2971
2972 3. The pcretest program used to cheat by referring to symbols in the library
2973 whose names begin with _pcre_. These are internal symbols that are not
2974 really supposed to be visible externally, and in some environments it is
2975 possible to suppress them. The cheating is now confined to including
2976 certain files from the library's source, which is a bit cleaner.
2977
2978 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2979 file's purpose clearer.
2980
2981 5. Reorganized pcre_ucp_findchar().
2982
2983
2984 Version 6.3 15-Aug-05
2985 ---------------------
2986
2987 1. The file libpcre.pc.in did not have general read permission in the tarball.
2988
2989 2. There were some problems when building without C++ support:
2990
2991 (a) If C++ support was not built, "make install" and "make test" still
2992 tried to test it.
2993
2994 (b) There were problems when the value of CXX was explicitly set. Some
2995 changes have been made to try to fix these, and ...
2996
2997 (c) --disable-cpp can now be used to explicitly disable C++ support.
2998
2999 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3000 backslash in a target when C++ was disabled. This confuses some
3001 versions of "make", apparently. Using an intermediate variable solves
3002 this. (Same for CPP_LOBJ.)
3003
3004 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3005 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3006 necessary on certain architectures.
3007
3008 4. Added a setting of -export-symbols-regex to the link command to remove
3009 those symbols that are exported in the C sense, but actually are local
3010 within the library, and not documented. Their names all begin with
3011 "_pcre_". This is not a perfect job, because (a) we have to except some
3012 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3013 available (and never for static libraries). I have made a note to try to
3014 find a way round (a) in the future.
3015
3016
3017 Version 6.2 01-Aug-05
3018 ---------------------
3019
3020 1. There was no test for integer overflow of quantifier values. A construction
3021 such as {1111111111111111} would give undefined results. What is worse, if
3022 a minimum quantifier for a parenthesized subpattern overflowed and became
3023 negative, the calculation of the memory size went wrong. This could have
3024 led to memory overwriting.
3025
3026 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3027
3028 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3029 operating environments where this matters.
3030
3031 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3032 PCRE options from within the C++ wrapper.
3033
3034 5. Named capturing subpatterns were not being correctly counted when a pattern
3035 was compiled. This caused two problems: (a) If there were more than 100
3036 such subpatterns, the calculation of the memory needed for the whole
3037 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3038 back references of the form \12, where the number was greater than 9, were
3039 not recognized as back references, even though there were sufficient
3040 previous subpatterns.
3041
3042 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3043 versions of gcc, e.g. 2.95.4.
3044
3045
3046 Version 6.1 21-Jun-05
3047 ---------------------
3048
3049 1. There was one reference to the variable "posix" in pcretest.c that was not
3050 surrounded by "#if !defined NOPOSIX".
3051
3052 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3053 the cross-check on the old pcre_info() function, for the benefit of the
3054 cut-down version of PCRE that is currently imported into Exim.
3055
3056 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3057 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3058 patterns that start (?i) but I don't think that matters. The use of (?i) is
3059 just an example; this all applies to the other options as well.
3060
3061 4. Since libtool seems to echo the compile commands it is issuing, the output
3062 from "make" can be reduced a bit by putting "@" in front of each libtool
3063 compile command.
3064
3065 5. Patch from the folks at Google for configure.in to be a bit more thorough
3066 in checking for a suitable C++ installation before trying to compile the
3067 C++ stuff. This should fix a reported problem when a compiler was present,
3068 but no suitable headers.
3069
3070 6. The man pages all had just "PCRE" as their title. I have changed them to
3071 be the relevant file name. I have also arranged that these names are
3072 retained in the file doc/pcre.txt, which is a concatenation in text format
3073 of all the man pages except the little individual ones for each function.
3074
3075 7. The NON-UNIX-USE file had not been updated for the different set of source
3076 files that come with release 6. I also added a few comments about the C++
3077 wrapper.
3078
3079
3080 Version 6.0 07-Jun-05
3081 ---------------------
3082
3083 1. Some minor internal re-organization to help with my DFA experiments.
3084
3085 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3086 didn't matter for the library itself when fully configured, but did matter
3087 when compiling without UCP support, or within Exim, where the ucp files are
3088 not imported.
3089
3090 3. Refactoring of the library code to split up the various functions into
3091 different source modules. The addition of the new DFA matching code (see
3092 below) to a single monolithic source would have made it really too
3093 unwieldy, quite apart from causing all the code to be include in a
3094 statically linked application, when only some functions are used. This is
3095 relevant even without the DFA addition now that patterns can be compiled in
3096 one application and matched in another.
3097
3098 The downside of splitting up is that there have to be some external
3099 functions and data tables that are used internally in different modules of
3100 the library but which are not part of the API. These have all had their
3101 names changed to start with "_pcre_" so that they are unlikely to clash
3102 with other external names.
3103
3104 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3105 a different (DFA) algorithm. Although it is slower than the original
3106 function, it does have some advantages for certain types of matching
3107 problem.
3108
3109 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3110 including restarting after a partial match.
3111
3112 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3113 defined when compiling for Windows was sent to me. I have put it into the
3114 code, though I have no means of testing or verifying it.
3115
3116 7. Added the pcre_refcount() auxiliary function.
3117
3118 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3119 match before or at the first newline in the subject string. In pcretest,
3120 the /f option on a pattern can be used to set this.
3121
3122 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3123 would behave wrongly. This has been present in PCRE since release 4.0.
3124
3125 10. A number of changes to the pcregrep command:
3126
3127 (a) Refactored how -x works; insert ^(...)$ instead of setting
3128 PCRE_ANCHORED and checking the length, in preparation for adding
3129 something similar for -w.
3130
3131 (b) Added the -w (match as a word) option.
3132
3133 (c) Refactored the way lines are read and buffered so as to have more
3134 than one at a time available.
3135
3136 (d) Implemented a pcregrep test script.
3137
3138 (e) Added the -M (multiline match) option. This allows patterns to match
3139 over several lines of the subject. The buffering ensures that at least
3140 8K, or the rest of the document (whichever is the shorter) is available
3141 for matching (and similarly the previous 8K for lookbehind assertions).
3142
3143 (f) Changed the --help output so that it now says
3144
3145 -w, --word-regex(p)
3146
3147 instead of two lines, one with "regex" and the other with "regexp"
3148 because that confused at least one person since the short forms are the
3149 same. (This required a bit of code, as the output is generated
3150 automatically from a table. It wasn't just a text change.)
3151
3152 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3153 option but starts with a hyphen. Could be a pattern or a path name
3154 starting with a hyphen, for instance.
3155
3156 (h) "-" can be given as a file name to represent stdin.
3157
3158 (i) When file names are being printed, "(standard input)" is used for
3159 the standard input, for compatibility with GNU grep. Previously
3160 "<stdin>" was used.
3161
3162 (j) The option --label=xxx can be used to supply a name to be used for
3163 stdin when file names are being printed. There is no short form.
3164
3165 (k) Re-factored the options decoding logic because we are going to add
3166 two more options that take data. Such options can now be given in four
3167 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3168
3169 (l) Added the -A, -B, and -C options for requesting that lines of context
3170 around matches be printed.
3171
3172 (m) Added the -L option to print the names of files that do not contain
3173 any matching lines, that is, the complement of -l.
3174
3175 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3176 continue to scan other files.
3177
3178 (o) The -s option was incorrectly implemented. For compatibility with other
3179 greps, it now suppresses the error message for a non-existent or non-
3180 accessible file (but not the return code). There is a new option called
3181 -q that suppresses the output of matching lines, which was what -s was
3182 previously doing.
3183
3184 (p) Added --include and --exclude options to specify files for inclusion
3185 and exclusion when recursing.
3186
3187 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3188 Hopefully, it now does.
3189
3190 12. Missing cast in pcre_study().
3191
3192 13. Added an "uninstall" target to the makefile.
3193
3194 14. Replaced "extern" in the function prototypes in Makefile.in with
3195 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3196 world, but is set differently for Windows.
3197
3198 15. Added a second compiling function called pcre_compile2(). The only
3199 difference is that it has an extra argument, which is a pointer to an
3200 integer error code. When there is a compile-time failure, this is set
3201 non-zero, in addition to the error test pointer being set to point to an
3202 error message. The new argument may be NULL if no error number is required
3203 (but then you may as well call pcre_compile(), which is now just a
3204 wrapper). This facility is provided because some applications need a
3205 numeric error indication, but it has also enabled me to tidy up the way
3206 compile-time errors are handled in the POSIX wrapper.
3207
3208 16. Added VPATH=.libs to the makefile; this should help when building with one
3209 prefix path and installing with another. (Or so I'm told by someone who
3210 knows more about this stuff than I do.)
3211
3212 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3213 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3214 match everything, including newlines. This is not POSIX-compatible, but
3215 somebody wanted the feature. From pcretest it can be activated by using
3216 both the P and the s flags.
3217
3218 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3219
3220 19. libpcre.pc was being incorrectly installed as executable.
3221
3222 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3223 it now also looks for '\r' so that it will work unmodified on Windows.
3224
3225 21. Added Google's contributed C++ wrapper to the distribution.
3226
3227 22. Added some untidy missing memory free() calls in pcretest, to keep
3228 Electric Fence happy when testing.
3229
3230
3231
3232 Version 5.0 13-Sep-04
3233 ---------------------
3234
3235 1. Internal change: literal characters are no longer packed up into items
3236 containing multiple characters in a single byte-string. Each character
3237 is now matched using a separate opcode. However, there may be more than one
3238 byte in the character in UTF-8 mode.
3239
3240 2. The pcre_callout_block structure has two new fields: pattern_position and
3241 next_item_length. These contain the offset in the pattern to the next match
3242 item, and its length, respectively.
3243
3244 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3245 insertion of callouts before each pattern item. Added the /C option to
3246 pcretest to make use of this.
3247
3248 4. On the advice of a Windows user, the lines
3249
3250 #if defined(_WIN32) || defined(WIN32)
3251 _setmode( _fileno( stdout ), 0x8000 );
3252 #endif /* defined(_WIN32) || defined(WIN32) */
3253
3254 have been added to the source of pcretest. This apparently does useful
3255 magic in relation to line terminators.
3256
3257 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3258 for the benefit of those environments where the "b" makes a difference.
3259
3260 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3261 to know about it. I have put a hack into configure.in that adds in code
3262 to set GCC=yes if CC=icc. This seems to end up at a point in the
3263 generated configure script that is early enough to affect the setting of
3264 compiler options, which is what is needed, but I have no means of testing
3265 whether it really works. (The user who reported this had patched the
3266 generated configure script, which of course I cannot do.)
3267
3268 LATER: After change 22 below (new libtool files), the configure script
3269 seems to know about icc (and also ecc). Therefore, I have commented out
3270 this hack in configure.in.
3271
3272 7. Added support for pkg-config (2 patches were sent in).
3273
3274 8. Negated POSIX character classes that used a combination of internal tables
3275 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3276 [[:^ascii]]. Typically, they would match almost any characters. The other
3277 POSIX classes were not broken in this way.
3278
3279 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3280 to find the match, as PCRE was deluded into thinking that the match had to
3281 start at the start point or following a newline. The same bug applied to
3282 patterns with negative forward assertions or any backward assertions
3283 preceding ".*" at the start, unless the pattern required a fixed first
3284 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3285
3286 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3287 starting at the last subject character, bytes beyond the end of the subject
3288 string were read.
3289
3290 11. Renamed the variable "class" as "classbits" to make life easier for C++
3291 users. (Previously there was a macro definition, but it apparently wasn't
3292 enough.)
3293
3294 12. Added the new field "tables" to the extra data so that tables can be passed
3295 in at exec time, or the internal tables can be re-selected. This allows
3296 a compiled regex to be saved and re-used at a later time by a different
3297 program that might have everything at different addresses.
3298
3299 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3300 -R library as well as a -L library.
3301
3302 14. The debugging options of pcretest (-d on the command line or D on a
3303 pattern) showed incorrect output for anything following an extended class
3304 that contained multibyte characters and which was followed by a quantifier.
3305
3306 15. Added optional support for general category Unicode character properties
3307 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3308 support. It adds about 90K to the size of the library. The meanings of the
3309 inbuilt class escapes such as \d and \s have NOT been changed.
3310
3311 16. Updated pcredemo.c to include calls to free() to release the memory for the
3312 compiled pattern.
3313
3314 17. The generated file chartables.c was being created in the source directory
3315 instead of in the building directory. This caused the build to fail if the
3316 source directory was different from the building directory, and was
3317 read-only.
3318
3319 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3320 file. No doubt somebody will tell me if they don't make sense... Also added
3321 Dan Mooney's comments about building on OpenVMS.
3322
3323 19. Added support for partial matching via the PCRE_PARTIAL option for
3324 pcre_exec() and the \P data escape in pcretest.
3325
3326 20. Extended pcretest with 3 new pattern features:
3327
3328 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3329 write the compiled pattern to the file whose name is "rest-of-line".
3330 This is a straight binary dump of the data, with the saved pointer to
3331 the character tables forced to be NULL. The study data, if any, is
3332 written too. After writing, pcretest reads a new pattern.
3333
3334 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3335 compiled pattern from the given file. There must not be any
3336 occurrences of "<" in the file name (pretty unlikely); if there are,
3337 pcretest will instead treat the initial "<" as a pattern delimiter.
3338 After reading in the pattern, pcretest goes on to read data lines as
3339 usual.
3340
3341 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3342 and 16-bit fields in a compiled pattern, to simulate a pattern that
3343 was compiled on a host of opposite endianness.
3344
3345 21. The pcre-exec() function can now cope with patterns that were compiled on
3346 hosts of opposite endianness, with this restriction:
3347
3348 As for any compiled expression that is saved and used later, the tables
3349 pointer field cannot be preserved; the extra_data field in the arguments
3350 to pcre_exec() should be used to pass in a tables address if a value
3351 other than the default internal tables were used at compile time.
3352
3353 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3354 now diagnosed as an error. Previously, most of the time, a negative number
3355 would have been treated as zero, but if in addition "ovector" was passed as
3356 NULL, a crash could occur.
3357
3358 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3359 new versions from the libtool 1.5 distribution (the last one is a copy of
3360 a file called libtool.m4). This seems to have fixed the need to patch
3361 "configure" to support Darwin 1.3 (which I used to do). However, I still
3362 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3363 workstation).
3364
3365 24. Changed the PCRE licence to be the more standard "BSD" licence.
3366
3367
3368 Version 4.5 01-Dec-03
3369 ---------------------
3370
3371 1. There has been some re-arrangement of the code for the match() function so
3372 that it can be compiled in a version that does not call itself recursively.
3373 Instead, it keeps those local variables that need separate instances for
3374 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3375 needs to "recurse". Keeping track of where control must go is done by means
3376 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3377 hide most of the details from the main code, and operates only if
3378 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3379 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3380 operating.
3381
3382 To make it easier for callers to provide specially tailored get/free
3383 functions for this usage, two new functions, pcre_stack_malloc, and
3384 pcre_stack_free, are used. They are always called in strict stacking order,
3385 and the size of block requested is always the same.
3386
3387 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3388 PCRE has been compiled to use the stack or the heap for recursion. The
3389 -C option of pcretest uses this to show which version is compiled.
3390
3391 A new data escape \S, is added to pcretest; it causes the amounts of store
3392 obtained and freed by both kinds of malloc/free at match time to be added
3393 to the output.
3394
3395 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3396 what's available on my current Linux desktop machine.
3397
3398 3. When matching a UTF-8 string, the test for a valid string at the start has
3399 been extended. If start_offset is not zero, PCRE now checks that it points
3400 to a byte that is the start of a UTF-8 character. If not, it returns
3401 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3402 this is necessary because there may be backward assertions in the pattern.
3403 When matching the same subject several times, it may save resources to use
3404 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3405
3406 4. The code for checking the validity of UTF-8 strings has been tightened so
3407 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3408 containing "overlong sequences".
3409
3410 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3411 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3412 should have been "&", but it just so happened that all the cases this let
3413 through by mistake were picked up later in the function.
3414
3415 6. I had used a variable called "isblank" - this is a C99 function, causing
3416 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3417
3418 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3419 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3420 so the version is shown; (c) stop "make test" echoing "./RunTest".
3421
3422 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3423
3424 9. The prototype for memmove() for systems that don't have it was using
3425 size_t, but the inclusion of the header that defines size_t was later. I've
3426 moved the #includes for the C headers earlier to avoid this.
3427
3428 10. Added some adjustments to the code to make it easier to compiler on certain
3429 special systems:
3430
3431 (a) Some "const" qualifiers were missing.
3432 (b) Added the macro EXPORT before all exported functions; by default this
3433 is defined to be empty.
3434 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3435 that it reads its output file name as an argument instead of writing
3436 to the standard output and assuming this can be redirected.
3437
3438 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3439 class containing characters with values greater than 255, PCRE compilation
3440 went into a loop.
3441
3442 12. A recursive reference to a subpattern that was within another subpattern
3443 that had a minimum quantifier of zero caused PCRE to crash. For example,
3444 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3445 recursion. If the recursively-called subpattern itself had a zero repeat,
3446 that was OK.
3447
3448 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3449 buffer into which it was copied (for escape processing) was still set at
3450 1024, so long lines caused crashes.
3451
3452 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3453 "internal error: code overflow...". This applied to any character class
3454 that was followed by a possessive quantifier.
3455
3456 15. Modified the Makefile to add libpcre.la as a prerequisite for
3457 libpcreposix.la because I was told this is needed for a parallel build to
3458 work.
3459
3460 16. If a pattern that contained .* following optional items at the start was
3461 studied, the wrong optimizing data was generated, leading to matching
3462 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3463 matching string must start with a or b or c. The correct conclusion for
3464 this pattern is that a match can start with any character.
3465
3466
3467 Version 4.4 13-Aug-03
3468 ---------------------
3469
3470 1. In UTF-8 mode, a character class containing characters with values between
3471 127 and 255 was not handled correctly if the compiled pattern was studied.
3472 In fixing this, I have also improved the studying algorithm for such
3473 classes (slightly).
3474
3475 2. Three internal functions had redundant arguments passed to them. Removal
3476 might give a very teeny performance improvement.
3477
3478 3. Documentation bug: the value of the capture_top field in a callout is *one
3479 more than* the number of the hightest numbered captured substring.
3480
3481 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3482 in incorrectly linking with a previously installed version. They now link
3483 explicitly with libpcre.la.
3484
3485 5. configure.in no longer needs to recognize Cygwin specially.
3486
3487 6. A problem in pcre.in for Windows platforms is fixed.
3488
3489 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3490 pcretest, it used to include the size of the study block as part of its
3491 output. Unfortunately, the structure contains a field that has a different
3492 size on different hardware architectures. This meant that the tests that
3493 showed this size failed. As the block is currently always of a fixed size,
3494 this information isn't actually particularly useful in pcretest output, so
3495 I have just removed it.
3496
3497 8. Three pre-processor statements accidentally did not start in column 1.
3498 Sadly, there are *still* compilers around that complain, even though
3499 standard C has not required this for well over a decade. Sigh.
3500
3501 9. In pcretest, the code for checking callouts passed small integers in the
3502 callout_data field, which is a void * field. However, some picky compilers
3503 complained about the casts involved for this on 64-bit systems. Now
3504 pcretest passes the address of the small integer instead, which should get
3505 rid of the warnings.
3506
3507 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3508 both compile and run time, and gives an error if an invalid UTF-8 sequence
3509 is found. There is a option for disabling this check in cases where the
3510 string is known to be correct and/or the maximum performance is wanted.
3511
3512 11. In response to a bug report, I changed one line in Makefile.in from
3513
3514 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3515 to
3516 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3517
3518 to look similar to other lines, but I have no way of telling whether this
3519 is the right thing to do, as I do not use Windows. No doubt I'll get told
3520 if it's wrong...
3521
3522
3523 Version 4.3 21-May-03
3524 ---------------------
3525
3526 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3527 Makefile.
3528
3529 2. Some refactoring to improve the quality of the code:
3530
3531 (i) The utf8_table... variables are now declared "const".
3532
3533 (ii) The code for \cx, which used the "case flipping" table to upper case
3534 lower case letters, now just substracts 32. This is ASCII-specific,
3535 but the whole concept of \cx is ASCII-specific, so it seems
3536 reasonable.
3537
3538 (iii) PCRE was using its character types table to recognize decimal and
3539 hexadecimal digits in the pattern. This is silly, because it handles
3540 only 0-9, a-f, and A-F, but the character types table is locale-
3541 specific, which means strange things might happen. A private
3542 table is now used for this - though it costs 256 bytes, a table is
3543 much faster than multiple explicit tests. Of course, the standard
3544 character types table is still used for matching digits in subject
3545 strings against \d.
3546
3547 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3548 ending in _t are). So I've renamed it as ESC_tee.
3549
3550 3. The first argument for regexec() in the POSIX wrapper should have been
3551 defined as "const".
3552
3553 4. Changed pcretest to use malloc() for its buffers so that they can be
3554 Electric Fenced for debugging.
3555
3556 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3557 to read one or more bytes before the start of the subject string. Often this
3558 had no effect on PCRE's behaviour, but in some circumstances it could
3559 provoke a segmentation fault.
3560
3561 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3562 to try to read one or more bytes before the start of the subject string.
3563
3564 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3565 UTF-8 support could misbehave in various ways if the subject string
3566 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3567 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3568 back over UTF-8 characters.)
3569
3570
3571 Version 4.2 14-Apr-03
3572 ---------------------
3573
3574 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3575
3576 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3577 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3578 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3579 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3580 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3581 and BUILD_EXEEXT
3582 Note: automatic setting of the BUILD variables is not yet working
3583 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3584 compile-time but not at link-time
3585 [LINK]: use for linking executables only
3586 make different versions for Windows and non-Windows
3587 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3588 libraries
3589 [LINK_FOR_BUILD]: new variable
3590 [OBJEXT]: use throughout
3591 [EXEEXT]: use throughout
3592 <winshared>: new target
3593 <wininstall>: new target
3594 <dftables.o>: use native compiler
3595 <dftables>: use native linker
3596 <install>: handle Windows platform correctly
3597 <clean>: ditto
3598 <check>: ditto
3599 copy DLL to top builddir before testing
3600
3601 As part of these changes, -no-undefined was removed again. This was reported
3602 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3603 in any case.
3604
3605 3. Some tidies to get rid of compiler warnings:
3606
3607 . In the match_data structure, match_limit was an unsigned long int, whereas
3608 match_call_count was an int. I've made them both unsigned long ints.
3609
3610 . In pcretest the fact that a const uschar * doesn't automatically cast to
3611 a void * provoked a warning.
3612
3613 . Turning on some more compiler warnings threw up some "shadow" variables
3614 and a few more missing casts.
3615
3616 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3617 option, a class that contained a single character with a value between 128
3618 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3619
3620 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3621 option, a class that contained several characters, but with at least one
3622 whose value was between 128 and 255 caused PCRE to crash.
3623
3624
3625 Version 4.1 12-Mar-03
3626 ---------------------
3627
3628 1. Compiling with gcc -pedantic found a couple of places where casts were
3629 needed, and a string in dftables.c that was longer than standard compilers are
3630 required to support.
3631
3632 2. Compiling with Sun's compiler found a few more places where the code could
3633 be tidied up in order to avoid warnings.
3634
3635 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3636 first of these names is deprecated in the latest Autoconf in favour of the name
3637 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3638 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3639 analogy I have changed it to CFLAGS_FOR_BUILD.
3640
3641 4. Added -no-undefined to the linking command in the Makefile, because this is
3642 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3643 linking step for the pcreposix library.
3644
3645 5. PCRE was failing to diagnose the case of two named groups with the same
3646 name.
3647
3648 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3649 literal character that is needed in the subject for a match, and scans along to
3650 ensure that it is present before embarking on the full matching process. This
3651 saves time in cases of nested unlimited repeats that are never going to match.
3652 Problem: the scan can take a lot of time if the subject is very long (e.g.
3653 megabytes), thus penalizing straightforward matches. It is now done only if the
3654 amount of subject to be scanned is less than 1000 bytes.
3655
3656 7. A lesser problem with the same optimization is that it was recording the
3657 first character of an anchored pattern as "needed", thus provoking a search
3658 right along the subject, even when the first match of the pattern was going to
3659 fail. The "needed" character is now not set for anchored patterns, unless it
3660 follows something in the pattern that is of non-fixed length. Thus, it still
3661 fulfils its original purpose of finding quick non-matches in cases of nested
3662 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3663
3664
3665 Version 4.0 17-Feb-03
3666 ---------------------
3667
3668 1. If a comment in an extended regex that started immediately after a meta-item
3669 extended to the end of string, PCRE compiled incorrect data. This could lead to
3670 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3671
3672 2. Moved to autoconf 2.53 and libtool 1.4.2.
3673
3674 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3675 the special perltest8 script is no longer needed - all the tests can be run
3676 from a single perltest script.
3677
3678 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3679 by \s. It has now been removed in PCRE. This means it isn't recognized as
3680 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3681 class [:space:] *does* include VT, thereby creating a mess.
3682
3683 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3684 space and tab.
3685
3686 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3687 its new features into the main test script, reducing the number of scripts.
3688
3689 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3690 were backward compatible, and made the (?i) apply to the whole pattern, as if
3691 /i were given. Now it behaves more logically, and applies the option setting
3692 only to what follows. PCRE has been changed to follow suit. However, if it
3693 finds options settings right at the start of the pattern, it extracts them into
3694 the global options, as before. Thus, they show up in the info data.
3695
3696 8. Added support for the \Q...\E escape sequence. Characters in between are
3697 treated as literals. This is slightly different from Perl in that $ and @ are
3698 also handled as literals inside the quotes. In Perl, they will cause variable
3699 interpolation. Note the following examples:
3700
3701 Pattern PCRE matches Perl matches
3702
3703 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3704 \Qabc\$xyz\E abc\$xyz abc\$xyz
3705 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3706
3707 For compatibility with Perl, \Q...\E sequences are recognized inside character
3708 classes as well as outside them.
3709
3710 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3711 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3712 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3713 signed/unsigned warnings.
3714
3715 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3716 option for pcretest, so I've replaced it by a simple function that does just
3717 that job.
3718
3719 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3720 "pcregrep -".
3721
3722 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3723 Java package. This provides some syntactic sugar for simple cases of what my
3724 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3725 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3726 item, you can use this simplified notation. Note that only makes sense with
3727 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3728 greediness, whatever the setting of the PCRE_UNGREEDY option.
3729
3730 13. A change of greediness default within a pattern was not taking effect at
3731 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3732 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3733 was abstracted outside.
3734
3735 14. PCRE now supports the \G assertion. It is true when the current matching
3736 position is at the start point of the match. This differs from \A when the
3737 starting offset is non-zero. Used with the /g option of pcretest (or similar
3738 code), it works in the same way as it does for Perl's /g option. If all
3739 alternatives of a regex begin with \G, the expression is anchored to the start
3740 match position, and the "anchored" flag is set in the compiled expression.
3741
3742 15. Some bugs concerning the handling of certain option changes within patterns
3743 have been fixed. These applied to options other than (?ims). For example,
3744 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3745 been the other way round. Some of this was related to change 7 above.
3746
3747 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3748 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3749 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3750 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3751
3752 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3753 mode. Unlike ".", it always matches newline, whatever the setting of
3754 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3755 assertions. Perl allows it, but it doesn't (in general) work because it can't
3756 calculate the length of the lookbehind. At least, that's the case for Perl
3757 5.8.0 - I've been told they are going to document that it doesn't work in
3758 future.
3759
3760 18. Added an error diagnosis for escapes that PCRE does not support: these are
3761 \L, \l, \N, \P, \p, \U, \u, and \X.
3762
3763 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3764 reading past the end of the pattern in cases such as /[abcd/.
3765
3766 20. PCRE was getting more memory than necessary for patterns with classes that
3767 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3768
3769 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3770 compiling PCRE for use with Virtual Pascal.
3771
3772 22. Small fix to the Makefile to make it work properly if the build is done
3773 outside the source tree.
3774
3775 23. Added a new extension: a condition to go with recursion. If a conditional
3776 subpattern starts with (?(R) the "true" branch is used if recursion has
3777 happened, whereas the "false" branch is used only at the top level.
3778
3779 24. When there was a very long string of literal characters (over 255 bytes
3780 without UTF support, over 250 bytes with UTF support), the computation of how
3781 much memory was required could be incorrect, leading to segfaults or other
3782 strange effects.
3783
3784 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3785 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3786 there was a subsequent back reference to those brackets. This meant that, for
3787 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3788 possible to check for precisely this case. All we can do is abandon the
3789 optimization if .* occurs inside capturing brackets when there are any back
3790 references whatsoever. (See below for a better fix that came later.)
3791
3792 26. The handling of the optimization for finding the first character of a
3793 non-anchored pattern, and for finding a character that is required later in the
3794 match were failing in some cases. This didn't break the matching; it just
3795 failed to optimize when it could. The way this is done has been re-implemented.
3796
3797 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3798
3799 28. Added a new feature that provides some of the functionality that Perl
3800 provides with (?{...}). The facility is termed a "callout". The way it is done
3801 in PCRE is for the caller to provide an optional function, by setting
3802 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3803 global variable. By default it is unset, which disables all calling out. To get
3804 the function called, the regex must include (?C) at appropriate points. This
3805 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3806 This provides a means of identifying different callout points. When PCRE
3807 reaches such a point in the regex, if pcre_callout has been set, the external
3808 function is called. It is provided with data in a structure called
3809 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3810 matching continues; if it returns a non-zero value, the match at the current
3811 point fails. However, backtracking will occur if possible. [This was changed
3812 later and other features added - see item 49 below.]
3813
3814 29. pcretest is upgraded to test the callout functionality. It provides a
3815 callout function that displays information. By default, it shows the start of
3816 the match and the current position in the text. There are some new data escapes
3817 to vary what happens:
3818
3819 \C+ in addition, show current contents of captured substrings
3820 \C- do not supply a callout function
3821 \C!n return 1 when callout number n is reached
3822 \C!n!m return 1 when callout number n is reached for the mth time
3823
3824 30. If pcregrep was called with the -l option and just a single file name, it
3825 output "<stdin>" if a match was found, instead of the file name.
3826
3827 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3828 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3829 pcre_exec(). This saves a malloc/free per call. The default value of
3830 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3831 when configuring.
3832
3833 32. The default maximum size of a compiled pattern is 64K. There have been a
3834 few cases of people hitting this limit. The code now uses macros to handle the
3835 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3836 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3837 configuring. Tests 2 and 5 work only with 2-byte links because they output
3838 debugging information about compiled patterns.
3839
3840 33. Internal code re-arrangements:
3841
3842 (a) Moved the debugging function for printing out a compiled regex into
3843 its own source file (printint.c) and used #include to pull it into
3844 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3845 separate copies.
3846
3847 (b) Defined the list of op-code names for debugging as a macro in
3848 internal.h so that it is next to the definition of the opcodes.
3849
3850 (c) Defined a table of op-code lengths for simpler skipping along compiled
3851 code. This is again a macro in internal.h so that it is next to the
3852 definition of the opcodes.
3853
3854 34. Added support for recursive calls to individual subpatterns, along the
3855 lines of Robin Houston's patch (but implemented somewhat differently).
3856
3857 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3858 allow it to read and process whole directories in Win32. This code was
3859 contributed by Lionel Fourquaux; it has not been tested by me.
3860
3861 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3862 used to name a group. Names consist of alphanumerics and underscores, and must
3863 be unique. Back references use the syntax (?P=name) and recursive calls use
3864 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3865 numbers. The function pcre_fullinfo() can be used after compilation to extract
3866 a name/number map. There are three relevant calls:
3867
3868 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3869 PCRE_INFO_NAMECOUNT yields the number of entries
3870 PCRE_INFO_NAMETABLE yields a pointer to the map.
3871
3872 The map is a vector of fixed-size entries. The size of each entry depends on
3873 the length of the longest name used. The first two bytes of each entry are the
3874 group number, most significant byte first. There follows the corresponding
3875 name, zero terminated. The names are in alphabetical order.
3876
3877 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3878 case instead of 255. Making it the same both with and without UTF-8 support
3879 means that the same test output works with both.
3880
3881 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3882 calling malloc() with a zero argument.
3883
3884 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3885 optimization. I've improved things by keeping a bitmap of backreferences with
3886 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3887 fact referenced, the optimization can be applied. It is unlikely that a
3888 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3889 the match to follow \n) will appear inside brackets with a number greater than
3890 31, but if it does, any back reference > 31 suppresses the optimization.
3891
3892 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3893 of disabling numbered capturing parentheses. Any opening parenthesis that is
3894 not followed by ? behaves as if it were followed by ?: but named parentheses
3895 can still be used for capturing (and they will acquire numbers in the usual
3896 way).
3897
3898 41. Redesigned the return codes from the match() function into yes/no/error so
3899 that errors can be passed back from deep inside the nested calls. A malloc
3900 failure while inside a recursive subpattern call now causes the
3901 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3902
3903 42. It is now possible to set a limit on the number of times the match()
3904 function is called in a call to pcre_exec(). This facility makes it possible to
3905 limit the amount of recursion and backtracking, though not in a directly
3906 obvious way, because the match() function is used in a number of different
3907 circumstances. The count starts from zero for each position in the subject
3908 string (for non-anchored patterns). The default limit is, for compatibility, a
3909 large number, namely 10 000 000. You can change this in two ways:
3910
3911 (a) When configuring PCRE before making, you can use --with-match-limit=n
3912 to set a default value for the compiled library.
3913
3914 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3915 a different value is set. See 45 below.
3916
3917 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3918
3919 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3920 of things that can be changed at compile time. The first argument specifies
3921 what is wanted and the second points to where the information is to be placed.
3922 The current list of available information is:
3923
3924 PCRE_CONFIG_UTF8
3925
3926 The output is an integer that is set to one if UTF-8 support is available;
3927 otherwise it is set to zero.
3928
3929 PCRE_CONFIG_NEWLINE
3930
3931 The output is an integer that it set to the value of the code that is used for
3932 newline. It is either LF (10) or CR (13).
3933
3934 PCRE_CONFIG_LINK_SIZE
3935
3936 The output is an integer that contains the number of bytes used for internal
3937 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3938
3939 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3940
3941 The output is an integer that contains the threshold above which the POSIX
3942 interface uses malloc() for output vectors. See item 31 above.
3943
3944 PCRE_CONFIG_MATCH_LIMIT
3945
3946 The output is an unsigned integer that contains the default limit of the number
3947 of match() calls in a pcre_exec() execution. See 42 above.
3948
3949 44. pcretest has been upgraded by the addition of the -C option. This causes it
3950 to extract all the available output from the new pcre_config() function, and to
3951 output it. The program then exits immediately.
3952
3953 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3954 order to support additional features. One way would have been to define
3955 pcre_exec2() (for example) with extra arguments, but this would not have been
3956 extensible, and would also have required all calls to the original function to
3957 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3958 is used for passing in "extra" data from pcre_study().
3959
3960 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3961 contains the following fields:
3962
3963 flags a bitmap indicating which of the following fields are set
3964 study_data opaque data from pcre_study()
3965 match_limit a way of specifying a limit on match() calls for a specific
3966 call to pcre_exec()
3967 callout_data data for callouts (see 49 below)
3968
3969 The flag bits are also defined in pcre.h, and are
3970
3971 PCRE_EXTRA_STUDY_DATA
3972 PCRE_EXTRA_MATCH_LIMIT
3973 PCRE_EXTRA_CALLOUT_DATA
3974
3975 The pcre_study() function now returns one of these new pcre_extra blocks, with
3976 the actual study data pointed to by the study_data field, and the
3977 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3978 before. That is, this change is entirely upwards-compatible and requires no
3979 change to existing code.
3980
3981 If you want to pass in additional data to pcre_exec(), you can either place it
3982 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3983 block.
3984
3985 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3986 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3987 times with different match limits, until it finds the minimum value needed for
3988 pcre_exec() to complete. The value is then output. This can be instructive; for
3989 most simple matches the number is quite small, but for pathological cases it
3990 gets very large very quickly.
3991
3992 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3993 returns the size of the data block pointed to by the study_data field in a
3994 pcre_extra block, that is, the value that was passed as the argument to
3995 pcre_malloc() when PCRE was getting memory in which to place the information
3996 created by pcre_study(). The fourth argument should point to a size_t variable.
3997 pcretest has been extended so that this information is shown after a successful
3998 pcre_study() call when information about the compiled regex is being displayed.
3999
4000 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4001 because what follows is always an absolute path. (Later: it turns out that this
4002 is more than cosmetic for MinGW, because it doesn't like empty path
4003 components.)
4004
4005 49. Some changes have been made to the callout feature (see 28 above):
4006
4007 (i) A callout function now has three choices for what it returns:
4008
4009 0 => success, carry on matching
4010 > 0 => failure at this point, but backtrack if possible
4011 < 0 => serious error, return this value from pcre_exec()
4012
4013 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4014 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4015 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4016 use by callout functions. It will never be used by PCRE itself.
4017
4018 (ii) The pcre_extra structure (see 45 above) has a void * field called
4019 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4020 pcre_callout_block structure has a field of the same name. The contents of
4021 the field passed in the pcre_extra structure are passed to the callout
4022 function in the corresponding field in the callout block. This makes it
4023 easier to use the same callout-containing regex from multiple threads. For
4024 testing, the pcretest program has a new data escape
4025
4026 \C*n pass the number n (may be negative) as callout_data
4027
4028 If the callout function in pcretest receives a non-zero value as
4029 callout_data, it returns that value.
4030
4031 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4032 there were some redundant $(CFLAGS) in commands that are now specified as
4033 $(LINK), which already includes $(CFLAGS).
4034
4035 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4036 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4037 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4038 one-byte characters throughout. Note that case-insensitive matching applies
4039 only to characters whose values are less than 256. PCRE doesn't support the
4040 notion of cases for higher-valued characters.
4041
4042 (i) A character class whose characters are all within 0-255 is handled as
4043 a bit map, and the map is inverted for negative classes. Previously, a
4044 character > 255 always failed to match such a class; however it should
4045 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4046
4047 (ii) A negated character class with a single character < 255 is coded as
4048 "not this character" (OP_NOT). This wasn't working properly when the test
4049 character was multibyte, either singly or repeated.
4050
4051 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4052 mode, for example: \x{100}{2,3}.
4053
4054 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4055 singly or repeated) now correctly test multibyte characters. However,
4056 PCRE doesn't recognize any characters with values greater than 255 as
4057 digits, spaces, or word characters. Such characters always match \D, \S,
4058 and \W, and never match \d, \s, or \w.
4059
4060 (v) Classes may now contain characters and character ranges with values
4061 greater than 255. For example: [ab\x{100}-\x{400}].
4062
4063 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4064 PCRE in UTF-8 mode.
4065
4066 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4067 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4068 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4069 value.)
4070
4071 53. The single man page has become too large. I have therefore split it up into
4072 a number of separate man pages. These also give rise to individual HTML pages;
4073 these are now put in a separate directory, and there is an index.html page that
4074 lists them all. Some hyperlinking between the pages has been installed.
4075
4076 54. Added convenience functions for handling named capturing parentheses.
4077
4078 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4079 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4080 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4081 are faulted.
4082
4083 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4084 calling configure. These values are used when compiling the dftables.c program
4085 which is run to generate the source of the default character tables. They
4086 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4087 you will need to set these values.
4088
4089 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4090
4091
4092 Version 3.9 02-Jan-02
4093 ---------------------
4094
4095 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4096
4097 2. If --disable-static was given, the building process failed when trying to
4098 build pcretest and pcregrep. (For some reason it was using libtool to compile
4099 them, which is not right, as they aren't part of the library.)
4100
4101
4102 Version 3.8 18-Dec-01
4103 ---------------------
4104
4105 1. The experimental UTF-8 code was completely screwed up. It was packing the
4106 bytes in the wrong order. How dumb can you get?
4107
4108
4109 Version 3.7 29-Oct-01
4110 ---------------------
4111
4112 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4113 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4114 this didn't happen under Solaris 8, where I normally test things.
4115
4116 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4117 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4118 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4119 this fix an hour or so after the initial 3.7 release.)
4120
4121
4122 Version 3.6 23-Oct-01
4123 ---------------------
4124
4125 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4126 offsets passed as NULL with zero offset count.
4127
4128 2. The config.guess and config.sub files had not been updated when I moved to
4129 the latest autoconf.
4130
4131
4132 Version 3.5 15-Aug-01
4133 ---------------------
4134
4135 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4136 had been forgotten.
4137
4138 2. By using declared but undefined structures, we can avoid using "void"
4139 definitions in pcre.h while keeping the internal definitions of the structures
4140 private.
4141
4142 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4143 user point of view, this means that both static and shared libraries are built
4144 by default, but this can be individually controlled. More of the work of
4145 handling this static/shared cases is now inside libtool instead of PCRE's make
4146 file.
4147
4148 4. The pcretest utility is now installed along with pcregrep because it is
4149 useful for users (to test regexs) and by doing this, it automatically gets
4150 relinked by libtool. The documentation has been turned into a man page, so
4151 there are now .1, .txt, and .html versions in /doc.
4152
4153 5. Upgrades to pcregrep:
4154 (i) Added long-form option names like gnu grep.
4155 (ii) Added --help to list all options with an explanatory phrase.
4156 (iii) Added -r, --recursive to recurse into sub-directories.
4157 (iv) Added -f, --file to read patterns from a file.
4158
4159 6. pcre_exec() was referring to its "code" argument before testing that
4160 argument for NULL (and giving an error if it was NULL).
4161
4162 7. Upgraded Makefile.in to allow for compiling in a different directory from
4163 the source directory.
4164
4165 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4166 options bits, the pointer it was passed was to an int instead of to an unsigned
4167 long int. This mattered only on 64-bit systems.
4168
4169 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4170 generated) instead of pcre.in, which it its source. Also made the same change
4171 in several of the .c files.
4172
4173 10. A new release of gcc defines printf() as a macro, which broke pcretest
4174 because it had an ifdef in the middle of a string argument for printf(). Fixed
4175 by using separate calls to printf().
4176
4177 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4178 script, to force use of CR or LF instead of \n in the source. On non-Unix
4179 systems, the value can be set in config.h.
4180
4181 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4182 absolute limit. Changed the text of the error message to make this clear, and
4183 likewise updated the man page.
4184
4185 13. The limit of 99 on the number of capturing subpatterns has been removed.
4186 The new limit is 65535, which I hope will not be a "real" limit.
4187
4188
4189 Version 3.4 22-Aug-00
4190 ---------------------
4191
4192 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4193
4194 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4195
4196
4197 Version 3.3 01-Aug-00
4198 ---------------------
4199
4200 1. If an octal character was given, but the value was greater than \377, it
4201 was not getting masked to the least significant bits, as documented. This could
4202 lead to crashes in some systems.
4203
4204 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4205 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4206
4207 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4208 These just pass their arguments on to (pcre_free)(), but they are provided
4209 because some uses of PCRE bind it to non-C systems that can call its functions,
4210 but cannot call free() or pcre_free() directly.
4211
4212 4. Add "make test" as a synonym for "make check". Corrected some comments in
4213 the Makefile.
4214
4215 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4216 Makefile.
4217
4218 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4219 command called pgrep for grepping around the active processes.
4220
4221 7. Added the beginnings of support for UTF-8 character strings.
4222
4223 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4224 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4225 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4226 out for the ar command.)
4227
4228
4229 Version 3.2 12-May-00
4230 ---------------------
4231
4232 This is purely a bug fixing release.
4233
4234 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4235 of ZA. This was just one example of several cases that could provoke this bug,
4236 which was introduced by change 9 of version 2.00. The code for breaking
4237 infinite loops after an iteration that matches an empty string was't working
4238 correctly.
4239
4240 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4241 when matched against abbab (for example). After matching an empty string, it
4242 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4243 caused it to match further down the string than it should.
4244
4245 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4246 was there because it doesn't seem to be needed, and it causes trouble on some
4247 systems, as it is not a Standard C header. It has been removed.
4248
4249 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4250 were reported on the Macintosh. The changes were from
4251
4252 while ((c = *(++ptr)) != 0 && c != '\n');
4253 to
4254 while ((c = *(++ptr)) != 0 && c != '\n') ;
4255
4256 Totally extraordinary, but if that's what it takes...
4257
4258 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4259 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4260 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4261 assumes the way PCRE uses memmove() (always moving upwards).
4262
4263 6. PCRE is being used in one environment where strchr() is not available. There
4264 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4265 faster code anyway.
4266
4267
4268 Version 3.1 09-Feb-00
4269 ---------------------
4270
4271 The only change in this release is the fixing of some bugs in Makefile.in for
4272 the "install" target:
4273
4274 (1) It was failing to install pcreposix.h.
4275
4276 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4277
4278
4279 Version 3.0 01-Feb-00
4280 ---------------------
4281
4282 1. Add support for the /+ modifier to perltest (to output $` like it does in
4283 pcretest).
4284
4285 2. Add support for the /g modifier to perltest.
4286
4287 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4288 matches null strings.
4289
4290 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4291 pattern. Perl treats empty patterns specially - it reuses the most recent
4292 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4293 effect.
4294
4295 5. The POSIX interface was broken in that it was just handing over the POSIX
4296 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4297 required a bigger vector, with some working space on the end. This means that
4298 the POSIX wrapper now has to get and free some memory, and copy the results.
4299
4300 6. Added some simple autoconf support, placing the test data and the
4301 documentation in separate directories, re-organizing some of the
4302 information files, and making it build pcre-config (a GNU standard). Also added
4303 libtool support for building PCRE as a shared library, which is now the
4304 default.
4305
4306 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4307 09 are not valid octal constants. Single digits will be used for minor values
4308 less than 10.
4309
4310 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4311 existing programs that set these in the POSIX interface can use PCRE without
4312 modification.
4313
4314 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4315 return all that pcre_info() returns, plus additional data. The pcre_info()
4316 function is retained for compatibility, but is considered to be obsolete.
4317
4318 10. Added experimental recursion feature (?R) to handle one common case that
4319 Perl 5.6 will be able to do with (?p{...}).
4320
4321 11. Added support for POSIX character classes like [:alpha:], which Perl is
4322 adopting.
4323
4324
4325 Version 2.08 31-Aug-99
4326 ----------------------
4327
4328 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4329 trying to match at the startoffset position, but instead was moving forward to
4330 the next newline as if a previous match had failed.
4331
4332 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4333 and could get into a loop if a null string was matched other than at the start
4334 of the subject.
4335
4336 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4337 be distinguished at compile time, and for completeness also added PCRE_DATE.
4338
4339 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4340 in GnuWin32 environments.
4341
4342
4343 Version 2.07 29-Jul-99
4344 ----------------------
4345
4346 1. The documentation is now supplied in plain text form and HTML as well as in
4347 the form of man page sources.
4348
4349 2. C++ compilers don't like assigning (void *) values to other pointer types.
4350 In particular this affects malloc(). Although there is no problem in Standard
4351 C, I've put in casts to keep C++ compilers happy.
4352
4353 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4354 should be (const char *).
4355
4356 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4357 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4358 However, I haven't made this a standard facility. The documentation doesn't
4359 mention it, and the Makefile doesn't support it.
4360
4361 5. The Makefile now contains an "install" target, with editable destinations at
4362 the top of the file. The pcretest program is not installed.
4363
4364 6. pgrep -V now gives the PCRE version number and date.
4365
4366 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4367 causing the entire string to be ignored, instead of just the last character.
4368
4369 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4370 non-matching string, it can take a very, very long time, even for strings of
4371 quite modest length, because of the nested recursion. PCRE now does better in
4372 some of these cases. It does this by remembering the last required literal
4373 character in the pattern, and pre-searching the subject to ensure it is present
4374 before running the real match. In other words, it applies a heuristic to detect
4375 some types of certain failure quickly, and in the above example, if presented
4376 with a string that has no trailing " it gives "no match" very quickly.
4377
4378 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4379 other alternatives are tried instead.
4380
4381
4382 Version 2.06 09-Jun-99
4383 ----------------------
4384
4385 1. Change pcretest's output for amount of store used to show just the code
4386 space, because the remainder (the data block) varies in size between 32-bit and
4387 64-bit systems.
4388
4389 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4390 start matching at. This allows lookbehinds to work when searching for multiple
4391 occurrences in a string.
4392
4393 3. Added additional options to pcretest for testing multiple occurrences:
4394
4395 /+ outputs the rest of the string that follows a match
4396 /g loops for multiple occurrences, using the new startoffset argument
4397 /G loops for multiple occurrences by passing an incremented pointer
4398
4399 4. PCRE wasn't doing the "first character" optimization for patterns starting
4400 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4401 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4402 the letter 'x'. On long subject strings, this gives a significant speed-up.
4403
4404
4405 Version 2.05 21-Apr-99
4406 ----------------------
4407
4408 1. Changed the type of magic_number from int to long int so that it works
4409 properly on 16-bit systems.
4410
4411 2. Fixed a bug which caused patterns starting with .* not to work correctly
4412 when the subject string contained newline characters. PCRE was assuming
4413 anchoring for such patterns in all cases, which is not correct because .* will
4414 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4415 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4416 must be retried after every newline in the subject.
4417
4418
4419 Version 2.04 18-Feb-99
4420 ----------------------
4421
4422 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4423 computation of the store needed to hold the pattern was incorrect (too large).
4424 If such patterns were nested a few deep, this could multiply and become a real
4425 problem.
4426
4427 2. Added /M option to pcretest to show the memory requirement of a specific
4428 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4429
4430 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4431 compiled in such a way that the backtracking after subsequent failure was
4432 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4433 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4434
4435
4436 Version 2.03 02-Feb-99
4437 ----------------------
4438
4439 1. Fixed typo and small mistake in man page.
4440
4441 2. Added 4th condition (GPL supersedes if conflict) and created separate
4442 LICENCE file containing the conditions.
4443
4444 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4445 Perl, that is the internal \ allows the delimiter to be included in the
4446 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4447 the final delimiter, add \ to the end of the pattern (to test the error).
4448
4449 4. Added the convenience functions for extracting substrings after a successful
4450 match. Updated pcretest to make it able to test these functions.
4451
4452
4453 Version 2.02 14-Jan-99
4454 ----------------------
4455
4456 1. Initialized the working variables associated with each extraction so that
4457 their saving and restoring doesn't refer to uninitialized store.
4458
4459 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4460 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4461 fix the problem.
4462
4463 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4464 calls, and wasn't printing the correct value for compiling calls. Increased the
4465 default value of LOOPREPEAT, and the number of significant figures in the
4466 times.
4467
4468 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4469
4470 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4471 a building problem on Windows NT with a FAT file system.
4472
4473
4474 Version 2.01 21-Oct-98
4475 ----------------------
4476
4477 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4478 to character tables built by pcre_maketables() in the current locale. If NULL
4479 is passed, the default tables are used.
4480
4481
4482 Version 2.00 24-Sep-98
4483 ----------------------
4484
4485 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4486 it any more.
4487
4488 2. Allow quantification of (?>) groups, and make it work correctly.
4489
4490 3. The first character computation wasn't working for (?>) groups.
4491
4492 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4493 end of the subject) and add 5.005's \z, which really does match only at the
4494 very end of the subject.
4495
4496 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4497
4498 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4499 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4500 localized options. All options to pcre_study() were also removed.
4501
4502 7. Add other new features from 5.005:
4503
4504 $(?<= positive lookbehind
4505 $(?<! negative lookbehind
4506 (?imsx-imsx) added the unsetting capability
4507 such a setting is global if at outer level; local otherwise
4508 (?imsx-imsx:) non-capturing groups with option setting
4509 (?(cond)re|re) conditional pattern matching
4510
4511 A backreference to itself in a repeated group matches the previous
4512 captured string.
4513
4514 8. General tidying up of studying (both automatic and via "study")
4515 consequential on the addition of new assertions.
4516
4517 9. As in 5.005, unlimited repeated groups that could match an empty substring
4518 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4519 runtime if any iteration does actually match an empty substring.
4520
4521 10. Include the RunTest script in the distribution.
4522
4523 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4524 discrepancies, some of which were old and were also with respect to 5.004. They
4525 have now been fixed.
4526
4527
4528 Version 1.09 28-Apr-98
4529 ----------------------
4530
4531 1. A negated single character class followed by a quantifier with a minimum
4532 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4533 program crashes, or just wrong answers. This did not apply to negated classes
4534 containing more than one character, or to minima other than one.
4535
4536
4537 Version 1.08 27-Mar-98
4538 ----------------------
4539
4540 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4541
4542 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4543 latter must appear before anything that relies on it in the pattern.
4544
4545
4546 Version 1.07 16-Feb-98
4547 ----------------------
4548
4549 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4550 repeat of a potentially empty string).
4551
4552
4553 Version 1.06 23-Jan-98
4554 ----------------------
4555
4556 1. Added Markus Oberhumer's little patches for C++.
4557
4558 2. Literal strings longer than 255 characters were broken.
4559
4560
4561 Version 1.05 23-Dec-97
4562 ----------------------
4563
4564 1. Negated character classes containing more than one character were failing if
4565 PCRE_CASELESS was set at run time.
4566
4567
4568 Version 1.04 19-Dec-97
4569 ----------------------
4570
4571 1. Corrected the man page, where some "const" qualifiers had been omitted.
4572
4573 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4574 input syntax.
4575
4576 3. Fixed memory leak which occurred when a regex with back references was
4577 matched with an offsets vector that wasn't big enough. The temporary memory
4578 that is used in this case wasn't being freed if the match failed.
4579
4580 4. Tidied pcretest to ensure it frees memory that it gets.
4581
4582 5. Temporary memory was being obtained in the case where the passed offsets
4583 vector was exactly big enough.
4584
4585 6. Corrected definition of offsetof() from change 5 below.
4586
4587 7. I had screwed up change 6 below and broken the rules for the use of
4588 setjmp(). Now fixed.
4589
4590
4591 Version 1.03 18-Dec-97
4592 ----------------------
4593
4594 1. A erroneous regex with a missing opening parenthesis was correctly
4595 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4596 on some systems.
4597
4598 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4599 it was reported that one broken compiler failed on the former because "code" is
4600 also an independent variable.
4601
4602 3. The erroneous regex a[]b caused an array overrun reference.
4603
4604 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4605 fail on data ending with that character. (It was going on too far, and checking
4606 the next character, typically a binary zero.) This was specific to the
4607 optimized code for single-character negative classes.
4608
4609 5. Added a contributed patch from the TIN world which does the following:
4610
4611 + Add an undef for memmove, in case the the system defines a macro for it.
4612
4613 + Add a definition of offsetof(), in case there isn't one. (I don't know
4614 the reason behind this - offsetof() is part of the ANSI standard - but
4615 it does no harm).
4616
4617 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4618 most of the places where whitespace preceded '#'. I have given up and
4619 allowed the remaining 2 cases to be at the margin.
4620
4621 + Rename some variables in pcre to eliminate shadowing. This seems very
4622 pedantic, but does no harm, of course.
4623
4624 6. Moved the call to setjmp() into its own function, to get rid of warnings
4625 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4626
4627 7. Constructs such as \d{8,} were compiling into the equivalent of
4628 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4629 outcome, but in this particular case used more store than had been allocated,
4630 which caused the bug to be discovered because it threw up an internal error.
4631
4632 8. The debugging code in both pcre and pcretest for outputting the compiled
4633 form of a regex was going wrong in the case of back references followed by
4634 curly-bracketed repeats.
4635
4636
4637 Version 1.02 12-Dec-97
4638 ----------------------
4639
4640 1. Typos in pcre.3 and comments in the source fixed.
4641
4642 2. Applied a contributed patch to get rid of places where it used to remove
4643 'const' from variables, and fixed some signed/unsigned and uninitialized
4644 variable warnings.
4645
4646 3. Added the "runtest" target to Makefile.
4647
4648 4. Set default compiler flag to -O2 rather than just -O.
4649
4650
4651 Version 1.01 19-Nov-97
4652 ----------------------
4653
4654 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4655 like /([ab]*)*/, that is, for classes with more than one character in them.
4656
4657 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4658 as /((?>a*))*/ (a PCRE_EXTRA facility).
4659
4660
4661 Version 1.00 18-Nov-97
4662 ----------------------
4663
4664 1. Added compile-time macros to support systems such as SunOS4 which don't have
4665 memmove() or strerror() but have other things that can be used instead.
4666
4667 2. Arranged that "make clean" removes the executables.
4668
4669
4670 Version 0.99 27-Oct-97
4671 ----------------------
4672
4673 1. Fixed bug in code for optimizing classes with only one character. It was
4674 initializing a 32-byte map regardless, which could cause it to run off the end
4675 of the memory it had got.
4676
4677 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4678
4679
4680 Version 0.98 22-Oct-97
4681 ----------------------
4682
4683 1. Fixed bug in code for handling temporary memory usage when there are more
4684 back references than supplied space in the ovector. This could cause segfaults.
4685
4686
4687 Version 0.97 21-Oct-97
4688 ----------------------
4689
4690 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4691
4692 2. Optimized negated single characters not to use a bit map.
4693
4694 3. Brought error texts together as macro definitions; clarified some of them;
4695 fixed one that was wrong - it said "range out of order" when it meant "invalid
4696 escape sequence".
4697
4698 4. Changed some char * arguments to const char *.
4699
4700 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4701
4702 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4703 pcretest.
4704
4705
4706 Version 0.96 16-Oct-97
4707 ----------------------
4708
4709 1. Added a simple "pgrep" utility to the distribution.
4710
4711 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4712 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4713 where "ddd" means "one or more decimal digits".
4714
4715 3. Fixed serious bug. If a pattern had a back reference, but the call to
4716 pcre_exec() didn't supply a large enough ovector to record the related
4717 identifying subpattern, the match always failed. PCRE now remembers the number
4718 of the largest back reference, and gets some temporary memory in which to save
4719 the offsets during matching if necessary, in order to ensure that
4720 backreferences always work.
4721
4722 4. Increased the compatibility with Perl in a number of ways:
4723
4724 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4725 to request this handling. The option can be set at compile or exec time.
4726
4727 (b) $ matches before a terminating newline by default; an option
4728 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4729 mode). The option can be set at compile or exec time.
4730
4731 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4732 the same as Perl's. If the decimal number it represents is less than 10
4733 or there aren't that many previous left capturing parentheses, an octal
4734 escape is read. Inside a character class, it's always an octal escape,
4735 even if it is a single digit.
4736
4737 (d) An escaped but undefined alphabetic character is taken as a literal,
4738 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4739 escapes.
4740
4741 (e) {0} is now permitted. (The previous item is removed from the compiled
4742 pattern).
4743
4744 5. Changed all the names of code files so that the basic parts are no longer
4745 than 10 characters, and abolished the teeny "globals.c" file.
4746
4747 6. Changed the handling of character classes; they are now done with a 32-byte
4748 bit map always.
4749
4750 7. Added the -d and /D options to pcretest to make it possible to look at the
4751 internals of compilation without having to recompile pcre.
4752
4753
4754 Version 0.95 23-Sep-97
4755 ----------------------
4756
4757 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4758 \x20 at the start of a run of normal characters. These were being treated as
4759 real characters, instead of the source characters being re-checked.
4760
4761
4762 Version 0.94 18-Sep-97
4763 ----------------------
4764
4765 1. The functions are now thread-safe, with the caveat that the global variables
4766 containing pointers to malloc() and free() or alternative functions are the
4767 same for all threads.
4768
4769 2. Get pcre_study() to generate a bitmap of initial characters for non-
4770 anchored patterns when this is possible, and use it if passed to pcre_exec().
4771
4772
4773 Version 0.93 15-Sep-97
4774 ----------------------
4775
4776 1. /(b)|(:+)/ was computing an incorrect first character.
4777
4778 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4779 but not actually doing anything yet.
4780
4781 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4782 as Perl does (e.g. [-az] or [az-]).
4783
4784 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4785 all possible positions.
4786
4787 5. Split up into different modules to avoid including unneeded functions in a
4788 compiled binary. However, compile and exec are still in one module. The "study"
4789 function is split off.
4790
4791 6. The character tables are now in a separate module whose source is generated
4792 by an auxiliary program - but can then be edited by hand if required. There are
4793 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4794 toupper() in the code.
4795
4796 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4797 make them global. Abolish the function for setting them, as the caller can now
4798 set them directly.
4799
4800
4801 Version 0.92 11-Sep-97
4802 ----------------------
4803
4804 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4805 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4806
4807 2. Caseless matching was not working in character classes if the characters in
4808 the pattern were in upper case.
4809
4810 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4811
4812 4. Make PCRE_ANCHORED public and accept as a compile option.
4813
4814 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4815 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4816 pass them.
4817
4818 6. Give an error if bad option bits passed at compile or run time.
4819
4820 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4821 pcretest to cause it to pass that flag.
4822
4823 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4824 options, and the first character, if set.
4825
4826 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4827
4828
4829 Version 0.91 10-Sep-97
4830 ----------------------
4831
4832 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4833 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4834
4835 2. PCRE was looping on encountering an indefinitely repeated back reference to
4836 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4837 Perl does - treats the match as successful.
4838
4839 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5