ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 977 - (show annotations)
Sun Jun 17 06:20:52 2012 UTC (9 years, 3 months ago) by zherczeg
File size: 231499 byte(s)
Optimizing single character iterators in JIT
1 ChangeLog for PCRE
2 ------------------
4 Version 8.31 02-June-2012
5 -------------------------
7 1. Fixing a wrong JIT test case and some compiler warnings.
9 2. Removed a bashism from the RunTest script.
11 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12 to unsigned type, result still unsigned" that was given by an MS compiler
13 on encountering the code "-sizeof(xxx)".
15 4. Partial matching support is added to the JIT compiler.
17 5. Fixed several bugs concerned with partial matching of items that consist
18 of more than one character:
20 (a) /^(..)\1/ did not partially match "aba" because checking references was
21 done on an "all or nothing" basis. This also applied to repeated
22 references.
24 (b) \R did not give a hard partial match if \r was found at the end of the
25 subject.
27 (c) \X did not give a hard partial match after matching one or more
28 characters at the end of the subject.
30 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31 a partial match for the string "\r".
33 (e) When newline was set to CRLF, the metacharacter "." did not recognize
34 a partial match for a CR character at the end of the subject string.
36 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37 running pcretest, the text "(JIT)" added to the output whenever JIT is
38 actually used to run the match.
40 7. Individual JIT compile options can be set in pcretest by following -s+[+]
41 or /S+[+] with a digit between 1 and 7.
43 8. OP_NOT now supports any UTF character not just single-byte ones.
45 9. (*MARK) control verb is now supported by the JIT compiler.
47 10. The command "./RunTest list" lists the available tests without actually
48 running any of them. (Because I keep forgetting what they all are.)
52 12. Applied a (slightly modified) user-supplied patch that improves performance
53 when the heap is used for recursion (compiled with --disable-stack-for-
54 recursion). Instead of malloc and free for each heap frame each time a
55 logical recursion happens, frames are retained on a chain and re-used where
56 possible. This sometimes gives as much as 30% improvement.
58 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59 call.
61 14. As documented, (*COMMIT) is now confined to within a positive assertion.
63 15. It is now possible to link pcretest with libedit as an alternative to
64 libreadline.
66 16. (*COMMIT) control verb is now supported by the JIT compiler.
68 17. The Unicode data tables have been updated to Unicode 6.1.0.
70 18. Added --file-list option to pcregrep.
72 19. Added binary file support to pcregrep, including the -a, --binary-files,
73 -I, and --text options.
75 20. The madvise function is renamed for posix_madvise for QNX compatibility
76 reasons. Fixed by Giuseppe D'Angelo.
78 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
80 22. Changed the option for creating symbolic links for 16-bit man pages from
81 -s to -sf so that re-installing does not cause issues.
83 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
85 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86 matches in certain environments (the workspace was not being correctly
87 retained). Also added to pcre_dfa_exec() a simple plausibility check on
88 some of the workspace data at the beginning of a restart.
90 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
91 was not doing so when it should - probably a typo introduced by SVN 528
92 (change 8.10/14).
94 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
95 \w+ when the character tables indicated that \x{c4} was a word character.
96 There were several related cases, all because the tests for doing a table
97 lookup were testing for characters less than 127 instead of 255.
99 27. If a pattern contains capturing parentheses that are not used in a match,
100 their slots in the ovector are set to -1. For those that are higher than
101 any matched groups, this happens at the end of processing. In the case when
102 there were back references that the ovector was too small to contain
103 (causing temporary malloc'd memory to be used during matching), and the
104 highest capturing number was not used, memory off the end of the ovector
105 was incorrectly being set to -1. (It was using the size of the temporary
106 memory instead of the true size.)
108 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
109 ovector size, it uses memory at the end of the block that it has got.
111 29. Check for an overlong MARK name and give an error at compile time. The
112 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
114 30. JIT compiler update.
116 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
117 Rill for the patch.
119 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
121 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
123 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
124 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
125 was enabled.
127 35. Improve JIT code generation for greedy plus quantifier.
129 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
130 1 to "aa" instead of to an empty string. The bug affected repeated groups
131 that could potentially match an empty string.
133 37. Optimizing single character iterators in JIT.
136 Version 8.30 04-February-2012
137 -----------------------------
139 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
140 name is defined in ctype.h.
142 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
143 only in quite long subpatterns.
145 3. Removed the function pcre_info(), which has been obsolete and deprecated
146 since it was replaced by pcre_fullinfo() in February 2000.
148 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
149 match a (*MARK), and the match failed at the start of the subject, a
150 reference to memory before the start of the subject could occur. This bug
151 was introduced by fix 17 of release 8.21.
153 5. A reference to an unset group with zero minimum repetition was giving
154 totally wrong answers (in non-JavaScript-compatibility mode). For example,
155 /(another)?(\1?)test/ matched against "hello world test". This bug was
156 introduced in release 8.13.
158 6. Add support for 16-bit character strings (a large amount of work involving
159 many changes and refactorings).
161 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
162 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
163 from a file.
165 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
166 rounding is not applied in this particular case).
168 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
169 if they appear, or are escaped, in patterns.
171 10. Get rid of a number of -Wunused-but-set-variable warnings.
173 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
174 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
175 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
176 also returns the mark "x". This bug applied to capturing parentheses,
177 non-capturing parentheses, and atomic parentheses. It also applied to some
178 assertions.
180 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
181 information out of configure.ac instead of relying on pcre.h.generic, which
182 is not stored in the repository.
184 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
185 -lreadline.
187 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
189 15. Applied Graycode's patch to put the top-level frame on the stack rather
190 than the heap when not using the stack for recursion. This gives a
191 performance improvement in many cases when recursion is not deep.
193 16. Experimental code added to "pcretest -C" to output the stack frame size.
196 Version 8.21 12-Dec-2011
197 ------------------------
199 1. Updating the JIT compiler.
201 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
202 are added as well.
204 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
205 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
206 calling _pcre_jit_exec. Some extra comments are added.
208 4. (*MARK) settings inside atomic groups that do not contain any capturing
209 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
210 was introduced by change 18 for 8.20.
212 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
213 ECMA-262 standard.
215 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
216 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
217 This bug was probably introduced by change 9 of 8.13.
219 7. While fixing 6 above, I noticed that a number of other items were being
220 incorrectly rejected as "not fixed length". This arose partly because newer
221 opcodes had not been added to the fixed-length checking code. I have (a)
222 corrected the bug and added tests for these items, and (b) arranged for an
223 error to occur if an unknown opcode is encountered while checking for fixed
224 length instead of just assuming "not fixed length". The items that were
225 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
226 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
227 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
229 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
230 being incorrectly compiled and would have given unpredicatble results.
232 9. A possessively repeated subpattern with minimum repeat count greater than
233 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
234 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
235 the first (A) could occur when it should not.
237 10. Add a cast and remove a redundant test from the code.
239 11. JIT should use pcre_malloc/pcre_free for allocation.
241 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
242 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
243 is anything other than /usr, -L is still shown).
245 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
247 14. Perl does not support \N without a following name in a [] class; PCRE now
248 also gives an error.
250 15. If a forward reference was repeated with an upper limit of around 2000,
251 it caused the error "internal error: overran compiling workspace". The
252 maximum number of forward references (including repeats) was limited by the
253 internal workspace, and dependent on the LINK_SIZE. The code has been
254 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
255 the default depends on LINK_SIZE. There is a new upper limit (for safety)
256 of around 200,000 forward references. While doing this, I also speeded up
257 the filling in of repeated forward references.
259 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
260 incorrectly expecting the subject to contain another "a" after the start.
262 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
263 in the match, the SKIP should be ignored. This was not happening; instead
264 the SKIP was being treated as NOMATCH. For patterns such as
265 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
266 tested.
268 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
269 now much more compatible with Perl, in particular in cases where the result
270 is a non-match for a non-anchored pattern. For example, if
271 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
272 "m", where previously it did not return a name. A side effect of this
273 change is that for partial matches, the last encountered mark name is
274 returned, as for non matches. A number of tests that were previously not
275 Perl-compatible have been moved into the Perl-compatible test files. The
276 refactoring has had the pleasing side effect of removing one argument from
277 the match() function, thus reducing its stack requirements.
279 19. If the /S+ option was used in pcretest to study a pattern using JIT,
280 subsequent uses of /S (without +) incorrectly behaved like /S+.
282 21. Retrieve executable code size support for the JIT compiler and fixing
283 some warnings.
285 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
286 not work when the shorter character appeared right at the end of the
287 subject string.
289 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
290 systems.
292 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
293 output it when the /M option is used in pcretest.
295 25. The CheckMan script was not being included in the distribution. Also, added
296 an explicit "perl" to run Perl scripts from the PrepareRelease script
297 because this is reportedly needed in Windows.
299 26. If study data was being save in a file and studying had not found a set of
300 "starts with" bytes for the pattern, the data written to the file (though
301 never used) was taken from uninitialized memory and so caused valgrind to
302 complain.
304 27. Updated RunTest.bat as provided by Sheri Pierce.
306 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
308 29. Computation of memory usage for the table of capturing group names was
309 giving an unnecessarily large value.
312 Version 8.20 21-Oct-2011
313 ------------------------
315 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
316 a POSIX class. After further experiments with Perl, which convinced me that
317 Perl has bugs and confusions, a closing square bracket is no longer allowed
318 in a POSIX name. This bug also affected patterns with classes that started
319 with full stops.
321 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
322 captured substring, but while checking the failing first alternative,
323 substring 1 is temporarily captured. If the output vector supplied to
324 pcre_exec() was not big enough for this capture, the yield of the function
325 was still zero ("insufficient space for captured substrings"). This cannot
326 be totally fixed without adding another stack variable, which seems a lot
327 of expense for a edge case. However, I have improved the situation in cases
328 such as /(a)(b)x|abc/ matched against "abc", where the return code
329 indicates that fewer than the maximum number of slots in the ovector have
330 been set.
332 3. Related to (2) above: when there are more back references in a pattern than
333 slots in the output vector, pcre_exec() uses temporary memory during
334 matching, and copies in the captures as far as possible afterwards. It was
335 using the entire output vector, but this conflicts with the specification
336 that only 2/3 is used for passing back captured substrings. Now it uses
337 only the first 2/3, for compatibility. This is, of course, another edge
338 case.
340 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
341 main code base, and can be used by building with --enable-jit. When this is
342 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
343 runtime --no-jit option is given.
345 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
346 ovector, the return from the function was zero, implying that there were
347 other matches that did not fit. The correct "exactly full" value is now
348 returned.
350 6. If a subpattern that was called recursively or as a subroutine contained
351 (*PRUNE) or any other control that caused it to give a non-standard return,
352 invalid errors such as "Error -26 (nested recursion at the same subject
353 position)" or even infinite loops could occur.
355 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
356 computing the minimum length on reaching *ACCEPT, and so ended up with the
357 wrong value of 1 rather than 0. Further investigation indicates that
358 computing a minimum subject length in the presence of *ACCEPT is difficult
359 (think back references, subroutine calls), and so I have changed the code
360 so that no minimum is registered for a pattern that contains *ACCEPT.
362 8. If (*THEN) was present in the first (true) branch of a conditional group,
363 it was not handled as intended. [But see 16 below.]
365 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
366 Sheri Pierce.
368 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
369 the first byte in a match must be "a".
371 11. Change 17 for 8.13 increased the recursion depth for patterns like
372 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
373 pattern contains any instances of (*THEN). If it does not, the old
374 optimizations are restored. It would be nice to do this on a per-group
375 basis, but at the moment that is not feasible.
377 12. In some environments, the output of pcretest -C is CRLF terminated. This
378 broke RunTest's code that checks for the link size. A single white space
379 character after the value is now allowed for.
381 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
382 For "fr", it uses the Windows-specific input and output files.
384 14. If (*THEN) appeared in a group that was called recursively or as a
385 subroutine, it did not work as intended. [But see next item.]
387 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
388 pattern fragments (but not containing any | characters). If A and B are
389 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
390 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
391 D. In other words, Perl considers parentheses that do not contain any |
392 characters to be part of a surrounding alternative, whereas PCRE was
393 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
394 differently. PCRE now behaves in the same way as Perl, except in the case
395 of subroutine/recursion calls such as (?1) which have in any case always
396 been different (but PCRE had them first :-).
398 16. Related to 15 above: Perl does not treat the | in a conditional group as
399 creating alternatives. Such a group is treated in the same way as an
400 ordinary group without any | characters when processing (*THEN). PCRE has
401 been changed to match Perl's behaviour.
403 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
404 RunGrepTest script failed.
406 18. Change 22 for version 13 caused atomic groups to use more stack. This is
407 inevitable for groups that contain captures, but it can lead to a lot of
408 stack use in large patterns. The old behaviour has been restored for atomic
409 groups that do not contain any capturing parentheses.
411 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
412 suppress the check for a minimum subject length at run time. (If it was
413 given to pcre_exec() or pcre_dfa_exec() it did work.)
415 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
416 fail to work when decoding hex characters in data strings in EBCDIC
417 environments.
419 21. It appears that in at least one Mac OS environment, the isxdigit() function
420 is implemented as a macro that evaluates to its argument more than once,
421 contravening the C 90 Standard (I haven't checked a later standard). There
422 was an instance in pcretest which caused it to go wrong when processing
423 \x{...} escapes in subject strings. The has been rewritten to avoid using
424 things like p++ in the argument of isxdigit().
427 Version 8.13 16-Aug-2011
428 ------------------------
430 1. The Unicode data tables have been updated to Unicode 6.0.0.
432 2. Two minor typos in pcre_internal.h have been fixed.
434 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
435 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
436 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
438 4. There were a number of related bugs in the code for matching backrefences
439 caselessly in UTF-8 mode when codes for the characters concerned were
440 different numbers of bytes. For example, U+023A and U+2C65 are an upper
441 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
442 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
443 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
444 2-byte code at the end of the subject (it thought there wasn't enough data
445 left).
447 5. Comprehensive information about what went wrong is now returned by
448 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
449 as the output vector has at least 2 elements. The offset of the start of
450 the failing character and a reason code are placed in the vector.
452 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
453 now returned is for the first byte of the failing character, instead of the
454 last byte inspected. This is an incompatible change, but I hope it is small
455 enough not to be a problem. It makes the returned offset consistent with
456 pcre_exec() and pcre_dfa_exec().
458 7. pcretest now gives a text phrase as well as the error number when
459 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
460 failure, the offset and reason code are output.
462 8. When \R was used with a maximizing quantifier it failed to skip backwards
463 over a \r\n pair if the subsequent match failed. Instead, it just skipped
464 back over a single character (\n). This seems wrong (because it treated the
465 two characters as a single entity when going forwards), conflicts with the
466 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
467 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
468 has been changed.
470 9. Some internal refactoring has changed the processing so that the handling
471 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
472 time (the PCRE_DOTALL option was changed this way some time ago: version
473 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
474 which was always a bit of a fudge. It also means that there is one less
475 argument for the match() function, which reduces its stack requirements
476 slightly. This change also fixes an incompatibility with Perl: the pattern
477 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
479 10. More internal refactoring has drastically reduced the number of recursive
480 calls to match() for possessively repeated groups such as (abc)++ when
481 using pcre_exec().
483 11. While implementing 10, a number of bugs in the handling of groups were
484 discovered and fixed:
486 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
487 (a|)*(?1) gave a compile-time internal error.
488 ((a|)+)+ did not notice that the outer group could match an empty string.
489 (^a|^)+ was not marked as anchored.
490 (.*a|.*)+ was not marked as matching at start or after a newline.
492 12. Yet more internal refactoring has removed another argument from the match()
493 function. Special calls to this function are now indicated by setting a
494 value in a variable in the "match data" data block.
496 13. Be more explicit in pcre_study() instead of relying on "default" for
497 opcodes that mean there is no starting character; this means that when new
498 ones are added and accidentally left out of pcre_study(), testing should
499 pick them up.
501 14. The -s option of pcretest has been documented for ages as being an old
502 synonym of -m (show memory usage). I have changed it to mean "force study
503 for every regex", that is, assume /S for every regex. This is similar to -i
504 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
505 using it. It makes it easier to run collections of tests with and without
506 study enabled, and thereby test pcre_study() more easily. All the standard
507 tests are now run with and without -s (but some patterns can be marked as
508 "never study" - see 20 below).
510 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
511 restoration of the capturing data to the outer values was not happening
512 correctly.
514 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
515 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
516 pattern had matched an empty string, and so incorrectly returned a no
517 match.
519 17. There was optimizing code for the last branch of non-capturing parentheses,
520 and also for the obeyed branch of a conditional subexpression, which used
521 tail recursion to cut down on stack usage. Unfortunately, now that there is
522 the possibility of (*THEN) occurring in these branches, tail recursion is
523 no longer possible because the return has to be checked for (*THEN). These
524 two optimizations have therefore been removed. [But see 8.20/11 above.]
526 18. If a pattern containing \R was studied, it was assumed that \R always
527 matched two bytes, thus causing the minimum subject length to be
528 incorrectly computed because \R can also match just one byte.
530 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
531 was incorrectly computed.
533 20. If /S is present twice on a test pattern in pcretest input, it now
534 *disables* studying, thereby overriding the use of -s on the command line
535 (see 14 above). This is necessary for one or two tests to keep the output
536 identical in both cases.
538 21. When (*ACCEPT) was used in an assertion that matched an empty string and
539 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
541 22. When an atomic group that contained a capturing parenthesis was
542 successfully matched, but the branch in which it appeared failed, the
543 capturing was not being forgotten if a higher numbered group was later
544 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
545 group 1 to "a", when in fact it should be unset. This applied to multi-
546 branched capturing and non-capturing groups, repeated or not, and also to
547 positive assertions (capturing in negative assertions does not happen
548 in PCRE) and also to nested atomic groups.
550 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
551 subject after a captured substring, to make it easier to tell which of a
552 number of identical substrings has been captured.
554 24. The way atomic groups are processed by pcre_exec() has been changed so that
555 if they are repeated, backtracking one repetition now resets captured
556 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
557 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
558 "aaa". Previously, it would have been "a". As part of this code
559 refactoring, the way recursive calls are handled has also been changed.
561 25. If an assertion condition captured any substrings, they were not passed
562 back unless some other capturing happened later. For example, if
563 (?(?=(a))a) was matched against "a", no capturing was returned.
565 26. When studying a pattern that contained subroutine calls or assertions,
566 the code for finding the minimum length of a possible match was handling
567 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
568 group 1 called group 2 while simultaneously a separate group 2 called group
569 1). A stack overflow occurred in this case. I have fixed this by limiting
570 the recursion depth to 10.
572 27. Updated RunTest.bat in the distribution to the version supplied by Tom
573 Fortmann. This supports explicit test numbers on the command line, and has
574 argument validation and error reporting.
576 28. An instance of \X with an unlimited repeat could fail if at any point the
577 first character it looked at was a mark character.
579 29. Some minor code refactoring concerning Unicode properties and scripts
580 should reduce the stack requirement of match() slightly.
582 30. Added the '=' option to pcretest to check the setting of unused capturing
583 slots at the end of the pattern, which are documented as being -1, but are
584 not included in the return count.
586 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
587 compiled something random. Now it gives a compile-time error (as does
588 Perl).
590 32. A *MARK encountered during the processing of a positive assertion is now
591 recorded and passed back (compatible with Perl).
593 33. If --only-matching or --colour was set on a pcregrep call whose pattern
594 had alternative anchored branches, the search for a second match in a line
595 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
596 matched the line "0102" twice. The same bug affected patterns that started
597 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
598 twice.
600 34. Previously, PCRE did not allow quantification of assertions. However, Perl
601 does, and because of capturing effects, quantifying parenthesized
602 assertions may at times be useful. Quantifiers are now allowed for
603 parenthesized assertions.
605 35. A minor code tidy in pcre_compile() when checking options for \R usage.
607 36. \g was being checked for fancy things in a character class, when it should
608 just be a literal "g".
610 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
611 appearance of a nested POSIX class supersedes an apparent external class.
612 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
613 unescaped square brackets may also appear as part of class names. For
614 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
615 more like Perl. (But see 8.20/1 above.)
617 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
618 was because it thought it was \N{name}, which is not supported).
620 39. Add minix to OS list not supporting the -S option in pcretest.
622 40. PCRE tries to detect cases of infinite recursion at compile time, but it
623 cannot analyze patterns in sufficient detail to catch mutual recursions
624 such as ((?1))((?2)). There is now a runtime test that gives an error if a
625 subgroup is called recursively as a subpattern for a second time at the
626 same position in the subject string. In previous releases this might have
627 been caught by the recursion limit, or it might have run out of stack.
629 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
630 happen only once. PCRE was, however incorrectly giving a compile time error
631 "recursive call could loop indefinitely" because it cannot analyze the
632 pattern in sufficient detail. The compile time test no longer happens when
633 PCRE is compiling a conditional subpattern, but actual runaway loops are
634 now caught at runtime (see 40 above).
636 42. It seems that Perl allows any characters other than a closing parenthesis
637 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
638 has been changed to be the same.
640 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
641 as not to get warnings when autogen.sh is called. Also changed
642 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
644 44. To help people who use pcregrep to scan files containing exceedingly long
645 lines, the following changes have been made:
647 (a) The default value of the buffer size parameter has been increased from
648 8K to 20K. (The actual buffer used is three times this size.)
650 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
651 PCRE is built.
653 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
654 to be set at run time.
656 (d) Numerical values in pcregrep options can be followed by K or M, for
657 example --buffer-size=50K.
659 (e) If a line being scanned overflows pcregrep's buffer, an error is now
660 given and the return code is set to 2.
662 45. Add a pointer to the latest mark to the callout data block.
664 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
665 partial match of an empty string instead of no match. This was specific to
666 the use of ".".
668 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
669 complete match instead of a partial match. This bug was dependent on both
670 the PCRE_UTF8 and PCRE_DOTALL options being set.
672 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
673 starting byte set, because \b was not being ignored.
676 Version 8.12 15-Jan-2011
677 ------------------------
679 1. Fixed some typos in the markup of the man pages, and wrote a script that
680 checks for such things as part of the documentation building process.
682 2. On a big-endian 64-bit system, pcregrep did not correctly process the
683 --match-limit and --recursion-limit options (added for 8.11). In
684 particular, this made one of the standard tests fail. (The integer value
685 went into the wrong half of a long int.)
687 3. If the --colour option was given to pcregrep with -v (invert match), it
688 did strange things, either producing crazy output, or crashing. It should,
689 of course, ignore a request for colour when reporting lines that do not
690 match.
692 4. Another pcregrep bug caused similar problems if --colour was specified with
693 -M (multiline) and the pattern match finished with a line ending.
695 5. In pcregrep, when a pattern that ended with a literal newline sequence was
696 matched in multiline mode, the following line was shown as part of the
697 match. This seems wrong, so I have changed it.
699 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
700 the check for further matches in the same line (so they could be coloured)
701 to overrun the end of the current line. If another match was found, it was
702 incorrectly shown (and then shown again when found in the next line).
704 7. If pcregrep was compiled under Windows, there was a reference to the
705 function pcregrep_exit() before it was defined. I am assuming this was
706 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
707 reported by a user. I've moved the definition above the reference.
710 Version 8.11 10-Dec-2010
711 ------------------------
713 1. (*THEN) was not working properly if there were untried alternatives prior
714 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
715 backtracked to try for "b" instead of moving to the next alternative branch
716 at the same level (in this case, to look for "c"). The Perl documentation
717 is clear that when (*THEN) is backtracked onto, it goes to the "next
718 alternative in the innermost enclosing group".
720 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
721 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
722 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
723 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
724 (*THEN).
726 3. If \s appeared in a character class, it removed the VT character from
727 the class, even if it had been included by some previous item, for example
728 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
729 of \s, but is part of the POSIX "space" class.)
731 4. A partial match never returns an empty string (because you can always
732 match an empty string at the end of the subject); however the checking for
733 an empty string was starting at the "start of match" point. This has been
734 changed to the "earliest inspected character" point, because the returned
735 data for a partial match starts at this character. This means that, for
736 example, /(?<=abc)def/ gives a partial match for the subject "abc"
737 (previously it gave "no match").
739 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
740 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
741 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
742 has an implication that the given string is incomplete (because a partial
743 match is preferred over a full match). For this reason, these items now
744 give a partial match in this situation. [Aside: previously, the one case
745 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
746 match rather than a full match, which was wrong by the old rules, but is
747 now correct.]
749 6. There was a bug in the handling of #-introduced comments, recognized when
750 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
751 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
752 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
753 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
754 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
755 places in pcre_compile().
757 7. Related to (6) above, when pcre_compile() was skipping #-introduced
758 comments when looking ahead for named forward references to subpatterns,
759 the only newline sequence it recognized was NL. It now handles newlines
760 according to the set newline convention.
762 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
763 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
764 cater for a lack of strerror(). These oversights have been fixed.
766 9. Added --match-limit and --recursion-limit to pcregrep.
768 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
770 11. When the -o option was used, pcregrep was setting a return code of 1, even
771 when matches were found, and --line-buffered was not being honoured.
773 12. Added an optional parentheses number to the -o and --only-matching options
774 of pcregrep.
776 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
777 can match an empty string. The code to do it in pcretest and pcredemo
778 needed fixing:
780 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
781 only one byte after an empty string match just before CRLF (this case
782 just got forgotten; "any" and "anycrlf" were OK).
784 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
785 mode when an empty string match preceded an ASCII character followed by
786 a non-ASCII character. (The code for advancing by one character rather
787 than one byte was nonsense.)
789 (c) The pcredemo.c sample program did not have any code at all to handle
790 the cases when CRLF is a valid newline sequence.
792 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
793 as a starting offset was within the subject string. There is now a new
794 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
795 negative or greater than the length of the string. In order to test this,
796 pcretest is extended to allow the setting of negative starting offsets.
798 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
799 starting offset points to the beginning of a UTF-8 character was
800 unnecessarily clumsy. I tidied it up.
802 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
803 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
805 17. Nobody had reported that the --include_dir option, which was added in
806 release 7.7 should have been called --include-dir (hyphen, not underscore)
807 for compatibility with GNU grep. I have changed it to --include-dir, but
808 left --include_dir as an undocumented synonym, and the same for
809 --exclude-dir, though that is not available in GNU grep, at least as of
810 release 2.5.4.
812 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
813 characters from a string of bytes) have been redefined so as not to use
814 loops, in order to improve performance in some environments. At the same
815 time, I abstracted some of the common code into auxiliary macros to save
816 repetition (this should not affect the compiled code).
818 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
819 compile-time error is now given if \c is not followed by an ASCII
820 character, that is, a byte less than 128. (In EBCDIC mode, the code is
821 different, and any byte value is allowed.)
823 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
824 START_OPTIMIZE option, which is now allowed at compile time - but just
825 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
826 to pcregrep and other applications that have no direct access to PCRE
827 options. The new /Y option in pcretest sets this option when calling
828 pcre_compile().
830 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
831 back references. Groups containing recursive back references were forced to
832 be atomic by that change, but in the case of named groups, the amount of
833 memory required was incorrectly computed, leading to "Failed: internal
834 error: code overflow". This has been fixed.
836 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
837 pcretest.c, to avoid build problems in some Borland environments.
840 Version 8.10 25-Jun-2010
841 ------------------------
843 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
844 THEN.
846 2. (*ACCEPT) was not working when inside an atomic group.
848 3. Inside a character class, \B is treated as a literal by default, but
849 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
850 causes the error). The code is unchanged, but I tidied the documentation.
852 4. Inside a character class, PCRE always treated \R and \X as literals,
853 whereas Perl faults them if its -w option is set. I have changed PCRE so
854 that it faults them when PCRE_EXTRA is set.
856 5. Added support for \N, which always matches any character other than
857 newline. (It is the same as "." when PCRE_DOTALL is not set.)
859 6. When compiling pcregrep with newer versions of gcc which may have
860 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
861 declared with attribute warn_unused_result" were given. Just casting the
862 result to (void) does not stop the warnings; a more elaborate fudge is
863 needed. I've used a macro to implement this.
865 7. Minor change to pcretest.c to avoid a compiler warning.
867 8. Added four artifical Unicode properties to help with an option to make
868 \s etc use properties (see next item). The new properties are: Xan
869 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
871 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
872 use Unicode properties. (*UCP) at the start of a pattern can be used to set
873 this option. Modified pcretest to add /W to test this facility. Added
874 REG_UCP to make it available via the POSIX interface.
876 10. Added --line-buffered to pcregrep.
878 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
879 studied, and the match started with a letter with a code point greater than
880 127 whose first byte was different to the first byte of the other case of
881 the letter, the other case of this starting letter was not recognized
882 (#976).
884 12. If a pattern that was studied started with a repeated Unicode property
885 test, for example, \p{Nd}+, there was the theoretical possibility of
886 setting up an incorrect bitmap of starting bytes, but fortunately it could
887 not have actually happened in practice until change 8 above was made (it
888 added property types that matched character-matching opcodes).
890 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
891 possible starting bytes for non-anchored patterns.
893 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
894 \R, and also a number of cases that involve Unicode properties, both
895 explicit and implicit when PCRE_UCP is set.
897 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
898 input, it could crash or give wrong results if characters with values
899 greater than 0xc0 were present in the subject string. (Detail: it assumed
900 UTF-8 input when processing these items.)
902 16. Added a lot of (int) casts to avoid compiler warnings in systems where
903 size_t is 64-bit (#991).
905 17. Added a check for running out of memory when PCRE is compiled with
906 --disable-stack-for-recursion (#990).
908 18. If the last data line in a file for pcretest does not have a newline on
909 the end, a newline was missing in the output.
911 19. The default pcre_chartables.c file recognizes only ASCII characters (values
912 less than 128) in its various bitmaps. However, there is a facility for
913 generating tables according to the current locale when PCRE is compiled. It
914 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
915 characters, are recognized by isspace() and therefore were getting set in
916 these tables, and indeed these tables seem to approximate to ISO 8859. This
917 caused a problem in UTF-8 mode when pcre_study() was used to create a list
918 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
919 which of course cannot start UTF-8 characters. I have changed the code so
920 that only real ASCII characters (less than 128) and the correct starting
921 bytes for UTF-8 encodings are set for characters greater than 127 when in
922 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
923 altogether.)
925 20. Added the /T option to pcretest so as to be able to run tests with non-
926 standard character tables, thus making it possible to include the tests
927 used for 19 above in the standard set of tests.
929 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
930 reference to a subpattern the other side of a comment that contains an
931 opening parenthesis caused either an internal compiling error, or a
932 reference to the wrong subpattern.
935 Version 8.02 19-Mar-2010
936 ------------------------
938 1. The Unicode data tables have been updated to Unicode 5.2.0.
940 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
941 configured.
943 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
944 original author of that file, following a query about its status.
946 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
947 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
949 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
950 quantifier applied to a forward-referencing subroutine call, could compile
951 incorrect code or give the error "internal error: previously-checked
952 referenced subpattern not found".
954 6. Both MS Visual Studio and Symbian OS have problems with initializing
955 variables to point to external functions. For these systems, therefore,
956 pcre_malloc etc. are now initialized to local functions that call the
957 relevant global functions.
959 7. There were two entries missing in the vectors called coptable and poptable
960 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
961 I've fixed the data, and added a kludgy way of testing at compile time that
962 the lengths are correct (equal to the number of opcodes).
964 8. Following on from 7, I added a similar kludge to check the length of the
965 eint vector in pcreposix.c.
967 9. Error texts for pcre_compile() are held as one long string to avoid too
968 much relocation at load time. To find a text, the string is searched,
969 counting zeros. There was no check for running off the end of the string,
970 which could happen if a new error number was added without updating the
971 string.
973 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
975 11. \K was not working if it appeared in an atomic group or in a group that
976 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
977 \K is "not well defined" if used in an assertion. PCRE now accepts it if
978 the assertion is positive, but not if it is negative.
980 12. Change 11 fortuitously reduced the size of the stack frame used in the
981 "match()" function of pcre_exec.c by one pointer. Forthcoming
982 implementation of support for (*MARK) will need an extra pointer on the
983 stack; I have reserved it now, so that the stack frame size does not
984 decrease.
986 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
987 item in branch that calls a recursion is a subroutine call - as in the
988 second branch in the above example - was incorrectly given the compile-
989 time error "recursive call could loop indefinitely" because pcre_compile()
990 was not correctly checking the subroutine for matching a non-empty string.
992 14. The checks for overrunning compiling workspace could trigger after an
993 overrun had occurred. This is a "should never occur" error, but it can be
994 triggered by pathological patterns such as hundreds of nested parentheses.
995 The checks now trigger 100 bytes before the end of the workspace.
997 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1000 Version 8.01 19-Jan-2010
1001 ------------------------
1003 1. If a pattern contained a conditional subpattern with only one branch (in
1004 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1005 computed the wrong minimum data length (which is of course zero for such
1006 subpatterns). This could cause incorrect "no match" results.
1008 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1009 the pattern is reset in the first branch, pcre_compile() failed with
1010 "internal error: code overflow at offset...". This happened only when
1011 the reset was to the original external option setting. (An optimization
1012 abstracts leading options settings into an external setting, which was the
1013 cause of this.)
1015 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1016 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1017 assertion pattern did not match (meaning that the assertion was true), it
1018 was incorrectly treated as false if the SKIP had been reached during the
1019 matching. This also applied to assertions used as conditions.
1021 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1022 assertion subpattern, including such a pattern used as a condition,
1023 unpredictable results occurred, instead of the error return
1026 5. The C++ GlobalReplace function was not working like Perl for the special
1027 situation when an empty string is matched. It now does the fancy magic
1028 stuff that is necessary.
1030 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1031 removed. (These were left over from very, very early versions of PCRE.)
1033 7. Some cosmetic changes to the code to make life easier when compiling it
1034 as part of something else:
1036 (a) Change DEBUG to PCRE_DEBUG.
1038 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1039 called "current" as "current_branch", to prevent a collision with the
1040 Linux macro when compiled as a kernel module.
1042 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1043 prevent a collision with the Linux macro when compiled as a kernel
1044 module.
1046 8. In pcre_compile() there are some checks for integer overflows that used to
1047 cast potentially large values to (double). This has been changed to that
1048 when building, a check for int64_t is made, and if it is found, it is used
1049 instead, thus avoiding the use of floating point arithmetic. (There is no
1050 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1051 double.
1053 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1054 2005 (difference between two addresses compared to an unsigned value).
1056 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1057 custom one, because of the following reported problem in Windows:
1059 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1060 under Win32.
1061 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1062 therefore missing the function definition.
1063 - The compiler thus generates a "C" signature for the test function.
1064 - The linker fails to find the "C" function.
1065 - PCRE fails to configure if asked to do so against libbz2.
1067 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1068 messages were output:
1070 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1071 rerunning libtoolize, to keep the correct libtool macros in-tree.
1072 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1074 I have done both of these things.
1076 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1077 most of the time, it *can* run out if it is given a pattern that contains a
1078 runaway infinite recursion. I updated the discussion in the pcrestack man
1079 page.
1081 13. Now that we have gone to the x.xx style of version numbers, the minor
1082 version may start with zero. Using 08 or 09 is a bad idea because users
1083 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1084 interpreted as invalid octal numbers. I've updated the previous comment in
1085 configure.ac, and also added a check that gives an error if 08 or 09 are
1086 used.
1088 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1089 causing partial matching to fail when the end of the subject matched \W
1090 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1092 15. There were some discrepancies between the declarations in pcre_internal.h
1093 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1094 their definitions. The declarations used "const uschar *" and the
1095 definitions used USPTR. Even though USPTR is normally defined as "const
1096 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1097 reported that: "This difference in casting confuses some C++ compilers, for
1098 example, SunCC recognizes above declarations as different functions and
1099 generates broken code for hbpcre." I have changed the declarations to use
1100 USPTR.
1102 16. GNU libtool is named differently on some systems. The autogen.sh script now
1103 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1104 (FreeBSD).
1106 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1107 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1108 comment: "Figure out how to create a longlong from a string: strtoll and
1109 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1110 instance, but it only takes 2 args instead of 3!"
1112 18. A subtle bug concerned with back references has been fixed by a change of
1113 specification, with a corresponding code fix. A pattern such as
1114 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1115 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1116 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1117 same bug. Such groups have to be quantified to be useful, or contained
1118 inside another quantified group. (If there's no repetition, the reference
1119 can never match.) The problem arises because, having left the group and
1120 moved on to the rest of the pattern, a later failure that backtracks into
1121 the group uses the captured value from the final iteration of the group
1122 rather than the correct earlier one. I have fixed this in PCRE by forcing
1123 any group that contains a reference to itself to be an atomic group; that
1124 is, there cannot be any backtracking into it once it has completed. This is
1125 similar to recursive and subroutine calls.
1128 Version 8.00 19-Oct-09
1129 ----------------------
1131 1. The table for translating pcre_compile() error codes into POSIX error codes
1132 was out-of-date, and there was no check on the pcre_compile() error code
1133 being within the table. This could lead to an OK return being given in
1134 error.
1136 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1137 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1138 in a Windows environment.
1140 3. The pcregrep --count option prints the count for each file even when it is
1141 zero, as does GNU grep. However, pcregrep was also printing all files when
1142 --files-with-matches was added. Now, when both options are given, it prints
1143 counts only for those files that have at least one match. (GNU grep just
1144 prints the file name in this circumstance, but including the count seems
1145 more useful - otherwise, why use --count?) Also ensured that the
1146 combination -clh just lists non-zero counts, with no names.
1148 4. The long form of the pcregrep -F option was incorrectly implemented as
1149 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1150 but it seems right to fix it, and I didn't think it was worth preserving
1151 the old behaviour.
1153 5. The command line items --regex=pattern and --regexp=pattern were not
1154 recognized by pcregrep, which required --regex pattern or --regexp pattern
1155 (with a space rather than an '='). The man page documented the '=' forms,
1156 which are compatible with GNU grep; these now work.
1158 6. No libpcreposix.pc file was created for pkg-config; there was just
1159 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1161 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1162 when UCP support is not needed, by modifying the Python script that
1163 generates it from Unicode data files. This should not matter if the module
1164 is correctly used as a library, but I received one complaint about 50K of
1165 unwanted data. My guess is that the person linked everything into his
1166 program rather than using a library. Anyway, it does no harm.
1168 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1169 was a minimum greater than 1 for a wide character in a possessive
1170 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1171 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1172 character. Chaos in the form of incorrect output or a compiling loop could
1173 result.
1175 9. The restrictions on what a pattern can contain when partial matching is
1176 requested for pcre_exec() have been removed. All patterns can now be
1177 partially matched by this function. In addition, if there are at least two
1178 slots in the offset vector, the offset of the earliest inspected character
1179 for the match and the offset of the end of the subject are set in them when
1180 PCRE_ERROR_PARTIAL is returned.
1182 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1183 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1184 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1185 and may be more useful for multi-segment matching.
1187 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1188 used to be given if ever the end of the subject was reached; now it is
1189 given only if matching could not proceed because another character was
1190 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1191 string "Z", which now yields "no match" instead of "partial match". In the
1192 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1193 final character ended with (*FAIL).
1195 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1196 if the pattern had a "must contain" character that was already found in the
1197 earlier partial match, unless partial matching was again requested. For
1198 example, with the pattern /dog.(body)?/, the "must contain" character is
1199 "g". If the first part-match was for the string "dog", restarting with
1200 "sbody" failed. This bug has been fixed.
1202 13. The string returned by pcre_dfa_exec() after a partial match has been
1203 changed so that it starts at the first inspected character rather than the
1204 first character of the match. This makes a difference only if the pattern
1205 starts with a lookbehind assertion or \b or \B (\K is not supported by
1206 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1207 matching functions compatible, and I think it's the right thing to do.
1209 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1210 so that the demonstration program is easily available in environments where
1211 PCRE has not been installed from source.
1213 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1214 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1215 library.
1217 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1218 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1219 is not the first non-POSIX option to be added. Clearly some people find
1220 these options useful.
1222 17. If a caller to the POSIX matching function regexec() passes a non-zero
1223 value for nmatch with a NULL value for pmatch, the value of
1224 nmatch is forced to zero.
1226 18. RunGrepTest did not have a test for the availability of the -u option of
1227 the diff command, as RunTest does. It now checks in the same way as
1228 RunTest, and also checks for the -b option.
1230 19. If an odd number of negated classes containing just a single character
1231 interposed, within parentheses, between a forward reference to a named
1232 subpattern and the definition of the subpattern, compilation crashed with
1233 an internal error, complaining that it could not find the referenced
1234 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1235 [The bug was that it was starting one character too far in when skipping
1236 over the character class, thus treating the ] as data rather than
1237 terminating the class. This meant it could skip too much.]
1239 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1240 /g option in pcretest when the pattern contains \K, which makes it possible
1241 to have an empty string match not at the start, even when the pattern is
1242 anchored. Updated pcretest and pcredemo to use this option.
1244 21. If the maximum number of capturing subpatterns in a recursion was greater
1245 than the maximum at the outer level, the higher number was returned, but
1246 with unset values at the outer level. The correct (outer level) value is
1247 now given.
1249 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1250 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1251 make it do so. The string so far is captured, making this feature
1252 compatible with Perl.
1254 23. The tests have been re-organized, adding tests 11 and 12, to make it
1255 possible to check the Perl 5.10 features against Perl 5.10.
1257 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1258 pattern matches a fixed length string. PCRE did not allow this; now it
1259 does. Neither allows recursion.
1261 25. I finally figured out how to implement a request to provide the minimum
1262 length of subject string that was needed in order to match a given pattern.
1263 (It was back references and recursion that I had previously got hung up
1264 on.) This code has now been added to pcre_study(); it finds a lower bound
1265 to the length of subject needed. It is not necessarily the greatest lower
1266 bound, but using it to avoid searching strings that are too short does give
1267 some useful speed-ups. The value is available to calling programs via
1268 pcre_fullinfo().
1270 26. While implementing 25, I discovered to my embarrassment that pcretest had
1271 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1272 study optimizations had never been tested with that matching function.
1273 Oops. What is worse, even when it was passed study data, there was a bug in
1274 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1275 were also very few tests of studied patterns with pcre_dfa_exec().
1277 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1278 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1279 on the other side of the coin, they are no longer allowed to have different
1280 names, because these cannot be distinguished in PCRE, and this has caused
1281 confusion. (This is a difference from Perl.)
1283 28. When duplicate subpattern names are present (necessarily with different
1284 numbers, as required by 27 above), and a test is made by name in a
1285 conditional pattern, either for a subpattern having been matched, or for
1286 recursion in such a pattern, all the associated numbered subpatterns are
1287 tested, and the overall condition is true if the condition is true for any
1288 one of them. This is the way Perl works, and is also more like the way
1289 testing by number works.
1292 Version 7.9 11-Apr-09
1293 ---------------------
1295 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1296 (pcretest), all targets were linked against these libraries. This included
1297 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1298 libraries. This caused unwanted dependencies to be created. This problem
1299 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1300 pcretest is linked with readline.
1302 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1303 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1304 moved outside it again, because FALSE and TRUE are already defined in AIX,
1305 but BOOL is not.
1307 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1308 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1310 4. The pcregrep documentation said spaces were inserted as well as colons (or
1311 hyphens) following file names and line numbers when outputting matching
1312 lines. This is not true; no spaces are inserted. I have also clarified the
1313 wording for the --colour (or --color) option.
1315 5. In pcregrep, when --colour was used with -o, the list of matching strings
1316 was not coloured; this is different to GNU grep, so I have changed it to be
1317 the same.
1319 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1320 each matching line was coloured. Now it goes on to look for further matches
1321 of any of the test patterns, which is the same behaviour as GNU grep.
1323 7. A pattern that could match an empty string could cause pcregrep to loop; it
1324 doesn't make sense to accept an empty string match in pcregrep, so I have
1325 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1326 seems to be how GNU grep behaves.
1328 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1329 start or after a newline", because the conditional assertion was not being
1330 correctly handled. The rule now is that both the assertion and what follows
1331 in the first alternative must satisfy the test.
1333 9. If auto-callout was enabled in a pattern with a conditional group whose
1334 condition was an assertion, PCRE could crash during matching, both with
1335 pcre_exec() and pcre_dfa_exec().
1337 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1338 used for matching.
1340 11. Unicode property support in character classes was not working for
1341 characters (bytes) greater than 127 when not in UTF-8 mode.
1343 12. Added the -M command line option to pcretest.
1345 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1347 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1349 16. Added comments and documentation about mis-use of no_arg in the C++
1350 wrapper.
1352 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1353 from Martin Jerabek that uses macro names for all relevant character and
1354 string constants.
1356 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1357 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1358 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1359 these, but not everybody uses configure.
1361 19. A conditional group that had only one branch was not being correctly
1362 recognized as an item that could match an empty string. This meant that an
1363 enclosing group might also not be so recognized, causing infinite looping
1364 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1365 with the subject "ab", where knowledge that the repeated group can match
1366 nothing is needed in order to break the loop.
1368 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1369 exec(), but without supplying a callout function, matching went wrong.
1371 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1372 leak if the size of the offset vector was greater than 30. When the vector
1373 is smaller, the saved offsets during recursion go onto a local stack
1374 vector, but for larger vectors malloc() is used. It was failing to free
1375 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1376 error, in fact).
1378 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1379 heapframe that is used only when UTF-8 support is enabled. This caused no
1380 problem, but was untidy.
1382 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1383 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1384 included within another project.
1386 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1387 slightly modified by me:
1389 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1390 not building pcregrep.
1392 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1393 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1395 25. Forward references, both numeric and by name, in patterns that made use of
1396 duplicate group numbers, could behave incorrectly or give incorrect errors,
1397 because when scanning forward to find the reference group, PCRE was not
1398 taking into account the duplicate group numbers. A pattern such as
1399 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1401 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1402 the feature of a custom pointer more persuasive (as requested by a user).
1404 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1405 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1406 pre-defined.
1408 28. Added support for (*UTF8) at the start of a pattern.
1410 29. Arrange for flags added by the "release type" setting in CMake to be shown
1411 in the configuration summary.
1414 Version 7.8 05-Sep-08
1415 ---------------------
1417 1. Replaced UCP searching code with optimized version as implemented for Ad
1418 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1419 stage table and inline lookup instead of a function, giving speed ups of 2
1420 to 5 times on some simple patterns that I tested. Permission was given to
1421 distribute the MultiStage2.py script that generates the tables (it's not in
1422 the tarball, but is in the Subversion repository).
1424 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1425 scripts.
1427 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1428 a group with a zero qualifier. The result of the study could be incorrect,
1429 or the function might crash, depending on the pattern.
1431 4. Caseless matching was not working for non-ASCII characters in back
1432 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1433 It now works when Unicode Property Support is available.
1435 5. In pcretest, an escape such as \x{de} in the data was always generating
1436 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1437 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1438 truncation.
1440 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1442 7. Added two (int) casts to pcregrep when printing the difference of two
1443 pointers, in case they are 64-bit values.
1445 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1446 test 2 if it fails.
1448 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1449 and a #define of that name to empty if it is not externally set. This is to
1450 allow users of MSVC to set it if necessary.
1452 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1453 the convenience functions in the pcre_get.c source file.
1455 11. An option change at the start of a pattern that had top-level alternatives
1456 could cause overwriting and/or a crash. This command provoked a crash in
1457 some environments:
1459 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1461 This potential security problem was recorded as CVE-2008-2371.
1463 12. For a pattern where the match had to start at the beginning or immediately
1464 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1465 pcre_dfa_exec() could read past the end of the passed subject if there was
1466 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1467 pcretest so that it places the subject at the end of its malloc-ed buffer.
1469 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1470 exec() might read past the end of the data buffer in UTF-8 mode.
1472 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1473 the data contained the byte 0x85 as part of a UTF-8 character within its
1474 first line. This applied both to normal and DFA matching.
1476 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1477 /^[^d]*?$/8 failed to match "abc".
1479 16. Added a missing copyright notice to pcrecpp_internal.h.
1481 17. Make it more clear in the documentation that values returned from
1482 pcre_exec() in ovector are byte offsets, not character counts.
1484 18. Tidied a few places to stop certain compilers from issuing warnings.
1486 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1487 supplied by Stefan Weber. I made a further small update for 7.8 because
1488 there is a change of source arrangements: the pcre_searchfuncs.c module is
1489 replaced by pcre_ucd.c.
1492 Version 7.7 07-May-08
1493 ---------------------
1495 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1496 a string to a long long, pretend we don't even have a long long." This is
1497 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1499 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1500 pre-7.6 versions, which defined a global no_arg variable instead of putting
1501 it in the RE class. (See also #8 below.)
1503 3. Remove a line of dead code, identified by coverity and reported by Nuno
1504 Lopes.
1506 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1508 (1) The include/exclude patterns were being applied to the whole pathnames
1509 of files, instead of just to the final components.
1511 (2) If there was more than one level of directory, the subdirectories were
1512 skipped unless they satisfied the include/exclude conditions. This is
1513 inconsistent with GNU grep (and could even be seen as contrary to the
1514 pcregrep specification - which I improved to make it absolutely clear).
1515 The action now is always to scan all levels of directory, and just
1516 apply the include/exclude patterns to regular files.
1518 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1519 --exclude_dir in the tests to avoid scanning .svn directories.
1521 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1522 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1523 doesn't support NULs in patterns.
1525 7. Added some missing "const"s to declarations of static tables in
1526 pcre_compile.c and pcre_dfa_exec.c.
1528 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1529 caused by fix #2 above. (Subsequently also a second patch to fix the
1530 first patch. And a third patch - this was a messy problem.)
1532 9. Applied Craig's patch to remove the use of push_back().
1534 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1535 matching function regexec().
1537 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1538 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1539 references. PCRE supports relative numbers with this syntax (I don't think
1540 Oniguruma does).
1542 12. Previously, a group with a zero repeat such as (...){0} was completely
1543 omitted from the compiled regex. However, this means that if the group
1544 was called as a subroutine from elsewhere in the pattern, things went wrong
1545 (an internal error was given). Such groups are now left in the compiled
1546 pattern, with a new opcode that causes them to be skipped at execution
1547 time.
1549 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1550 to the way PCRE behaves:
1552 (a) A lone ] character is dis-allowed (Perl treats it as data).
1554 (b) A back reference to an unmatched subpattern matches an empty string
1555 (Perl fails the current match path).
1557 (c) A data ] in a character class must be notated as \] because if the
1558 first data character in a class is ], it defines an empty class. (In
1559 Perl it is not possible to have an empty class.) The empty class []
1560 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1561 The negative empty class [^] matches any one character, independently
1562 of the DOTALL setting.
1564 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1565 non-existent subpattern following a character class starting with ']' and
1566 containing () gave an internal compiling error instead of "reference to
1567 non-existent subpattern". Fortunately, when the pattern did exist, the
1568 compiled code was correct. (When scanning forwards to check for the
1569 existencd of the subpattern, it was treating the data ']' as terminating
1570 the class, so got the count wrong. When actually compiling, the reference
1571 was subsequently set up correctly.)
1573 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1574 it was being rejected as not supported by pcre_dfa_exec(), even though
1575 other assertions are supported. I have made pcre_dfa_exec() support
1576 (*FAIL).
1578 16. The implementation of 13c above involved the invention of a new opcode,
1579 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1580 cannot be changed at match time, I realized I could make a small
1581 improvement to matching performance by compiling OP_ALLANY instead of
1582 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1583 on the OP_ANY path.
1585 17. Compiling pcretest on Windows with readline support failed without the
1586 following two fixes: (1) Make the unistd.h include conditional on
1587 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1589 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1590 ncurses library to be included for pcretest when ReadLine support is
1591 requested, but also to allow for it to be overridden. This patch came from
1592 Daniel Bergström.
1594 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1595 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1596 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1597 spotting this.
1600 Version 7.6 28-Jan-08
1601 ---------------------
1603 1. A character class containing a very large number of characters with
1604 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1605 overflow.
1607 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1608 HAVE_LONG_LONG is not defined.
1610 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1611 bring them up to date and include new features. This patch includes:
1613 - Fixed PH's badly added libz and libbz2 support.
1614 - Fixed a problem with static linking.
1615 - Added pcredemo. [But later removed - see 7 below.]
1616 - Fixed dftables problem and added an option.
1617 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1619 - Added readline support for pcretest.
1620 - Added an listing of the option settings after cmake has run.
1622 4. A user submitted a patch to Makefile that makes it easy to create
1623 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1624 Makefile.am that cause it to include this special target, without
1625 affecting anything else. Note that the same mingw target plus all
1626 the other distribution libraries and programs are now supported
1627 when configuring with CMake (see 6 below) instead of with
1628 Configure/Make.
1630 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1631 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1632 exported in the Windows port". It has not yet been confirmed that the patch
1633 solves the problem, but it does no harm.
1635 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1636 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1637 with CMake, and also correct the comment about stack recursion.
1639 7. Remove the automatic building of pcredemo from the ./configure system and
1640 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1641 of a program that users should build themselves after PCRE is installed, so
1642 building it automatically is not really right. What is more, it gave
1643 trouble in some build environments.
1645 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1648 Version 7.5 10-Jan-08
1649 ---------------------
1651 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1652 values in parens when parsing an RE using the C++ wrapper."
1654 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1655 Characters greater than 255 were excluded from the class instead of being
1656 included.
1658 3. The same bug as (2) above applied to negated POSIX classes such as
1659 [:^space:].
1661 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1662 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1663 I have changed it.
1665 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1666 first named subpattern) and a construct such as (?&a) would reference the
1667 first named subpattern whose name started with "a" (in other words, the
1668 length check was missing). Both these problems are fixed. "Subpattern name
1669 expected" is now given for (?&) (a zero-length name), and this patch also
1670 makes it give the same error for \k'' (previously it complained that that
1671 was a reference to a non-existent subpattern).
1673 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1674 this is right because (?- can be followed by option settings as well as by
1675 digits. I have, however, made the messages clearer.
1677 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1678 than the number used in the conditional) now cause a compile-time error.
1679 This is actually not compatible with Perl, which accepts such patterns, but
1680 treats the conditional as always being FALSE (as PCRE used to), but it
1681 seems to me that giving a diagnostic is better.
1683 8. Change "alphameric" to the more common word "alphanumeric" in comments
1684 and messages.
1686 9. Fix two occurrences of "backslash" in comments that should have been
1687 "backspace".
1689 10. Remove two redundant lines of code that can never be obeyed (their function
1690 was moved elsewhere).
1692 11. The program that makes PCRE's Unicode character property table had a bug
1693 which caused it to generate incorrect table entries for sequences of
1694 characters that have the same character type, but are in different scripts.
1695 It amalgamated them into a single range, with the script of the first of
1696 them. In other words, some characters were in the wrong script. There were
1697 thirteen such cases, affecting characters in the following ranges:
1699 U+002b0 - U+002c1
1700 U+0060c - U+0060d
1701 U+0061e - U+00612
1702 U+0064b - U+0065e
1703 U+0074d - U+0076d
1704 U+01800 - U+01805
1705 U+01d00 - U+01d77
1706 U+01d9b - U+01dbf
1707 U+0200b - U+0200f
1708 U+030fc - U+030fe
1709 U+03260 - U+0327f
1710 U+0fb46 - U+0fbb1
1711 U+10450 - U+1049d
1713 12. The -o option (show only the matching part of a line) for pcregrep was not
1714 compatible with GNU grep in that, if there was more than one match in a
1715 line, it showed only the first of them. It now behaves in the same way as
1716 GNU grep.
1718 13. If the -o and -v options were combined for pcregrep, it printed a blank
1719 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1720 does the same. The return code can be used to tell if there were any
1721 non-matching lines.
1723 14. Added --file-offsets and --line-offsets to pcregrep.
1725 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1726 infinitely looping recursion. The bug was that positive lookaheads were not
1727 being skipped when checking for a possible empty match (negative lookaheads
1728 and both kinds of lookbehind were skipped).
1730 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1731 inclusion of <windows.h> to before rather than after the definition of
1732 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1734 17. Specifying a possessive quantifier with a specific limit for a Unicode
1735 character property caused pcre_compile() to compile bad code, which led at
1736 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1737 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1738 caused the error; without that there was no problem.
1740 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1742 19. Added --enable-pcretest-libreadline.
1744 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1745 RE::GlobalReplace(). As a result, the number of replacements returned was
1746 double what it should be. I removed one of the increments, but Craig sent a
1747 later patch that removed the other one (the right fix) and added unit tests
1748 that check the return values (which was not done before).
1750 21. Several CMake things:
1752 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1753 the names libpcre and libpcreposix, not just pcre and pcreposix.
1755 (2) The above change means that pcretest and pcregrep are now correctly
1756 linked with the newly-built libraries, not previously installed ones.
1760 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1761 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1762 UTF-8 newline character). The key issue is that the pattern starts .*;
1763 this means that the match must be either at the beginning, or after a
1764 newline. The bug was in the code for advancing after a failed match and
1765 checking that the new position followed a newline. It was not taking
1766 account of UTF-8 characters correctly.
1768 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1769 character classes. PCRE was not treating the sequence [:...:] as a
1770 character class unless the ... were all letters. Perl, however, seems to
1771 allow any characters between [: and :], though of course it rejects as
1772 unknown any "names" that contain non-letters, because all the known class
1773 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1774 for example, whereas PCRE did not - it did not recognize a POSIX character
1775 class. This seemed a bit dangerous, so the code has been changed to be
1776 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1777 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1778 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1779 Perl does, and where it didn't before.
1781 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1782 Windows environments %n is disabled by default.
1785 Version 7.4 21-Sep-07
1786 ---------------------
1788 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1789 means that a class such as [\s] counted as "explicit reference to CR or
1790 LF". That isn't really right - the whole point of the change was to try to
1791 help when there was an actual mention of one of the two characters. So now
1792 the change happens only if \r or \n (or a literal CR or LF) character is
1793 encountered.
1795 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1796 of both had grown to the point where there were only 3 bits left.
1797 Fortunately, there was spare space in the data structure, and so I have
1798 moved the internal flags into a new 16-bit field to free up more option
1799 bits.
1801 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1802 but did not set the internal JCHANGED flag - either of these is enough to
1803 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1804 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1805 start sets both bits.
1807 4. Added options (at build time, compile time, exec time) to change \R from
1808 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1810 5. doc/pcresyntax.html was missing from the distribution.
1812 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1813 compatibility, even though it is no longer used.
1815 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1816 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1817 windows.h file is present (where different names are used). [This was
1818 reversed later after testing - see 16 below.]
1820 8. Changed all #include <config.h> to #include "config.h". There were also
1821 some further <pcre.h> cases that I changed to "pcre.h".
1823 9. When pcregrep was used with the --colour option, it missed the line ending
1824 sequence off the lines that it output.
1826 10. It was pointed out to me that arrays of string pointers cause lots of
1827 relocations when a shared library is dynamically loaded. A technique of
1828 using a single long string with a table of offsets can drastically reduce
1829 these. I have refactored PCRE in four places to do this. The result is
1830 dramatic:
1832 Originally: 290
1833 After changing UCP table: 187
1834 After changing error message table: 43
1835 After changing table of "verbs" 36
1836 After changing table of Posix names 22
1838 Thanks to the folks working on Gregex for glib for this insight.
1840 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1841 unicode-properties was also set.
1843 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1845 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1846 checked only for CRLF.
1848 14. Added casts to pcretest.c to avoid compiler warnings.
1850 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1852 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1853 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1854 entirely. This removes changes made in 7 above.
1856 17. The CMake files have been updated, and there is now more information about
1857 building with CMake in the NON-UNIX-USE document.
1860 Version 7.3 28-Aug-07
1861 ---------------------
1863 1. In the rejigging of the build system that eventually resulted in 7.1, the
1864 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1865 brackets there is not right, since it causes compilers to look for an
1866 installed pcre.h, not the version that is in the source that is being
1867 compiled (which of course may be different). I have changed it back to:
1869 #include "pcre.h"
1871 I have a vague recollection that the change was concerned with compiling in
1872 different directories, but in the new build system, that is taken care of
1873 by the VPATH setting the Makefile.
1875 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1876 when the subject happened to end in the byte 0x85 (e.g. if the last
1877 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1878 characters but of course it shouldn't be taken as a newline when it is part
1879 of another character. The bug was that, for an unlimited repeat of . in
1880 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1881 characters when looking for a newline.
1883 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1885 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1886 in debug output.
1888 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1889 long printing in the pcrecpp unittest when running under MinGW.
1891 6. ESC_K was left out of the EBCDIC table.
1893 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1894 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1895 limit also applies to "virtual nesting" when a pattern is recursive, and in
1896 this case 1000 isn't so big. I have been able to remove this limit at the
1897 expense of backing off one optimization in certain circumstances. Normally,
1898 when pcre_exec() would call its internal match() function recursively and
1899 immediately return the result unconditionally, it uses a "tail recursion"
1900 feature to save stack. However, when a subpattern that can match an empty
1901 string has an unlimited repetition quantifier, it no longer makes this
1902 optimization. That gives it a stack frame in which to save the data for
1903 checking that an empty string has been matched. Previously this was taken
1904 from the 1000-entry workspace that had been reserved. So now there is no
1905 explicit limit, but more stack is used.
1907 8. Applied Daniel's patches to solve problems with the import/export magic
1908 syntax that is required for Windows, and which was going wrong for the
1909 pcreposix and pcrecpp parts of the library. These were overlooked when this
1910 problem was solved for the main library.
1912 9. There were some crude static tests to avoid integer overflow when computing
1913 the size of patterns that contain repeated groups with explicit upper
1914 limits. As the maximum quantifier is 65535, the maximum group length was
1915 set at 30,000 so that the product of these two numbers did not overflow a
1916 32-bit integer. However, it turns out that people want to use groups that
1917 are longer than 30,000 bytes (though not repeat them that many times).
1918 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1919 made it possible to implement the integer overflow checks in a much more
1920 dynamic way, which I have now done. The artificial limitation on group
1921 length has been removed - we now have only the limit on the total length of
1922 the compiled pattern, which depends on the LINK_SIZE setting.
1924 10. Fixed a bug in the documentation for get/copy named substring when
1925 duplicate names are permitted. If none of the named substrings are set, the
1926 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1927 empty string.
1929 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1930 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1931 because the ] is interpreted as the first data character and the
1932 terminating ] is not found. PCRE has been made compatible with Perl in this
1933 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1934 cause memory overwriting.
1936 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1937 string has been matched (to stop an infinite loop). It was not recognizing
1938 a conditional subpattern that could match an empty string if that
1939 subpattern was within another subpattern. For example, it looped when
1940 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1941 condition was not nested. This bug has been fixed.
1943 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1944 past the start of the subject in the presence of bytes with the top bit
1945 set, for example "\x8aBCD".
1947 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1948 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1950 14. Optimized (?!) to (*FAIL).
1952 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1953 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1954 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1955 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1956 does: it's just the validity check that is more restrictive.
1958 16. Inserted checks for integer overflows during escape sequence (backslash)
1959 processing, and also fixed erroneous offset values for syntax errors during
1960 backslash processing.
1962 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1963 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1965 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1966 caused an overrun.
1968 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1969 something other than just ASCII characters) inside a group that had an
1970 unlimited repeat caused a loop at compile time (while checking to see
1971 whether the group could match an empty string).
1973 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1974 [\P{Any}] did so. (Error in the code for printing property names.)
1976 21. An orphan \E inside a character class could cause a crash.
1978 22. A repeated capturing bracket such as (A)? could cause a wild memory
1979 reference during compilation.
1981 23. There are several functions in pcre_compile() that scan along a compiled
1982 expression for various reasons (e.g. to see if it's fixed length for look
1983 behind). There were bugs in these functions when a repeated \p or \P was
1984 present in the pattern. These operators have additional parameters compared
1985 with \d, etc, and these were not being taken into account when moving along
1986 the compiled data. Specifically:
1988 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1989 length.
1991 (b) An item such as \pL+ within a repeated group could cause crashes or
1992 loops.
1994 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1995 "reference to non-existent subpattern" error.
1997 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1999 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2000 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2002 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2004 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2005 character were causing crashes (broken optimization).
2007 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2008 \p or \P) caused a compile-time loop.
2010 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2011 break. For example, the unstudied pattern [\r\n]A does not match the string
2012 "\r\nA" because change 7.0/46 below moves the current point on by two
2013 characters after failing to match at the start. However, the pattern \nA
2014 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2015 the same is true. There doesn't seem any very clean way out of this, but
2016 what I have chosen to do makes the common cases work: PCRE now takes note
2017 of whether there can be an explicit match for \r or \n anywhere in the
2018 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2019 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2020 pattern has explicit CR or LF references.
2022 29. Added (*CR) etc for changing newline setting at start of pattern.
2025 Version 7.2 19-Jun-07
2026 ---------------------
2028 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2029 which is apparently normally available under Windows.
2031 2. Re-jig the pcregrep tests with different newline settings in an attempt
2032 to make them independent of the local environment's newline setting.
2034 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2036 4. Some of the "internals" tests were previously cut out when the link size
2037 was not 2, because the output contained actual offsets. The recent new
2038 "Z" feature of pcretest means that these can be cut out, making the tests
2039 usable with all link sizes.
2041 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2042 stack recursion. This gives a massive performance boost under BSD, but just
2043 a small improvement under Linux. However, it saves one field in the frame
2044 in all cases.
2046 6. Added more features from the forthcoming Perl 5.10:
2048 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2049 recursion call. It refers to the nth most recently opened parentheses.
2051 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2052 to be opened parentheses.
2054 (c) Conditions that refer to capturing parentheses can be specified
2055 relatively, for example, (?(-2)... or (?(+3)...
2057 (d) \K resets the start of the current match so that everything before
2058 is not part of it.
2060 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2062 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2063 reference syntax.
2065 (g) (?| introduces a group in which the numbering of parentheses in each
2066 alternative starts with the same number.
2068 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2070 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2073 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2074 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2075 for detecting groups that can match an empty string.
2077 9. A pattern with a very large number of alternatives (more than several
2078 hundred) was running out of internal workspace during the pre-compile
2079 phase, where pcre_compile() figures out how much memory will be needed. A
2080 bit of new cunning has reduced the workspace needed for groups with
2081 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2082 workspace instead of running out of the 4096 that are available.
2084 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2086 11. Applied patch from Google to remove an optimization that didn't quite work.
2087 The report of the bug said:
2089 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2090 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2091 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2093 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2094 it matched the wrong number of bytes.
2097 Version 7.1 24-Apr-07
2098 ---------------------
2100 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2101 that is more "standard", making use of automake and other Autotools. There
2102 is some re-arrangement of the files and adjustment of comments consequent
2103 on this.
2105 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2106 for recursive directory scanning broke on some systems because the files
2107 are not scanned in any specific order and on different systems the order
2108 was different. A call to "sort" has been inserted into RunGrepTest for the
2109 approprate test as a short-term fix. In the longer term there may be an
2110 alternative.
2112 3. I had an email from Eric Raymond about problems translating some of PCRE's
2113 man pages to HTML (despite the fact that I distribute HTML pages, some
2114 people do their own conversions for various reasons). The problems
2115 concerned the use of low-level troff macros .br and .in. I have therefore
2116 removed all such uses from the man pages (some were redundant, some could
2117 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2118 HTML has been updated to handle .nf/.fi and to complain if it encounters
2119 .br or .in.
2121 4. Updated comments in configure.ac that get placed in config.h.in and also
2122 arranged for config.h to be included in the distribution, with the name
2123 config.h.generic, for the benefit of those who have to compile without
2124 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2126 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2127 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2128 makevp.bat for the current PCRE, using the additional files
2129 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2131 6. A Windows user reported a minor discrepancy with test 2, which turned out
2132 to be caused by a trailing space on an input line that had got lost in his
2133 copy. The trailing space was an accident, so I've just removed it.
2135 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2136 that is needed.
2138 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2139 as "const" (a) because they are and (b) because it helps the PHP
2140 maintainers who have recently made a script to detect big data structures
2141 in the php code that should be moved to the .rodata section. I remembered
2142 to update Builducptable as well, so it won't revert if ucptable.h is ever
2143 re-created.
2145 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2146 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2147 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2148 support is not required. This saves 1.5-2K of code, which is important in
2149 some applications.
2151 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2152 so as not to refer to the tables, even though these functions will never be
2153 called when UTF-8 support is disabled. Otherwise there are problems with a
2154 shared library.
2156 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2158 (a) It was defining its arguments as char * instead of void *.
2160 (b) It was assuming that all moves were upwards in memory; this was true
2161 a long time ago when I wrote it, but is no longer the case.
2163 The emulated memove() is provided for those environments that have neither
2164 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2165 is clearly not the case, as these two bugs were recently reported.
2167 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2168 and Detrail to create the HTML documentation, the .txt form of the man
2169 pages, and it removes trailing spaces from listed files. It also creates
2170 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2171 case, it wraps all the #defines with #ifndefs. This script should be run
2172 before "make dist".
2174 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2175 with Unicode property support.
2177 (a) For a maximizing quantifier, if the two different cases of the
2178 character were of different lengths in their UTF-8 codings (there are
2179 some cases like this - I found 11), and the matching function had to
2180 back up over a mixture of the two cases, it incorrectly assumed they
2181 were both the same length.
2183 (b) When PCRE was configured to use the heap rather than the stack for
2184 recursion during matching, it was not correctly preserving the data for
2185 the other case of a UTF-8 character when checking ahead for a match
2186 while processing a minimizing repeat. If the check also involved
2187 matching a wide character, but failed, corruption could cause an
2188 erroneous result when trying to check for a repeat of the original
2189 character.
2191 13. Some tidying changes to the testing mechanism:
2193 (a) The RunTest script now detects the internal link size and whether there
2194 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2195 values substituted by "configure". (The RunGrepTest script already did
2196 this for UTF-8.) The configure.ac script no longer substitutes the
2197 relevant variables.
2199 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2200 with length and offset values. This means that the output is different
2201 for different internal link sizes. Test 2 is skipped for link sizes
2202 other than 2 because of this, bypassing the problem. Unfortunately,
2203 there was also a test in test 3 (the locale tests) that used /B and
2204 failed for link sizes other than 2. Rather than cut the whole test out,
2205 I have added a new /Z option to pcretest that replaces the length and
2206 offset values with spaces. This is now used to make test 3 independent
2207 of link size. (Test 2 will be tidied up later.)
2209 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2210 segmentation fault instead of returning the appropriate error message.
2212 15. In multiline mode when the newline sequence was set to "any", the pattern
2213 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2214 This doesn't seem right; it now treats the CRLF combination as the line
2215 ending, and so does not match in that case. It's only a pattern such as ^$
2216 that would hit this one: something like ^ABC$ would have failed after \r
2217 and then tried again after \r\n.
2219 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2220 in an attempt to make files that differ only in their line terminators
2221 compare equal. This works on Linux.
2223 17. Under certain error circumstances pcregrep might try to free random memory
2224 as it exited. This is now fixed, thanks to valgrind.
2226 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2227 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2228 was because its rules for how to advance for /g after matching an empty
2229 string at the end of a line did not allow for this case. They now check for
2230 it specially.
2232 20. pcretest is supposed to handle patterns and data of any length, by
2233 extending its buffers when necessary. It was getting this wrong when the
2234 buffer for a data line had to be extended.
2236 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2237 CRLF as a newline sequence.
2239 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2240 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2241 I have nevertheless tidied it up.
2243 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2245 24. Added a man page for pcre-config.
2248 Version 7.0 19-Dec-06
2249 ---------------------
2251 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2252 moving to gcc 4.1.1.
2254 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2255 sys/time.h, which is documented as needed for this function. It doesn't
2256 seem to matter on Linux, but it showed up on some releases of OS X.
2258 3. It seems that there are systems where bytes whose values are greater than
2259 127 match isprint() in the "C" locale. The "C" locale should be the
2260 default when a C program starts up. In most systems, only ASCII printing
2261 characters match isprint(). This difference caused the output from pcretest
2262 to vary, making some of the tests fail. I have changed pcretest so that:
2264 (a) When it is outputting text in the compiled version of a pattern, bytes
2265 other than 32-126 are always shown as hex escapes.
2267 (b) When it is outputting text that is a matched part of a subject string,
2268 it does the same, unless a different locale has been set for the match
2269 (using the /L modifier). In this case, it uses isprint() to decide.
2271 4. Fixed a major bug that caused incorrect computation of the amount of memory
2272 required for a compiled pattern when options that changed within the
2273 pattern affected the logic of the preliminary scan that determines the
2274 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2275 that the computed length was too small. The symptoms of this bug were
2276 either the PCRE error "internal error: code overflow" from pcre_compile(),
2277 or a glibc crash with a message such as "pcretest: free(): invalid next
2278 size (fast)". Examples of patterns that provoked this bug (shown in
2279 pcretest format) are:
2281 /(?-x: )/x
2282 /(?x)(?-x: \s*#\s*)/
2283 /((?i)[\x{c0}])/8
2284 /(?i:[\x{c0}])/8
2286 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2287 is now done differently.
2289 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2290 wrapper classes; (b) implement a new function in the C++ scanner that is
2291 more efficient than the old way of doing things because it avoids levels of
2292 recursion in the regex matching; (c) add a paragraph to the documentation
2293 for the FullMatch() function.
2295 6. The escape sequence \n was being treated as whatever was defined as
2296 "newline". Not only was this contrary to the documentation, which states
2297 that \n is character 10 (hex 0A), but it also went horribly wrong when
2298 "newline" was defined as CRLF. This has been fixed.
2300 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2301 was being set to -1 for the "end of line" case (supposedly a value that no
2302 character can have). Though this value is never used (the check for end of
2303 line is "zero bytes in current character"), it caused compiler complaints.
2304 I've changed it to 0xffffffff.
2306 8. In pcre_version.c, the version string was being built by a sequence of
2307 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2308 string (as it is for production releases) called a macro with an empty
2309 argument. The C standard says the result of this is undefined. The gcc
2310 compiler treats it as an empty string (which was what was wanted) but it is
2311 reported that Visual C gives an error. The source has been hacked around to
2312 avoid this problem.
2314 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2315 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2316 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2317 of them did).
2319 10. Originally, pcretest opened its input and output without "b"; then I was
2320 told that "b" was needed in some environments, so it was added for release
2321 5.0 to both the input and output. (It makes no difference on Unix-like
2322 systems.) Later I was told that it is wrong for the input on Windows. I've
2323 now abstracted the modes into two macros, to make it easier to fiddle with
2324 them, and removed "b" from the input mode under Windows.
2326 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2328 12. Added -help and --help to pcretest as an official way of being reminded
2329 of the options.
2331 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2332 and pcrecpp.cc because they annoy compilers at high warning levels.
2334 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2336 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2337 scripts are not C programs :-) and which was not noticed because it works
2338 on Linux.
2340 16. pcretest is supposed to handle any length of pattern and data line (as one
2341 line or as a continued sequence of lines) by extending its input buffer if
2342 necessary. This feature was broken for very long pattern lines, leading to
2343 a string of junk being passed to pcre_compile() if the pattern was longer
2344 than about 50K.
2346 17. I have done a major re-factoring of the way pcre_compile() computes the
2347 amount of memory needed for a compiled pattern. Previously, there was code
2348 that made a preliminary scan of the pattern in order to do this. That was
2349 OK when PCRE was new, but as the facilities have expanded, it has become
2350 harder and harder to keep it in step with the real compile phase, and there
2351 have been a number of bugs (see for example, 4 above). I have now found a
2352 cunning way of running the real compile function in a "fake" mode that
2353 enables it to compute how much memory it would need, while actually only
2354 ever using a few hundred bytes of working memory and without too many
2355 tests of the mode. This should make future maintenance and development
2356 easier. A side effect of this work is that the limit of 200 on the nesting
2357 depth of parentheses has been removed (though this was never a serious
2358 limitation, I suspect). However, there is a downside: pcre_compile() now
2359 runs more slowly than before (30% or more, depending on the pattern). I
2360 hope this isn't a big issue. There is no effect on runtime performance.
2362 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2363 newline (only possible for the last line of a file) and it was a
2364 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2366 19. Added additional timing features to pcretest. (1) The -tm option now times
2367 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2368 separate command line item, by a number that specifies the number of
2369 repeats to use when timing. The default is 50000; this gives better
2370 precision, but takes uncomfortably long for very large patterns.
2372 20. Extended pcre_study() to be more clever in cases where a branch of a
2373 subpattern has no definite first character. For example, (a*|b*)[cd] would
2374 previously give no result from pcre_study(). Now it recognizes that the
2375 first character must be a, b, c, or d.
2377 21. There was an incorrect error "recursive call could loop indefinitely" if
2378 a subpattern (or the entire pattern) that was being tested for matching an
2379 empty string contained only one non-empty item after a nested subpattern.
2380 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2381 incorrectly, because the \d was being skipped in the check.
2383 22. The pcretest program now has a new pattern option /B and a command line
2384 option -b, which is equivalent to adding /B to every pattern. This causes
2385 it to show the compiled bytecode, without the additional information that
2386 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2387 is the same as /B/I).
2389 23. A new optimization is now able automatically to treat some sequences such
2390 as a*b as a*+b. More specifically, if something simple (such as a character
2391 or a simple class like \d) has an unlimited quantifier, and is followed by
2392 something that cannot possibly match the quantified thing, the quantifier
2393 is automatically "possessified".
2395 24. A recursive reference to a subpattern whose number was greater than 39
2396 went wrong under certain circumstances in UTF-8 mode. This bug could also
2397 have affected the operation of pcre_study().
2399 25. Realized that a little bit of performance could be had by replacing
2400 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2402 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2404 27. Possessive quantifiers such as a++ were previously implemented by turning
2405 them into atomic groups such as ($>a+). Now they have their own opcodes,
2406 which improves performance. This includes the automatically created ones
2407 from 23 above.
2409 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2410 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2411 the first matched character to be a colon. This applied both to named and
2412 numbered groups.
2414 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2416 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2417 building PCRE on Windows easier, so I have included it in the distribution.
2419 31. There is now a check in pcretest against a ridiculously large number being
2420 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2421 loop, the loop is abandoned.
2423 32. Forward references to subpatterns in conditions such as (?(2)...) where
2424 subpattern 2 is defined later cause pcre_compile() to search forwards in
2425 the pattern for the relevant set of parentheses. This search went wrong
2426 when there were unescaped parentheses in a character class, parentheses
2427 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2429 33. "Subroutine" calls and backreferences were previously restricted to
2430 referencing subpatterns earlier in the regex. This restriction has now
2431 been removed.
2433 34. Added a number of extra features that are going to be in Perl 5.10. On the
2434 whole, these are just syntactic alternatives for features that PCRE had
2435 previously implemented using the Python syntax or my own invention. The
2436 other formats are all retained for compatibility.
2438 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2439 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2440 also .NET compatible.
2442 (b) A recursion or subroutine call to a named group can now be defined as
2443 (?&name) as well as (?P>name).
2445 (c) A backreference to a named group can now be defined as \k<name> or
2446 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2447 5.10, are also .NET compatible.
2449 (d) A conditional reference to a named group can now use the syntax
2450 (?(<name>) or (?('name') as well as (?(name).
2452 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2453 groups (named and numbered) that are never evaluated inline, but can be
2454 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2455 is always false. There may be only one alternative in such a group.
2457 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2458 as the simple (?(R). The condition is true only if the most recent
2459 recursion is that of the given number or name. It does not search out
2460 through the entire recursion stack.
2462 (g) The escape \gN or \g{N} has been added, where N is a positive or
2463 negative number, specifying an absolute or relative reference.
2465 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2466 some "unreachable code" warnings.
2468 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2469 things, this adds five new scripts.
2471 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2472 There were also incompatibilities regarding the handling of \Q..\E inside
2473 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2474 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2476 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2477 matches an empty string, and forcibly breaks the loop. There were bugs in
2478 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2479 against aaaa the result was just "a" rather than "aaaa", for example. Two
2480 separate and independent bugs (that affected different cases) have been
2481 fixed.
2483 39. Refactored the code to abolish the use of different opcodes for small
2484 capturing bracket numbers. This is a tidy that I avoided doing when I
2485 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2486 The new approach is not only tidier, it makes it possible to reduce the
2487 memory needed to fix the previous bug (38).
2489 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2490 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2491 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2492 mode.
2494 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2495 report.
2497 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2498 copy construction and assignment in the C++ wrapper.
2500 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2501 couple of bugs that could have given wrong results in the "--newline=crlf"
2502 case.
2504 44. Added a number of casts and did some reorganization of signed/unsigned int
2505 variables following suggestions from Dair Grant. Also renamed the variable
2506 "this" as "item" because it is a C++ keyword.
2508 45. Arranged for dftables to add
2510 #include "pcre_internal.h"
2512 to pcre_chartables.c because without it, gcc 4.x may remove the array
2513 definition from the final binary if PCRE is built into a static library and
2514 dead code stripping is activated.
2516 46. For an unanchored pattern, if a match attempt fails at the start of a
2517 newline sequence, and the newline setting is CRLF or ANY, and the next two
2518 characters are CRLF, advance by two characters instead of one.
2521 Version 6.7 04-Jul-06
2522 ---------------------
2524 1. In order to handle tests when input lines are enormously long, pcretest has
2525 been re-factored so that it automatically extends its buffers when
2526 necessary. The code is crude, but this _is_ just a test program. The
2527 default size has been increased from 32K to 50K.
2529 2. The code in pcre_study() was using the value of the re argument before
2530 testing it for NULL. (Of course, in any sensible call of the function, it
2531 won't be NULL.)
2533 3. The memmove() emulation function in pcre_internal.h, which is used on
2534 systems that lack both memmove() and bcopy() - that is, hardly ever -
2535 was missing a "static" storage class specifier.
2537 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2538 containing an extended class (one that cannot be represented by a bitmap
2539 because it contains high-valued characters or Unicode property items, e.g.
2540 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2541 pattern, but PCRE should not loop if you do not (it no longer does).
2542 [Detail: two cases were found: (a) a repeated subpattern containing an
2543 extended class; (b) a recursive reference to a subpattern that followed a
2544 previous extended class. It wasn't skipping over the extended class
2545 correctly when UTF-8 mode was not set.]
2547 5. A negated single-character class was not being recognized as fixed-length
2548 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2549 compile error "lookbehind assertion is not fixed length".
2551 6. The RunPerlTest auxiliary script was showing an unexpected difference
2552 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2553 write a Perl script that can interpret lines of an input file either as
2554 byte characters or as UTF-8, which is what "perltest" was being required to
2555 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2556 can't do is switch easily at run time between having the "use utf8;" pragma
2557 or not. In the end, I fudged it by using the RunPerlTest script to insert
2558 "use utf8;" explicitly for the UTF-8 tests.
2560 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2561 the end of the subject string, contrary to the documentation and to what
2562 Perl does. This was true of both matching functions. Now it matches only at
2563 the start of the subject and immediately after *internal* newlines.
2565 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2566 a pointer to an int instead of a pointer to an unsigned long int. This
2567 caused problems on 64-bit systems.
2569 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2570 instance of the 'standard' template library not being so standard".
2572 10. There was no check on the number of named subpatterns nor the maximum
2573 length of a subpattern name. The product of these values is used to compute
2574 the size of the memory block for a compiled pattern. By supplying a very
2575 long subpattern name and a large number of named subpatterns, the size
2576 computation could be caused to overflow. This is now prevented by limiting
2577 the length of names to 32 characters, and the number of named subpatterns
2578 to 10,000.
2580 11. Subpatterns that are repeated with specific counts have to be replicated in
2581 the compiled pattern. The size of memory for this was computed from the
2582 length of the subpattern and the repeat count. The latter is limited to
2583 65535, but there was no limit on the former, meaning that integer overflow
2584 could in principle occur. The compiled length of a repeated subpattern is
2585 now limited to 30,000 bytes in order to prevent this.
2587 12. Added the optional facility to have named substrings with the same name.
2589 13. Added the ability to use a named substring as a condition, using the
2590 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2591 are numbers (not recommended). Forward references are permitted.
2593 14. Added forward references in named backreferences (if you see what I mean).
2595 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2596 pattern could run off the end of the subject. For example, the pattern
2597 "(?s)(.{1,5})"8 did this with the subject "ab".
2599 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2600 PCRE_CASELESS was set when matching characters that were quantified with ?
2601 or *.
2603 17. A character class other than a single negated character that had a minimum
2604 but no maximum quantifier - for example [ab]{6,} - was not handled
2605 correctly by pce_dfa_exec(). It would match only one character.
2607 18. A valid (though odd) pattern that looked like a POSIX character
2608 class but used an invalid character after [ (for example [[,abc,]]) caused
2609 pcre_compile() to give the error "Failed: internal error: code overflow" or
2610 in some cases to crash with a glibc free() error. This could even happen if
2611 the pattern terminated after [[ but there just happened to be a sequence of
2612 letters, a binary zero, and a closing ] in the memory that followed.
2614 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2615 over the years. Originally (before any Unicode support), just the bottom 8
2616 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2617 output from "man perlunicode" includes this:
2619 The regular expression compiler produces polymorphic opcodes. That
2620 is, the pattern adapts to the data and automatically switches to
2621 the Unicode character scheme when presented with Unicode data--or
2622 instead uses a traditional byte scheme when presented with byte
2623 data.
2625 Sadly, a wide octal escape does not cause a switch, and in a string with
2626 no other multibyte characters, these octal escapes are treated as before.
2627 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2628 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2629 Unicode string.
2631 I have not perpetrated such confusion in PCRE. Up till now, it took just
2632 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2633 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2634 translate to the appropriate multibyte character.
2636 29. Applied some refactoring to reduce the number of warnings from Microsoft
2637 and Borland compilers. This has included removing the fudge introduced
2638 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2639 a warning about an unused variable.
2641 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2642 characters since release 4.0, because Perl (from release 5.004) does not.
2643 [Or at least, is documented not to: some releases seem to be in conflict
2644 with the documentation.] However, when a pattern was studied with
2645 pcre_study() and all its branches started with \s, PCRE still included VT
2646 as a possible starting character. Of course, this did no harm; it just
2647 caused an unnecessary match attempt.
2649 22. Removed a now-redundant internal flag bit that recorded the fact that case
2650 dependency changed within the pattern. This was once needed for "required
2651 byte" processing, but is no longer used. This recovers a now-scarce options
2652 bit. Also moved the least significant internal flag bit to the most-
2653 significant bit of the word, which was not previously used (hangover from
2654 the days when it was an int rather than a uint) to free up another bit for
2655 the future.
2657 23. Added support for CRLF line endings as well as CR and LF. As well as the
2658 default being selectable at build time, it can now be changed at runtime
2659 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2660 specify that it is scanning data with non-default line endings.
2662 24. Changed the definition of CXXLINK to make it agree with the definition of
2663 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2665 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2666 recursions. This makes a big different to stack usage for some patterns.
2668 26. If a subpattern containing a named recursion or subroutine reference such
2669 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2670 the space required for the compiled pattern went wrong and gave too small a
2671 value. Depending on the environment, this could lead to "Failed: internal
2672 error: code overflow at offset 49" or "glibc detected double free or
2673 corruption" errors.
2675 27. Applied patches from Google (a) to support the new newline modes and (b) to
2676 advance over multibyte UTF-8 characters in GlobalReplace.
2678 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2679 difference for some implementation of PCRE in some Windows version.
2681 29. Added some extra testing facilities to pcretest:
2683 \q<number> in a data line sets the "match limit" value
2684 \Q<number> in a data line sets the "match recursion limt" value
2685 -S <number> sets the stack size, where <number> is in megabytes
2687 The -S option isn't available for Windows.
2690 Version 6.6 06-Feb-06
2691 ---------------------
2693 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2694 in pcreposix.h. I have copied the definition from pcre.h.
2696 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2697 because pcre.h is no longer a built file.
2699 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2700 not normally included in the compiled code.
2703 Version 6.5 01-Feb-06
2704 ---------------------
2706 1. When using the partial match feature with pcre_dfa_exec(), it was not
2707 anchoring the second and subsequent partial matches at the new starting
2708 point. This could lead to incorrect results. For example, with the pattern
2709 /1234/, partially matching against "123" and then "a4" gave a match.
2711 2. Changes to pcregrep:
2713 (a) All non-match returns from pcre_exec() were being treated as failures
2714 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2715 error message is output. Some extra information is given for the
2717 probably the only errors that are likely to be caused by users (by
2718 specifying a regex that has nested indefinite repeats, for instance).
2719 If there are more than 20 of these errors, pcregrep is abandoned.
2721 (b) A binary zero was treated as data while matching, but terminated the
2722 output line if it was written out. This has been fixed: binary zeroes
2723 are now no different to any other data bytes.
2725 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2726 used to set a locale for matching. The --locale=xxxx long option has
2727 been added (no short equivalent) to specify a locale explicitly on the
2728 pcregrep command, overriding the environment variables.
2730 (d) When -B was used with -n, some line numbers in the output were one less
2731 than they should have been.
2733 (e) Added the -o (--only-matching) option.
2735 (f) If -A or -C was used with -c (count only), some lines of context were
2736 accidentally printed for the final match.
2738 (g) Added the -H (--with-filename) option.
2740 (h) The combination of options -rh failed to suppress file names for files
2741 that were found from directory arguments.
2743 (i) Added the -D (--devices) and -d (--directories) options.
2745 (j) Added the -F (--fixed-strings) option.
2747 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2749 (l) Added the --colo(u)r option.
2751 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2752 is not present by default.
2754 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2755 items such as (?R) or (?1), when the recursion could match a number of
2756 alternatives. If it matched one of the alternatives, but subsequently,
2757 outside the recursion, there was a failure, the code tried to back up into
2758 the recursion. However, because of the way PCRE is implemented, this is not
2759 possible, and the result was an incorrect result from the match.
2761 In order to prevent this happening, the specification of recursion has
2762 been changed so that all such subpatterns are automatically treated as
2763 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2765 4. I had overlooked the fact that, in some locales, there are characters for
2766 which isalpha() is true but neither isupper() nor islower() are true. In
2767 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2768 and ordfeminine) are like this. This affected the treatment of \w and \W
2769 when they appeared in character classes, but not when they appeared outside
2770 a character class. The bit map for "word" characters is now created
2771 separately from the results of isalnum() instead of just taking it from the
2772 upper, lower, and digit maps. (Plus the underscore character, of course.)
2774 5. The above bug also affected the handling of POSIX character classes such as
2775 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2776 permanent tables. Instead, the bit maps for such a class were previously
2777 created as the appropriate unions of the upper, lower, and digit bitmaps.
2778 Now they are created by subtraction from the [[:word:]] class, which has
2779 its own bitmap.
2781 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2782 It is created by subtracting the vertical space characters (\x09, \x0a,
2783 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2784 subtraction was done in the overall bitmap for a character class, meaning
2785 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2786 be recognized. This bug has been fixed.
2788 7. Patches from the folks at Google:
2790 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2791 real life, but is still worth protecting against".
2793 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2794 regular expressions".
2796 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2797 have it.
2799 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2800 "configure" and the latter not, in order to fix a problem somebody had
2801 with compiling the Arg class on HP-UX.
2803 (e) Improve the error-handling of the C++ wrapper a little bit.
2805 (f) New tests for checking recursion limiting.
2807 8. The pcre_memmove() function, which is used only if the environment does not
2808 have a standard memmove() function (and is therefore rarely compiled),
2809 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2810 returning a result (though PCRE never actually uses the result).
2812 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2813 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2814 returned instead of calling malloc() with an overflowing number that would
2815 most likely cause subsequent chaos.
2817 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2819 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2820 with this option is matched, the nmatch and pmatch options of regexec() are
2821 ignored.
2823 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2824 provided in case anyone wants to the the POSIX interface with UTF-8
2825 strings.
2827 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2828 C++ linking (needed for some HP-UX environments).
2830 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2831 (unused parameter) and in the pcre_printint() function (omitted "default"
2832 switch label when the default is to do nothing).
2834 15. Added some code to make it possible, when PCRE is compiled as a C++
2835 library, to replace subject pointers for pcre_exec() with a smart pointer
2836 class, thus making it possible to process discontinuous strings.
2838 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2839 much the same function. They were added by different people who were trying
2840 to make PCRE easy to compile on non-Unix systems. It has been suggested
2841 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2842 for compiling on Windows systems. I have therefore replaced it with
2843 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2844 defaults to "extern" for C or "extern C" for C++, which works fine on
2845 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2846 SCOPE with something explicit in config.h. In addition:
2848 (a) pcreposix.h still had just "extern" instead of either of these macros;
2849 I have replaced it with PCRE_DATA_SCOPE.
2851 (b) Functions such as _pcre_xclass(), which are internal to the library,
2852 but external in the C sense, all had PCRE_EXPORT in their definitions.
2853 This is apparently wrong for the Windows case, so I have removed it.
2854 (It makes no difference on Unix-like systems.)
2856 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2857 of recursive calls to match(). This is different to MATCH_LIMIT because
2858 that limits the total number of calls to match(), not all of which increase
2859 the depth of recursion. Limiting the recursion depth limits the amount of
2860 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2861 when PCRE is compiled, and changed at run time. A patch from Google adds
2862 this functionality to the C++ interface.
2864 18. Changes to the handling of Unicode character properties:
2866 (a) Updated the table to Unicode 4.1.0.
2868 (b) Recognize characters that are not in the table as "Cn" (undefined).
2870 (c) I revised the way the table is implemented to a much improved format
2871 which includes recognition of ranges. It now supports the ranges that
2872 are defined in UnicodeData.txt, and it also amalgamates other
2873 characters into ranges. This has reduced the number of entries in the
2874 table from around 16,000 to around 3,000, thus reducing its size
2875 considerably. I realized I did not need to use a tree structure after
2876 all - a binary chop search is just as efficient. Having reduced the
2877 number of entries, I extended their size from 6 bytes to 8 bytes to
2878 allow for more data.
2880 (d) Added support for Unicode script names via properties such as \p{Han}.
2882 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2883 matching that character.
2885 20. When matching a repeated Unicode property with a minimum greater than zero,
2886 (for example \pL{2,}), PCRE could look past the end of the subject if it
2887 reached it while seeking the minimum number of characters. This could
2888 happen only if some of the characters were more than one byte long, because
2889 there is a check for at least the minimum number of bytes.
2891 21. Refactored the implementation of \p and \P so as to be more general, to
2892 allow for more different types of property in future. This has changed the
2893 compiled form incompatibly. Anybody with saved compiled patterns that use
2894 \p or \P will have to recompile them.
2896 22. Added "Any" and "L&" to the supported property types.
2898 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2899 but give a compile time error if the value is greater than 0xff.
2901 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2902 accidentally not being installed or uninstalled.
2904 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2905 made were to insert the current release number. This seemed silly, because
2906 it made things harder for people building PCRE on systems that don't run
2907 "configure". I have turned pcre.h into a distributed file, no longer built
2908 by "configure", with the version identification directly included. There is
2909 no longer a pcre.h.in file.
2911 However, this change necessitated a change to the pcre-config script as
2912 well. It is built from pcre-config.in, and one of the substitutions was the
2913 release number. I have updated configure.ac so that ./configure now finds
2914 the release number by grepping pcre.h.
2916 26. Added the ability to run the tests under valgrind.
2919 Version 6.4 05-Sep-05
2920 ---------------------
2922 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2923 "--" to be printed when multiple files were scanned, even when none of the
2924 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2925 consider it to be a bug, and have restored the previous behaviour.
2927 2. A couple of code tidies to get rid of compiler warnings.
2929 3. The pcretest program used to cheat by referring to symbols in the library
2930 whose names begin with _pcre_. These are internal symbols that are not
2931 really supposed to be visible externally, and in some environments it is
2932 possible to suppress them. The cheating is now confined to including
2933 certain files from the library's source, which is a bit cleaner.
2935 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2936 file's purpose clearer.
2938 5. Reorganized pcre_ucp_findchar().
2941 Version 6.3 15-Aug-05
2942 ---------------------
2944 1. The file libpcre.pc.in did not have general read permission in the tarball.
2946 2. There were some problems when building without C++ support:
2948 (a) If C++ support was not built, "make install" and "make test" still
2949 tried to test it.
2951 (b) There were problems when the value of CXX was explicitly set. Some
2952 changes have been made to try to fix these, and ...
2954 (c) --disable-cpp can now be used to explicitly disable C++ support.
2956 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2957 backslash in a target when C++ was disabled. This confuses some
2958 versions of "make", apparently. Using an intermediate variable solves
2959 this. (Same for CPP_LOBJ.)
2961 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2962 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2963 necessary on certain architectures.
2965 4. Added a setting of -export-symbols-regex to the link command to remove
2966 those symbols that are exported in the C sense, but actually are local
2967 within the library, and not documented. Their names all begin with
2968 "_pcre_". This is not a perfect job, because (a) we have to except some
2969 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2970 available (and never for static libraries). I have made a note to try to
2971 find a way round (a) in the future.
2974 Version 6.2 01-Aug-05
2975 ---------------------
2977 1. There was no test for integer overflow of quantifier values. A construction
2978 such as {1111111111111111} would give undefined results. What is worse, if
2979 a minimum quantifier for a parenthesized subpattern overflowed and became
2980 negative, the calculation of the memory size went wrong. This could have
2981 led to memory overwriting.
2983 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2985 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2986 operating environments where this matters.
2988 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2989 PCRE options from within the C++ wrapper.
2991 5. Named capturing subpatterns were not being correctly counted when a pattern
2992 was compiled. This caused two problems: (a) If there were more than 100
2993 such subpatterns, the calculation of the memory needed for the whole
2994 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2995 back references of the form \12, where the number was greater than 9, were
2996 not recognized as back references, even though there were sufficient
2997 previous subpatterns.
2999 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3000 versions of gcc, e.g. 2.95.4.
3003 Version 6.1 21-Jun-05
3004 ---------------------
3006 1. There was one reference to the variable "posix" in pcretest.c that was not
3007 surrounded by "#if !defined NOPOSIX".
3009 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3010 the cross-check on the old pcre_info() function, for the benefit of the
3011 cut-down version of PCRE that is currently imported into Exim.
3013 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3014 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3015 patterns that start (?i) but I don't think that matters. The use of (?i) is
3016 just an example; this all applies to the other options as well.
3018 4. Since libtool seems to echo the compile commands it is issuing, the output
3019 from "make" can be reduced a bit by putting "@" in front of each libtool
3020 compile command.
3022 5. Patch from the folks at Google for configure.in to be a bit more thorough
3023 in checking for a suitable C++ installation before trying to compile the
3024 C++ stuff. This should fix a reported problem when a compiler was present,
3025 but no suitable headers.
3027 6. The man pages all had just "PCRE" as their title. I have changed them to
3028 be the relevant file name. I have also arranged that these names are
3029 retained in the file doc/pcre.txt, which is a concatenation in text format
3030 of all the man pages except the little individual ones for each function.
3032 7. The NON-UNIX-USE file had not been updated for the different set of source
3033 files that come with release 6. I also added a few comments about the C++
3034 wrapper.
3037 Version 6.0 07-Jun-05
3038 ---------------------
3040 1. Some minor internal re-organization to help with my DFA experiments.
3042 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3043 didn't matter for the library itself when fully configured, but did matter
3044 when compiling without UCP support, or within Exim, where the ucp files are
3045 not imported.
3047 3. Refactoring of the library code to split up the various functions into
3048 different source modules. The addition of the new DFA matching code (see
3049 below) to a single monolithic source would have made it really too
3050 unwieldy, quite apart from causing all the code to be include in a
3051 statically linked application, when only some functions are used. This is
3052 relevant even without the DFA addition now that patterns can be compiled in
3053 one application and matched in another.
3055 The downside of splitting up is that there have to be some external
3056 functions and data tables that are used internally in different modules of
3057 the library but which are not part of the API. These have all had their
3058 names changed to start with "_pcre_" so that they are unlikely to clash
3059 with other external names.
3061 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3062 a different (DFA) algorithm. Although it is slower than the original
3063 function, it does have some advantages for certain types of matching
3064 problem.
3066 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3067 including restarting after a partial match.
3069 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3070 defined when compiling for Windows was sent to me. I have put it into the
3071 code, though I have no means of testing or verifying it.
3073 7. Added the pcre_refcount() auxiliary function.
3075 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3076 match before or at the first newline in the subject string. In pcretest,
3077 the /f option on a pattern can be used to set this.
3079 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3080 would behave wrongly. This has been present in PCRE since release 4.0.
3082 10. A number of changes to the pcregrep command:
3084 (a) Refactored how -x works; insert ^(...)$ instead of setting
3085 PCRE_ANCHORED and checking the length, in preparation for adding
3086 something similar for -w.
3088 (b) Added the -w (match as a word) option.
3090 (c) Refactored the way lines are read and buffered so as to have more
3091 than one at a time available.
3093 (d) Implemented a pcregrep test script.
3095 (e) Added the -M (multiline match) option. This allows patterns to match
3096 over several lines of the subject. The buffering ensures that at least
3097 8K, or the rest of the document (whichever is the shorter) is available
3098 for matching (and similarly the previous 8K for lookbehind assertions).
3100 (f) Changed the --help output so that it now says
3102 -w, --word-regex(p)
3104 instead of two lines, one with "regex" and the other with "regexp"
3105 because that confused at least one person since the short forms are the
3106 same. (This required a bit of code, as the output is generated
3107 automatically from a table. It wasn't just a text change.)
3109 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3110 option but starts with a hyphen. Could be a pattern or a path name
3111 starting with a hyphen, for instance.
3113 (h) "-" can be given as a file name to represent stdin.
3115 (i) When file names are being printed, "(standard input)" is used for
3116 the standard input, for compatibility with GNU grep. Previously
3117 "<stdin>" was used.
3119 (j) The option --label=xxx can be used to supply a name to be used for
3120 stdin when file names are being printed. There is no short form.
3122 (k) Re-factored the options decoding logic because we are going to add
3123 two more options that take data. Such options can now be given in four
3124 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3126 (l) Added the -A, -B, and -C options for requesting that lines of context
3127 around matches be printed.
3129 (m) Added the -L option to print the names of files that do not contain
3130 any matching lines, that is, the complement of -l.
3132 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3133 continue to scan other files.
3135 (o) The -s option was incorrectly implemented. For compatibility with other
3136 greps, it now suppresses the error message for a non-existent or non-
3137 accessible file (but not the return code). There is a new option called
3138 -q that suppresses the output of matching lines, which was what -s was
3139 previously doing.
3141 (p) Added --include and --exclude options to specify files for inclusion
3142 and exclusion when recursing.
3144 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3145 Hopefully, it now does.
3147 12. Missing cast in pcre_study().
3149 13. Added an "uninstall" target to the makefile.
3151 14. Replaced "extern" in the function prototypes in Makefile.in with
3152 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3153 world, but is set differently for Windows.
3155 15. Added a second compiling function called pcre_compile2(). The only
3156 difference is that it has an extra argument, which is a pointer to an
3157 integer error code. When there is a compile-time failure, this is set
3158 non-zero, in addition to the error test pointer being set to point to an
3159 error message. The new argument may be NULL if no error number is required
3160 (but then you may as well call pcre_compile(), which is now just a
3161 wrapper). This facility is provided because some applications need a
3162 numeric error indication, but it has also enabled me to tidy up the way
3163 compile-time errors are handled in the POSIX wrapper.
3165 16. Added VPATH=.libs to the makefile; this should help when building with one
3166 prefix path and installing with another. (Or so I'm told by someone who
3167 knows more about this stuff than I do.)
3169 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3170 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3171 match everything, including newlines. This is not POSIX-compatible, but
3172 somebody wanted the feature. From pcretest it can be activated by using
3173 both the P and the s flags.
3175 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3177 19. libpcre.pc was being incorrectly installed as executable.
3179 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3180 it now also looks for '\r' so that it will work unmodified on Windows.
3182 21. Added Google's contributed C++ wrapper to the distribution.
3184 22. Added some untidy missing memory free() calls in pcretest, to keep
3185 Electric Fence happy when testing.
3189 Version 5.0 13-Sep-04
3190 ---------------------
3192 1. Internal change: literal characters are no longer packed up into items
3193 containing multiple characters in a single byte-string. Each character
3194 is now matched using a separate opcode. However, there may be more than one
3195 byte in the character in UTF-8 mode.
3197 2. The pcre_callout_block structure has two new fields: pattern_position and
3198 next_item_length. These contain the offset in the pattern to the next match
3199 item, and its length, respectively.
3201 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3202 insertion of callouts before each pattern item. Added the /C option to
3203 pcretest to make use of this.
3205 4. On the advice of a Windows user, the lines
3207 #if defined(_WIN32) || defined(WIN32)
3208 _setmode( _fileno( stdout ), 0x8000 );
3209 #endif /* defined(_WIN32) || defined(WIN32) */
3211 have been added to the source of pcretest. This apparently does useful
3212 magic in relation to line terminators.
3214 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3215 for the benefit of those environments where the "b" makes a difference.
3217 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3218 to know about it. I have put a hack into configure.in that adds in code
3219 to set GCC=yes if CC=icc. This seems to end up at a point in the
3220 generated configure script that is early enough to affect the setting of
3221 compiler options, which is what is needed, but I have no means of testing
3222 whether it really works. (The user who reported this had patched the
3223 generated configure script, which of course I cannot do.)
3225 LATER: After change 22 below (new libtool files), the configure script
3226 seems to know about icc (and also ecc). Therefore, I have commented out
3227 this hack in configure.in.
3229 7. Added support for pkg-config (2 patches were sent in).
3231 8. Negated POSIX character classes that used a combination of internal tables
3232 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3233 [[:^ascii]]. Typically, they would match almost any characters. The other
3234 POSIX classes were not broken in this way.
3236 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3237 to find the match, as PCRE was deluded into thinking that the match had to
3238 start at the start point or following a newline. The same bug applied to
3239 patterns with negative forward assertions or any backward assertions
3240 preceding ".*" at the start, unless the pattern required a fixed first
3241 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3243 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3244 starting at the last subject character, bytes beyond the end of the subject
3245 string were read.
3247 11. Renamed the variable "class" as "classbits" to make life easier for C++
3248 users. (Previously there was a macro definition, but it apparently wasn't
3249 enough.)
3251 12. Added the new field "tables" to the extra data so that tables can be passed
3252 in at exec time, or the internal tables can be re-selected. This allows
3253 a compiled regex to be saved and re-used at a later time by a different
3254 program that might have everything at different addresses.
3256 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3257 -R library as well as a -L library.
3259 14. The debugging options of pcretest (-d on the command line or D on a
3260 pattern) showed incorrect output for anything following an extended class
3261 that contained multibyte characters and which was followed by a quantifier.
3263 15. Added optional support for general category Unicode character properties
3264 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3265 support. It adds about 90K to the size of the library. The meanings of the
3266 inbuilt class escapes such as \d and \s have NOT been changed.
3268 16. Updated pcredemo.c to include calls to free() to release the memory for the
3269 compiled pattern.
3271 17. The generated file chartables.c was being created in the source directory
3272 instead of in the building directory. This caused the build to fail if the
3273 source directory was different from the building directory, and was
3274 read-only.
3276 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3277 file. No doubt somebody will tell me if they don't make sense... Also added
3278 Dan Mooney's comments about building on OpenVMS.
3280 19. Added support for partial matching via the PCRE_PARTIAL option for
3281 pcre_exec() and the \P data escape in pcretest.
3283 20. Extended pcretest with 3 new pattern features:
3285 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3286 write the compiled pattern to the file whose name is "rest-of-line".
3287 This is a straight binary dump of the data, with the saved pointer to
3288 the character tables forced to be NULL. The study data, if any, is
3289 written too. After writing, pcretest reads a new pattern.
3291 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3292 compiled pattern from the given file. There must not be any
3293 occurrences of "<" in the file name (pretty unlikely); if there are,
3294 pcretest will instead treat the initial "<" as a pattern delimiter.
3295 After reading in the pattern, pcretest goes on to read data lines as
3296 usual.
3298 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3299 and 16-bit fields in a compiled pattern, to simulate a pattern that
3300 was compiled on a host of opposite endianness.
3302 21. The pcre-exec() function can now cope with patterns that were compiled on
3303 hosts of opposite endianness, with this restriction:
3305 As for any compiled expression that is saved and used later, the tables
3306 pointer field cannot be preserved; the extra_data field in the arguments
3307 to pcre_exec() should be used to pass in a tables address if a value
3308 other than the default internal tables were used at compile time.
3310 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3311 now diagnosed as an error. Previously, most of the time, a negative number
3312 would have been treated as zero, but if in addition "ovector" was passed as
3313 NULL, a crash could occur.
3315 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3316 new versions from the libtool 1.5 distribution (the last one is a copy of
3317 a file called libtool.m4). This seems to have fixed the need to patch
3318 "configure" to support Darwin 1.3 (which I used to do). However, I still
3319 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3320 workstation).
3322 24. Changed the PCRE licence to be the more standard "BSD" licence.
3325 Version 4.5 01-Dec-03
3326 ---------------------
3328 1. There has been some re-arrangement of the code for the match() function so
3329 that it can be compiled in a version that does not call itself recursively.
3330 Instead, it keeps those local variables that need separate instances for
3331 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3332 needs to "recurse". Keeping track of where control must go is done by means
3333 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3334 hide most of the details from the main code, and operates only if
3335 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3336 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3337 operating.
3339 To make it easier for callers to provide specially tailored get/free
3340 functions for this usage, two new functions, pcre_stack_malloc, and
3341 pcre_stack_free, are used. They are always called in strict stacking order,
3342 and the size of block requested is always the same.
3344 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3345 PCRE has been compiled to use the stack or the heap for recursion. The
3346 -C option of pcretest uses this to show which version is compiled.
3348 A new data escape \S, is added to pcretest; it causes the amounts of store
3349 obtained and freed by both kinds of malloc/free at match time to be added
3350 to the output.
3352 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3353 what's available on my current Linux desktop machine.
3355 3. When matching a UTF-8 string, the test for a valid string at the start has
3356 been extended. If start_offset is not zero, PCRE now checks that it points
3357 to a byte that is the start of a UTF-8 character. If not, it returns
3358 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3359 this is necessary because there may be backward assertions in the pattern.
3360 When matching the same subject several times, it may save resources to use
3361 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3363 4. The code for checking the validity of UTF-8 strings has been tightened so
3364 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3365 containing "overlong sequences".
3367 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3368 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3369 should have been "&", but it just so happened that all the cases this let
3370 through by mistake were picked up later in the function.
3372 6. I had used a variable called "isblank" - this is a C99 function, causing
3373 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3375 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3376 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3377 so the version is shown; (c) stop "make test" echoing "./RunTest".
3379 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3381 9. The prototype for memmove() for systems that don't have it was using
3382 size_t, but the inclusion of the header that defines size_t was later. I've
3383 moved the #includes for the C headers earlier to avoid this.
3385 10. Added some adjustments to the code to make it easier to compiler on certain
3386 special systems:
3388 (a) Some "const" qualifiers were missing.
3389 (b) Added the macro EXPORT before all exported functions; by default this
3390 is defined to be empty.
3391 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3392 that it reads its output file name as an argument instead of writing
3393 to the standard output and assuming this can be redirected.
3395 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3396 class containing characters with values greater than 255, PCRE compilation
3397 went into a loop.
3399 12. A recursive reference to a subpattern that was within another subpattern
3400 that had a minimum quantifier of zero caused PCRE to crash. For example,
3401 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3402 recursion. If the recursively-called subpattern itself had a zero repeat,
3403 that was OK.
3405 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3406 buffer into which it was copied (for escape processing) was still set at
3407 1024, so long lines caused crashes.
3409 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3410 "internal error: code overflow...". This applied to any character class
3411 that was followed by a possessive quantifier.
3413 15. Modified the Makefile to add libpcre.la as a prerequisite for
3414 libpcreposix.la because I was told this is needed for a parallel build to
3415 work.
3417 16. If a pattern that contained .* following optional items at the start was
3418 studied, the wrong optimizing data was generated, leading to matching
3419 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3420 matching string must start with a or b or c. The correct conclusion for
3421 this pattern is that a match can start with any character.
3424 Version 4.4 13-Aug-03
3425 ---------------------
3427 1. In UTF-8 mode, a character class containing characters with values between
3428 127 and 255 was not handled correctly if the compiled pattern was studied.
3429 In fixing this, I have also improved the studying algorithm for such
3430 classes (slightly).
3432 2. Three internal functions had redundant arguments passed to them. Removal
3433 might give a very teeny performance improvement.
3435 3. Documentation bug: the value of the capture_top field in a callout is *one
3436 more than* the number of the hightest numbered captured substring.
3438 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3439 in incorrectly linking with a previously installed version. They now link
3440 explicitly with libpcre.la.
3442 5. configure.in no longer needs to recognize Cygwin specially.
3444 6. A problem in pcre.in for Windows platforms is fixed.
3446 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3447 pcretest, it used to include the size of the study block as part of its
3448 output. Unfortunately, the structure contains a field that has a different
3449 size on different hardware architectures. This meant that the tests that
3450 showed this size failed. As the block is currently always of a fixed size,
3451 this information isn't actually particularly useful in pcretest output, so
3452 I have just removed it.
3454 8. Three pre-processor statements accidentally did not start in column 1.
3455 Sadly, there are *still* compilers around that complain, even though
3456 standard C has not required this for well over a decade. Sigh.
3458 9. In pcretest, the code for checking callouts passed small integers in the
3459 callout_data field, which is a void * field. However, some picky compilers
3460 complained about the casts involved for this on 64-bit systems. Now
3461 pcretest passes the address of the small integer instead, which should get
3462 rid of the warnings.
3464 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3465 both compile and run time, and gives an error if an invalid UTF-8 sequence
3466 is found. There is a option for disabling this check in cases where the
3467 string is known to be correct and/or the maximum performance is wanted.
3469 11. In response to a bug report, I changed one line in Makefile.in from
3471 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3472 to
3473 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3475 to look similar to other lines, but I have no way of telling whether this
3476 is the right thing to do, as I do not use Windows. No doubt I'll get told
3477 if it's wrong...
3480 Version 4.3 21-May-03
3481 ---------------------
3483 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3484 Makefile.
3486 2. Some refactoring to improve the quality of the code:
3488 (i) The utf8_table... variables are now declared "const".
3490 (ii) The code for \cx, which used the "case flipping" table to upper case
3491 lower case letters, now just substracts 32. This is ASCII-specific,
3492 but the whole concept of \cx is ASCII-specific, so it seems
3493 reasonable.
3495 (iii) PCRE was using its character types table to recognize decimal and
3496 hexadecimal digits in the pattern. This is silly, because it handles
3497 only 0-9, a-f, and A-F, but the character types table is locale-
3498 specific, which means strange things might happen. A private
3499 table is now used for this - though it costs 256 bytes, a table is
3500 much faster than multiple explicit tests. Of course, the standard
3501 character types table is still used for matching digits in subject
3502 strings against \d.
3504 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3505 ending in _t are). So I've renamed it as ESC_tee.
3507 3. The first argument for regexec() in the POSIX wrapper should have been
3508 defined as "const".
3510 4. Changed pcretest to use malloc() for its buffers so that they can be
3511 Electric Fenced for debugging.
3513 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3514 to read one or more bytes before the start of the subject string. Often this
3515 had no effect on PCRE's behaviour, but in some circumstances it could
3516 provoke a segmentation fault.
3518 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3519 to try to read one or more bytes before the start of the subject string.
3521 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3522 UTF-8 support could misbehave in various ways if the subject string
3523 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3524 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3525 back over UTF-8 characters.)
3528 Version 4.2 14-Apr-03
3529 ---------------------
3531 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3533 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3534 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3535 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3536 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3537 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3539 Note: automatic setting of the BUILD variables is not yet working
3540 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3541 compile-time but not at link-time
3542 [LINK]: use for linking executables only
3543 make different versions for Windows and non-Windows
3544 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3545 libraries
3546 [LINK_FOR_BUILD]: new variable
3547 [OBJEXT]: use throughout
3548 [EXEEXT]: use throughout
3549 <winshared>: new target
3550 <wininstall>: new target
3551 <dftables.o>: use native compiler
3552 <dftables>: use native linker
3553 <install>: handle Windows platform correctly
3554 <clean>: ditto
3555 <check>: ditto
3556 copy DLL to top builddir before testing
3558 As part of these changes, -no-undefined was removed again. This was reported
3559 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3560 in any case.
3562 3. Some tidies to get rid of compiler warnings:
3564 . In the match_data structure, match_limit was an unsigned long int, whereas
3565 match_call_count was an int. I've made them both unsigned long ints.
3567 . In pcretest the fact that a const uschar * doesn't automatically cast to
3568 a void * provoked a warning.
3570 . Turning on some more compiler warnings threw up some "shadow" variables
3571 and a few more missing casts.
3573 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3574 option, a class that contained a single character with a value between 128
3575 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3577 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3578 option, a class that contained several characters, but with at least one
3579 whose value was between 128 and 255 caused PCRE to crash.
3582 Version 4.1 12-Mar-03
3583 ---------------------
3585 1. Compiling with gcc -pedantic found a couple of places where casts were
3586 needed, and a string in dftables.c that was longer than standard compilers are
3587 required to support.
3589 2. Compiling with Sun's compiler found a few more places where the code could
3590 be tidied up in order to avoid warnings.
3592 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3593 first of these names is deprecated in the latest Autoconf in favour of the name
3594 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3595 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3596 analogy I have changed it to CFLAGS_FOR_BUILD.
3598 4. Added -no-undefined to the linking command in the Makefile, because this is
3599 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3600 linking step for the pcreposix library.
3602 5. PCRE was failing to diagnose the case of two named groups with the same
3603 name.
3605 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3606 literal character that is needed in the subject for a match, and scans along to
3607 ensure that it is present before embarking on the full matching process. This
3608 saves time in cases of nested unlimited repeats that are never going to match.
3609 Problem: the scan can take a lot of time if the subject is very long (e.g.
3610 megabytes), thus penalizing straightforward matches. It is now done only if the
3611 amount of subject to be scanned is less than 1000 bytes.
3613 7. A lesser problem with the same optimization is that it was recording the
3614 first character of an anchored pattern as "needed", thus provoking a search
3615 right along the subject, even when the first match of the pattern was going to
3616 fail. The "needed" character is now not set for anchored patterns, unless it
3617 follows something in the pattern that is of non-fixed length. Thus, it still
3618 fulfils its original purpose of finding quick non-matches in cases of nested
3619 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3622 Version 4.0 17-Feb-03
3623 ---------------------
3625 1. If a comment in an extended regex that started immediately after a meta-item
3626 extended to the end of string, PCRE compiled incorrect data. This could lead to
3627 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3629 2. Moved to autoconf 2.53 and libtool 1.4.2.
3631 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3632 the special perltest8 script is no longer needed - all the tests can be run
3633 from a single perltest script.
3635 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3636 by \s. It has now been removed in PCRE. This means it isn't recognized as
3637 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3638 class [:space:] *does* include VT, thereby creating a mess.
3640 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3641 space and tab.
3643 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3644 its new features into the main test script, reducing the number of scripts.
3646 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3647 were backward compatible, and made the (?i) apply to the whole pattern, as if
3648 /i were given. Now it behaves more logically, and applies the option setting
3649 only to what follows. PCRE has been changed to follow suit. However, if it
3650 finds options settings right at the start of the pattern, it extracts them into
3651 the global options, as before. Thus, they show up in the info data.
3653 8. Added support for the \Q...\E escape sequence. Characters in between are
3654 treated as literals. This is slightly different from Perl in that $ and @ are
3655 also handled as literals inside the quotes. In Perl, they will cause variable
3656 interpolation. Note the following examples:
3658 Pattern PCRE matches Perl matches
3660 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3661 \Qabc\$xyz\E abc\$xyz abc\$xyz
3662 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3664 For compatibility with Perl, \Q...\E sequences are recognized inside character
3665 classes as well as outside them.
3667 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3668 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3669 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3670 signed/unsigned warnings.
3672 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3673 option for pcretest, so I've replaced it by a simple function that does just
3674 that job.
3676 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3677 "pcregrep -".
3679 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3680 Java package. This provides some syntactic sugar for simple cases of what my
3681 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3682 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3683 item, you can use this simplified notation. Note that only makes sense with
3684 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3685 greediness, whatever the setting of the PCRE_UNGREEDY option.
3687 13. A change of greediness default within a pattern was not taking effect at
3688 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3689 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3690 was abstracted outside.
3692 14. PCRE now supports the \G assertion. It is true when the current matching
3693 position is at the start point of the match. This differs from \A when the
3694 starting offset is non-zero. Used with the /g option of pcretest (or similar
3695 code), it works in the same way as it does for Perl's /g option. If all
3696 alternatives of a regex begin with \G, the expression is anchored to the start
3697 match position, and the "anchored" flag is set in the compiled expression.
3699 15. Some bugs concerning the handling of certain option changes within patterns
3700 have been fixed. These applied to options other than (?ims). For example,
3701 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3702 been the other way round. Some of this was related to change 7 above.
3704 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3705 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3706 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3707 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3709 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3710 mode. Unlike ".", it always matches newline, whatever the setting of
3711 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3712 assertions. Perl allows it, but it doesn't (in general) work because it can't
3713 calculate the length of the lookbehind. At least, that's the case for Perl
3714 5.8.0 - I've been told they are going to document that it doesn't work in
3715 future.
3717 18. Added an error diagnosis for escapes that PCRE does not support: these are
3718 \L, \l, \N, \P, \p, \U, \u, and \X.
3720 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3721 reading past the end of the pattern in cases such as /[abcd/.
3723 20. PCRE was getting more memory than necessary for patterns with classes that
3724 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3726 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3727 compiling PCRE for use with Virtual Pascal.
3729 22. Small fix to the Makefile to make it work properly if the build is done
3730 outside the source tree.
3732 23. Added a new extension: a condition to go with recursion. If a conditional
3733 subpattern starts with (?(R) the "true" branch is used if recursion has
3734 happened, whereas the "false" branch is used only at the top level.
3736 24. When there was a very long string of literal characters (over 255 bytes
3737 without UTF support, over 250 bytes with UTF support), the computation of how
3738 much memory was required could be incorrect, leading to segfaults or other
3739 strange effects.
3741 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3742 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3743 there was a subsequent back reference to those brackets. This meant that, for
3744 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3745 possible to check for precisely this case. All we can do is abandon the
3746 optimization if .* occurs inside capturing brackets when there are any back
3747 references whatsoever. (See below for a better fix that came later.)
3749 26. The handling of the optimization for finding the first character of a
3750 non-anchored pattern, and for finding a character that is required later in the
3751 match were failing in some cases. This didn't break the matching; it just
3752 failed to optimize when it could. The way this is done has been re-implemented.
3754 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3756 28. Added a new feature that provides some of the functionality that Perl
3757 provides with (?{...}). The facility is termed a "callout". The way it is done
3758 in PCRE is for the caller to provide an optional function, by setting
3759 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3760 global variable. By default it is unset, which disables all calling out. To get
3761 the function called, the regex must include (?C) at appropriate points. This
3762 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3763 This provides a means of identifying different callout points. When PCRE
3764 reaches such a point in the regex, if pcre_callout has been set, the external
3765 function is called. It is provided with data in a structure called
3766 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3767 matching continues; if it returns a non-zero value, the match at the current
3768 point fails. However, backtracking will occur if possible. [This was changed
3769 later and other features added - see item 49 below.]
3771 29. pcretest is upgraded to test the callout functionality. It provides a
3772 callout function that displays information. By default, it shows the start of
3773 the match and the current position in the text. There are some new data escapes
3774 to vary what happens:
3776 \C+ in addition, show current contents of captured substrings
3777 \C- do not supply a callout function
3778 \C!n return 1 when callout number n is reached
3779 \C!n!m return 1 when callout number n is reached for the mth time
3781 30. If pcregrep was called with the -l option and just a single file name, it
3782 output "<stdin>" if a match was found, instead of the file name.
3784 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3785 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3786 pcre_exec(). This saves a malloc/free per call. The default value of
3787 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3788 when configuring.
3790 32. The default maximum size of a compiled pattern is 64K. There have been a
3791 few cases of people hitting this limit. The code now uses macros to handle the
3792 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3793 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3794 configuring. Tests 2 and 5 work only with 2-byte links because they output
3795 debugging information about compiled patterns.
3797 33. Internal code re-arrangements:
3799 (a) Moved the debugging function for printing out a compiled regex into
3800 its own source file (printint.c) and used #include to pull it into
3801 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3802 separate copies.
3804 (b) Defined the list of op-code names for debugging as a macro in
3805 internal.h so that it is next to the definition of the opcodes.
3807 (c) Defined a table of op-code lengths for simpler skipping along compiled
3808 code. This is again a macro in internal.h so that it is next to the
3809 definition of the opcodes.
3811 34. Added support for recursive calls to individual subpatterns, along the
3812 lines of Robin Houston's patch (but implemented somewhat differently).
3814 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3815 allow it to read and process whole directories in Win32. This code was
3816 contributed by Lionel Fourquaux; it has not been tested by me.
3818 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3819 used to name a group. Names consist of alphanumerics and underscores, and must
3820 be unique. Back references use the syntax (?P=name) and recursive calls use
3821 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3822 numbers. The function pcre_fullinfo() can be used after compilation to extract
3823 a name/number map. There are three relevant calls:
3825 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3826 PCRE_INFO_NAMECOUNT yields the number of entries
3827 PCRE_INFO_NAMETABLE yields a pointer to the map.
3829 The map is a vector of fixed-size entries. The size of each entry depends on
3830 the length of the longest name used. The first two bytes of each entry are the
3831 group number, most significant byte first. There follows the corresponding
3832 name, zero terminated. The names are in alphabetical order.
3834 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3835 case instead of 255. Making it the same both with and without UTF-8 support
3836 means that the same test output works with both.
3838 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3839 calling malloc() with a zero argument.
3841 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3842 optimization. I've improved things by keeping a bitmap of backreferences with
3843 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3844 fact referenced, the optimization can be applied. It is unlikely that a
3845 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3846 the match to follow \n) will appear inside brackets with a number greater than
3847 31, but if it does, any back reference > 31 suppresses the optimization.
3849 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3850 of disabling numbered capturing parentheses. Any opening parenthesis that is
3851 not followed by ? behaves as if it were followed by ?: but named parentheses
3852 can still be used for capturing (and they will acquire numbers in the usual
3853 way).
3855 41. Redesigned the return codes from the match() function into yes/no/error so
3856 that errors can be passed back from deep inside the nested calls. A malloc
3857 failure while inside a recursive subpattern call now causes the
3858 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3860 42. It is now possible to set a limit on the number of times the match()
3861 function is called in a call to pcre_exec(). This facility makes it possible to
3862 limit the amount of recursion and backtracking, though not in a directly
3863 obvious way, because the match() function is used in a number of different
3864 circumstances. The count starts from zero for each position in the subject
3865 string (for non-anchored patterns). The default limit is, for compatibility, a
3866 large number, namely 10 000 000. You can change this in two ways:
3868 (a) When configuring PCRE before making, you can use --with-match-limit=n
3869 to set a default value for the compiled library.
3871 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3872 a different value is set. See 45 below.
3874 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3876 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3877 of things that can be changed at compile time. The first argument specifies
3878 what is wanted and the second points to where the information is to be placed.
3879 The current list of available information is:
3883 The output is an integer that is set to one if UTF-8 support is available;
3884 otherwise it is set to zero.
3888 The output is an integer that it set to the value of the code that is used for
3889 newline. It is either LF (10) or CR (13).
3893 The output is an integer that contains the number of bytes used for internal
3894 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3898 The output is an integer that contains the threshold above which the POSIX
3899 interface uses malloc() for output vectors. See item 31 above.
3903 The output is an unsigned integer that contains the default limit of the number
3904 of match() calls in a pcre_exec() execution. See 42 above.
3906 44. pcretest has been upgraded by the addition of the -C option. This causes it
3907 to extract all the available output from the new pcre_config() function, and to
3908 output it. The program then exits immediately.
3910 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3911 order to support additional features. One way would have been to define
3912 pcre_exec2() (for example) with extra arguments, but this would not have been
3913 extensible, and would also have required all calls to the original function to
3914 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3915 is used for passing in "extra" data from pcre_study().
3917 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3918 contains the following fields:
3920 flags a bitmap indicating which of the following fields are set
3921 study_data opaque data from pcre_study()
3922 match_limit a way of specifying a limit on match() calls for a specific
3923 call to pcre_exec()
3924 callout_data data for callouts (see 49 below)
3926 The flag bits are also defined in pcre.h, and are
3932 The pcre_study() function now returns one of these new pcre_extra blocks, with
3933 the actual study data pointed to by the study_data field, and the
3934 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3935 before. That is, this change is entirely upwards-compatible and requires no
3936 change to existing code.
3938 If you want to pass in additional data to pcre_exec(), you can either place it
3939 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3940 block.
3942 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3943 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3944 times with different match limits, until it finds the minimum value needed for
3945 pcre_exec() to complete. The value is then output. This can be instructive; for
3946 most simple matches the number is quite small, but for pathological cases it
3947 gets very large very quickly.
3949 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3950 returns the size of the data block pointed to by the study_data field in a
3951 pcre_extra block, that is, the value that was passed as the argument to
3952 pcre_malloc() when PCRE was getting memory in which to place the information
3953 created by pcre_study(). The fourth argument should point to a size_t variable.
3954 pcretest has been extended so that this information is shown after a successful
3955 pcre_study() call when information about the compiled regex is being displayed.
3957 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3958 because what follows is always an absolute path. (Later: it turns out that this
3959 is more than cosmetic for MinGW, because it doesn't like empty path
3960 components.)
3962 49. Some changes have been made to the callout feature (see 28 above):
3964 (i) A callout function now has three choices for what it returns:
3966 0 => success, carry on matching
3967 > 0 => failure at this point, but backtrack if possible
3968 < 0 => serious error, return this value from pcre_exec()
3970 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3971 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3972 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3973 use by callout functions. It will never be used by PCRE itself.
3975 (ii) The pcre_extra structure (see 45 above) has a void * field called
3976 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3977 pcre_callout_block structure has a field of the same name. The contents of
3978 the field passed in the pcre_extra structure are passed to the callout
3979 function in the corresponding field in the callout block. This makes it
3980 easier to use the same callout-containing regex from multiple threads. For
3981 testing, the pcretest program has a new data escape
3983 \C*n pass the number n (may be negative) as callout_data
3985 If the callout function in pcretest receives a non-zero value as
3986 callout_data, it returns that value.
3988 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3989 there were some redundant $(CFLAGS) in commands that are now specified as
3990 $(LINK), which already includes $(CFLAGS).
3992 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3993 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3994 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3995 one-byte characters throughout. Note that case-insensitive matching applies
3996 only to characters whose values are less than 256. PCRE doesn't support the
3997 notion of cases for higher-valued characters.
3999 (i) A character class whose characters are all within 0-255 is handled as
4000 a bit map, and the map is inverted for negative classes. Previously, a
4001 character > 255 always failed to match such a class; however it should
4002 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4004 (ii) A negated character class with a single character < 255 is coded as
4005 "not this character" (OP_NOT). This wasn't working properly when the test
4006 character was multibyte, either singly or repeated.
4008 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4009 mode, for example: \x{100}{2,3}.
4011 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4012 singly or repeated) now correctly test multibyte characters. However,
4013 PCRE doesn't recognize any characters with values greater than 255 as
4014 digits, spaces, or word characters. Such characters always match \D, \S,
4015 and \W, and never match \d, \s, or \w.
4017 (v) Classes may now contain characters and character ranges with values
4018 greater than 255. For example: [ab\x{100}-\x{400}].
4020 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4021 PCRE in UTF-8 mode.
4023 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4024 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4025 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4026 value.)
4028 53. The single man page has become too large. I have therefore split it up into
4029 a number of separate man pages. These also give rise to individual HTML pages;
4030 these are now put in a separate directory, and there is an index.html page that
4031 lists them all. Some hyperlinking between the pages has been installed.
4033 54. Added convenience functions for handling named capturing parentheses.
4035 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4036 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4037 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4038 are faulted.
4040 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4041 calling configure. These values are used when compiling the dftables.c program
4042 which is run to generate the source of the default character tables. They
4043 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4044 you will need to set these values.
4046 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4049 Version 3.9 02-Jan-02
4050 ---------------------
4052 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4054 2. If --disable-static was given, the building process failed when trying to
4055 build pcretest and pcregrep. (For some reason it was using libtool to compile
4056 them, which is not right, as they aren't part of the library.)
4059 Version 3.8 18-Dec-01
4060 ---------------------
4062 1. The experimental UTF-8 code was completely screwed up. It was packing the
4063 bytes in the wrong order. How dumb can you get?
4066 Version 3.7 29-Oct-01
4067 ---------------------
4069 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4070 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4071 this didn't happen under Solaris 8, where I normally test things.
4073 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4074 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4075 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4076 this fix an hour or so after the initial 3.7 release.)
4079 Version 3.6 23-Oct-01
4080 ---------------------
4082 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4083 offsets passed as NULL with zero offset count.
4085 2. The config.guess and config.sub files had not been updated when I moved to
4086 the latest autoconf.
4089 Version 3.5 15-Aug-01
4090 ---------------------
4092 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4093 had been forgotten.
4095 2. By using declared but undefined structures, we can avoid using "void"
4096 definitions in pcre.h while keeping the internal definitions of the structures
4097 private.
4099 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4100 user point of view, this means that both static and shared libraries are built
4101 by default, but this can be individually controlled. More of the work of
4102 handling this static/shared cases is now inside libtool instead of PCRE's make
4103 file.
4105 4. The pcretest utility is now installed along with pcregrep because it is
4106 useful for users (to test regexs) and by doing this, it automatically gets
4107 relinked by libtool. The documentation has been turned into a man page, so
4108 there are now .1, .txt, and .html versions in /doc.
4110 5. Upgrades to pcregrep:
4111 (i) Added long-form option names like gnu grep.
4112 (ii) Added --help to list all options with an explanatory phrase.
4113 (iii) Added -r, --recursive to recurse into sub-directories.
4114 (iv) Added -f, --file to read patterns from a file.
4116 6. pcre_exec() was referring to its "code" argument before testing that
4117 argument for NULL (and giving an error if it was NULL).
4119 7. Upgraded Makefile.in to allow for compiling in a different directory from
4120 the source directory.
4122 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4123 options bits, the pointer it was passed was to an int instead of to an unsigned
4124 long int. This mattered only on 64-bit systems.
4126 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4127 generated) instead of pcre.in, which it its source. Also made the same change
4128 in several of the .c files.
4130 10. A new release of gcc defines printf() as a macro, which broke pcretest
4131 because it had an ifdef in the middle of a string argument for printf(). Fixed
4132 by using separate calls to printf().
4134 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4135 script, to force use of CR or LF instead of \n in the source. On non-Unix
4136 systems, the value can be set in config.h.
4138 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4139 absolute limit. Changed the text of the error message to make this clear, and
4140 likewise updated the man page.
4142 13. The limit of 99 on the number of capturing subpatterns has been removed.
4143 The new limit is 65535, which I hope will not be a "real" limit.
4146 Version 3.4 22-Aug-00
4147 ---------------------
4149 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4151 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4154 Version 3.3 01-Aug-00
4155 ---------------------
4157 1. If an octal character was given, but the value was greater than \377, it
4158 was not getting masked to the least significant bits, as documented. This could
4159 lead to crashes in some systems.
4161 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4162 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4164 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4165 These just pass their arguments on to (pcre_free)(), but they are provided
4166 because some uses of PCRE bind it to non-C systems that can call its functions,
4167 but cannot call free() or pcre_free() directly.
4169 4. Add "make test" as a synonym for "make check". Corrected some comments in
4170 the Makefile.
4172 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4173 Makefile.
4175 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4176 command called pgrep for grepping around the active processes.
4178 7. Added the beginnings of support for UTF-8 character strings.
4180 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4181 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4182 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4183 out for the ar command.)
4186 Version 3.2 12-May-00
4187 ---------------------
4189 This is purely a bug fixing release.
4191 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4192 of ZA. This was just one example of several cases that could provoke this bug,
4193 which was introduced by change 9 of version 2.00. The code for breaking
4194 infinite loops after an iteration that matches an empty string was't working
4195 correctly.
4197 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4198 when matched against abbab (for example). After matching an empty string, it
4199 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4200 caused it to match further down the string than it should.
4202 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4203 was there because it doesn't seem to be needed, and it causes trouble on some
4204 systems, as it is not a Standard C header. It has been removed.
4206 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4207 were reported on the Macintosh. The changes were from
4209 while ((c = *(++ptr)) != 0 && c != '\n');
4210 to
4211 while ((c = *(++ptr)) != 0 && c != '\n') ;
4213 Totally extraordinary, but if that's what it takes...
4215 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4216 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4217 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4218 assumes the way PCRE uses memmove() (always moving upwards).
4220 6. PCRE is being used in one environment where strchr() is not available. There
4221 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4222 faster code anyway.
4225 Version 3.1 09-Feb-00
4226 ---------------------
4228 The only change in this release is the fixing of some bugs in Makefile.in for
4229 the "install" target:
4231 (1) It was failing to install pcreposix.h.
4233 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4236 Version 3.0 01-Feb-00
4237 ---------------------
4239 1. Add support for the /+ modifier to perltest (to output $` like it does in
4240 pcretest).
4242 2. Add support for the /g modifier to perltest.
4244 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4245 matches null strings.
4247 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4248 pattern. Perl treats empty patterns specially - it reuses the most recent
4249 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4250 effect.
4252 5. The POSIX interface was broken in that it was just handing over the POSIX
4253 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4254 required a bigger vector, with some working space on the end. This means that
4255 the POSIX wrapper now has to get and free some memory, and copy the results.
4257 6. Added some simple autoconf support, placing the test data and the
4258 documentation in separate directories, re-organizing some of the
4259 information files, and making it build pcre-config (a GNU standard). Also added
4260 libtool support for building PCRE as a shared library, which is now the
4261 default.
4263 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4264 09 are not valid octal constants. Single digits will be used for minor values
4265 less than 10.
4267 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4268 existing programs that set these in the POSIX interface can use PCRE without
4269 modification.
4271 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4272 return all that pcre_info() returns, plus additional data. The pcre_info()
4273 function is retained for compatibility, but is considered to be obsolete.
4275 10. Added experimental recursion feature (?R) to handle one common case that
4276 Perl 5.6 will be able to do with (?p{...}).
4278 11. Added support for POSIX character classes like [:alpha:], which Perl is
4279 adopting.
4282 Version 2.08 31-Aug-99
4283 ----------------------
4285 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4286 trying to match at the startoffset position, but instead was moving forward to
4287 the next newline as if a previous match had failed.
4289 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4290 and could get into a loop if a null string was matched other than at the start
4291 of the subject.
4293 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4294 be distinguished at compile time, and for completeness also added PCRE_DATE.
4296 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4297 in GnuWin32 environments.
4300 Version 2.07 29-Jul-99
4301 ----------------------
4303 1. The documentation is now supplied in plain text form and HTML as well as in
4304 the form of man page sources.
4306 2. C++ compilers don't like assigning (void *) values to other pointer types.
4307 In particular this affects malloc(). Although there is no problem in Standard
4308 C, I've put in casts to keep C++ compilers happy.
4310 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4311 should be (const char *).
4313 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4314 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4315 However, I haven't made this a standard facility. The documentation doesn't
4316 mention it, and the Makefile doesn't support it.
4318 5. The Makefile now contains an "install" target, with editable destinations at
4319 the top of the file. The pcretest program is not installed.
4321 6. pgrep -V now gives the PCRE version number and date.
4323 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4324 causing the entire string to be ignored, instead of just the last character.
4326 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4327 non-matching string, it can take a very, very long time, even for strings of
4328 quite modest length, because of the nested recursion. PCRE now does better in
4329 some of these cases. It does this by remembering the last required literal
4330 character in the pattern, and pre-searching the subject to ensure it is present
4331 before running the real match. In other words, it applies a heuristic to detect
4332 some types of certain failure quickly, and in the above example, if presented
4333 with a string that has no trailing " it gives "no match" very quickly.
4335 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4336 other alternatives are tried instead.
4339 Version 2.06 09-Jun-99
4340 ----------------------
4342 1. Change pcretest's output for amount of store used to show just the code
4343 space, because the remainder (the data block) varies in size between 32-bit and
4344 64-bit systems.
4346 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4347 start matching at. This allows lookbehinds to work when searching for multiple
4348 occurrences in a string.
4350 3. Added additional options to pcretest for testing multiple occurrences:
4352 /+ outputs the rest of the string that follows a match
4353 /g loops for multiple occurrences, using the new startoffset argument
4354 /G loops for multiple occurrences by passing an incremented pointer
4356 4. PCRE wasn't doing the "first character" optimization for patterns starting
4357 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4358 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4359 the letter 'x'. On long subject strings, this gives a significant speed-up.
4362 Version 2.05 21-Apr-99
4363 ----------------------
4365 1. Changed the type of magic_number from int to long int so that it works
4366 properly on 16-bit systems.
4368 2. Fixed a bug which caused patterns starting with .* not to work correctly
4369 when the subject string contained newline characters. PCRE was assuming
4370 anchoring for such patterns in all cases, which is not correct because .* will
4371 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4372 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4373 must be retried after every newline in the subject.
4376 Version 2.04 18-Feb-99
4377 ----------------------
4379 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4380 computation of the store needed to hold the pattern was incorrect (too large).
4381 If such patterns were nested a few deep, this could multiply and become a real
4382 problem.
4384 2. Added /M option to pcretest to show the memory requirement of a specific
4385 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4387 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4388 compiled in such a way that the backtracking after subsequent failure was
4389 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4390 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4393 Version 2.03 02-Feb-99
4394 ----------------------
4396 1. Fixed typo and small mistake in man page.
4398 2. Added 4th condition (GPL supersedes if conflict) and created separate
4399 LICENCE file containing the conditions.
4401 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4402 Perl, that is the internal \ allows the delimiter to be included in the
4403 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4404 the final delimiter, add \ to the end of the pattern (to test the error).
4406 4. Added the convenience functions for extracting substrings after a successful
4407 match. Updated pcretest to make it able to test these functions.
4410 Version 2.02 14-Jan-99
4411 ----------------------
4413 1. Initialized the working variables associated with each extraction so that
4414 their saving and restoring doesn't refer to uninitialized store.
4416 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4417 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4418 fix the problem.
4420 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4421 calls, and wasn't printing the correct value for compiling calls. Increased the
4422 default value of LOOPREPEAT, and the number of significant figures in the
4423 times.
4425 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4427 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4428 a building problem on Windows NT with a FAT file system.
4431 Version 2.01 21-Oct-98
4432 ----------------------
4434 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4435 to character tables built by pcre_maketables() in the current locale. If NULL
4436 is passed, the default tables are used.
4439 Version 2.00 24-Sep-98
4440 ----------------------
4442 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4443 it any more.
4445 2. Allow quantification of (?>) groups, and make it work correctly.
4447 3. The first character computation wasn't working for (?>) groups.
4449 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4450 end of the subject) and add 5.005's \z, which really does match only at the
4451 very end of the subject.
4453 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4455 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4456 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4457 localized options. All options to pcre_study() were also removed.
4459 7. Add other new features from 5.005:
4461 $(?<= positive lookbehind
4462 $(?<! negative lookbehind
4463 (?imsx-imsx) added the unsetting capability
4464 such a setting is global if at outer level; local otherwise
4465 (?imsx-imsx:) non-capturing groups with option setting
4466 (?(cond)re|re) conditional pattern matching
4468 A backreference to itself in a repeated group matches the previous
4469 captured string.
4471 8. General tidying up of studying (both automatic and via "study")
4472 consequential on the addition of new assertions.
4474 9. As in 5.005, unlimited repeated groups that could match an empty substring
4475 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4476 runtime if any iteration does actually match an empty substring.
4478 10. Include the RunTest script in the distribution.
4480 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4481 discrepancies, some of which were old and were also with respect to 5.004. They
4482 have now been fixed.
4485 Version 1.09 28-Apr-98
4486 ----------------------
4488 1. A negated single character class followed by a quantifier with a minimum
4489 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4490 program crashes, or just wrong answers. This did not apply to negated classes
4491 containing more than one character, or to minima other than one.
4494 Version 1.08 27-Mar-98
4495 ----------------------
4497 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4499 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4500 latter must appear before anything that relies on it in the pattern.
4503 Version 1.07 16-Feb-98
4504 ----------------------
4506 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4507 repeat of a potentially empty string).
4510 Version 1.06 23-Jan-98
4511 ----------------------
4513 1. Added Markus Oberhumer's little patches for C++.
4515 2. Literal strings longer than 255 characters were broken.
4518 Version 1.05 23-Dec-97
4519 ----------------------
4521 1. Negated character classes containing more than one character were failing if
4522 PCRE_CASELESS was set at run time.
4525 Version 1.04 19-Dec-97
4526 ----------------------
4528 1. Corrected the man page, where some "const" qualifiers had been omitted.
4530 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4531 input syntax.
4533 3. Fixed memory leak which occurred when a regex with back references was
4534 matched with an offsets vector that wasn't big enough. The temporary memory
4535 that is used in this case wasn't being freed if the match failed.
4537 4. Tidied pcretest to ensure it frees memory that it gets.
4539 5. Temporary memory was being obtained in the case where the passed offsets
4540 vector was exactly big enough.
4542 6. Corrected definition of offsetof() from change 5 below.
4544 7. I had screwed up change 6 below and broken the rules for the use of
4545 setjmp(). Now fixed.
4548 Version 1.03 18-Dec-97
4549 ----------------------
4551 1. A erroneous regex with a missing opening parenthesis was correctly
4552 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4553 on some systems.
4555 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4556 it was reported that one broken compiler failed on the former because "code" is
4557 also an independent variable.
4559 3. The erroneous regex a[]b caused an array overrun reference.
4561 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4562 fail on data ending with that character. (It was going on too far, and checking
4563 the next character, typically a binary zero.) This was specific to the
4564 optimized code for single-character negative classes.
4566 5. Added a contributed patch from the TIN world which does the following:
4568 + Add an undef for memmove, in case the the system defines a macro for it.
4570 + Add a definition of offsetof(), in case there isn't one. (I don't know
4571 the reason behind this - offsetof() is part of the ANSI standard - but
4572 it does no harm).
4574 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4575 most of the places where whitespace preceded '#'. I have given up and
4576 allowed the remaining 2 cases to be at the margin.
4578 + Rename some variables in pcre to eliminate shadowing. This seems very
4579 pedantic, but does no harm, of course.
4581 6. Moved the call to setjmp() into its own function, to get rid of warnings
4582 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4584 7. Constructs such as \d{8,} were compiling into the equivalent of
4585 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4586 outcome, but in this particular case used more store than had been allocated,
4587 which caused the bug to be discovered because it threw up an internal error.
4589 8. The debugging code in both pcre and pcretest for outputting the compiled
4590 form of a regex was going wrong in the case of back references followed by
4591 curly-bracketed repeats.
4594 Version 1.02 12-Dec-97
4595 ----------------------
4597 1. Typos in pcre.3 and comments in the source fixed.
4599 2. Applied a contributed patch to get rid of places where it used to remove
4600 'const' from variables, and fixed some signed/unsigned and uninitialized
4601 variable warnings.
4603 3. Added the "runtest" target to Makefile.
4605 4. Set default compiler flag to -O2 rather than just -O.
4608 Version 1.01 19-Nov-97
4609 ----------------------
4611 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4612 like /([ab]*)*/, that is, for classes with more than one character in them.
4614 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4615 as /((?>a*))*/ (a PCRE_EXTRA facility).
4618 Version 1.00 18-Nov-97
4619 ----------------------
4621 1. Added compile-time macros to support systems such as SunOS4 which don't have
4622 memmove() or strerror() but have other things that can be used instead.
4624 2. Arranged that "make clean" removes the executables.
4627 Version 0.99 27-Oct-97
4628 ----------------------
4630 1. Fixed bug in code for optimizing classes with only one character. It was
4631 initializing a 32-byte map regardless, which could cause it to run off the end
4632 of the memory it had got.
4634 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4637 Version 0.98 22-Oct-97
4638 ----------------------
4640 1. Fixed bug in code for handling temporary memory usage when there are more
4641 back references than supplied space in the ovector. This could cause segfaults.
4644 Version 0.97 21-Oct-97
4645 ----------------------
4647 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4649 2. Optimized negated single characters not to use a bit map.
4651 3. Brought error texts together as macro definitions; clarified some of them;
4652 fixed one that was wrong - it said "range out of order" when it meant "invalid
4653 escape sequence".
4655 4. Changed some char * arguments to const char *.
4657 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4659 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4660 pcretest.
4663 Version 0.96 16-Oct-97
4664 ----------------------
4666 1. Added a simple "pgrep" utility to the distribution.
4668 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4669 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4670 where "ddd" means "one or more decimal digits".
4672 3. Fixed serious bug. If a pattern had a back reference, but the call to
4673 pcre_exec() didn't supply a large enough ovector to record the related
4674 identifying subpattern, the match always failed. PCRE now remembers the number
4675 of the largest back reference, and gets some temporary memory in which to save
4676 the offsets during matching if necessary, in order to ensure that
4677 backreferences always work.
4679 4. Increased the compatibility with Perl in a number of ways:
4681 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4682 to request this handling. The option can be set at compile or exec time.
4684 (b) $ matches before a terminating newline by default; an option
4685 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4686 mode). The option can be set at compile or exec time.
4688 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4689 the same as Perl's. If the decimal number it represents is less than 10
4690 or there aren't that many previous left capturing parentheses, an octal
4691 escape is read. Inside a character class, it's always an octal escape,
4692 even if it is a single digit.
4694 (d) An escaped but undefined alphabetic character is taken as a literal,
4695 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4696 escapes.
4698 (e) {0} is now permitted. (The previous item is removed from the compiled
4699 pattern).
4701 5. Changed all the names of code files so that the basic parts are no longer
4702 than 10 characters, and abolished the teeny "globals.c" file.
4704 6. Changed the handling of character classes; they are now done with a 32-byte
4705 bit map always.
4707 7. Added the -d and /D options to pcretest to make it possible to look at the
4708 internals of compilation without having to recompile pcre.
4711 Version 0.95 23-Sep-97
4712 ----------------------
4714 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4715 \x20 at the start of a run of normal characters. These were being treated as
4716 real characters, instead of the source characters being re-checked.
4719 Version 0.94 18-Sep-97
4720 ----------------------
4722 1. The functions are now thread-safe, with the caveat that the global variables
4723 containing pointers to malloc() and free() or alternative functions are the
4724 same for all threads.
4726 2. Get pcre_study() to generate a bitmap of initial characters for non-
4727 anchored patterns when this is possible, and use it if passed to pcre_exec().
4730 Version 0.93 15-Sep-97
4731 ----------------------
4733 1. /(b)|(:+)/ was computing an incorrect first character.
4735 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4736 but not actually doing anything yet.
4738 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4739 as Perl does (e.g. [-az] or [az-]).
4741 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4742 all possible positions.
4744 5. Split up into different modules to avoid including unneeded functions in a
4745 compiled binary. However, compile and exec are still in one module. The "study"
4746 function is split off.
4748 6. The character tables are now in a separate module whose source is generated
4749 by an auxiliary program - but can then be edited by hand if required. There are
4750 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4751 toupper() in the code.
4753 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4754 make them global. Abolish the function for setting them, as the caller can now
4755 set them directly.
4758 Version 0.92 11-Sep-97
4759 ----------------------
4761 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4762 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4764 2. Caseless matching was not working in character classes if the characters in
4765 the pattern were in upper case.
4767 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4769 4. Make PCRE_ANCHORED public and accept as a compile option.
4771 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4772 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4773 pass them.
4775 6. Give an error if bad option bits passed at compile or run time.
4777 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4778 pcretest to cause it to pass that flag.
4780 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4781 options, and the first character, if set.
4783 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4786 Version 0.91 10-Sep-97
4787 ----------------------
4789 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4790 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4792 2. PCRE was looping on encountering an indefinitely repeated back reference to
4793 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4794 Perl does - treats the match as successful.
4796 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5