ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1038 - (show annotations)
Wed Sep 12 09:03:23 2012 UTC (7 years, 5 months ago) by zherczeg
File size: 235738 byte(s)
Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.32
5 ------------
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
13 3. Single character iterator optimizations in the JIT compiler.
15 4. Improved JIT compiler optimizations for character ranges.
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
31 (a) There is now no limit to the number of patterns to be matched.
33 (b) An error is given if a pattern is too long.
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
38 (d) --exclude-from and --include-from (multiple use) have been added.
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
43 (f) Multiple uses of -f and --file-list are now supported.
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
49 (h) The documentation has been revised and clarified in places.
51 9. Improve the matching speed of capturing brackets.
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
61 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62 structure could go wrong in environments where size_t is not the same size
63 as int.
65 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66 set.
68 15. The EBCDIC support had decayed; later updates to the code had included
69 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70 general tidy up of EBCDIC-related issues, and the documentation was also
71 not quite right. There is now a test that can be run on ASCII systems to
72 check some of the EBCDIC-related things (but is it not a full test).
74 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75 in a small tidy to the code.
77 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
80 Version 8.31 06-July-2012
81 -------------------------
83 1. Fixing a wrong JIT test case and some compiler warnings.
85 2. Removed a bashism from the RunTest script.
87 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
88 to unsigned type, result still unsigned" that was given by an MS compiler
89 on encountering the code "-sizeof(xxx)".
91 4. Partial matching support is added to the JIT compiler.
93 5. Fixed several bugs concerned with partial matching of items that consist
94 of more than one character:
96 (a) /^(..)\1/ did not partially match "aba" because checking references was
97 done on an "all or nothing" basis. This also applied to repeated
98 references.
100 (b) \R did not give a hard partial match if \r was found at the end of the
101 subject.
103 (c) \X did not give a hard partial match after matching one or more
104 characters at the end of the subject.
106 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
107 a partial match for the string "\r".
109 (e) When newline was set to CRLF, the metacharacter "." did not recognize
110 a partial match for a CR character at the end of the subject string.
112 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
113 running pcretest, the text "(JIT)" added to the output whenever JIT is
114 actually used to run the match.
116 7. Individual JIT compile options can be set in pcretest by following -s+[+]
117 or /S+[+] with a digit between 1 and 7.
119 8. OP_NOT now supports any UTF character not just single-byte ones.
121 9. (*MARK) control verb is now supported by the JIT compiler.
123 10. The command "./RunTest list" lists the available tests without actually
124 running any of them. (Because I keep forgetting what they all are.)
128 12. Applied a (slightly modified) user-supplied patch that improves performance
129 when the heap is used for recursion (compiled with --disable-stack-for-
130 recursion). Instead of malloc and free for each heap frame each time a
131 logical recursion happens, frames are retained on a chain and re-used where
132 possible. This sometimes gives as much as 30% improvement.
134 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
135 call.
137 14. As documented, (*COMMIT) is now confined to within a positive assertion.
139 15. It is now possible to link pcretest with libedit as an alternative to
140 libreadline.
142 16. (*COMMIT) control verb is now supported by the JIT compiler.
144 17. The Unicode data tables have been updated to Unicode 6.1.0.
146 18. Added --file-list option to pcregrep.
148 19. Added binary file support to pcregrep, including the -a, --binary-files,
149 -I, and --text options.
151 20. The madvise function is renamed for posix_madvise for QNX compatibility
152 reasons. Fixed by Giuseppe D'Angelo.
154 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
156 22. Changed the option for creating symbolic links for 16-bit man pages from
157 -s to -sf so that re-installing does not cause issues.
159 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
161 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
162 matches in certain environments (the workspace was not being correctly
163 retained). Also added to pcre_dfa_exec() a simple plausibility check on
164 some of the workspace data at the beginning of a restart.
166 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
167 was not doing so when it should - probably a typo introduced by SVN 528
168 (change 8.10/14).
170 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
171 \w+ when the character tables indicated that \x{c4} was a word character.
172 There were several related cases, all because the tests for doing a table
173 lookup were testing for characters less than 127 instead of 255.
175 27. If a pattern contains capturing parentheses that are not used in a match,
176 their slots in the ovector are set to -1. For those that are higher than
177 any matched groups, this happens at the end of processing. In the case when
178 there were back references that the ovector was too small to contain
179 (causing temporary malloc'd memory to be used during matching), and the
180 highest capturing number was not used, memory off the end of the ovector
181 was incorrectly being set to -1. (It was using the size of the temporary
182 memory instead of the true size.)
184 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
185 ovector size, it uses memory at the end of the block that it has got.
187 29. Check for an overlong MARK name and give an error at compile time. The
188 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
190 30. JIT compiler update.
192 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
193 Rill for the patch.
195 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
197 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
199 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
200 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
201 was enabled.
203 35. Improve JIT code generation for greedy plus quantifier.
205 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
206 1 to "aa" instead of to an empty string. The bug affected repeated groups
207 that could potentially match an empty string.
209 37. Optimizing single character iterators in JIT.
211 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
212 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
213 codepoints that are too big for the mode are faulted, and in a UTF mode,
214 disallowed codepoints are also faulted.
216 39. If PCRE was compiled with UTF support, in three places in the DFA
217 matcher there was code that should only have been obeyed in UTF mode, but
218 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
219 processing when bytes with values greater than 127 were present. In 16-bit
220 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
221 both cases the values are those that cannot be the first data item in a UTF
222 character. The three items that might have provoked this were recursions,
223 possessively repeated groups, and atomic groups.
225 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
226 and pcregrep, because some OS require shared objects to be explicitly
227 passed to ld, causing the link step to fail if they are not.
229 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
230 mode, patterns that started with \h* or \R* might be incorrectly matched.
233 Version 8.30 04-February-2012
234 -----------------------------
236 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
237 name is defined in ctype.h.
239 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
240 only in quite long subpatterns.
242 3. Removed the function pcre_info(), which has been obsolete and deprecated
243 since it was replaced by pcre_fullinfo() in February 2000.
245 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
246 match a (*MARK), and the match failed at the start of the subject, a
247 reference to memory before the start of the subject could occur. This bug
248 was introduced by fix 17 of release 8.21.
250 5. A reference to an unset group with zero minimum repetition was giving
251 totally wrong answers (in non-JavaScript-compatibility mode). For example,
252 /(another)?(\1?)test/ matched against "hello world test". This bug was
253 introduced in release 8.13.
255 6. Add support for 16-bit character strings (a large amount of work involving
256 many changes and refactorings).
258 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
259 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
260 from a file.
262 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
263 rounding is not applied in this particular case).
265 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
266 if they appear, or are escaped, in patterns.
268 10. Get rid of a number of -Wunused-but-set-variable warnings.
270 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
271 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
272 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
273 also returns the mark "x". This bug applied to capturing parentheses,
274 non-capturing parentheses, and atomic parentheses. It also applied to some
275 assertions.
277 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
278 information out of configure.ac instead of relying on pcre.h.generic, which
279 is not stored in the repository.
281 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
282 -lreadline.
284 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
286 15. Applied Graycode's patch to put the top-level frame on the stack rather
287 than the heap when not using the stack for recursion. This gives a
288 performance improvement in many cases when recursion is not deep.
290 16. Experimental code added to "pcretest -C" to output the stack frame size.
293 Version 8.21 12-Dec-2011
294 ------------------------
296 1. Updating the JIT compiler.
298 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
299 are added as well.
301 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
302 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
303 calling _pcre_jit_exec. Some extra comments are added.
305 4. (*MARK) settings inside atomic groups that do not contain any capturing
306 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
307 was introduced by change 18 for 8.20.
309 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
310 ECMA-262 standard.
312 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
313 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
314 This bug was probably introduced by change 9 of 8.13.
316 7. While fixing 6 above, I noticed that a number of other items were being
317 incorrectly rejected as "not fixed length". This arose partly because newer
318 opcodes had not been added to the fixed-length checking code. I have (a)
319 corrected the bug and added tests for these items, and (b) arranged for an
320 error to occur if an unknown opcode is encountered while checking for fixed
321 length instead of just assuming "not fixed length". The items that were
322 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
323 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
324 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
326 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
327 being incorrectly compiled and would have given unpredicatble results.
329 9. A possessively repeated subpattern with minimum repeat count greater than
330 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
331 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
332 the first (A) could occur when it should not.
334 10. Add a cast and remove a redundant test from the code.
336 11. JIT should use pcre_malloc/pcre_free for allocation.
338 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
339 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
340 is anything other than /usr, -L is still shown).
342 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
344 14. Perl does not support \N without a following name in a [] class; PCRE now
345 also gives an error.
347 15. If a forward reference was repeated with an upper limit of around 2000,
348 it caused the error "internal error: overran compiling workspace". The
349 maximum number of forward references (including repeats) was limited by the
350 internal workspace, and dependent on the LINK_SIZE. The code has been
351 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
352 the default depends on LINK_SIZE. There is a new upper limit (for safety)
353 of around 200,000 forward references. While doing this, I also speeded up
354 the filling in of repeated forward references.
356 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
357 incorrectly expecting the subject to contain another "a" after the start.
359 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
360 in the match, the SKIP should be ignored. This was not happening; instead
361 the SKIP was being treated as NOMATCH. For patterns such as
362 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
363 tested.
365 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
366 now much more compatible with Perl, in particular in cases where the result
367 is a non-match for a non-anchored pattern. For example, if
368 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
369 "m", where previously it did not return a name. A side effect of this
370 change is that for partial matches, the last encountered mark name is
371 returned, as for non matches. A number of tests that were previously not
372 Perl-compatible have been moved into the Perl-compatible test files. The
373 refactoring has had the pleasing side effect of removing one argument from
374 the match() function, thus reducing its stack requirements.
376 19. If the /S+ option was used in pcretest to study a pattern using JIT,
377 subsequent uses of /S (without +) incorrectly behaved like /S+.
379 21. Retrieve executable code size support for the JIT compiler and fixing
380 some warnings.
382 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
383 not work when the shorter character appeared right at the end of the
384 subject string.
386 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
387 systems.
389 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
390 output it when the /M option is used in pcretest.
392 25. The CheckMan script was not being included in the distribution. Also, added
393 an explicit "perl" to run Perl scripts from the PrepareRelease script
394 because this is reportedly needed in Windows.
396 26. If study data was being save in a file and studying had not found a set of
397 "starts with" bytes for the pattern, the data written to the file (though
398 never used) was taken from uninitialized memory and so caused valgrind to
399 complain.
401 27. Updated RunTest.bat as provided by Sheri Pierce.
403 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
405 29. Computation of memory usage for the table of capturing group names was
406 giving an unnecessarily large value.
409 Version 8.20 21-Oct-2011
410 ------------------------
412 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
413 a POSIX class. After further experiments with Perl, which convinced me that
414 Perl has bugs and confusions, a closing square bracket is no longer allowed
415 in a POSIX name. This bug also affected patterns with classes that started
416 with full stops.
418 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
419 captured substring, but while checking the failing first alternative,
420 substring 1 is temporarily captured. If the output vector supplied to
421 pcre_exec() was not big enough for this capture, the yield of the function
422 was still zero ("insufficient space for captured substrings"). This cannot
423 be totally fixed without adding another stack variable, which seems a lot
424 of expense for a edge case. However, I have improved the situation in cases
425 such as /(a)(b)x|abc/ matched against "abc", where the return code
426 indicates that fewer than the maximum number of slots in the ovector have
427 been set.
429 3. Related to (2) above: when there are more back references in a pattern than
430 slots in the output vector, pcre_exec() uses temporary memory during
431 matching, and copies in the captures as far as possible afterwards. It was
432 using the entire output vector, but this conflicts with the specification
433 that only 2/3 is used for passing back captured substrings. Now it uses
434 only the first 2/3, for compatibility. This is, of course, another edge
435 case.
437 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
438 main code base, and can be used by building with --enable-jit. When this is
439 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
440 runtime --no-jit option is given.
442 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
443 ovector, the return from the function was zero, implying that there were
444 other matches that did not fit. The correct "exactly full" value is now
445 returned.
447 6. If a subpattern that was called recursively or as a subroutine contained
448 (*PRUNE) or any other control that caused it to give a non-standard return,
449 invalid errors such as "Error -26 (nested recursion at the same subject
450 position)" or even infinite loops could occur.
452 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
453 computing the minimum length on reaching *ACCEPT, and so ended up with the
454 wrong value of 1 rather than 0. Further investigation indicates that
455 computing a minimum subject length in the presence of *ACCEPT is difficult
456 (think back references, subroutine calls), and so I have changed the code
457 so that no minimum is registered for a pattern that contains *ACCEPT.
459 8. If (*THEN) was present in the first (true) branch of a conditional group,
460 it was not handled as intended. [But see 16 below.]
462 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
463 Sheri Pierce.
465 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
466 the first byte in a match must be "a".
468 11. Change 17 for 8.13 increased the recursion depth for patterns like
469 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
470 pattern contains any instances of (*THEN). If it does not, the old
471 optimizations are restored. It would be nice to do this on a per-group
472 basis, but at the moment that is not feasible.
474 12. In some environments, the output of pcretest -C is CRLF terminated. This
475 broke RunTest's code that checks for the link size. A single white space
476 character after the value is now allowed for.
478 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
479 For "fr", it uses the Windows-specific input and output files.
481 14. If (*THEN) appeared in a group that was called recursively or as a
482 subroutine, it did not work as intended. [But see next item.]
484 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
485 pattern fragments (but not containing any | characters). If A and B are
486 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
487 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
488 D. In other words, Perl considers parentheses that do not contain any |
489 characters to be part of a surrounding alternative, whereas PCRE was
490 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
491 differently. PCRE now behaves in the same way as Perl, except in the case
492 of subroutine/recursion calls such as (?1) which have in any case always
493 been different (but PCRE had them first :-).
495 16. Related to 15 above: Perl does not treat the | in a conditional group as
496 creating alternatives. Such a group is treated in the same way as an
497 ordinary group without any | characters when processing (*THEN). PCRE has
498 been changed to match Perl's behaviour.
500 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
501 RunGrepTest script failed.
503 18. Change 22 for version 13 caused atomic groups to use more stack. This is
504 inevitable for groups that contain captures, but it can lead to a lot of
505 stack use in large patterns. The old behaviour has been restored for atomic
506 groups that do not contain any capturing parentheses.
508 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
509 suppress the check for a minimum subject length at run time. (If it was
510 given to pcre_exec() or pcre_dfa_exec() it did work.)
512 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
513 fail to work when decoding hex characters in data strings in EBCDIC
514 environments.
516 21. It appears that in at least one Mac OS environment, the isxdigit() function
517 is implemented as a macro that evaluates to its argument more than once,
518 contravening the C 90 Standard (I haven't checked a later standard). There
519 was an instance in pcretest which caused it to go wrong when processing
520 \x{...} escapes in subject strings. The has been rewritten to avoid using
521 things like p++ in the argument of isxdigit().
524 Version 8.13 16-Aug-2011
525 ------------------------
527 1. The Unicode data tables have been updated to Unicode 6.0.0.
529 2. Two minor typos in pcre_internal.h have been fixed.
531 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
532 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
533 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
535 4. There were a number of related bugs in the code for matching backrefences
536 caselessly in UTF-8 mode when codes for the characters concerned were
537 different numbers of bytes. For example, U+023A and U+2C65 are an upper
538 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
539 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
540 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
541 2-byte code at the end of the subject (it thought there wasn't enough data
542 left).
544 5. Comprehensive information about what went wrong is now returned by
545 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
546 as the output vector has at least 2 elements. The offset of the start of
547 the failing character and a reason code are placed in the vector.
549 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
550 now returned is for the first byte of the failing character, instead of the
551 last byte inspected. This is an incompatible change, but I hope it is small
552 enough not to be a problem. It makes the returned offset consistent with
553 pcre_exec() and pcre_dfa_exec().
555 7. pcretest now gives a text phrase as well as the error number when
556 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
557 failure, the offset and reason code are output.
559 8. When \R was used with a maximizing quantifier it failed to skip backwards
560 over a \r\n pair if the subsequent match failed. Instead, it just skipped
561 back over a single character (\n). This seems wrong (because it treated the
562 two characters as a single entity when going forwards), conflicts with the
563 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
564 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
565 has been changed.
567 9. Some internal refactoring has changed the processing so that the handling
568 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
569 time (the PCRE_DOTALL option was changed this way some time ago: version
570 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
571 which was always a bit of a fudge. It also means that there is one less
572 argument for the match() function, which reduces its stack requirements
573 slightly. This change also fixes an incompatibility with Perl: the pattern
574 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
576 10. More internal refactoring has drastically reduced the number of recursive
577 calls to match() for possessively repeated groups such as (abc)++ when
578 using pcre_exec().
580 11. While implementing 10, a number of bugs in the handling of groups were
581 discovered and fixed:
583 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
584 (a|)*(?1) gave a compile-time internal error.
585 ((a|)+)+ did not notice that the outer group could match an empty string.
586 (^a|^)+ was not marked as anchored.
587 (.*a|.*)+ was not marked as matching at start or after a newline.
589 12. Yet more internal refactoring has removed another argument from the match()
590 function. Special calls to this function are now indicated by setting a
591 value in a variable in the "match data" data block.
593 13. Be more explicit in pcre_study() instead of relying on "default" for
594 opcodes that mean there is no starting character; this means that when new
595 ones are added and accidentally left out of pcre_study(), testing should
596 pick them up.
598 14. The -s option of pcretest has been documented for ages as being an old
599 synonym of -m (show memory usage). I have changed it to mean "force study
600 for every regex", that is, assume /S for every regex. This is similar to -i
601 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
602 using it. It makes it easier to run collections of tests with and without
603 study enabled, and thereby test pcre_study() more easily. All the standard
604 tests are now run with and without -s (but some patterns can be marked as
605 "never study" - see 20 below).
607 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
608 restoration of the capturing data to the outer values was not happening
609 correctly.
611 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
612 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
613 pattern had matched an empty string, and so incorrectly returned a no
614 match.
616 17. There was optimizing code for the last branch of non-capturing parentheses,
617 and also for the obeyed branch of a conditional subexpression, which used
618 tail recursion to cut down on stack usage. Unfortunately, now that there is
619 the possibility of (*THEN) occurring in these branches, tail recursion is
620 no longer possible because the return has to be checked for (*THEN). These
621 two optimizations have therefore been removed. [But see 8.20/11 above.]
623 18. If a pattern containing \R was studied, it was assumed that \R always
624 matched two bytes, thus causing the minimum subject length to be
625 incorrectly computed because \R can also match just one byte.
627 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
628 was incorrectly computed.
630 20. If /S is present twice on a test pattern in pcretest input, it now
631 *disables* studying, thereby overriding the use of -s on the command line
632 (see 14 above). This is necessary for one or two tests to keep the output
633 identical in both cases.
635 21. When (*ACCEPT) was used in an assertion that matched an empty string and
636 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
638 22. When an atomic group that contained a capturing parenthesis was
639 successfully matched, but the branch in which it appeared failed, the
640 capturing was not being forgotten if a higher numbered group was later
641 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
642 group 1 to "a", when in fact it should be unset. This applied to multi-
643 branched capturing and non-capturing groups, repeated or not, and also to
644 positive assertions (capturing in negative assertions does not happen
645 in PCRE) and also to nested atomic groups.
647 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
648 subject after a captured substring, to make it easier to tell which of a
649 number of identical substrings has been captured.
651 24. The way atomic groups are processed by pcre_exec() has been changed so that
652 if they are repeated, backtracking one repetition now resets captured
653 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
654 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
655 "aaa". Previously, it would have been "a". As part of this code
656 refactoring, the way recursive calls are handled has also been changed.
658 25. If an assertion condition captured any substrings, they were not passed
659 back unless some other capturing happened later. For example, if
660 (?(?=(a))a) was matched against "a", no capturing was returned.
662 26. When studying a pattern that contained subroutine calls or assertions,
663 the code for finding the minimum length of a possible match was handling
664 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
665 group 1 called group 2 while simultaneously a separate group 2 called group
666 1). A stack overflow occurred in this case. I have fixed this by limiting
667 the recursion depth to 10.
669 27. Updated RunTest.bat in the distribution to the version supplied by Tom
670 Fortmann. This supports explicit test numbers on the command line, and has
671 argument validation and error reporting.
673 28. An instance of \X with an unlimited repeat could fail if at any point the
674 first character it looked at was a mark character.
676 29. Some minor code refactoring concerning Unicode properties and scripts
677 should reduce the stack requirement of match() slightly.
679 30. Added the '=' option to pcretest to check the setting of unused capturing
680 slots at the end of the pattern, which are documented as being -1, but are
681 not included in the return count.
683 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
684 compiled something random. Now it gives a compile-time error (as does
685 Perl).
687 32. A *MARK encountered during the processing of a positive assertion is now
688 recorded and passed back (compatible with Perl).
690 33. If --only-matching or --colour was set on a pcregrep call whose pattern
691 had alternative anchored branches, the search for a second match in a line
692 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
693 matched the line "0102" twice. The same bug affected patterns that started
694 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
695 twice.
697 34. Previously, PCRE did not allow quantification of assertions. However, Perl
698 does, and because of capturing effects, quantifying parenthesized
699 assertions may at times be useful. Quantifiers are now allowed for
700 parenthesized assertions.
702 35. A minor code tidy in pcre_compile() when checking options for \R usage.
704 36. \g was being checked for fancy things in a character class, when it should
705 just be a literal "g".
707 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
708 appearance of a nested POSIX class supersedes an apparent external class.
709 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
710 unescaped square brackets may also appear as part of class names. For
711 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
712 more like Perl. (But see 8.20/1 above.)
714 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
715 was because it thought it was \N{name}, which is not supported).
717 39. Add minix to OS list not supporting the -S option in pcretest.
719 40. PCRE tries to detect cases of infinite recursion at compile time, but it
720 cannot analyze patterns in sufficient detail to catch mutual recursions
721 such as ((?1))((?2)). There is now a runtime test that gives an error if a
722 subgroup is called recursively as a subpattern for a second time at the
723 same position in the subject string. In previous releases this might have
724 been caught by the recursion limit, or it might have run out of stack.
726 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
727 happen only once. PCRE was, however incorrectly giving a compile time error
728 "recursive call could loop indefinitely" because it cannot analyze the
729 pattern in sufficient detail. The compile time test no longer happens when
730 PCRE is compiling a conditional subpattern, but actual runaway loops are
731 now caught at runtime (see 40 above).
733 42. It seems that Perl allows any characters other than a closing parenthesis
734 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
735 has been changed to be the same.
737 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
738 as not to get warnings when autogen.sh is called. Also changed
739 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
741 44. To help people who use pcregrep to scan files containing exceedingly long
742 lines, the following changes have been made:
744 (a) The default value of the buffer size parameter has been increased from
745 8K to 20K. (The actual buffer used is three times this size.)
747 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
748 PCRE is built.
750 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
751 to be set at run time.
753 (d) Numerical values in pcregrep options can be followed by K or M, for
754 example --buffer-size=50K.
756 (e) If a line being scanned overflows pcregrep's buffer, an error is now
757 given and the return code is set to 2.
759 45. Add a pointer to the latest mark to the callout data block.
761 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
762 partial match of an empty string instead of no match. This was specific to
763 the use of ".".
765 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
766 complete match instead of a partial match. This bug was dependent on both
767 the PCRE_UTF8 and PCRE_DOTALL options being set.
769 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
770 starting byte set, because \b was not being ignored.
773 Version 8.12 15-Jan-2011
774 ------------------------
776 1. Fixed some typos in the markup of the man pages, and wrote a script that
777 checks for such things as part of the documentation building process.
779 2. On a big-endian 64-bit system, pcregrep did not correctly process the
780 --match-limit and --recursion-limit options (added for 8.11). In
781 particular, this made one of the standard tests fail. (The integer value
782 went into the wrong half of a long int.)
784 3. If the --colour option was given to pcregrep with -v (invert match), it
785 did strange things, either producing crazy output, or crashing. It should,
786 of course, ignore a request for colour when reporting lines that do not
787 match.
789 4. Another pcregrep bug caused similar problems if --colour was specified with
790 -M (multiline) and the pattern match finished with a line ending.
792 5. In pcregrep, when a pattern that ended with a literal newline sequence was
793 matched in multiline mode, the following line was shown as part of the
794 match. This seems wrong, so I have changed it.
796 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
797 the check for further matches in the same line (so they could be coloured)
798 to overrun the end of the current line. If another match was found, it was
799 incorrectly shown (and then shown again when found in the next line).
801 7. If pcregrep was compiled under Windows, there was a reference to the
802 function pcregrep_exit() before it was defined. I am assuming this was
803 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
804 reported by a user. I've moved the definition above the reference.
807 Version 8.11 10-Dec-2010
808 ------------------------
810 1. (*THEN) was not working properly if there were untried alternatives prior
811 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
812 backtracked to try for "b" instead of moving to the next alternative branch
813 at the same level (in this case, to look for "c"). The Perl documentation
814 is clear that when (*THEN) is backtracked onto, it goes to the "next
815 alternative in the innermost enclosing group".
817 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
818 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
819 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
820 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
821 (*THEN).
823 3. If \s appeared in a character class, it removed the VT character from
824 the class, even if it had been included by some previous item, for example
825 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
826 of \s, but is part of the POSIX "space" class.)
828 4. A partial match never returns an empty string (because you can always
829 match an empty string at the end of the subject); however the checking for
830 an empty string was starting at the "start of match" point. This has been
831 changed to the "earliest inspected character" point, because the returned
832 data for a partial match starts at this character. This means that, for
833 example, /(?<=abc)def/ gives a partial match for the subject "abc"
834 (previously it gave "no match").
836 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
837 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
838 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
839 has an implication that the given string is incomplete (because a partial
840 match is preferred over a full match). For this reason, these items now
841 give a partial match in this situation. [Aside: previously, the one case
842 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
843 match rather than a full match, which was wrong by the old rules, but is
844 now correct.]
846 6. There was a bug in the handling of #-introduced comments, recognized when
847 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
848 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
849 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
850 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
851 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
852 places in pcre_compile().
854 7. Related to (6) above, when pcre_compile() was skipping #-introduced
855 comments when looking ahead for named forward references to subpatterns,
856 the only newline sequence it recognized was NL. It now handles newlines
857 according to the set newline convention.
859 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
860 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
861 cater for a lack of strerror(). These oversights have been fixed.
863 9. Added --match-limit and --recursion-limit to pcregrep.
865 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
867 11. When the -o option was used, pcregrep was setting a return code of 1, even
868 when matches were found, and --line-buffered was not being honoured.
870 12. Added an optional parentheses number to the -o and --only-matching options
871 of pcregrep.
873 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
874 can match an empty string. The code to do it in pcretest and pcredemo
875 needed fixing:
877 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
878 only one byte after an empty string match just before CRLF (this case
879 just got forgotten; "any" and "anycrlf" were OK).
881 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
882 mode when an empty string match preceded an ASCII character followed by
883 a non-ASCII character. (The code for advancing by one character rather
884 than one byte was nonsense.)
886 (c) The pcredemo.c sample program did not have any code at all to handle
887 the cases when CRLF is a valid newline sequence.
889 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
890 as a starting offset was within the subject string. There is now a new
891 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
892 negative or greater than the length of the string. In order to test this,
893 pcretest is extended to allow the setting of negative starting offsets.
895 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
896 starting offset points to the beginning of a UTF-8 character was
897 unnecessarily clumsy. I tidied it up.
899 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
900 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
902 17. Nobody had reported that the --include_dir option, which was added in
903 release 7.7 should have been called --include-dir (hyphen, not underscore)
904 for compatibility with GNU grep. I have changed it to --include-dir, but
905 left --include_dir as an undocumented synonym, and the same for
906 --exclude-dir, though that is not available in GNU grep, at least as of
907 release 2.5.4.
909 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
910 characters from a string of bytes) have been redefined so as not to use
911 loops, in order to improve performance in some environments. At the same
912 time, I abstracted some of the common code into auxiliary macros to save
913 repetition (this should not affect the compiled code).
915 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
916 compile-time error is now given if \c is not followed by an ASCII
917 character, that is, a byte less than 128. (In EBCDIC mode, the code is
918 different, and any byte value is allowed.)
920 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
921 START_OPTIMIZE option, which is now allowed at compile time - but just
922 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
923 to pcregrep and other applications that have no direct access to PCRE
924 options. The new /Y option in pcretest sets this option when calling
925 pcre_compile().
927 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
928 back references. Groups containing recursive back references were forced to
929 be atomic by that change, but in the case of named groups, the amount of
930 memory required was incorrectly computed, leading to "Failed: internal
931 error: code overflow". This has been fixed.
933 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
934 pcretest.c, to avoid build problems in some Borland environments.
937 Version 8.10 25-Jun-2010
938 ------------------------
940 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
941 THEN.
943 2. (*ACCEPT) was not working when inside an atomic group.
945 3. Inside a character class, \B is treated as a literal by default, but
946 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
947 causes the error). The code is unchanged, but I tidied the documentation.
949 4. Inside a character class, PCRE always treated \R and \X as literals,
950 whereas Perl faults them if its -w option is set. I have changed PCRE so
951 that it faults them when PCRE_EXTRA is set.
953 5. Added support for \N, which always matches any character other than
954 newline. (It is the same as "." when PCRE_DOTALL is not set.)
956 6. When compiling pcregrep with newer versions of gcc which may have
957 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
958 declared with attribute warn_unused_result" were given. Just casting the
959 result to (void) does not stop the warnings; a more elaborate fudge is
960 needed. I've used a macro to implement this.
962 7. Minor change to pcretest.c to avoid a compiler warning.
964 8. Added four artifical Unicode properties to help with an option to make
965 \s etc use properties (see next item). The new properties are: Xan
966 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
968 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
969 use Unicode properties. (*UCP) at the start of a pattern can be used to set
970 this option. Modified pcretest to add /W to test this facility. Added
971 REG_UCP to make it available via the POSIX interface.
973 10. Added --line-buffered to pcregrep.
975 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
976 studied, and the match started with a letter with a code point greater than
977 127 whose first byte was different to the first byte of the other case of
978 the letter, the other case of this starting letter was not recognized
979 (#976).
981 12. If a pattern that was studied started with a repeated Unicode property
982 test, for example, \p{Nd}+, there was the theoretical possibility of
983 setting up an incorrect bitmap of starting bytes, but fortunately it could
984 not have actually happened in practice until change 8 above was made (it
985 added property types that matched character-matching opcodes).
987 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
988 possible starting bytes for non-anchored patterns.
990 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
991 \R, and also a number of cases that involve Unicode properties, both
992 explicit and implicit when PCRE_UCP is set.
994 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
995 input, it could crash or give wrong results if characters with values
996 greater than 0xc0 were present in the subject string. (Detail: it assumed
997 UTF-8 input when processing these items.)
999 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1000 size_t is 64-bit (#991).
1002 17. Added a check for running out of memory when PCRE is compiled with
1003 --disable-stack-for-recursion (#990).
1005 18. If the last data line in a file for pcretest does not have a newline on
1006 the end, a newline was missing in the output.
1008 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1009 less than 128) in its various bitmaps. However, there is a facility for
1010 generating tables according to the current locale when PCRE is compiled. It
1011 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1012 characters, are recognized by isspace() and therefore were getting set in
1013 these tables, and indeed these tables seem to approximate to ISO 8859. This
1014 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1015 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1016 which of course cannot start UTF-8 characters. I have changed the code so
1017 that only real ASCII characters (less than 128) and the correct starting
1018 bytes for UTF-8 encodings are set for characters greater than 127 when in
1019 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1020 altogether.)
1022 20. Added the /T option to pcretest so as to be able to run tests with non-
1023 standard character tables, thus making it possible to include the tests
1024 used for 19 above in the standard set of tests.
1026 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1027 reference to a subpattern the other side of a comment that contains an
1028 opening parenthesis caused either an internal compiling error, or a
1029 reference to the wrong subpattern.
1032 Version 8.02 19-Mar-2010
1033 ------------------------
1035 1. The Unicode data tables have been updated to Unicode 5.2.0.
1037 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1038 configured.
1040 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1041 original author of that file, following a query about its status.
1043 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1044 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1046 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1047 quantifier applied to a forward-referencing subroutine call, could compile
1048 incorrect code or give the error "internal error: previously-checked
1049 referenced subpattern not found".
1051 6. Both MS Visual Studio and Symbian OS have problems with initializing
1052 variables to point to external functions. For these systems, therefore,
1053 pcre_malloc etc. are now initialized to local functions that call the
1054 relevant global functions.
1056 7. There were two entries missing in the vectors called coptable and poptable
1057 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1058 I've fixed the data, and added a kludgy way of testing at compile time that
1059 the lengths are correct (equal to the number of opcodes).
1061 8. Following on from 7, I added a similar kludge to check the length of the
1062 eint vector in pcreposix.c.
1064 9. Error texts for pcre_compile() are held as one long string to avoid too
1065 much relocation at load time. To find a text, the string is searched,
1066 counting zeros. There was no check for running off the end of the string,
1067 which could happen if a new error number was added without updating the
1068 string.
1070 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1072 11. \K was not working if it appeared in an atomic group or in a group that
1073 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1074 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1075 the assertion is positive, but not if it is negative.
1077 12. Change 11 fortuitously reduced the size of the stack frame used in the
1078 "match()" function of pcre_exec.c by one pointer. Forthcoming
1079 implementation of support for (*MARK) will need an extra pointer on the
1080 stack; I have reserved it now, so that the stack frame size does not
1081 decrease.
1083 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1084 item in branch that calls a recursion is a subroutine call - as in the
1085 second branch in the above example - was incorrectly given the compile-
1086 time error "recursive call could loop indefinitely" because pcre_compile()
1087 was not correctly checking the subroutine for matching a non-empty string.
1089 14. The checks for overrunning compiling workspace could trigger after an
1090 overrun had occurred. This is a "should never occur" error, but it can be
1091 triggered by pathological patterns such as hundreds of nested parentheses.
1092 The checks now trigger 100 bytes before the end of the workspace.
1094 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1097 Version 8.01 19-Jan-2010
1098 ------------------------
1100 1. If a pattern contained a conditional subpattern with only one branch (in
1101 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1102 computed the wrong minimum data length (which is of course zero for such
1103 subpatterns). This could cause incorrect "no match" results.
1105 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1106 the pattern is reset in the first branch, pcre_compile() failed with
1107 "internal error: code overflow at offset...". This happened only when
1108 the reset was to the original external option setting. (An optimization
1109 abstracts leading options settings into an external setting, which was the
1110 cause of this.)
1112 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1113 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1114 assertion pattern did not match (meaning that the assertion was true), it
1115 was incorrectly treated as false if the SKIP had been reached during the
1116 matching. This also applied to assertions used as conditions.
1118 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1119 assertion subpattern, including such a pattern used as a condition,
1120 unpredictable results occurred, instead of the error return
1123 5. The C++ GlobalReplace function was not working like Perl for the special
1124 situation when an empty string is matched. It now does the fancy magic
1125 stuff that is necessary.
1127 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1128 removed. (These were left over from very, very early versions of PCRE.)
1130 7. Some cosmetic changes to the code to make life easier when compiling it
1131 as part of something else:
1133 (a) Change DEBUG to PCRE_DEBUG.
1135 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1136 called "current" as "current_branch", to prevent a collision with the
1137 Linux macro when compiled as a kernel module.
1139 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1140 prevent a collision with the Linux macro when compiled as a kernel
1141 module.
1143 8. In pcre_compile() there are some checks for integer overflows that used to
1144 cast potentially large values to (double). This has been changed to that
1145 when building, a check for int64_t is made, and if it is found, it is used
1146 instead, thus avoiding the use of floating point arithmetic. (There is no
1147 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1148 double.
1150 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1151 2005 (difference between two addresses compared to an unsigned value).
1153 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1154 custom one, because of the following reported problem in Windows:
1156 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1157 under Win32.
1158 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1159 therefore missing the function definition.
1160 - The compiler thus generates a "C" signature for the test function.
1161 - The linker fails to find the "C" function.
1162 - PCRE fails to configure if asked to do so against libbz2.
1164 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1165 messages were output:
1167 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1168 rerunning libtoolize, to keep the correct libtool macros in-tree.
1169 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1171 I have done both of these things.
1173 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1174 most of the time, it *can* run out if it is given a pattern that contains a
1175 runaway infinite recursion. I updated the discussion in the pcrestack man
1176 page.
1178 13. Now that we have gone to the x.xx style of version numbers, the minor
1179 version may start with zero. Using 08 or 09 is a bad idea because users
1180 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1181 interpreted as invalid octal numbers. I've updated the previous comment in
1182 configure.ac, and also added a check that gives an error if 08 or 09 are
1183 used.
1185 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1186 causing partial matching to fail when the end of the subject matched \W
1187 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1189 15. There were some discrepancies between the declarations in pcre_internal.h
1190 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1191 their definitions. The declarations used "const uschar *" and the
1192 definitions used USPTR. Even though USPTR is normally defined as "const
1193 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1194 reported that: "This difference in casting confuses some C++ compilers, for
1195 example, SunCC recognizes above declarations as different functions and
1196 generates broken code for hbpcre." I have changed the declarations to use
1197 USPTR.
1199 16. GNU libtool is named differently on some systems. The autogen.sh script now
1200 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1201 (FreeBSD).
1203 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1204 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1205 comment: "Figure out how to create a longlong from a string: strtoll and
1206 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1207 instance, but it only takes 2 args instead of 3!"
1209 18. A subtle bug concerned with back references has been fixed by a change of
1210 specification, with a corresponding code fix. A pattern such as
1211 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1212 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1213 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1214 same bug. Such groups have to be quantified to be useful, or contained
1215 inside another quantified group. (If there's no repetition, the reference
1216 can never match.) The problem arises because, having left the group and
1217 moved on to the rest of the pattern, a later failure that backtracks into
1218 the group uses the captured value from the final iteration of the group
1219 rather than the correct earlier one. I have fixed this in PCRE by forcing
1220 any group that contains a reference to itself to be an atomic group; that
1221 is, there cannot be any backtracking into it once it has completed. This is
1222 similar to recursive and subroutine calls.
1225 Version 8.00 19-Oct-09
1226 ----------------------
1228 1. The table for translating pcre_compile() error codes into POSIX error codes
1229 was out-of-date, and there was no check on the pcre_compile() error code
1230 being within the table. This could lead to an OK return being given in
1231 error.
1233 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1234 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1235 in a Windows environment.
1237 3. The pcregrep --count option prints the count for each file even when it is
1238 zero, as does GNU grep. However, pcregrep was also printing all files when
1239 --files-with-matches was added. Now, when both options are given, it prints
1240 counts only for those files that have at least one match. (GNU grep just
1241 prints the file name in this circumstance, but including the count seems
1242 more useful - otherwise, why use --count?) Also ensured that the
1243 combination -clh just lists non-zero counts, with no names.
1245 4. The long form of the pcregrep -F option was incorrectly implemented as
1246 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1247 but it seems right to fix it, and I didn't think it was worth preserving
1248 the old behaviour.
1250 5. The command line items --regex=pattern and --regexp=pattern were not
1251 recognized by pcregrep, which required --regex pattern or --regexp pattern
1252 (with a space rather than an '='). The man page documented the '=' forms,
1253 which are compatible with GNU grep; these now work.
1255 6. No libpcreposix.pc file was created for pkg-config; there was just
1256 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1258 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1259 when UCP support is not needed, by modifying the Python script that
1260 generates it from Unicode data files. This should not matter if the module
1261 is correctly used as a library, but I received one complaint about 50K of
1262 unwanted data. My guess is that the person linked everything into his
1263 program rather than using a library. Anyway, it does no harm.
1265 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1266 was a minimum greater than 1 for a wide character in a possessive
1267 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1268 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1269 character. Chaos in the form of incorrect output or a compiling loop could
1270 result.
1272 9. The restrictions on what a pattern can contain when partial matching is
1273 requested for pcre_exec() have been removed. All patterns can now be
1274 partially matched by this function. In addition, if there are at least two
1275 slots in the offset vector, the offset of the earliest inspected character
1276 for the match and the offset of the end of the subject are set in them when
1277 PCRE_ERROR_PARTIAL is returned.
1279 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1280 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1281 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1282 and may be more useful for multi-segment matching.
1284 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1285 used to be given if ever the end of the subject was reached; now it is
1286 given only if matching could not proceed because another character was
1287 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1288 string "Z", which now yields "no match" instead of "partial match". In the
1289 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1290 final character ended with (*FAIL).
1292 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1293 if the pattern had a "must contain" character that was already found in the
1294 earlier partial match, unless partial matching was again requested. For
1295 example, with the pattern /dog.(body)?/, the "must contain" character is
1296 "g". If the first part-match was for the string "dog", restarting with
1297 "sbody" failed. This bug has been fixed.
1299 13. The string returned by pcre_dfa_exec() after a partial match has been
1300 changed so that it starts at the first inspected character rather than the
1301 first character of the match. This makes a difference only if the pattern
1302 starts with a lookbehind assertion or \b or \B (\K is not supported by
1303 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1304 matching functions compatible, and I think it's the right thing to do.
1306 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1307 so that the demonstration program is easily available in environments where
1308 PCRE has not been installed from source.
1310 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1311 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1312 library.
1314 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1315 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1316 is not the first non-POSIX option to be added. Clearly some people find
1317 these options useful.
1319 17. If a caller to the POSIX matching function regexec() passes a non-zero
1320 value for nmatch with a NULL value for pmatch, the value of
1321 nmatch is forced to zero.
1323 18. RunGrepTest did not have a test for the availability of the -u option of
1324 the diff command, as RunTest does. It now checks in the same way as
1325 RunTest, and also checks for the -b option.
1327 19. If an odd number of negated classes containing just a single character
1328 interposed, within parentheses, between a forward reference to a named
1329 subpattern and the definition of the subpattern, compilation crashed with
1330 an internal error, complaining that it could not find the referenced
1331 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1332 [The bug was that it was starting one character too far in when skipping
1333 over the character class, thus treating the ] as data rather than
1334 terminating the class. This meant it could skip too much.]
1336 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1337 /g option in pcretest when the pattern contains \K, which makes it possible
1338 to have an empty string match not at the start, even when the pattern is
1339 anchored. Updated pcretest and pcredemo to use this option.
1341 21. If the maximum number of capturing subpatterns in a recursion was greater
1342 than the maximum at the outer level, the higher number was returned, but
1343 with unset values at the outer level. The correct (outer level) value is
1344 now given.
1346 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1347 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1348 make it do so. The string so far is captured, making this feature
1349 compatible with Perl.
1351 23. The tests have been re-organized, adding tests 11 and 12, to make it
1352 possible to check the Perl 5.10 features against Perl 5.10.
1354 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1355 pattern matches a fixed length string. PCRE did not allow this; now it
1356 does. Neither allows recursion.
1358 25. I finally figured out how to implement a request to provide the minimum
1359 length of subject string that was needed in order to match a given pattern.
1360 (It was back references and recursion that I had previously got hung up
1361 on.) This code has now been added to pcre_study(); it finds a lower bound
1362 to the length of subject needed. It is not necessarily the greatest lower
1363 bound, but using it to avoid searching strings that are too short does give
1364 some useful speed-ups. The value is available to calling programs via
1365 pcre_fullinfo().
1367 26. While implementing 25, I discovered to my embarrassment that pcretest had
1368 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1369 study optimizations had never been tested with that matching function.
1370 Oops. What is worse, even when it was passed study data, there was a bug in
1371 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1372 were also very few tests of studied patterns with pcre_dfa_exec().
1374 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1375 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1376 on the other side of the coin, they are no longer allowed to have different
1377 names, because these cannot be distinguished in PCRE, and this has caused
1378 confusion. (This is a difference from Perl.)
1380 28. When duplicate subpattern names are present (necessarily with different
1381 numbers, as required by 27 above), and a test is made by name in a
1382 conditional pattern, either for a subpattern having been matched, or for
1383 recursion in such a pattern, all the associated numbered subpatterns are
1384 tested, and the overall condition is true if the condition is true for any
1385 one of them. This is the way Perl works, and is also more like the way
1386 testing by number works.
1389 Version 7.9 11-Apr-09
1390 ---------------------
1392 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1393 (pcretest), all targets were linked against these libraries. This included
1394 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1395 libraries. This caused unwanted dependencies to be created. This problem
1396 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1397 pcretest is linked with readline.
1399 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1400 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1401 moved outside it again, because FALSE and TRUE are already defined in AIX,
1402 but BOOL is not.
1404 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1405 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1407 4. The pcregrep documentation said spaces were inserted as well as colons (or
1408 hyphens) following file names and line numbers when outputting matching
1409 lines. This is not true; no spaces are inserted. I have also clarified the
1410 wording for the --colour (or --color) option.
1412 5. In pcregrep, when --colour was used with -o, the list of matching strings
1413 was not coloured; this is different to GNU grep, so I have changed it to be
1414 the same.
1416 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1417 each matching line was coloured. Now it goes on to look for further matches
1418 of any of the test patterns, which is the same behaviour as GNU grep.
1420 7. A pattern that could match an empty string could cause pcregrep to loop; it
1421 doesn't make sense to accept an empty string match in pcregrep, so I have
1422 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1423 seems to be how GNU grep behaves.
1425 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1426 start or after a newline", because the conditional assertion was not being
1427 correctly handled. The rule now is that both the assertion and what follows
1428 in the first alternative must satisfy the test.
1430 9. If auto-callout was enabled in a pattern with a conditional group whose
1431 condition was an assertion, PCRE could crash during matching, both with
1432 pcre_exec() and pcre_dfa_exec().
1434 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1435 used for matching.
1437 11. Unicode property support in character classes was not working for
1438 characters (bytes) greater than 127 when not in UTF-8 mode.
1440 12. Added the -M command line option to pcretest.
1442 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1444 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1446 16. Added comments and documentation about mis-use of no_arg in the C++
1447 wrapper.
1449 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1450 from Martin Jerabek that uses macro names for all relevant character and
1451 string constants.
1453 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1454 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1455 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1456 these, but not everybody uses configure.
1458 19. A conditional group that had only one branch was not being correctly
1459 recognized as an item that could match an empty string. This meant that an
1460 enclosing group might also not be so recognized, causing infinite looping
1461 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1462 with the subject "ab", where knowledge that the repeated group can match
1463 nothing is needed in order to break the loop.
1465 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1466 exec(), but without supplying a callout function, matching went wrong.
1468 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1469 leak if the size of the offset vector was greater than 30. When the vector
1470 is smaller, the saved offsets during recursion go onto a local stack
1471 vector, but for larger vectors malloc() is used. It was failing to free
1472 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1473 error, in fact).
1475 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1476 heapframe that is used only when UTF-8 support is enabled. This caused no
1477 problem, but was untidy.
1479 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1480 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1481 included within another project.
1483 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1484 slightly modified by me:
1486 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1487 not building pcregrep.
1489 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1490 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1492 25. Forward references, both numeric and by name, in patterns that made use of
1493 duplicate group numbers, could behave incorrectly or give incorrect errors,
1494 because when scanning forward to find the reference group, PCRE was not
1495 taking into account the duplicate group numbers. A pattern such as
1496 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1498 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1499 the feature of a custom pointer more persuasive (as requested by a user).
1501 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1502 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1503 pre-defined.
1505 28. Added support for (*UTF8) at the start of a pattern.
1507 29. Arrange for flags added by the "release type" setting in CMake to be shown
1508 in the configuration summary.
1511 Version 7.8 05-Sep-08
1512 ---------------------
1514 1. Replaced UCP searching code with optimized version as implemented for Ad
1515 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1516 stage table and inline lookup instead of a function, giving speed ups of 2
1517 to 5 times on some simple patterns that I tested. Permission was given to
1518 distribute the MultiStage2.py script that generates the tables (it's not in
1519 the tarball, but is in the Subversion repository).
1521 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1522 scripts.
1524 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1525 a group with a zero qualifier. The result of the study could be incorrect,
1526 or the function might crash, depending on the pattern.
1528 4. Caseless matching was not working for non-ASCII characters in back
1529 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1530 It now works when Unicode Property Support is available.
1532 5. In pcretest, an escape such as \x{de} in the data was always generating
1533 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1534 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1535 truncation.
1537 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1539 7. Added two (int) casts to pcregrep when printing the difference of two
1540 pointers, in case they are 64-bit values.
1542 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1543 test 2 if it fails.
1545 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1546 and a #define of that name to empty if it is not externally set. This is to
1547 allow users of MSVC to set it if necessary.
1549 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1550 the convenience functions in the pcre_get.c source file.
1552 11. An option change at the start of a pattern that had top-level alternatives
1553 could cause overwriting and/or a crash. This command provoked a crash in
1554 some environments:
1556 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1558 This potential security problem was recorded as CVE-2008-2371.
1560 12. For a pattern where the match had to start at the beginning or immediately
1561 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1562 pcre_dfa_exec() could read past the end of the passed subject if there was
1563 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1564 pcretest so that it places the subject at the end of its malloc-ed buffer.
1566 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1567 exec() might read past the end of the data buffer in UTF-8 mode.
1569 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1570 the data contained the byte 0x85 as part of a UTF-8 character within its
1571 first line. This applied both to normal and DFA matching.
1573 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1574 /^[^d]*?$/8 failed to match "abc".
1576 16. Added a missing copyright notice to pcrecpp_internal.h.
1578 17. Make it more clear in the documentation that values returned from
1579 pcre_exec() in ovector are byte offsets, not character counts.
1581 18. Tidied a few places to stop certain compilers from issuing warnings.
1583 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1584 supplied by Stefan Weber. I made a further small update for 7.8 because
1585 there is a change of source arrangements: the pcre_searchfuncs.c module is
1586 replaced by pcre_ucd.c.
1589 Version 7.7 07-May-08
1590 ---------------------
1592 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1593 a string to a long long, pretend we don't even have a long long." This is
1594 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1596 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1597 pre-7.6 versions, which defined a global no_arg variable instead of putting
1598 it in the RE class. (See also #8 below.)
1600 3. Remove a line of dead code, identified by coverity and reported by Nuno
1601 Lopes.
1603 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1605 (1) The include/exclude patterns were being applied to the whole pathnames
1606 of files, instead of just to the final components.
1608 (2) If there was more than one level of directory, the subdirectories were
1609 skipped unless they satisfied the include/exclude conditions. This is
1610 inconsistent with GNU grep (and could even be seen as contrary to the
1611 pcregrep specification - which I improved to make it absolutely clear).
1612 The action now is always to scan all levels of directory, and just
1613 apply the include/exclude patterns to regular files.
1615 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1616 --exclude_dir in the tests to avoid scanning .svn directories.
1618 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1619 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1620 doesn't support NULs in patterns.
1622 7. Added some missing "const"s to declarations of static tables in
1623 pcre_compile.c and pcre_dfa_exec.c.
1625 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1626 caused by fix #2 above. (Subsequently also a second patch to fix the
1627 first patch. And a third patch - this was a messy problem.)
1629 9. Applied Craig's patch to remove the use of push_back().
1631 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1632 matching function regexec().
1634 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1635 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1636 references. PCRE supports relative numbers with this syntax (I don't think
1637 Oniguruma does).
1639 12. Previously, a group with a zero repeat such as (...){0} was completely
1640 omitted from the compiled regex. However, this means that if the group
1641 was called as a subroutine from elsewhere in the pattern, things went wrong
1642 (an internal error was given). Such groups are now left in the compiled
1643 pattern, with a new opcode that causes them to be skipped at execution
1644 time.
1646 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1647 to the way PCRE behaves:
1649 (a) A lone ] character is dis-allowed (Perl treats it as data).
1651 (b) A back reference to an unmatched subpattern matches an empty string
1652 (Perl fails the current match path).
1654 (c) A data ] in a character class must be notated as \] because if the
1655 first data character in a class is ], it defines an empty class. (In
1656 Perl it is not possible to have an empty class.) The empty class []
1657 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1658 The negative empty class [^] matches any one character, independently
1659 of the DOTALL setting.
1661 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1662 non-existent subpattern following a character class starting with ']' and
1663 containing () gave an internal compiling error instead of "reference to
1664 non-existent subpattern". Fortunately, when the pattern did exist, the
1665 compiled code was correct. (When scanning forwards to check for the
1666 existencd of the subpattern, it was treating the data ']' as terminating
1667 the class, so got the count wrong. When actually compiling, the reference
1668 was subsequently set up correctly.)
1670 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1671 it was being rejected as not supported by pcre_dfa_exec(), even though
1672 other assertions are supported. I have made pcre_dfa_exec() support
1673 (*FAIL).
1675 16. The implementation of 13c above involved the invention of a new opcode,
1676 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1677 cannot be changed at match time, I realized I could make a small
1678 improvement to matching performance by compiling OP_ALLANY instead of
1679 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1680 on the OP_ANY path.
1682 17. Compiling pcretest on Windows with readline support failed without the
1683 following two fixes: (1) Make the unistd.h include conditional on
1684 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1686 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1687 ncurses library to be included for pcretest when ReadLine support is
1688 requested, but also to allow for it to be overridden. This patch came from
1689 Daniel Bergström.
1691 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1692 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1693 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1694 spotting this.
1697 Version 7.6 28-Jan-08
1698 ---------------------
1700 1. A character class containing a very large number of characters with
1701 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1702 overflow.
1704 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1705 HAVE_LONG_LONG is not defined.
1707 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1708 bring them up to date and include new features. This patch includes:
1710 - Fixed PH's badly added libz and libbz2 support.
1711 - Fixed a problem with static linking.
1712 - Added pcredemo. [But later removed - see 7 below.]
1713 - Fixed dftables problem and added an option.
1714 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1716 - Added readline support for pcretest.
1717 - Added an listing of the option settings after cmake has run.
1719 4. A user submitted a patch to Makefile that makes it easy to create
1720 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1721 Makefile.am that cause it to include this special target, without
1722 affecting anything else. Note that the same mingw target plus all
1723 the other distribution libraries and programs are now supported
1724 when configuring with CMake (see 6 below) instead of with
1725 Configure/Make.
1727 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1728 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1729 exported in the Windows port". It has not yet been confirmed that the patch
1730 solves the problem, but it does no harm.
1732 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1733 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1734 with CMake, and also correct the comment about stack recursion.
1736 7. Remove the automatic building of pcredemo from the ./configure system and
1737 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1738 of a program that users should build themselves after PCRE is installed, so
1739 building it automatically is not really right. What is more, it gave
1740 trouble in some build environments.
1742 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1745 Version 7.5 10-Jan-08
1746 ---------------------
1748 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1749 values in parens when parsing an RE using the C++ wrapper."
1751 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1752 Characters greater than 255 were excluded from the class instead of being
1753 included.
1755 3. The same bug as (2) above applied to negated POSIX classes such as
1756 [:^space:].
1758 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1759 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1760 I have changed it.
1762 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1763 first named subpattern) and a construct such as (?&a) would reference the
1764 first named subpattern whose name started with "a" (in other words, the
1765 length check was missing). Both these problems are fixed. "Subpattern name
1766 expected" is now given for (?&) (a zero-length name), and this patch also
1767 makes it give the same error for \k'' (previously it complained that that
1768 was a reference to a non-existent subpattern).
1770 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1771 this is right because (?- can be followed by option settings as well as by
1772 digits. I have, however, made the messages clearer.
1774 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1775 than the number used in the conditional) now cause a compile-time error.
1776 This is actually not compatible with Perl, which accepts such patterns, but
1777 treats the conditional as always being FALSE (as PCRE used to), but it
1778 seems to me that giving a diagnostic is better.
1780 8. Change "alphameric" to the more common word "alphanumeric" in comments
1781 and messages.
1783 9. Fix two occurrences of "backslash" in comments that should have been
1784 "backspace".
1786 10. Remove two redundant lines of code that can never be obeyed (their function
1787 was moved elsewhere).
1789 11. The program that makes PCRE's Unicode character property table had a bug
1790 which caused it to generate incorrect table entries for sequences of
1791 characters that have the same character type, but are in different scripts.
1792 It amalgamated them into a single range, with the script of the first of
1793 them. In other words, some characters were in the wrong script. There were
1794 thirteen such cases, affecting characters in the following ranges:
1796 U+002b0 - U+002c1
1797 U+0060c - U+0060d
1798 U+0061e - U+00612
1799 U+0064b - U+0065e
1800 U+0074d - U+0076d
1801 U+01800 - U+01805
1802 U+01d00 - U+01d77
1803 U+01d9b - U+01dbf
1804 U+0200b - U+0200f
1805 U+030fc - U+030fe
1806 U+03260 - U+0327f
1807 U+0fb46 - U+0fbb1
1808 U+10450 - U+1049d
1810 12. The -o option (show only the matching part of a line) for pcregrep was not
1811 compatible with GNU grep in that, if there was more than one match in a
1812 line, it showed only the first of them. It now behaves in the same way as
1813 GNU grep.
1815 13. If the -o and -v options were combined for pcregrep, it printed a blank
1816 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1817 does the same. The return code can be used to tell if there were any
1818 non-matching lines.
1820 14. Added --file-offsets and --line-offsets to pcregrep.
1822 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1823 infinitely looping recursion. The bug was that positive lookaheads were not
1824 being skipped when checking for a possible empty match (negative lookaheads
1825 and both kinds of lookbehind were skipped).
1827 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1828 inclusion of <windows.h> to before rather than after the definition of
1829 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1831 17. Specifying a possessive quantifier with a specific limit for a Unicode
1832 character property caused pcre_compile() to compile bad code, which led at
1833 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1834 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1835 caused the error; without that there was no problem.
1837 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1839 19. Added --enable-pcretest-libreadline.
1841 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1842 RE::GlobalReplace(). As a result, the number of replacements returned was
1843 double what it should be. I removed one of the increments, but Craig sent a
1844 later patch that removed the other one (the right fix) and added unit tests
1845 that check the return values (which was not done before).
1847 21. Several CMake things:
1849 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1850 the names libpcre and libpcreposix, not just pcre and pcreposix.
1852 (2) The above change means that pcretest and pcregrep are now correctly
1853 linked with the newly-built libraries, not previously installed ones.
1857 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1858 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1859 UTF-8 newline character). The key issue is that the pattern starts .*;
1860 this means that the match must be either at the beginning, or after a
1861 newline. The bug was in the code for advancing after a failed match and
1862 checking that the new position followed a newline. It was not taking
1863 account of UTF-8 characters correctly.
1865 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1866 character classes. PCRE was not treating the sequence [:...:] as a
1867 character class unless the ... were all letters. Perl, however, seems to
1868 allow any characters between [: and :], though of course it rejects as
1869 unknown any "names" that contain non-letters, because all the known class
1870 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1871 for example, whereas PCRE did not - it did not recognize a POSIX character
1872 class. This seemed a bit dangerous, so the code has been changed to be
1873 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1874 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1875 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1876 Perl does, and where it didn't before.
1878 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1879 Windows environments %n is disabled by default.
1882 Version 7.4 21-Sep-07
1883 ---------------------
1885 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1886 means that a class such as [\s] counted as "explicit reference to CR or
1887 LF". That isn't really right - the whole point of the change was to try to
1888 help when there was an actual mention of one of the two characters. So now
1889 the change happens only if \r or \n (or a literal CR or LF) character is
1890 encountered.
1892 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1893 of both had grown to the point where there were only 3 bits left.
1894 Fortunately, there was spare space in the data structure, and so I have
1895 moved the internal flags into a new 16-bit field to free up more option
1896 bits.
1898 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1899 but did not set the internal JCHANGED flag - either of these is enough to
1900 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1901 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1902 start sets both bits.
1904 4. Added options (at build time, compile time, exec time) to change \R from
1905 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1907 5. doc/pcresyntax.html was missing from the distribution.
1909 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1910 compatibility, even though it is no longer used.
1912 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1913 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1914 windows.h file is present (where different names are used). [This was
1915 reversed later after testing - see 16 below.]
1917 8. Changed all #include <config.h> to #include "config.h". There were also
1918 some further <pcre.h> cases that I changed to "pcre.h".
1920 9. When pcregrep was used with the --colour option, it missed the line ending
1921 sequence off the lines that it output.
1923 10. It was pointed out to me that arrays of string pointers cause lots of
1924 relocations when a shared library is dynamically loaded. A technique of
1925 using a single long string with a table of offsets can drastically reduce
1926 these. I have refactored PCRE in four places to do this. The result is
1927 dramatic:
1929 Originally: 290
1930 After changing UCP table: 187
1931 After changing error message table: 43
1932 After changing table of "verbs" 36
1933 After changing table of Posix names 22
1935 Thanks to the folks working on Gregex for glib for this insight.
1937 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1938 unicode-properties was also set.
1940 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1942 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1943 checked only for CRLF.
1945 14. Added casts to pcretest.c to avoid compiler warnings.
1947 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1949 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1950 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1951 entirely. This removes changes made in 7 above.
1953 17. The CMake files have been updated, and there is now more information about
1954 building with CMake in the NON-UNIX-USE document.
1957 Version 7.3 28-Aug-07
1958 ---------------------
1960 1. In the rejigging of the build system that eventually resulted in 7.1, the
1961 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1962 brackets there is not right, since it causes compilers to look for an
1963 installed pcre.h, not the version that is in the source that is being
1964 compiled (which of course may be different). I have changed it back to:
1966 #include "pcre.h"
1968 I have a vague recollection that the change was concerned with compiling in
1969 different directories, but in the new build system, that is taken care of
1970 by the VPATH setting the Makefile.
1972 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1973 when the subject happened to end in the byte 0x85 (e.g. if the last
1974 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1975 characters but of course it shouldn't be taken as a newline when it is part
1976 of another character. The bug was that, for an unlimited repeat of . in
1977 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1978 characters when looking for a newline.
1980 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1982 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1983 in debug output.
1985 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1986 long printing in the pcrecpp unittest when running under MinGW.
1988 6. ESC_K was left out of the EBCDIC table.
1990 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1991 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1992 limit also applies to "virtual nesting" when a pattern is recursive, and in
1993 this case 1000 isn't so big. I have been able to remove this limit at the
1994 expense of backing off one optimization in certain circumstances. Normally,
1995 when pcre_exec() would call its internal match() function recursively and
1996 immediately return the result unconditionally, it uses a "tail recursion"
1997 feature to save stack. However, when a subpattern that can match an empty
1998 string has an unlimited repetition quantifier, it no longer makes this
1999 optimization. That gives it a stack frame in which to save the data for
2000 checking that an empty string has been matched. Previously this was taken
2001 from the 1000-entry workspace that had been reserved. So now there is no
2002 explicit limit, but more stack is used.
2004 8. Applied Daniel's patches to solve problems with the import/export magic
2005 syntax that is required for Windows, and which was going wrong for the
2006 pcreposix and pcrecpp parts of the library. These were overlooked when this
2007 problem was solved for the main library.
2009 9. There were some crude static tests to avoid integer overflow when computing
2010 the size of patterns that contain repeated groups with explicit upper
2011 limits. As the maximum quantifier is 65535, the maximum group length was
2012 set at 30,000 so that the product of these two numbers did not overflow a
2013 32-bit integer. However, it turns out that people want to use groups that
2014 are longer than 30,000 bytes (though not repeat them that many times).
2015 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2016 made it possible to implement the integer overflow checks in a much more
2017 dynamic way, which I have now done. The artificial limitation on group
2018 length has been removed - we now have only the limit on the total length of
2019 the compiled pattern, which depends on the LINK_SIZE setting.
2021 10. Fixed a bug in the documentation for get/copy named substring when
2022 duplicate names are permitted. If none of the named substrings are set, the
2023 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2024 empty string.
2026 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2027 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2028 because the ] is interpreted as the first data character and the
2029 terminating ] is not found. PCRE has been made compatible with Perl in this
2030 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2031 cause memory overwriting.
2033 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2034 string has been matched (to stop an infinite loop). It was not recognizing
2035 a conditional subpattern that could match an empty string if that
2036 subpattern was within another subpattern. For example, it looped when
2037 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2038 condition was not nested. This bug has been fixed.
2040 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2041 past the start of the subject in the presence of bytes with the top bit
2042 set, for example "\x8aBCD".
2044 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2045 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2047 14. Optimized (?!) to (*FAIL).
2049 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2050 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2051 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2052 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2053 does: it's just the validity check that is more restrictive.
2055 16. Inserted checks for integer overflows during escape sequence (backslash)
2056 processing, and also fixed erroneous offset values for syntax errors during
2057 backslash processing.
2059 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2060 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2062 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2063 caused an overrun.
2065 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2066 something other than just ASCII characters) inside a group that had an
2067 unlimited repeat caused a loop at compile time (while checking to see
2068 whether the group could match an empty string).
2070 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2071 [\P{Any}] did so. (Error in the code for printing property names.)
2073 21. An orphan \E inside a character class could cause a crash.
2075 22. A repeated capturing bracket such as (A)? could cause a wild memory
2076 reference during compilation.
2078 23. There are several functions in pcre_compile() that scan along a compiled
2079 expression for various reasons (e.g. to see if it's fixed length for look
2080 behind). There were bugs in these functions when a repeated \p or \P was
2081 present in the pattern. These operators have additional parameters compared
2082 with \d, etc, and these were not being taken into account when moving along
2083 the compiled data. Specifically:
2085 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2086 length.
2088 (b) An item such as \pL+ within a repeated group could cause crashes or
2089 loops.
2091 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2092 "reference to non-existent subpattern" error.
2094 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2096 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2097 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2099 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2101 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2102 character were causing crashes (broken optimization).
2104 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2105 \p or \P) caused a compile-time loop.
2107 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2108 break. For example, the unstudied pattern [\r\n]A does not match the string
2109 "\r\nA" because change 7.0/46 below moves the current point on by two
2110 characters after failing to match at the start. However, the pattern \nA
2111 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2112 the same is true. There doesn't seem any very clean way out of this, but
2113 what I have chosen to do makes the common cases work: PCRE now takes note
2114 of whether there can be an explicit match for \r or \n anywhere in the
2115 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2116 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2117 pattern has explicit CR or LF references.
2119 29. Added (*CR) etc for changing newline setting at start of pattern.
2122 Version 7.2 19-Jun-07
2123 ---------------------
2125 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2126 which is apparently normally available under Windows.
2128 2. Re-jig the pcregrep tests with different newline settings in an attempt
2129 to make them independent of the local environment's newline setting.
2131 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2133 4. Some of the "internals" tests were previously cut out when the link size
2134 was not 2, because the output contained actual offsets. The recent new
2135 "Z" feature of pcretest means that these can be cut out, making the tests
2136 usable with all link sizes.
2138 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2139 stack recursion. This gives a massive performance boost under BSD, but just
2140 a small improvement under Linux. However, it saves one field in the frame
2141 in all cases.
2143 6. Added more features from the forthcoming Perl 5.10:
2145 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2146 recursion call. It refers to the nth most recently opened parentheses.
2148 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2149 to be opened parentheses.
2151 (c) Conditions that refer to capturing parentheses can be specified
2152 relatively, for example, (?(-2)... or (?(+3)...
2154 (d) \K resets the start of the current match so that everything before
2155 is not part of it.
2157 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2159 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2160 reference syntax.
2162 (g) (?| introduces a group in which the numbering of parentheses in each
2163 alternative starts with the same number.
2165 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2167 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2170 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2171 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2172 for detecting groups that can match an empty string.
2174 9. A pattern with a very large number of alternatives (more than several
2175 hundred) was running out of internal workspace during the pre-compile
2176 phase, where pcre_compile() figures out how much memory will be needed. A
2177 bit of new cunning has reduced the workspace needed for groups with
2178 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2179 workspace instead of running out of the 4096 that are available.
2181 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2183 11. Applied patch from Google to remove an optimization that didn't quite work.
2184 The report of the bug said:
2186 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2187 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2188 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2190 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2191 it matched the wrong number of bytes.
2194 Version 7.1 24-Apr-07
2195 ---------------------
2197 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2198 that is more "standard", making use of automake and other Autotools. There
2199 is some re-arrangement of the files and adjustment of comments consequent
2200 on this.
2202 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2203 for recursive directory scanning broke on some systems because the files
2204 are not scanned in any specific order and on different systems the order
2205 was different. A call to "sort" has been inserted into RunGrepTest for the
2206 approprate test as a short-term fix. In the longer term there may be an
2207 alternative.
2209 3. I had an email from Eric Raymond about problems translating some of PCRE's
2210 man pages to HTML (despite the fact that I distribute HTML pages, some
2211 people do their own conversions for various reasons). The problems
2212 concerned the use of low-level troff macros .br and .in. I have therefore
2213 removed all such uses from the man pages (some were redundant, some could
2214 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2215 HTML has been updated to handle .nf/.fi and to complain if it encounters
2216 .br or .in.
2218 4. Updated comments in configure.ac that get placed in config.h.in and also
2219 arranged for config.h to be included in the distribution, with the name
2220 config.h.generic, for the benefit of those who have to compile without
2221 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2223 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2224 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2225 makevp.bat for the current PCRE, using the additional files
2226 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2228 6. A Windows user reported a minor discrepancy with test 2, which turned out
2229 to be caused by a trailing space on an input line that had got lost in his
2230 copy. The trailing space was an accident, so I've just removed it.
2232 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2233 that is needed.
2235 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2236 as "const" (a) because they are and (b) because it helps the PHP
2237 maintainers who have recently made a script to detect big data structures
2238 in the php code that should be moved to the .rodata section. I remembered
2239 to update Builducptable as well, so it won't revert if ucptable.h is ever
2240 re-created.
2242 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2243 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2244 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2245 support is not required. This saves 1.5-2K of code, which is important in
2246 some applications.
2248 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2249 so as not to refer to the tables, even though these functions will never be
2250 called when UTF-8 support is disabled. Otherwise there are problems with a
2251 shared library.
2253 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2255 (a) It was defining its arguments as char * instead of void *.
2257 (b) It was assuming that all moves were upwards in memory; this was true
2258 a long time ago when I wrote it, but is no longer the case.
2260 The emulated memove() is provided for those environments that have neither
2261 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2262 is clearly not the case, as these two bugs were recently reported.
2264 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2265 and Detrail to create the HTML documentation, the .txt form of the man
2266 pages, and it removes trailing spaces from listed files. It also creates
2267 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2268 case, it wraps all the #defines with #ifndefs. This script should be run
2269 before "make dist".
2271 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2272 with Unicode property support.
2274 (a) For a maximizing quantifier, if the two different cases of the
2275 character were of different lengths in their UTF-8 codings (there are
2276 some cases like this - I found 11), and the matching function had to
2277 back up over a mixture of the two cases, it incorrectly assumed they
2278 were both the same length.
2280 (b) When PCRE was configured to use the heap rather than the stack for
2281 recursion during matching, it was not correctly preserving the data for
2282 the other case of a UTF-8 character when checking ahead for a match
2283 while processing a minimizing repeat. If the check also involved
2284 matching a wide character, but failed, corruption could cause an
2285 erroneous result when trying to check for a repeat of the original
2286 character.
2288 13. Some tidying changes to the testing mechanism:
2290 (a) The RunTest script now detects the internal link size and whether there
2291 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2292 values substituted by "configure". (The RunGrepTest script already did
2293 this for UTF-8.) The configure.ac script no longer substitutes the
2294 relevant variables.
2296 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2297 with length and offset values. This means that the output is different
2298 for different internal link sizes. Test 2 is skipped for link sizes
2299 other than 2 because of this, bypassing the problem. Unfortunately,
2300 there was also a test in test 3 (the locale tests) that used /B and
2301 failed for link sizes other than 2. Rather than cut the whole test out,
2302 I have added a new /Z option to pcretest that replaces the length and
2303 offset values with spaces. This is now used to make test 3 independent
2304 of link size. (Test 2 will be tidied up later.)
2306 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2307 segmentation fault instead of returning the appropriate error message.
2309 15. In multiline mode when the newline sequence was set to "any", the pattern
2310 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2311 This doesn't seem right; it now treats the CRLF combination as the line
2312 ending, and so does not match in that case. It's only a pattern such as ^$
2313 that would hit this one: something like ^ABC$ would have failed after \r
2314 and then tried again after \r\n.
2316 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2317 in an attempt to make files that differ only in their line terminators
2318 compare equal. This works on Linux.
2320 17. Under certain error circumstances pcregrep might try to free random memory
2321 as it exited. This is now fixed, thanks to valgrind.
2323 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2324 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2325 was because its rules for how to advance for /g after matching an empty
2326 string at the end of a line did not allow for this case. They now check for
2327 it specially.
2329 20. pcretest is supposed to handle patterns and data of any length, by
2330 extending its buffers when necessary. It was getting this wrong when the
2331 buffer for a data line had to be extended.
2333 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2334 CRLF as a newline sequence.
2336 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2337 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2338 I have nevertheless tidied it up.
2340 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2342 24. Added a man page for pcre-config.
2345 Version 7.0 19-Dec-06
2346 ---------------------
2348 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2349 moving to gcc 4.1.1.
2351 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2352 sys/time.h, which is documented as needed for this function. It doesn't
2353 seem to matter on Linux, but it showed up on some releases of OS X.
2355 3. It seems that there are systems where bytes whose values are greater than
2356 127 match isprint() in the "C" locale. The "C" locale should be the
2357 default when a C program starts up. In most systems, only ASCII printing
2358 characters match isprint(). This difference caused the output from pcretest
2359 to vary, making some of the tests fail. I have changed pcretest so that:
2361 (a) When it is outputting text in the compiled version of a pattern, bytes
2362 other than 32-126 are always shown as hex escapes.
2364 (b) When it is outputting text that is a matched part of a subject string,
2365 it does the same, unless a different locale has been set for the match
2366 (using the /L modifier). In this case, it uses isprint() to decide.
2368 4. Fixed a major bug that caused incorrect computation of the amount of memory
2369 required for a compiled pattern when options that changed within the
2370 pattern affected the logic of the preliminary scan that determines the
2371 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2372 that the computed length was too small. The symptoms of this bug were
2373 either the PCRE error "internal error: code overflow" from pcre_compile(),
2374 or a glibc crash with a message such as "pcretest: free(): invalid next
2375 size (fast)". Examples of patterns that provoked this bug (shown in
2376 pcretest format) are:
2378 /(?-x: )/x
2379 /(?x)(?-x: \s*#\s*)/
2380 /((?i)[\x{c0}])/8
2381 /(?i:[\x{c0}])/8
2383 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2384 is now done differently.
2386 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2387 wrapper classes; (b) implement a new function in the C++ scanner that is
2388 more efficient than the old way of doing things because it avoids levels of
2389 recursion in the regex matching; (c) add a paragraph to the documentation
2390 for the FullMatch() function.
2392 6. The escape sequence \n was being treated as whatever was defined as
2393 "newline". Not only was this contrary to the documentation, which states
2394 that \n is character 10 (hex 0A), but it also went horribly wrong when
2395 "newline" was defined as CRLF. This has been fixed.
2397 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2398 was being set to -1 for the "end of line" case (supposedly a value that no
2399 character can have). Though this value is never used (the check for end of
2400 line is "zero bytes in current character"), it caused compiler complaints.
2401 I've changed it to 0xffffffff.
2403 8. In pcre_version.c, the version string was being built by a sequence of
2404 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2405 string (as it is for production releases) called a macro with an empty
2406 argument. The C standard says the result of this is undefined. The gcc
2407 compiler treats it as an empty string (which was what was wanted) but it is
2408 reported that Visual C gives an error. The source has been hacked around to
2409 avoid this problem.
2411 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2412 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2413 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2414 of them did).
2416 10. Originally, pcretest opened its input and output without "b"; then I was
2417 told that "b" was needed in some environments, so it was added for release
2418 5.0 to both the input and output. (It makes no difference on Unix-like
2419 systems.) Later I was told that it is wrong for the input on Windows. I've
2420 now abstracted the modes into two macros, to make it easier to fiddle with
2421 them, and removed "b" from the input mode under Windows.
2423 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2425 12. Added -help and --help to pcretest as an official way of being reminded
2426 of the options.
2428 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2429 and pcrecpp.cc because they annoy compilers at high warning levels.
2431 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2433 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2434 scripts are not C programs :-) and which was not noticed because it works
2435 on Linux.
2437 16. pcretest is supposed to handle any length of pattern and data line (as one
2438 line or as a continued sequence of lines) by extending its input buffer if
2439 necessary. This feature was broken for very long pattern lines, leading to
2440 a string of junk being passed to pcre_compile() if the pattern was longer
2441 than about 50K.
2443 17. I have done a major re-factoring of the way pcre_compile() computes the
2444 amount of memory needed for a compiled pattern. Previously, there was code
2445 that made a preliminary scan of the pattern in order to do this. That was
2446 OK when PCRE was new, but as the facilities have expanded, it has become
2447 harder and harder to keep it in step with the real compile phase, and there
2448 have been a number of bugs (see for example, 4 above). I have now found a
2449 cunning way of running the real compile function in a "fake" mode that
2450 enables it to compute how much memory it would need, while actually only
2451 ever using a few hundred bytes of working memory and without too many
2452 tests of the mode. This should make future maintenance and development
2453 easier. A side effect of this work is that the limit of 200 on the nesting
2454 depth of parentheses has been removed (though this was never a serious
2455 limitation, I suspect). However, there is a downside: pcre_compile() now
2456 runs more slowly than before (30% or more, depending on the pattern). I
2457 hope this isn't a big issue. There is no effect on runtime performance.
2459 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2460 newline (only possible for the last line of a file) and it was a
2461 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2463 19. Added additional timing features to pcretest. (1) The -tm option now times
2464 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2465 separate command line item, by a number that specifies the number of
2466 repeats to use when timing. The default is 50000; this gives better
2467 precision, but takes uncomfortably long for very large patterns.
2469 20. Extended pcre_study() to be more clever in cases where a branch of a
2470 subpattern has no definite first character. For example, (a*|b*)[cd] would
2471 previously give no result from pcre_study(). Now it recognizes that the
2472 first character must be a, b, c, or d.
2474 21. There was an incorrect error "recursive call could loop indefinitely" if
2475 a subpattern (or the entire pattern) that was being tested for matching an
2476 empty string contained only one non-empty item after a nested subpattern.
2477 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2478 incorrectly, because the \d was being skipped in the check.
2480 22. The pcretest program now has a new pattern option /B and a command line
2481 option -b, which is equivalent to adding /B to every pattern. This causes
2482 it to show the compiled bytecode, without the additional information that
2483 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2484 is the same as /B/I).
2486 23. A new optimization is now able automatically to treat some sequences such
2487 as a*b as a*+b. More specifically, if something simple (such as a character
2488 or a simple class like \d) has an unlimited quantifier, and is followed by
2489 something that cannot possibly match the quantified thing, the quantifier
2490 is automatically "possessified".
2492 24. A recursive reference to a subpattern whose number was greater than 39
2493 went wrong under certain circumstances in UTF-8 mode. This bug could also
2494 have affected the operation of pcre_study().
2496 25. Realized that a little bit of performance could be had by replacing
2497 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2499 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2501 27. Possessive quantifiers such as a++ were previously implemented by turning
2502 them into atomic groups such as ($>a+). Now they have their own opcodes,
2503 which improves performance. This includes the automatically created ones
2504 from 23 above.
2506 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2507 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2508 the first matched character to be a colon. This applied both to named and
2509 numbered groups.
2511 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2513 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2514 building PCRE on Windows easier, so I have included it in the distribution.
2516 31. There is now a check in pcretest against a ridiculously large number being
2517 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2518 loop, the loop is abandoned.
2520 32. Forward references to subpatterns in conditions such as (?(2)...) where
2521 subpattern 2 is defined later cause pcre_compile() to search forwards in
2522 the pattern for the relevant set of parentheses. This search went wrong
2523 when there were unescaped parentheses in a character class, parentheses
2524 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2526 33. "Subroutine" calls and backreferences were previously restricted to
2527 referencing subpatterns earlier in the regex. This restriction has now
2528 been removed.
2530 34. Added a number of extra features that are going to be in Perl 5.10. On the
2531 whole, these are just syntactic alternatives for features that PCRE had
2532 previously implemented using the Python syntax or my own invention. The
2533 other formats are all retained for compatibility.
2535 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2536 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2537 also .NET compatible.
2539 (b) A recursion or subroutine call to a named group can now be defined as
2540 (?&name) as well as (?P>name).
2542 (c) A backreference to a named group can now be defined as \k<name> or
2543 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2544 5.10, are also .NET compatible.
2546 (d) A conditional reference to a named group can now use the syntax
2547 (?(<name>) or (?('name') as well as (?(name).
2549 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2550 groups (named and numbered) that are never evaluated inline, but can be
2551 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2552 is always false. There may be only one alternative in such a group.
2554 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2555 as the simple (?(R). The condition is true only if the most recent
2556 recursion is that of the given number or name. It does not search out
2557 through the entire recursion stack.
2559 (g) The escape \gN or \g{N} has been added, where N is a positive or
2560 negative number, specifying an absolute or relative reference.
2562 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2563 some "unreachable code" warnings.
2565 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2566 things, this adds five new scripts.
2568 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2569 There were also incompatibilities regarding the handling of \Q..\E inside
2570 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2571 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2573 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2574 matches an empty string, and forcibly breaks the loop. There were bugs in
2575 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2576 against aaaa the result was just "a" rather than "aaaa", for example. Two
2577 separate and independent bugs (that affected different cases) have been
2578 fixed.
2580 39. Refactored the code to abolish the use of different opcodes for small
2581 capturing bracket numbers. This is a tidy that I avoided doing when I
2582 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2583 The new approach is not only tidier, it makes it possible to reduce the
2584 memory needed to fix the previous bug (38).
2586 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2587 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2588 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2589 mode.
2591 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2592 report.
2594 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2595 copy construction and assignment in the C++ wrapper.
2597 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2598 couple of bugs that could have given wrong results in the "--newline=crlf"
2599 case.
2601 44. Added a number of casts and did some reorganization of signed/unsigned int
2602 variables following suggestions from Dair Grant. Also renamed the variable
2603 "this" as "item" because it is a C++ keyword.
2605 45. Arranged for dftables to add
2607 #include "pcre_internal.h"
2609 to pcre_chartables.c because without it, gcc 4.x may remove the array
2610 definition from the final binary if PCRE is built into a static library and
2611 dead code stripping is activated.
2613 46. For an unanchored pattern, if a match attempt fails at the start of a
2614 newline sequence, and the newline setting is CRLF or ANY, and the next two
2615 characters are CRLF, advance by two characters instead of one.
2618 Version 6.7 04-Jul-06
2619 ---------------------
2621 1. In order to handle tests when input lines are enormously long, pcretest has
2622 been re-factored so that it automatically extends its buffers when
2623 necessary. The code is crude, but this _is_ just a test program. The
2624 default size has been increased from 32K to 50K.
2626 2. The code in pcre_study() was using the value of the re argument before
2627 testing it for NULL. (Of course, in any sensible call of the function, it
2628 won't be NULL.)
2630 3. The memmove() emulation function in pcre_internal.h, which is used on
2631 systems that lack both memmove() and bcopy() - that is, hardly ever -
2632 was missing a "static" storage class specifier.
2634 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2635 containing an extended class (one that cannot be represented by a bitmap
2636 because it contains high-valued characters or Unicode property items, e.g.
2637 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2638 pattern, but PCRE should not loop if you do not (it no longer does).
2639 [Detail: two cases were found: (a) a repeated subpattern containing an
2640 extended class; (b) a recursive reference to a subpattern that followed a
2641 previous extended class. It wasn't skipping over the extended class
2642 correctly when UTF-8 mode was not set.]
2644 5. A negated single-character class was not being recognized as fixed-length
2645 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2646 compile error "lookbehind assertion is not fixed length".
2648 6. The RunPerlTest auxiliary script was showing an unexpected difference
2649 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2650 write a Perl script that can interpret lines of an input file either as
2651 byte characters or as UTF-8, which is what "perltest" was being required to
2652 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2653 can't do is switch easily at run time between having the "use utf8;" pragma
2654 or not. In the end, I fudged it by using the RunPerlTest script to insert
2655 "use utf8;" explicitly for the UTF-8 tests.
2657 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2658 the end of the subject string, contrary to the documentation and to what
2659 Perl does. This was true of both matching functions. Now it matches only at
2660 the start of the subject and immediately after *internal* newlines.
2662 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2663 a pointer to an int instead of a pointer to an unsigned long int. This
2664 caused problems on 64-bit systems.
2666 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2667 instance of the 'standard' template library not being so standard".
2669 10. There was no check on the number of named subpatterns nor the maximum
2670 length of a subpattern name. The product of these values is used to compute
2671 the size of the memory block for a compiled pattern. By supplying a very
2672 long subpattern name and a large number of named subpatterns, the size
2673 computation could be caused to overflow. This is now prevented by limiting
2674 the length of names to 32 characters, and the number of named subpatterns
2675 to 10,000.
2677 11. Subpatterns that are repeated with specific counts have to be replicated in
2678 the compiled pattern. The size of memory for this was computed from the
2679 length of the subpattern and the repeat count. The latter is limited to
2680 65535, but there was no limit on the former, meaning that integer overflow
2681 could in principle occur. The compiled length of a repeated subpattern is
2682 now limited to 30,000 bytes in order to prevent this.
2684 12. Added the optional facility to have named substrings with the same name.
2686 13. Added the ability to use a named substring as a condition, using the
2687 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2688 are numbers (not recommended). Forward references are permitted.
2690 14. Added forward references in named backreferences (if you see what I mean).
2692 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2693 pattern could run off the end of the subject. For example, the pattern
2694 "(?s)(.{1,5})"8 did this with the subject "ab".
2696 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2697 PCRE_CASELESS was set when matching characters that were quantified with ?
2698 or *.
2700 17. A character class other than a single negated character that had a minimum
2701 but no maximum quantifier - for example [ab]{6,} - was not handled
2702 correctly by pce_dfa_exec(). It would match only one character.
2704 18. A valid (though odd) pattern that looked like a POSIX character
2705 class but used an invalid character after [ (for example [[,abc,]]) caused
2706 pcre_compile() to give the error "Failed: internal error: code overflow" or
2707 in some cases to crash with a glibc free() error. This could even happen if
2708 the pattern terminated after [[ but there just happened to be a sequence of
2709 letters, a binary zero, and a closing ] in the memory that followed.
2711 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2712 over the years. Originally (before any Unicode support), just the bottom 8
2713 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2714 output from "man perlunicode" includes this:
2716 The regular expression compiler produces polymorphic opcodes. That
2717 is, the pattern adapts to the data and automatically switches to
2718 the Unicode character scheme when presented with Unicode data--or
2719 instead uses a traditional byte scheme when presented with byte
2720 data.
2722 Sadly, a wide octal escape does not cause a switch, and in a string with
2723 no other multibyte characters, these octal escapes are treated as before.
2724 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2725 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2726 Unicode string.
2728 I have not perpetrated such confusion in PCRE. Up till now, it took just
2729 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2730 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2731 translate to the appropriate multibyte character.
2733 29. Applied some refactoring to reduce the number of warnings from Microsoft
2734 and Borland compilers. This has included removing the fudge introduced
2735 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2736 a warning about an unused variable.
2738 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2739 characters since release 4.0, because Perl (from release 5.004) does not.
2740 [Or at least, is documented not to: some releases seem to be in conflict
2741 with the documentation.] However, when a pattern was studied with
2742 pcre_study() and all its branches started with \s, PCRE still included VT
2743 as a possible starting character. Of course, this did no harm; it just
2744 caused an unnecessary match attempt.
2746 22. Removed a now-redundant internal flag bit that recorded the fact that case
2747 dependency changed within the pattern. This was once needed for "required
2748 byte" processing, but is no longer used. This recovers a now-scarce options
2749 bit. Also moved the least significant internal flag bit to the most-
2750 significant bit of the word, which was not previously used (hangover from
2751 the days when it was an int rather than a uint) to free up another bit for
2752 the future.
2754 23. Added support for CRLF line endings as well as CR and LF. As well as the
2755 default being selectable at build time, it can now be changed at runtime
2756 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2757 specify that it is scanning data with non-default line endings.
2759 24. Changed the definition of CXXLINK to make it agree with the definition of
2760 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2762 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2763 recursions. This makes a big different to stack usage for some patterns.
2765 26. If a subpattern containing a named recursion or subroutine reference such
2766 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2767 the space required for the compiled pattern went wrong and gave too small a
2768 value. Depending on the environment, this could lead to "Failed: internal
2769 error: code overflow at offset 49" or "glibc detected double free or
2770 corruption" errors.
2772 27. Applied patches from Google (a) to support the new newline modes and (b) to
2773 advance over multibyte UTF-8 characters in GlobalReplace.
2775 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2776 difference for some implementation of PCRE in some Windows version.
2778 29. Added some extra testing facilities to pcretest:
2780 \q<number> in a data line sets the "match limit" value
2781 \Q<number> in a data line sets the "match recursion limt" value
2782 -S <number> sets the stack size, where <number> is in megabytes
2784 The -S option isn't available for Windows.
2787 Version 6.6 06-Feb-06
2788 ---------------------
2790 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2791 in pcreposix.h. I have copied the definition from pcre.h.
2793 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2794 because pcre.h is no longer a built file.
2796 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2797 not normally included in the compiled code.
2800 Version 6.5 01-Feb-06
2801 ---------------------
2803 1. When using the partial match feature with pcre_dfa_exec(), it was not
2804 anchoring the second and subsequent partial matches at the new starting
2805 point. This could lead to incorrect results. For example, with the pattern
2806 /1234/, partially matching against "123" and then "a4" gave a match.
2808 2. Changes to pcregrep:
2810 (a) All non-match returns from pcre_exec() were being treated as failures
2811 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2812 error message is output. Some extra information is given for the
2814 probably the only errors that are likely to be caused by users (by
2815 specifying a regex that has nested indefinite repeats, for instance).
2816 If there are more than 20 of these errors, pcregrep is abandoned.
2818 (b) A binary zero was treated as data while matching, but terminated the
2819 output line if it was written out. This has been fixed: binary zeroes
2820 are now no different to any other data bytes.
2822 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2823 used to set a locale for matching. The --locale=xxxx long option has
2824 been added (no short equivalent) to specify a locale explicitly on the
2825 pcregrep command, overriding the environment variables.
2827 (d) When -B was used with -n, some line numbers in the output were one less
2828 than they should have been.
2830 (e) Added the -o (--only-matching) option.
2832 (f) If -A or -C was used with -c (count only), some lines of context were
2833 accidentally printed for the final match.
2835 (g) Added the -H (--with-filename) option.
2837 (h) The combination of options -rh failed to suppress file names for files
2838 that were found from directory arguments.
2840 (i) Added the -D (--devices) and -d (--directories) options.
2842 (j) Added the -F (--fixed-strings) option.
2844 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2846 (l) Added the --colo(u)r option.
2848 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2849 is not present by default.
2851 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2852 items such as (?R) or (?1), when the recursion could match a number of
2853 alternatives. If it matched one of the alternatives, but subsequently,
2854 outside the recursion, there was a failure, the code tried to back up into
2855 the recursion. However, because of the way PCRE is implemented, this is not
2856 possible, and the result was an incorrect result from the match.
2858 In order to prevent this happening, the specification of recursion has
2859 been changed so that all such subpatterns are automatically treated as
2860 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2862 4. I had overlooked the fact that, in some locales, there are characters for
2863 which isalpha() is true but neither isupper() nor islower() are true. In
2864 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2865 and ordfeminine) are like this. This affected the treatment of \w and \W
2866 when they appeared in character classes, but not when they appeared outside
2867 a character class. The bit map for "word" characters is now created
2868 separately from the results of isalnum() instead of just taking it from the
2869 upper, lower, and digit maps. (Plus the underscore character, of course.)
2871 5. The above bug also affected the handling of POSIX character classes such as
2872 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2873 permanent tables. Instead, the bit maps for such a class were previously
2874 created as the appropriate unions of the upper, lower, and digit bitmaps.
2875 Now they are created by subtraction from the [[:word:]] class, which has
2876 its own bitmap.
2878 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2879 It is created by subtracting the vertical space characters (\x09, \x0a,
2880 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2881 subtraction was done in the overall bitmap for a character class, meaning
2882 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2883 be recognized. This bug has been fixed.
2885 7. Patches from the folks at Google:
2887 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2888 real life, but is still worth protecting against".
2890 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2891 regular expressions".
2893 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2894 have it.
2896 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2897 "configure" and the latter not, in order to fix a problem somebody had
2898 with compiling the Arg class on HP-UX.
2900 (e) Improve the error-handling of the C++ wrapper a little bit.
2902 (f) New tests for checking recursion limiting.
2904 8. The pcre_memmove() function, which is used only if the environment does not
2905 have a standard memmove() function (and is therefore rarely compiled),
2906 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2907 returning a result (though PCRE never actually uses the result).
2909 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2910 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2911 returned instead of calling malloc() with an overflowing number that would
2912 most likely cause subsequent chaos.
2914 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2916 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2917 with this option is matched, the nmatch and pmatch options of regexec() are
2918 ignored.
2920 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2921 provided in case anyone wants to the the POSIX interface with UTF-8
2922 strings.
2924 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2925 C++ linking (needed for some HP-UX environments).
2927 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2928 (unused parameter) and in the pcre_printint() function (omitted "default"
2929 switch label when the default is to do nothing).
2931 15. Added some code to make it possible, when PCRE is compiled as a C++
2932 library, to replace subject pointers for pcre_exec() with a smart pointer
2933 class, thus making it possible to process discontinuous strings.
2935 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2936 much the same function. They were added by different people who were trying
2937 to make PCRE easy to compile on non-Unix systems. It has been suggested
2938 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2939 for compiling on Windows systems. I have therefore replaced it with
2940 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2941 defaults to "extern" for C or "extern C" for C++, which works fine on
2942 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2943 SCOPE with something explicit in config.h. In addition:
2945 (a) pcreposix.h still had just "extern" instead of either of these macros;
2946 I have replaced it with PCRE_DATA_SCOPE.
2948 (b) Functions such as _pcre_xclass(), which are internal to the library,
2949 but external in the C sense, all had PCRE_EXPORT in their definitions.
2950 This is apparently wrong for the Windows case, so I have removed it.
2951 (It makes no difference on Unix-like systems.)
2953 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2954 of recursive calls to match(). This is different to MATCH_LIMIT because
2955 that limits the total number of calls to match(), not all of which increase
2956 the depth of recursion. Limiting the recursion depth limits the amount of
2957 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2958 when PCRE is compiled, and changed at run time. A patch from Google adds
2959 this functionality to the C++ interface.
2961 18. Changes to the handling of Unicode character properties:
2963 (a) Updated the table to Unicode 4.1.0.
2965 (b) Recognize characters that are not in the table as "Cn" (undefined).
2967 (c) I revised the way the table is implemented to a much improved format
2968 which includes recognition of ranges. It now supports the ranges that
2969 are defined in UnicodeData.txt, and it also amalgamates other
2970 characters into ranges. This has reduced the number of entries in the
2971 table from around 16,000 to around 3,000, thus reducing its size
2972 considerably. I realized I did not need to use a tree structure after
2973 all - a binary chop search is just as efficient. Having reduced the
2974 number of entries, I extended their size from 6 bytes to 8 bytes to
2975 allow for more data.
2977 (d) Added support for Unicode script names via properties such as \p{Han}.
2979 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2980 matching that character.
2982 20. When matching a repeated Unicode property with a minimum greater than zero,
2983 (for example \pL{2,}), PCRE could look past the end of the subject if it
2984 reached it while seeking the minimum number of characters. This could
2985 happen only if some of the characters were more than one byte long, because
2986 there is a check for at least the minimum number of bytes.
2988 21. Refactored the implementation of \p and \P so as to be more general, to
2989 allow for more different types of property in future. This has changed the
2990 compiled form incompatibly. Anybody with saved compiled patterns that use
2991 \p or \P will have to recompile them.
2993 22. Added "Any" and "L&" to the supported property types.
2995 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2996 but give a compile time error if the value is greater than 0xff.
2998 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2999 accidentally not being installed or uninstalled.
3001 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3002 made were to insert the current release number. This seemed silly, because
3003 it made things harder for people building PCRE on systems that don't run
3004 "configure". I have turned pcre.h into a distributed file, no longer built
3005 by "configure", with the version identification directly included. There is
3006 no longer a pcre.h.in file.
3008 However, this change necessitated a change to the pcre-config script as
3009 well. It is built from pcre-config.in, and one of the substitutions was the
3010 release number. I have updated configure.ac so that ./configure now finds
3011 the release number by grepping pcre.h.
3013 26. Added the ability to run the tests under valgrind.
3016 Version 6.4 05-Sep-05
3017 ---------------------
3019 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3020 "--" to be printed when multiple files were scanned, even when none of the
3021 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3022 consider it to be a bug, and have restored the previous behaviour.
3024 2. A couple of code tidies to get rid of compiler warnings.
3026 3. The pcretest program used to cheat by referring to symbols in the library
3027 whose names begin with _pcre_. These are internal symbols that are not
3028 really supposed to be visible externally, and in some environments it is
3029 possible to suppress them. The cheating is now confined to including
3030 certain files from the library's source, which is a bit cleaner.
3032 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3033 file's purpose clearer.
3035 5. Reorganized pcre_ucp_findchar().
3038 Version 6.3 15-Aug-05
3039 ---------------------
3041 1. The file libpcre.pc.in did not have general read permission in the tarball.
3043 2. There were some problems when building without C++ support:
3045 (a) If C++ support was not built, "make install" and "make test" still
3046 tried to test it.
3048 (b) There were problems when the value of CXX was explicitly set. Some
3049 changes have been made to try to fix these, and ...
3051 (c) --disable-cpp can now be used to explicitly disable C++ support.
3053 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3054 backslash in a target when C++ was disabled. This confuses some
3055 versions of "make", apparently. Using an intermediate variable solves
3056 this. (Same for CPP_LOBJ.)
3058 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3059 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3060 necessary on certain architectures.
3062 4. Added a setting of -export-symbols-regex to the link command to remove
3063 those symbols that are exported in the C sense, but actually are local
3064 within the library, and not documented. Their names all begin with
3065 "_pcre_". This is not a perfect job, because (a) we have to except some
3066 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3067 available (and never for static libraries). I have made a note to try to
3068 find a way round (a) in the future.
3071 Version 6.2 01-Aug-05
3072 ---------------------
3074 1. There was no test for integer overflow of quantifier values. A construction
3075 such as {1111111111111111} would give undefined results. What is worse, if
3076 a minimum quantifier for a parenthesized subpattern overflowed and became
3077 negative, the calculation of the memory size went wrong. This could have
3078 led to memory overwriting.
3080 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3082 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3083 operating environments where this matters.
3085 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3086 PCRE options from within the C++ wrapper.
3088 5. Named capturing subpatterns were not being correctly counted when a pattern
3089 was compiled. This caused two problems: (a) If there were more than 100
3090 such subpatterns, the calculation of the memory needed for the whole
3091 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3092 back references of the form \12, where the number was greater than 9, were
3093 not recognized as back references, even though there were sufficient
3094 previous subpatterns.
3096 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3097 versions of gcc, e.g. 2.95.4.
3100 Version 6.1 21-Jun-05
3101 ---------------------
3103 1. There was one reference to the variable "posix" in pcretest.c that was not
3104 surrounded by "#if !defined NOPOSIX".
3106 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3107 the cross-check on the old pcre_info() function, for the benefit of the
3108 cut-down version of PCRE that is currently imported into Exim.
3110 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3111 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3112 patterns that start (?i) but I don't think that matters. The use of (?i) is
3113 just an example; this all applies to the other options as well.
3115 4. Since libtool seems to echo the compile commands it is issuing, the output
3116 from "make" can be reduced a bit by putting "@" in front of each libtool
3117 compile command.
3119 5. Patch from the folks at Google for configure.in to be a bit more thorough
3120 in checking for a suitable C++ installation before trying to compile the
3121 C++ stuff. This should fix a reported problem when a compiler was present,
3122 but no suitable headers.
3124 6. The man pages all had just "PCRE" as their title. I have changed them to
3125 be the relevant file name. I have also arranged that these names are
3126 retained in the file doc/pcre.txt, which is a concatenation in text format
3127 of all the man pages except the little individual ones for each function.
3129 7. The NON-UNIX-USE file had not been updated for the different set of source
3130 files that come with release 6. I also added a few comments about the C++
3131 wrapper.
3134 Version 6.0 07-Jun-05
3135 ---------------------
3137 1. Some minor internal re-organization to help with my DFA experiments.
3139 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3140 didn't matter for the library itself when fully configured, but did matter
3141 when compiling without UCP support, or within Exim, where the ucp files are
3142 not imported.
3144 3. Refactoring of the library code to split up the various functions into
3145 different source modules. The addition of the new DFA matching code (see
3146 below) to a single monolithic source would have made it really too
3147 unwieldy, quite apart from causing all the code to be include in a
3148 statically linked application, when only some functions are used. This is
3149 relevant even without the DFA addition now that patterns can be compiled in
3150 one application and matched in another.
3152 The downside of splitting up is that there have to be some external
3153 functions and data tables that are used internally in different modules of
3154 the library but which are not part of the API. These have all had their
3155 names changed to start with "_pcre_" so that they are unlikely to clash
3156 with other external names.
3158 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3159 a different (DFA) algorithm. Although it is slower than the original
3160 function, it does have some advantages for certain types of matching
3161 problem.
3163 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3164 including restarting after a partial match.
3166 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3167 defined when compiling for Windows was sent to me. I have put it into the
3168 code, though I have no means of testing or verifying it.
3170 7. Added the pcre_refcount() auxiliary function.
3172 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3173 match before or at the first newline in the subject string. In pcretest,
3174 the /f option on a pattern can be used to set this.
3176 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3177 would behave wrongly. This has been present in PCRE since release 4.0.
3179 10. A number of changes to the pcregrep command:
3181 (a) Refactored how -x works; insert ^(...)$ instead of setting
3182 PCRE_ANCHORED and checking the length, in preparation for adding
3183 something similar for -w.
3185 (b) Added the -w (match as a word) option.
3187 (c) Refactored the way lines are read and buffered so as to have more
3188 than one at a time available.
3190 (d) Implemented a pcregrep test script.
3192 (e) Added the -M (multiline match) option. This allows patterns to match
3193 over several lines of the subject. The buffering ensures that at least
3194 8K, or the rest of the document (whichever is the shorter) is available
3195 for matching (and similarly the previous 8K for lookbehind assertions).
3197 (f) Changed the --help output so that it now says
3199 -w, --word-regex(p)
3201 instead of two lines, one with "regex" and the other with "regexp"
3202 because that confused at least one person since the short forms are the
3203 same. (This required a bit of code, as the output is generated
3204 automatically from a table. It wasn't just a text change.)
3206 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3207 option but starts with a hyphen. Could be a pattern or a path name
3208 starting with a hyphen, for instance.
3210 (h) "-" can be given as a file name to represent stdin.
3212 (i) When file names are being printed, "(standard input)" is used for
3213 the standard input, for compatibility with GNU grep. Previously
3214 "<stdin>" was used.
3216 (j) The option --label=xxx can be used to supply a name to be used for
3217 stdin when file names are being printed. There is no short form.
3219 (k) Re-factored the options decoding logic because we are going to add
3220 two more options that take data. Such options can now be given in four
3221 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3223 (l) Added the -A, -B, and -C options for requesting that lines of context
3224 around matches be printed.
3226 (m) Added the -L option to print the names of files that do not contain
3227 any matching lines, that is, the complement of -l.
3229 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3230 continue to scan other files.
3232 (o) The -s option was incorrectly implemented. For compatibility with other
3233 greps, it now suppresses the error message for a non-existent or non-
3234 accessible file (but not the return code). There is a new option called
3235 -q that suppresses the output of matching lines, which was what -s was
3236 previously doing.
3238 (p) Added --include and --exclude options to specify files for inclusion
3239 and exclusion when recursing.
3241 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3242 Hopefully, it now does.
3244 12. Missing cast in pcre_study().
3246 13. Added an "uninstall" target to the makefile.
3248 14. Replaced "extern" in the function prototypes in Makefile.in with
3249 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3250 world, but is set differently for Windows.
3252 15. Added a second compiling function called pcre_compile2(). The only
3253 difference is that it has an extra argument, which is a pointer to an
3254 integer error code. When there is a compile-time failure, this is set
3255 non-zero, in addition to the error test pointer being set to point to an
3256 error message. The new argument may be NULL if no error number is required
3257 (but then you may as well call pcre_compile(), which is now just a
3258 wrapper). This facility is provided because some applications need a
3259 numeric error indication, but it has also enabled me to tidy up the way
3260 compile-time errors are handled in the POSIX wrapper.
3262 16. Added VPATH=.libs to the makefile; this should help when building with one
3263 prefix path and installing with another. (Or so I'm told by someone who
3264 knows more about this stuff than I do.)
3266 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3267 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3268 match everything, including newlines. This is not POSIX-compatible, but
3269 somebody wanted the feature. From pcretest it can be activated by using
3270 both the P and the s flags.
3272 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3274 19. libpcre.pc was being incorrectly installed as executable.
3276 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3277 it now also looks for '\r' so that it will work unmodified on Windows.
3279 21. Added Google's contributed C++ wrapper to the distribution.
3281 22. Added some untidy missing memory free() calls in pcretest, to keep
3282 Electric Fence happy when testing.
3286 Version 5.0 13-Sep-04
3287 ---------------------
3289 1. Internal change: literal characters are no longer packed up into items
3290 containing multiple characters in a single byte-string. Each character
3291 is now matched using a separate opcode. However, there may be more than one
3292 byte in the character in UTF-8 mode.
3294 2. The pcre_callout_block structure has two new fields: pattern_position and
3295 next_item_length. These contain the offset in the pattern to the next match
3296 item, and its length, respectively.
3298 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3299 insertion of callouts before each pattern item. Added the /C option to
3300 pcretest to make use of this.
3302 4. On the advice of a Windows user, the lines
3304 #if defined(_WIN32) || defined(WIN32)
3305 _setmode( _fileno( stdout ), 0x8000 );
3306 #endif /* defined(_WIN32) || defined(WIN32) */
3308 have been added to the source of pcretest. This apparently does useful
3309 magic in relation to line terminators.
3311 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3312 for the benefit of those environments where the "b" makes a difference.
3314 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3315 to know about it. I have put a hack into configure.in that adds in code
3316 to set GCC=yes if CC=icc. This seems to end up at a point in the
3317 generated configure script that is early enough to affect the setting of
3318 compiler options, which is what is needed, but I have no means of testing
3319 whether it really works. (The user who reported this had patched the
3320 generated configure script, which of course I cannot do.)
3322 LATER: After change 22 below (new libtool files), the configure script
3323 seems to know about icc (and also ecc). Therefore, I have commented out
3324 this hack in configure.in.
3326 7. Added support for pkg-config (2 patches were sent in).
3328 8. Negated POSIX character classes that used a combination of internal tables
3329 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3330 [[:^ascii]]. Typically, they would match almost any characters. The other
3331 POSIX classes were not broken in this way.
3333 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3334 to find the match, as PCRE was deluded into thinking that the match had to
3335 start at the start point or following a newline. The same bug applied to
3336 patterns with negative forward assertions or any backward assertions
3337 preceding ".*" at the start, unless the pattern required a fixed first
3338 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3340 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3341 starting at the last subject character, bytes beyond the end of the subject
3342 string were read.
3344 11. Renamed the variable "class" as "classbits" to make life easier for C++
3345 users. (Previously there was a macro definition, but it apparently wasn't
3346 enough.)
3348 12. Added the new field "tables" to the extra data so that tables can be passed
3349 in at exec time, or the internal tables can be re-selected. This allows
3350 a compiled regex to be saved and re-used at a later time by a different
3351 program that might have everything at different addresses.
3353 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3354 -R library as well as a -L library.
3356 14. The debugging options of pcretest (-d on the command line or D on a
3357 pattern) showed incorrect output for anything following an extended class
3358 that contained multibyte characters and which was followed by a quantifier.
3360 15. Added optional support for general category Unicode character properties
3361 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3362 support. It adds about 90K to the size of the library. The meanings of the
3363 inbuilt class escapes such as \d and \s have NOT been changed.
3365 16. Updated pcredemo.c to include calls to free() to release the memory for the
3366 compiled pattern.
3368 17. The generated file chartables.c was being created in the source directory
3369 instead of in the building directory. This caused the build to fail if the
3370 source directory was different from the building directory, and was
3371 read-only.
3373 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3374 file. No doubt somebody will tell me if they don't make sense... Also added
3375 Dan Mooney's comments about building on OpenVMS.
3377 19. Added support for partial matching via the PCRE_PARTIAL option for
3378 pcre_exec() and the \P data escape in pcretest.
3380 20. Extended pcretest with 3 new pattern features:
3382 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3383 write the compiled pattern to the file whose name is "rest-of-line".
3384 This is a straight binary dump of the data, with the saved pointer to
3385 the character tables forced to be NULL. The study data, if any, is
3386 written too. After writing, pcretest reads a new pattern.
3388 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3389 compiled pattern from the given file. There must not be any
3390 occurrences of "<" in the file name (pretty unlikely); if there are,
3391 pcretest will instead treat the initial "<" as a pattern delimiter.
3392 After reading in the pattern, pcretest goes on to read data lines as
3393 usual.
3395 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3396 and 16-bit fields in a compiled pattern, to simulate a pattern that
3397 was compiled on a host of opposite endianness.
3399 21. The pcre-exec() function can now cope with patterns that were compiled on
3400 hosts of opposite endianness, with this restriction:
3402 As for any compiled expression that is saved and used later, the tables
3403 pointer field cannot be preserved; the extra_data field in the arguments
3404 to pcre_exec() should be used to pass in a tables address if a value
3405 other than the default internal tables were used at compile time.
3407 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3408 now diagnosed as an error. Previously, most of the time, a negative number
3409 would have been treated as zero, but if in addition "ovector" was passed as
3410 NULL, a crash could occur.
3412 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3413 new versions from the libtool 1.5 distribution (the last one is a copy of
3414 a file called libtool.m4). This seems to have fixed the need to patch
3415 "configure" to support Darwin 1.3 (which I used to do). However, I still
3416 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3417 workstation).
3419 24. Changed the PCRE licence to be the more standard "BSD" licence.
3422 Version 4.5 01-Dec-03
3423 ---------------------
3425 1. There has been some re-arrangement of the code for the match() function so
3426 that it can be compiled in a version that does not call itself recursively.
3427 Instead, it keeps those local variables that need separate instances for
3428 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3429 needs to "recurse". Keeping track of where control must go is done by means
3430 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3431 hide most of the details from the main code, and operates only if
3432 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3433 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3434 operating.
3436 To make it easier for callers to provide specially tailored get/free
3437 functions for this usage, two new functions, pcre_stack_malloc, and
3438 pcre_stack_free, are used. They are always called in strict stacking order,
3439 and the size of block requested is always the same.
3441 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3442 PCRE has been compiled to use the stack or the heap for recursion. The
3443 -C option of pcretest uses this to show which version is compiled.
3445 A new data escape \S, is added to pcretest; it causes the amounts of store
3446 obtained and freed by both kinds of malloc/free at match time to be added
3447 to the output.
3449 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3450 what's available on my current Linux desktop machine.
3452 3. When matching a UTF-8 string, the test for a valid string at the start has
3453 been extended. If start_offset is not zero, PCRE now checks that it points
3454 to a byte that is the start of a UTF-8 character. If not, it returns
3455 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3456 this is necessary because there may be backward assertions in the pattern.
3457 When matching the same subject several times, it may save resources to use
3458 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3460 4. The code for checking the validity of UTF-8 strings has been tightened so
3461 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3462 containing "overlong sequences".
3464 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3465 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3466 should have been "&", but it just so happened that all the cases this let
3467 through by mistake were picked up later in the function.
3469 6. I had used a variable called "isblank" - this is a C99 function, causing
3470 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3472 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3473 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3474 so the version is shown; (c) stop "make test" echoing "./RunTest".
3476 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3478 9. The prototype for memmove() for systems that don't have it was using
3479 size_t, but the inclusion of the header that defines size_t was later. I've
3480 moved the #includes for the C headers earlier to avoid this.
3482 10. Added some adjustments to the code to make it easier to compiler on certain
3483 special systems:
3485 (a) Some "const" qualifiers were missing.
3486 (b) Added the macro EXPORT before all exported functions; by default this
3487 is defined to be empty.
3488 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3489 that it reads its output file name as an argument instead of writing
3490 to the standard output and assuming this can be redirected.
3492 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3493 class containing characters with values greater than 255, PCRE compilation
3494 went into a loop.
3496 12. A recursive reference to a subpattern that was within another subpattern
3497 that had a minimum quantifier of zero caused PCRE to crash. For example,
3498 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3499 recursion. If the recursively-called subpattern itself had a zero repeat,
3500 that was OK.
3502 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3503 buffer into which it was copied (for escape processing) was still set at
3504 1024, so long lines caused crashes.
3506 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3507 "internal error: code overflow...". This applied to any character class
3508 that was followed by a possessive quantifier.
3510 15. Modified the Makefile to add libpcre.la as a prerequisite for
3511 libpcreposix.la because I was told this is needed for a parallel build to
3512 work.
3514 16. If a pattern that contained .* following optional items at the start was
3515 studied, the wrong optimizing data was generated, leading to matching
3516 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3517 matching string must start with a or b or c. The correct conclusion for
3518 this pattern is that a match can start with any character.
3521 Version 4.4 13-Aug-03
3522 ---------------------
3524 1. In UTF-8 mode, a character class containing characters with values between
3525 127 and 255 was not handled correctly if the compiled pattern was studied.
3526 In fixing this, I have also improved the studying algorithm for such
3527 classes (slightly).
3529 2. Three internal functions had redundant arguments passed to them. Removal
3530 might give a very teeny performance improvement.
3532 3. Documentation bug: the value of the capture_top field in a callout is *one
3533 more than* the number of the hightest numbered captured substring.
3535 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3536 in incorrectly linking with a previously installed version. They now link
3537 explicitly with libpcre.la.
3539 5. configure.in no longer needs to recognize Cygwin specially.
3541 6. A problem in pcre.in for Windows platforms is fixed.
3543 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3544 pcretest, it used to include the size of the study block as part of its
3545 output. Unfortunately, the structure contains a field that has a different
3546 size on different hardware architectures. This meant that the tests that
3547 showed this size failed. As the block is currently always of a fixed size,
3548 this information isn't actually particularly useful in pcretest output, so
3549 I have just removed it.
3551 8. Three pre-processor statements accidentally did not start in column 1.
3552 Sadly, there are *still* compilers around that complain, even though
3553 standard C has not required this for well over a decade. Sigh.
3555 9. In pcretest, the code for checking callouts passed small integers in the
3556 callout_data field, which is a void * field. However, some picky compilers
3557 complained about the casts involved for this on 64-bit systems. Now
3558 pcretest passes the address of the small integer instead, which should get
3559 rid of the warnings.
3561 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3562 both compile and run time, and gives an error if an invalid UTF-8 sequence
3563 is found. There is a option for disabling this check in cases where the
3564 string is known to be correct and/or the maximum performance is wanted.
3566 11. In response to a bug report, I changed one line in Makefile.in from
3568 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3569 to
3570 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3572 to look similar to other lines, but I have no way of telling whether this
3573 is the right thing to do, as I do not use Windows. No doubt I'll get told
3574 if it's wrong...
3577 Version 4.3 21-May-03
3578 ---------------------
3580 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3581 Makefile.
3583 2. Some refactoring to improve the quality of the code:
3585 (i) The utf8_table... variables are now declared "const".
3587 (ii) The code for \cx, which used the "case flipping" table to upper case
3588 lower case letters, now just substracts 32. This is ASCII-specific,
3589 but the whole concept of \cx is ASCII-specific, so it seems
3590 reasonable.
3592 (iii) PCRE was using its character types table to recognize decimal and
3593 hexadecimal digits in the pattern. This is silly, because it handles
3594 only 0-9, a-f, and A-F, but the character types table is locale-
3595 specific, which means strange things might happen. A private
3596 table is now used for this - though it costs 256 bytes, a table is
3597 much faster than multiple explicit tests. Of course, the standard
3598 character types table is still used for matching digits in subject
3599 strings against \d.
3601 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3602 ending in _t are). So I've renamed it as ESC_tee.
3604 3. The first argument for regexec() in the POSIX wrapper should have been
3605 defined as "const".
3607 4. Changed pcretest to use malloc() for its buffers so that they can be
3608 Electric Fenced for debugging.
3610 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3611 to read one or more bytes before the start of the subject string. Often this
3612 had no effect on PCRE's behaviour, but in some circumstances it could
3613 provoke a segmentation fault.
3615 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3616 to try to read one or more bytes before the start of the subject string.
3618 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3619 UTF-8 support could misbehave in various ways if the subject string
3620 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3621 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3622 back over UTF-8 characters.)
3625 Version 4.2 14-Apr-03
3626 ---------------------
3628 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3630 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3631 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3632 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3633 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3634 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3636 Note: automatic setting of the BUILD variables is not yet working
3637 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3638 compile-time but not at link-time
3639 [LINK]: use for linking executables only
3640 make different versions for Windows and non-Windows
3641 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3642 libraries
3643 [LINK_FOR_BUILD]: new variable
3644 [OBJEXT]: use throughout
3645 [EXEEXT]: use throughout
3646 <winshared>: new target
3647 <wininstall>: new target
3648 <dftables.o>: use native compiler
3649 <dftables>: use native linker
3650 <install>: handle Windows platform correctly
3651 <clean>: ditto
3652 <check>: ditto
3653 copy DLL to top builddir before testing
3655 As part of these changes, -no-undefined was removed again. This was reported
3656 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3657 in any case.
3659 3. Some tidies to get rid of compiler warnings:
3661 . In the match_data structure, match_limit was an unsigned long int, whereas
3662 match_call_count was an int. I've made them both unsigned long ints.
3664 . In pcretest the fact that a const uschar * doesn't automatically cast to
3665 a void * provoked a warning.
3667 . Turning on some more compiler warnings threw up some "shadow" variables
3668 and a few more missing casts.
3670 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3671 option, a class that contained a single character with a value between 128
3672 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3674 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3675 option, a class that contained several characters, but with at least one
3676 whose value was between 128 and 255 caused PCRE to crash.
3679 Version 4.1 12-Mar-03
3680 ---------------------
3682 1. Compiling with gcc -pedantic found a couple of places where casts were
3683 needed, and a string in dftables.c that was longer than standard compilers are
3684 required to support.
3686 2. Compiling with Sun's compiler found a few more places where the code could
3687 be tidied up in order to avoid warnings.
3689 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3690 first of these names is deprecated in the latest Autoconf in favour of the name
3691 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3692 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3693 analogy I have changed it to CFLAGS_FOR_BUILD.
3695 4. Added -no-undefined to the linking command in the Makefile, because this is
3696 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3697 linking step for the pcreposix library.
3699 5. PCRE was failing to diagnose the case of two named groups with the same
3700 name.
3702 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3703 literal character that is needed in the subject for a match, and scans along to
3704 ensure that it is present before embarking on the full matching process. This
3705 saves time in cases of nested unlimited repeats that are never going to match.
3706 Problem: the scan can take a lot of time if the subject is very long (e.g.
3707 megabytes), thus penalizing straightforward matches. It is now done only if the
3708 amount of subject to be scanned is less than 1000 bytes.
3710 7. A lesser problem with the same optimization is that it was recording the
3711 first character of an anchored pattern as "needed", thus provoking a search
3712 right along the subject, even when the first match of the pattern was going to
3713 fail. The "needed" character is now not set for anchored patterns, unless it
3714 follows something in the pattern that is of non-fixed length. Thus, it still
3715 fulfils its original purpose of finding quick non-matches in cases of nested
3716 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3719 Version 4.0 17-Feb-03
3720 ---------------------
3722 1. If a comment in an extended regex that started immediately after a meta-item
3723 extended to the end of string, PCRE compiled incorrect data. This could lead to
3724 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3726 2. Moved to autoconf 2.53 and libtool 1.4.2.
3728 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3729 the special perltest8 script is no longer needed - all the tests can be run
3730 from a single perltest script.
3732 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3733 by \s. It has now been removed in PCRE. This means it isn't recognized as
3734 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3735 class [:space:] *does* include VT, thereby creating a mess.
3737 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3738 space and tab.
3740 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3741 its new features into the main test script, reducing the number of scripts.
3743 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3744 were backward compatible, and made the (?i) apply to the whole pattern, as if
3745 /i were given. Now it behaves more logically, and applies the option setting
3746 only to what follows. PCRE has been changed to follow suit. However, if it
3747 finds options settings right at the start of the pattern, it extracts them into
3748 the global options, as before. Thus, they show up in the info data.
3750 8. Added support for the \Q...\E escape sequence. Characters in between are
3751 treated as literals. This is slightly different from Perl in that $ and @ are
3752 also handled as literals inside the quotes. In Perl, they will cause variable
3753 interpolation. Note the following examples:
3755 Pattern PCRE matches Perl matches
3757 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3758 \Qabc\$xyz\E abc\$xyz abc\$xyz
3759 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3761 For compatibility with Perl, \Q...\E sequences are recognized inside character
3762 classes as well as outside them.
3764 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3765 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3766 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3767 signed/unsigned warnings.
3769 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3770 option for pcretest, so I've replaced it by a simple function that does just
3771 that job.
3773 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3774 "pcregrep -".
3776 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3777 Java package. This provides some syntactic sugar for simple cases of what my
3778 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3779 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3780 item, you can use this simplified notation. Note that only makes sense with
3781 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3782 greediness, whatever the setting of the PCRE_UNGREEDY option.
3784 13. A change of greediness default within a pattern was not taking effect at
3785 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3786 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3787 was abstracted outside.
3789 14. PCRE now supports the \G assertion. It is true when the current matching
3790 position is at the start point of the match. This differs from \A when the
3791 starting offset is non-zero. Used with the /g option of pcretest (or similar
3792 code), it works in the same way as it does for Perl's /g option. If all
3793 alternatives of a regex begin with \G, the expression is anchored to the start
3794 match position, and the "anchored" flag is set in the compiled expression.
3796 15. Some bugs concerning the handling of certain option changes within patterns
3797 have been fixed. These applied to options other than (?ims). For example,
3798 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3799 been the other way round. Some of this was related to change 7 above.
3801 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3802 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3803 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3804 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3806 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3807 mode. Unlike ".", it always matches newline, whatever the setting of
3808 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3809 assertions. Perl allows it, but it doesn't (in general) work because it can't
3810 calculate the length of the lookbehind. At least, that's the case for Perl
3811 5.8.0 - I've been told they are going to document that it doesn't work in
3812 future.
3814 18. Added an error diagnosis for escapes that PCRE does not support: these are
3815 \L, \l, \N, \P, \p, \U, \u, and \X.
3817 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3818 reading past the end of the pattern in cases such as /[abcd/.
3820 20. PCRE was getting more memory than necessary for patterns with classes that
3821 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3823 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3824 compiling PCRE for use with Virtual Pascal.
3826 22. Small fix to the Makefile to make it work properly if the build is done
3827 outside the source tree.
3829 23. Added a new extension: a condition to go with recursion. If a conditional
3830 subpattern starts with (?(R) the "true" branch is used if recursion has
3831 happened, whereas the "false" branch is used only at the top level.
3833 24. When there was a very long string of literal characters (over 255 bytes
3834 without UTF support, over 250 bytes with UTF support), the computation of how
3835 much memory was required could be incorrect, leading to segfaults or other
3836 strange effects.
3838 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3839 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3840 there was a subsequent back reference to those brackets. This meant that, for
3841 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3842 possible to check for precisely this case. All we can do is abandon the
3843 optimization if .* occurs inside capturing brackets when there are any back
3844 references whatsoever. (See below for a better fix that came later.)
3846 26. The handling of the optimization for finding the first character of a
3847 non-anchored pattern, and for finding a character that is required later in the
3848 match were failing in some cases. This didn't break the matching; it just
3849 failed to optimize when it could. The way this is done has been re-implemented.
3851 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3853 28. Added a new feature that provides some of the functionality that Perl
3854 provides with (?{...}). The facility is termed a "callout". The way it is done
3855 in PCRE is for the caller to provide an optional function, by setting
3856 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3857 global variable. By default it is unset, which disables all calling out. To get
3858 the function called, the regex must include (?C) at appropriate points. This
3859 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3860 This provides a means of identifying different callout points. When PCRE
3861 reaches such a point in the regex, if pcre_callout has been set, the external
3862 function is called. It is provided with data in a structure called
3863 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3864 matching continues; if it returns a non-zero value, the match at the current
3865 point fails. However, backtracking will occur if possible. [This was changed
3866 later and other features added - see item 49 below.]
3868 29. pcretest is upgraded to test the callout functionality. It provides a
3869 callout function that displays information. By default, it shows the start of
3870 the match and the current position in the text. There are some new data escapes
3871 to vary what happens:
3873 \C+ in addition, show current contents of captured substrings
3874 \C- do not supply a callout function
3875 \C!n return 1 when callout number n is reached
3876 \C!n!m return 1 when callout number n is reached for the mth time
3878 30. If pcregrep was called with the -l option and just a single file name, it
3879 output "<stdin>" if a match was found, instead of the file name.
3881 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3882 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3883 pcre_exec(). This saves a malloc/free per call. The default value of
3884 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3885 when configuring.
3887 32. The default maximum size of a compiled pattern is 64K. There have been a
3888 few cases of people hitting this limit. The code now uses macros to handle the
3889 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3890 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3891 configuring. Tests 2 and 5 work only with 2-byte links because they output
3892 debugging information about compiled patterns.
3894 33. Internal code re-arrangements:
3896 (a) Moved the debugging function for printing out a compiled regex into
3897 its own source file (printint.c) and used #include to pull it into
3898 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3899 separate copies.
3901 (b) Defined the list of op-code names for debugging as a macro in
3902 internal.h so that it is next to the definition of the opcodes.
3904 (c) Defined a table of op-code lengths for simpler skipping along compiled
3905 code. This is again a macro in internal.h so that it is next to the
3906 definition of the opcodes.
3908 34. Added support for recursive calls to individual subpatterns, along the
3909 lines of Robin Houston's patch (but implemented somewhat differently).
3911 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3912 allow it to read and process whole directories in Win32. This code was
3913 contributed by Lionel Fourquaux; it has not been tested by me.
3915 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3916 used to name a group. Names consist of alphanumerics and underscores, and must
3917 be unique. Back references use the syntax (?P=name) and recursive calls use
3918 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3919 numbers. The function pcre_fullinfo() can be used after compilation to extract
3920 a name/number map. There are three relevant calls:
3922 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3923 PCRE_INFO_NAMECOUNT yields the number of entries
3924 PCRE_INFO_NAMETABLE yields a pointer to the map.
3926 The map is a vector of fixed-size entries. The size of each entry depends on
3927 the length of the longest name used. The first two bytes of each entry are the
3928 group number, most significant byte first. There follows the corresponding
3929 name, zero terminated. The names are in alphabetical order.
3931 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3932 case instead of 255. Making it the same both with and without UTF-8 support
3933 means that the same test output works with both.
3935 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3936 calling malloc() with a zero argument.
3938 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3939 optimization. I've improved things by keeping a bitmap of backreferences with
3940 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3941 fact referenced, the optimization can be applied. It is unlikely that a
3942 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3943 the match to follow \n) will appear inside brackets with a number greater than
3944 31, but if it does, any back reference > 31 suppresses the optimization.
3946 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3947 of disabling numbered capturing parentheses. Any opening parenthesis that is
3948 not followed by ? behaves as if it were followed by ?: but named parentheses
3949 can still be used for capturing (and they will acquire numbers in the usual
3950 way).
3952 41. Redesigned the return codes from the match() function into yes/no/error so
3953 that errors can be passed back from deep inside the nested calls. A malloc
3954 failure while inside a recursive subpattern call now causes the
3955 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3957 42. It is now possible to set a limit on the number of times the match()
3958 function is called in a call to pcre_exec(). This facility makes it possible to
3959 limit the amount of recursion and backtracking, though not in a directly
3960 obvious way, because the match() function is used in a number of different
3961 circumstances. The count starts from zero for each position in the subject
3962 string (for non-anchored patterns). The default limit is, for compatibility, a
3963 large number, namely 10 000 000. You can change this in two ways:
3965 (a) When configuring PCRE before making, you can use --with-match-limit=n
3966 to set a default value for the compiled library.
3968 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3969 a different value is set. See 45 below.
3971 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3973 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3974 of things that can be changed at compile time. The first argument specifies
3975 what is wanted and the second points to where the information is to be placed.
3976 The current list of available information is:
3980 The output is an integer that is set to one if UTF-8 support is available;
3981 otherwise it is set to zero.
3985 The output is an integer that it set to the value of the code that is used for
3986 newline. It is either LF (10) or CR (13).
3990 The output is an integer that contains the number of bytes used for internal
3991 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3995 The output is an integer that contains the threshold above which the POSIX
3996 interface uses malloc() for output vectors. See item 31 above.
4000 The output is an unsigned integer that contains the default limit of the number
4001 of match() calls in a pcre_exec() execution. See 42 above.
4003 44. pcretest has been upgraded by the addition of the -C option. This causes it
4004 to extract all the available output from the new pcre_config() function, and to
4005 output it. The program then exits immediately.
4007 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4008 order to support additional features. One way would have been to define
4009 pcre_exec2() (for example) with extra arguments, but this would not have been
4010 extensible, and would also have required all calls to the original function to
4011 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4012 is used for passing in "extra" data from pcre_study().
4014 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4015 contains the following fields:
4017 flags a bitmap indicating which of the following fields are set
4018 study_data opaque data from pcre_study()
4019 match_limit a way of specifying a limit on match() calls for a specific
4020 call to pcre_exec()
4021 callout_data data for callouts (see 49 below)
4023 The flag bits are also defined in pcre.h, and are
4029 The pcre_study() function now returns one of these new pcre_extra blocks, with
4030 the actual study data pointed to by the study_data field, and the
4031 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4032 before. That is, this change is entirely upwards-compatible and requires no
4033 change to existing code.
4035 If you want to pass in additional data to pcre_exec(), you can either place it
4036 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4037 block.
4039 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4040 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4041 times with different match limits, until it finds the minimum value needed for
4042 pcre_exec() to complete. The value is then output. This can be instructive; for
4043 most simple matches the number is quite small, but for pathological cases it
4044 gets very large very quickly.
4046 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4047 returns the size of the data block pointed to by the study_data field in a
4048 pcre_extra block, that is, the value that was passed as the argument to
4049 pcre_malloc() when PCRE was getting memory in which to place the information
4050 created by pcre_study(). The fourth argument should point to a size_t variable.
4051 pcretest has been extended so that this information is shown after a successful
4052 pcre_study() call when information about the compiled regex is being displayed.
4054 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4055 because what follows is always an absolute path. (Later: it turns out that this
4056 is more than cosmetic for MinGW, because it doesn't like empty path
4057 components.)
4059 49. Some changes have been made to the callout feature (see 28 above):
4061 (i) A callout function now has three choices for what it returns:
4063 0 => success, carry on matching
4064 > 0 => failure at this point, but backtrack if possible
4065 < 0 => serious error, return this value from pcre_exec()
4067 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4068 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4069 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4070 use by callout functions. It will never be used by PCRE itself.
4072 (ii) The pcre_extra structure (see 45 above) has a void * field called
4073 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4074 pcre_callout_block structure has a field of the same name. The contents of
4075 the field passed in the pcre_extra structure are passed to the callout
4076 function in the corresponding field in the callout block. This makes it
4077 easier to use the same callout-containing regex from multiple threads. For
4078 testing, the pcretest program has a new data escape
4080 \C*n pass the number n (may be negative) as callout_data
4082 If the callout function in pcretest receives a non-zero value as
4083 callout_data, it returns that value.
4085 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4086 there were some redundant $(CFLAGS) in commands that are now specified as
4087 $(LINK), which already includes $(CFLAGS).
4089 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4090 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4091 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4092 one-byte characters throughout. Note that case-insensitive matching applies
4093 only to characters whose values are less than 256. PCRE doesn't support the
4094 notion of cases for higher-valued characters.
4096 (i) A character class whose characters are all within 0-255 is handled as
4097 a bit map, and the map is inverted for negative classes. Previously, a
4098 character > 255 always failed to match such a class; however it should
4099 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4101 (ii) A negated character class with a single character < 255 is coded as
4102 "not this character" (OP_NOT). This wasn't working properly when the test
4103 character was multibyte, either singly or repeated.
4105 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4106 mode, for example: \x{100}{2,3}.
4108 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4109 singly or repeated) now correctly test multibyte characters. However,
4110 PCRE doesn't recognize any characters with values greater than 255 as
4111 digits, spaces, or word characters. Such characters always match \D, \S,
4112 and \W, and never match \d, \s, or \w.
4114 (v) Classes may now contain characters and character ranges with values
4115 greater than 255. For example: [ab\x{100}-\x{400}].
4117 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4118 PCRE in UTF-8 mode.
4120 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4121 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4122 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4123 value.)
4125 53. The single man page has become too large. I have therefore split it up into
4126 a number of separate man pages. These also give rise to individual HTML pages;
4127 these are now put in a separate directory, and there is an index.html page that
4128 lists them all. Some hyperlinking between the pages has been installed.
4130 54. Added convenience functions for handling named capturing parentheses.
4132 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4133 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4134 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4135 are faulted.
4137 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4138 calling configure. These values are used when compiling the dftables.c program
4139 which is run to generate the source of the default character tables. They
4140 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4141 you will need to set these values.
4143 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4146 Version 3.9 02-Jan-02
4147 ---------------------
4149 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4151 2. If --disable-static was given, the building process failed when trying to
4152 build pcretest and pcregrep. (For some reason it was using libtool to compile
4153 them, which is not right, as they aren't part of the library.)
4156 Version 3.8 18-Dec-01
4157 ---------------------
4159 1. The experimental UTF-8 code was completely screwed up. It was packing the
4160 bytes in the wrong order. How dumb can you get?
4163 Version 3.7 29-Oct-01
4164 ---------------------
4166 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4167 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4168 this didn't happen under Solaris 8, where I normally test things.
4170 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4171 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4172 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4173 this fix an hour or so after the initial 3.7 release.)
4176 Version 3.6 23-Oct-01
4177 ---------------------
4179 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4180 offsets passed as NULL with zero offset count.
4182 2. The config.guess and config.sub files had not been updated when I moved to
4183 the latest autoconf.
4186 Version 3.5 15-Aug-01
4187 ---------------------
4189 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4190 had been forgotten.
4192 2. By using declared but undefined structures, we can avoid using "void"
4193 definitions in pcre.h while keeping the internal definitions of the structures
4194 private.
4196 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4197 user point of view, this means that both static and shared libraries are built
4198 by default, but this can be individually controlled. More of the work of
4199 handling this static/shared cases is now inside libtool instead of PCRE's make
4200 file.
4202 4. The pcretest utility is now installed along with pcregrep because it is
4203 useful for users (to test regexs) and by doing this, it automatically gets
4204 relinked by libtool. The documentation has been turned into a man page, so
4205 there are now .1, .txt, and .html versions in /doc.
4207 5. Upgrades to pcregrep:
4208 (i) Added long-form option names like gnu grep.
4209 (ii) Added --help to list all options with an explanatory phrase.
4210 (iii) Added -r, --recursive to recurse into sub-directories.
4211 (iv) Added -f, --file to read patterns from a file.
4213 6. pcre_exec() was referring to its "code" argument before testing that
4214 argument for NULL (and giving an error if it was NULL).
4216 7. Upgraded Makefile.in to allow for compiling in a different directory from
4217 the source directory.
4219 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4220 options bits, the pointer it was passed was to an int instead of to an unsigned
4221 long int. This mattered only on 64-bit systems.
4223 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4224 generated) instead of pcre.in, which it its source. Also made the same change
4225 in several of the .c files.
4227 10. A new release of gcc defines printf() as a macro, which broke pcretest
4228 because it had an ifdef in the middle of a string argument for printf(). Fixed
4229 by using separate calls to printf().
4231 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4232 script, to force use of CR or LF instead of \n in the source. On non-Unix
4233 systems, the value can be set in config.h.
4235 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4236 absolute limit. Changed the text of the error message to make this clear, and
4237 likewise updated the man page.
4239 13. The limit of 99 on the number of capturing subpatterns has been removed.
4240 The new limit is 65535, which I hope will not be a "real" limit.
4243 Version 3.4 22-Aug-00
4244 ---------------------
4246 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4248 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4251 Version 3.3 01-Aug-00
4252 ---------------------
4254 1. If an octal character was given, but the value was greater than \377, it
4255 was not getting masked to the least significant bits, as documented. This could
4256 lead to crashes in some systems.
4258 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4259 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4261 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4262 These just pass their arguments on to (pcre_free)(), but they are provided
4263 because some uses of PCRE bind it to non-C systems that can call its functions,
4264 but cannot call free() or pcre_free() directly.
4266 4. Add "make test" as a synonym for "make check". Corrected some comments in
4267 the Makefile.
4269 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4270 Makefile.
4272 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4273 command called pgrep for grepping around the active processes.
4275 7. Added the beginnings of support for UTF-8 character strings.
4277 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4278 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4279 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4280 out for the ar command.)
4283 Version 3.2 12-May-00
4284 ---------------------
4286 This is purely a bug fixing release.
4288 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4289 of ZA. This was just one example of several cases that could provoke this bug,
4290 which was introduced by change 9 of version 2.00. The code for breaking
4291 infinite loops after an iteration that matches an empty string was't working
4292 correctly.
4294 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4295 when matched against abbab (for example). After matching an empty string, it
4296 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4297 caused it to match further down the string than it should.
4299 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4300 was there because it doesn't seem to be needed, and it causes trouble on some
4301 systems, as it is not a Standard C header. It has been removed.
4303 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4304 were reported on the Macintosh. The changes were from
4306 while ((c = *(++ptr)) != 0 && c != '\n');
4307 to
4308 while ((c = *(++ptr)) != 0 && c != '\n') ;
4310 Totally extraordinary, but if that's what it takes...
4312 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4313 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4314 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4315 assumes the way PCRE uses memmove() (always moving upwards).
4317 6. PCRE is being used in one environment where strchr() is not available. There
4318 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4319 faster code anyway.
4322 Version 3.1 09-Feb-00
4323 ---------------------
4325 The only change in this release is the fixing of some bugs in Makefile.in for
4326 the "install" target:
4328 (1) It was failing to install pcreposix.h.
4330 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4333 Version 3.0 01-Feb-00
4334 ---------------------
4336 1. Add support for the /+ modifier to perltest (to output $` like it does in
4337 pcretest).
4339 2. Add support for the /g modifier to perltest.
4341 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4342 matches null strings.
4344 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4345 pattern. Perl treats empty patterns specially - it reuses the most recent
4346 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4347 effect.
4349 5. The POSIX interface was broken in that it was just handing over the POSIX
4350 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4351 required a bigger vector, with some working space on the end. This means that
4352 the POSIX wrapper now has to get and free some memory, and copy the results.
4354 6. Added some simple autoconf support, placing the test data and the
4355 documentation in separate directories, re-organizing some of the
4356 information files, and making it build pcre-config (a GNU standard). Also added
4357 libtool support for building PCRE as a shared library, which is now the
4358 default.
4360 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4361 09 are not valid octal constants. Single digits will be used for minor values
4362 less than 10.
4364 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4365 existing programs that set these in the POSIX interface can use PCRE without
4366 modification.
4368 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4369 return all that pcre_info() returns, plus additional data. The pcre_info()
4370 function is retained for compatibility, but is considered to be obsolete.
4372 10. Added experimental recursion feature (?R) to handle one common case that
4373 Perl 5.6 will be able to do with (?p{...}).
4375 11. Added support for POSIX character classes like [:alpha:], which Perl is
4376 adopting.
4379 Version 2.08 31-Aug-99
4380 ----------------------
4382 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4383 trying to match at the startoffset position, but instead was moving forward to
4384 the next newline as if a previous match had failed.
4386 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4387 and could get into a loop if a null string was matched other than at the start
4388 of the subject.
4390 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4391 be distinguished at compile time, and for completeness also added PCRE_DATE.
4393 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4394 in GnuWin32 environments.
4397 Version 2.07 29-Jul-99
4398 ----------------------
4400 1. The documentation is now supplied in plain text form and HTML as well as in
4401 the form of man page sources.
4403 2. C++ compilers don't like assigning (void *) values to other pointer types.
4404 In particular this affects malloc(). Although there is no problem in Standard
4405 C, I've put in casts to keep C++ compilers happy.
4407 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4408 should be (const char *).
4410 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4411 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4412 However, I haven't made this a standard facility. The documentation doesn't
4413 mention it, and the Makefile doesn't support it.
4415 5. The Makefile now contains an "install" target, with editable destinations at
4416 the top of the file. The pcretest program is not installed.
4418 6. pgrep -V now gives the PCRE version number and date.
4420 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4421 causing the entire string to be ignored, instead of just the last character.
4423 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4424 non-matching string, it can take a very, very long time, even for strings of
4425 quite modest length, because of the nested recursion. PCRE now does better in
4426 some of these cases. It does this by remembering the last required literal
4427 character in the pattern, and pre-searching the subject to ensure it is present
4428 before running the real match. In other words, it applies a heuristic to detect
4429 some types of certain failure quickly, and in the above example, if presented
4430 with a string that has no trailing " it gives "no match" very quickly.
4432 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4433 other alternatives are tried instead.
4436 Version 2.06 09-Jun-99
4437 ----------------------
4439 1. Change pcretest's output for amount of store used to show just the code
4440 space, because the remainder (the data block) varies in size between 32-bit and
4441 64-bit systems.
4443 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4444 start matching at. This allows lookbehinds to work when searching for multiple
4445 occurrences in a string.
4447 3. Added additional options to pcretest for testing multiple occurrences:
4449 /+ outputs the rest of the string that follows a match
4450 /g loops for multiple occurrences, using the new startoffset argument
4451 /G loops for multiple occurrences by passing an incremented pointer
4453 4. PCRE wasn't doing the "first character" optimization for patterns starting
4454 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4455 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4456 the letter 'x'. On long subject strings, this gives a significant speed-up.
4459 Version 2.05 21-Apr-99
4460 ----------------------
4462 1. Changed the type of magic_number from int to long int so that it works
4463 properly on 16-bit systems.
4465 2. Fixed a bug which caused patterns starting with .* not to work correctly
4466 when the subject string contained newline characters. PCRE was assuming
4467 anchoring for such patterns in all cases, which is not correct because .* will
4468 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4469 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4470 must be retried after every newline in the subject.
4473 Version 2.04 18-Feb-99
4474 ----------------------
4476 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4477 computation of the store needed to hold the pattern was incorrect (too large).
4478 If such patterns were nested a few deep, this could multiply and become a real
4479 problem.
4481 2. Added /M option to pcretest to show the memory requirement of a specific
4482 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4484 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4485 compiled in such a way that the backtracking after subsequent failure was
4486 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4487 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4490 Version 2.03 02-Feb-99
4491 ----------------------
4493 1. Fixed typo and small mistake in man page.
4495 2. Added 4th condition (GPL supersedes if conflict) and created separate
4496 LICENCE file containing the conditions.
4498 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4499 Perl, that is the internal \ allows the delimiter to be included in the
4500 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4501 the final delimiter, add \ to the end of the pattern (to test the error).
4503 4. Added the convenience functions for extracting substrings after a successful
4504 match. Updated pcretest to make it able to test these functions.
4507 Version 2.02 14-Jan-99
4508 ----------------------
4510 1. Initialized the working variables associated with each extraction so that
4511 their saving and restoring doesn't refer to uninitialized store.
4513 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4514 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4515 fix the problem.
4517 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4518 calls, and wasn't printing the correct value for compiling calls. Increased the
4519 default value of LOOPREPEAT, and the number of significant figures in the
4520 times.
4522 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4524 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4525 a building problem on Windows NT with a FAT file system.
4528 Version 2.01 21-Oct-98
4529 ----------------------
4531 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4532 to character tables built by pcre_maketables() in the current locale. If NULL
4533 is passed, the default tables are used.
4536 Version 2.00 24-Sep-98
4537 ----------------------
4539 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4540 it any more.
4542 2. Allow quantification of (?>) groups, and make it work correctly.
4544 3. The first character computation wasn't working for (?>) groups.
4546 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4547 end of the subject) and add 5.005's \z, which really does match only at the
4548 very end of the subject.
4550 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4552 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4553 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4554 localized options. All options to pcre_study() were also removed.
4556 7. Add other new features from 5.005:
4558 $(?<= positive lookbehind
4559 $(?<! negative lookbehind
4560 (?imsx-imsx) added the unsetting capability
4561 such a setting is global if at outer level; local otherwise
4562 (?imsx-imsx:) non-capturing groups with option setting
4563 (?(cond)re|re) conditional pattern matching
4565 A backreference to itself in a repeated group matches the previous
4566 captured string.
4568 8. General tidying up of studying (both automatic and via "study")
4569 consequential on the addition of new assertions.
4571 9. As in 5.005, unlimited repeated groups that could match an empty substring
4572 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4573 runtime if any iteration does actually match an empty substring.
4575 10. Include the RunTest script in the distribution.
4577 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4578 discrepancies, some of which were old and were also with respect to 5.004. They
4579 have now been fixed.
4582 Version 1.09 28-Apr-98
4583 ----------------------
4585 1. A negated single character class followed by a quantifier with a minimum
4586 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4587 program crashes, or just wrong answers. This did not apply to negated classes
4588 containing more than one character, or to minima other than one.
4591 Version 1.08 27-Mar-98
4592 ----------------------
4594 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4596 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4597 latter must appear before anything that relies on it in the pattern.
4600 Version 1.07 16-Feb-98
4601 ----------------------
4603 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4604 repeat of a potentially empty string).
4607 Version 1.06 23-Jan-98
4608 ----------------------
4610 1. Added Markus Oberhumer's little patches for C++.
4612 2. Literal strings longer than 255 characters were broken.
4615 Version 1.05 23-Dec-97
4616 ----------------------
4618 1. Negated character classes containing more than one character were failing if
4619 PCRE_CASELESS was set at run time.
4622 Version 1.04 19-Dec-97
4623 ----------------------
4625 1. Corrected the man page, where some "const" qualifiers had been omitted.
4627 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4628 input syntax.
4630 3. Fixed memory leak which occurred when a regex with back references was
4631 matched with an offsets vector that wasn't big enough. The temporary memory
4632 that is used in this case wasn't being freed if the match failed.
4634 4. Tidied pcretest to ensure it frees memory that it gets.
4636 5. Temporary memory was being obtained in the case where the passed offsets
4637 vector was exactly big enough.
4639 6. Corrected definition of offsetof() from change 5 below.
4641 7. I had screwed up change 6 below and broken the rules for the use of
4642 setjmp(). Now fixed.
4645 Version 1.03 18-Dec-97
4646 ----------------------
4648 1. A erroneous regex with a missing opening parenthesis was correctly
4649 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4650 on some systems.
4652 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4653 it was reported that one broken compiler failed on the former because "code" is
4654 also an independent variable.
4656 3. The erroneous regex a[]b caused an array overrun reference.
4658 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4659 fail on data ending with that character. (It was going on too far, and checking
4660 the next character, typically a binary zero.) This was specific to the
4661 optimized code for single-character negative classes.
4663 5. Added a contributed patch from the TIN world which does the following:
4665 + Add an undef for memmove, in case the the system defines a macro for it.
4667 + Add a definition of offsetof(), in case there isn't one. (I don't know
4668 the reason behind this - offsetof() is part of the ANSI standard - but
4669 it does no harm).
4671 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4672 most of the places where whitespace preceded '#'. I have given up and
4673 allowed the remaining 2 cases to be at the margin.
4675 + Rename some variables in pcre to eliminate shadowing. This seems very
4676 pedantic, but does no harm, of course.
4678 6. Moved the call to setjmp() into its own function, to get rid of warnings
4679 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4681 7. Constructs such as \d{8,} were compiling into the equivalent of
4682 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4683 outcome, but in this particular case used more store than had been allocated,
4684 which caused the bug to be discovered because it threw up an internal error.
4686 8. The debugging code in both pcre and pcretest for outputting the compiled
4687 form of a regex was going wrong in the case of back references followed by
4688 curly-bracketed repeats.
4691 Version 1.02 12-Dec-97
4692 ----------------------
4694 1. Typos in pcre.3 and comments in the source fixed.
4696 2. Applied a contributed patch to get rid of places where it used to remove
4697 'const' from variables, and fixed some signed/unsigned and uninitialized
4698 variable warnings.
4700 3. Added the "runtest" target to Makefile.
4702 4. Set default compiler flag to -O2 rather than just -O.
4705 Version 1.01 19-Nov-97
4706 ----------------------
4708 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4709 like /([ab]*)*/, that is, for classes with more than one character in them.
4711 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4712 as /((?>a*))*/ (a PCRE_EXTRA facility).
4715 Version 1.00 18-Nov-97
4716 ----------------------
4718 1. Added compile-time macros to support systems such as SunOS4 which don't have
4719 memmove() or strerror() but have other things that can be used instead.
4721 2. Arranged that "make clean" removes the executables.
4724 Version 0.99 27-Oct-97
4725 ----------------------
4727 1. Fixed bug in code for optimizing classes with only one character. It was
4728 initializing a 32-byte map regardless, which could cause it to run off the end
4729 of the memory it had got.
4731 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4734 Version 0.98 22-Oct-97
4735 ----------------------
4737 1. Fixed bug in code for handling temporary memory usage when there are more
4738 back references than supplied space in the ovector. This could cause segfaults.
4741 Version 0.97 21-Oct-97
4742 ----------------------
4744 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4746 2. Optimized negated single characters not to use a bit map.
4748 3. Brought error texts together as macro definitions; clarified some of them;
4749 fixed one that was wrong - it said "range out of order" when it meant "invalid
4750 escape sequence".
4752 4. Changed some char * arguments to const char *.
4754 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4756 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4757 pcretest.
4760 Version 0.96 16-Oct-97
4761 ----------------------
4763 1. Added a simple "pgrep" utility to the distribution.
4765 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4766 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4767 where "ddd" means "one or more decimal digits".
4769 3. Fixed serious bug. If a pattern had a back reference, but the call to
4770 pcre_exec() didn't supply a large enough ovector to record the related
4771 identifying subpattern, the match always failed. PCRE now remembers the number
4772 of the largest back reference, and gets some temporary memory in which to save
4773 the offsets during matching if necessary, in order to ensure that
4774 backreferences always work.
4776 4. Increased the compatibility with Perl in a number of ways:
4778 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4779 to request this handling. The option can be set at compile or exec time.
4781 (b) $ matches before a terminating newline by default; an option
4782 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4783 mode). The option can be set at compile or exec time.
4785 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4786 the same as Perl's. If the decimal number it represents is less than 10
4787 or there aren't that many previous left capturing parentheses, an octal
4788 escape is read. Inside a character class, it's always an octal escape,
4789 even if it is a single digit.
4791 (d) An escaped but undefined alphabetic character is taken as a literal,
4792 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4793 escapes.
4795 (e) {0} is now permitted. (The previous item is removed from the compiled
4796 pattern).
4798 5. Changed all the names of code files so that the basic parts are no longer
4799 than 10 characters, and abolished the teeny "globals.c" file.
4801 6. Changed the handling of character classes; they are now done with a 32-byte
4802 bit map always.
4804 7. Added the -d and /D options to pcretest to make it possible to look at the
4805 internals of compilation without having to recompile pcre.
4808 Version 0.95 23-Sep-97
4809 ----------------------
4811 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4812 \x20 at the start of a run of normal characters. These were being treated as
4813 real characters, instead of the source characters being re-checked.
4816 Version 0.94 18-Sep-97
4817 ----------------------
4819 1. The functions are now thread-safe, with the caveat that the global variables
4820 containing pointers to malloc() and free() or alternative functions are the
4821 same for all threads.
4823 2. Get pcre_study() to generate a bitmap of initial characters for non-
4824 anchored patterns when this is possible, and use it if passed to pcre_exec().
4827 Version 0.93 15-Sep-97
4828 ----------------------
4830 1. /(b)|(:+)/ was computing an incorrect first character.
4832 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4833 but not actually doing anything yet.
4835 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4836 as Perl does (e.g. [-az] or [az-]).
4838 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4839 all possible positions.
4841 5. Split up into different modules to avoid including unneeded functions in a
4842 compiled binary. However, compile and exec are still in one module. The "study"
4843 function is split off.
4845 6. The character tables are now in a separate module whose source is generated
4846 by an auxiliary program - but can then be edited by hand if required. There are
4847 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4848 toupper() in the code.
4850 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4851 make them global. Abolish the function for setting them, as the caller can now
4852 set them directly.
4855 Version 0.92 11-Sep-97
4856 ----------------------
4858 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4859 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4861 2. Caseless matching was not working in character classes if the characters in
4862 the pattern were in upper case.
4864 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4866 4. Make PCRE_ANCHORED public and accept as a compile option.
4868 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4869 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4870 pass them.
4872 6. Give an error if bad option bits passed at compile or run time.
4874 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4875 pcretest to cause it to pass that flag.
4877 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4878 options, and the first character, if set.
4880 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4883 Version 0.91 10-Sep-97
4884 ----------------------
4886 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4887 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4889 2. PCRE was looping on encountering an indefinitely repeated back reference to
4890 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4891 Perl does - treats the match as successful.
4893 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5