ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1196 - (show annotations)
Thu Nov 1 17:41:36 2012 UTC (7 years, 1 month ago) by ph10
File size: 239266 byte(s)
GCC visibility patch.
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.
79 18. If the --only-matching (-o) option in pcregrep is specified multiple
80 times, each one causes appropriate output. For example, -o1 -o2 outputs the
81 substrings matched by the 1st and 2nd capturing parentheses. A separating
82 string can be specified by --om-separator (default empty).
84 19. Improving the first n character searches.
86 20. Turn case lists for horizontal and vertical white space into macros so that
87 they are defined only once.
89 21. This set of changes together give more compatible Unicode case-folding
90 behaviour for characters that have more than one other case when UCP
91 support is available.
93 (a) The Unicode property table now has offsets into a new table of sets of
94 three or more characters that are case-equivalent. The MultiStage2.py
95 script that generates these tables (the pcre_ucd.c file) now scans
96 CaseFolding.txt instead of UnicodeData.txt for character case
97 information.
99 (b) The code for adding characters or ranges of characters to a character
100 class has been abstracted into a generalized function that also handles
101 case-independence. In UTF-mode with UCP support, this uses the new data
102 to handle characters with more than one other case.
104 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105 whose other case is greater than 256 are now correctly matched
106 caselessly. Previously, the high codepoint matched the low one, but not
107 vice versa.
109 (d) The processing of \h, \H, \v, and \ in character classes now makes use
110 of the new class addition function, using character lists defined as
111 macros alongside the case definitions of 20 above.
113 (e) Caseless back references now work with characters that have more than
114 one other case.
116 (f) General caseless matching of characters with more than one other case
117 is supported.
119 22. Unicode character properties were updated from Unicode 6.2.0
121 23. Improved CMake support under Windows. Patch by Daniel Richard G.
123 24. Add support for 32-bit character strings, and UTF-32
125 25. Major JIT compiler update (code refactoring and bugfixing).
126 Experimental Sparc 32 support is added.
128 26. Applied a modified version of Daniel Richard G's patch to create
129 pcre.h.generic and config.h.generic by "make" instead of in the
130 PrepareRelease script.
132 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
133 pcre_compile.c when checking for a zero character.
135 28. Introducing a native interface for JIT. Through this interface, the compiled
136 machine code can be directly executed. The purpose of this interface is to
137 provide fast pattern matching, so several sanity checks are not performed.
138 However, feature tests are still performed. The new interface provides
139 1.4x speedup compared to the old one.
141 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
142 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
143 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
145 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
146 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
147 the "old" RFC 2279). Instead, it ended up passing a negative length to
148 pcre_exec().
150 31. Add support for GCC's visibility feature to hide internal functions.
153 Version 8.31 06-July-2012
154 -------------------------
156 1. Fixing a wrong JIT test case and some compiler warnings.
158 2. Removed a bashism from the RunTest script.
160 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
161 to unsigned type, result still unsigned" that was given by an MS compiler
162 on encountering the code "-sizeof(xxx)".
164 4. Partial matching support is added to the JIT compiler.
166 5. Fixed several bugs concerned with partial matching of items that consist
167 of more than one character:
169 (a) /^(..)\1/ did not partially match "aba" because checking references was
170 done on an "all or nothing" basis. This also applied to repeated
171 references.
173 (b) \R did not give a hard partial match if \r was found at the end of the
174 subject.
176 (c) \X did not give a hard partial match after matching one or more
177 characters at the end of the subject.
179 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
180 a partial match for the string "\r".
182 (e) When newline was set to CRLF, the metacharacter "." did not recognize
183 a partial match for a CR character at the end of the subject string.
185 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
186 running pcretest, the text "(JIT)" added to the output whenever JIT is
187 actually used to run the match.
189 7. Individual JIT compile options can be set in pcretest by following -s+[+]
190 or /S+[+] with a digit between 1 and 7.
192 8. OP_NOT now supports any UTF character not just single-byte ones.
194 9. (*MARK) control verb is now supported by the JIT compiler.
196 10. The command "./RunTest list" lists the available tests without actually
197 running any of them. (Because I keep forgetting what they all are.)
201 12. Applied a (slightly modified) user-supplied patch that improves performance
202 when the heap is used for recursion (compiled with --disable-stack-for-
203 recursion). Instead of malloc and free for each heap frame each time a
204 logical recursion happens, frames are retained on a chain and re-used where
205 possible. This sometimes gives as much as 30% improvement.
207 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
208 call.
210 14. As documented, (*COMMIT) is now confined to within a positive assertion.
212 15. It is now possible to link pcretest with libedit as an alternative to
213 libreadline.
215 16. (*COMMIT) control verb is now supported by the JIT compiler.
217 17. The Unicode data tables have been updated to Unicode 6.1.0.
219 18. Added --file-list option to pcregrep.
221 19. Added binary file support to pcregrep, including the -a, --binary-files,
222 -I, and --text options.
224 20. The madvise function is renamed for posix_madvise for QNX compatibility
225 reasons. Fixed by Giuseppe D'Angelo.
227 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
229 22. Changed the option for creating symbolic links for 16-bit man pages from
230 -s to -sf so that re-installing does not cause issues.
232 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
234 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
235 matches in certain environments (the workspace was not being correctly
236 retained). Also added to pcre_dfa_exec() a simple plausibility check on
237 some of the workspace data at the beginning of a restart.
239 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
240 was not doing so when it should - probably a typo introduced by SVN 528
241 (change 8.10/14).
243 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
244 \w+ when the character tables indicated that \x{c4} was a word character.
245 There were several related cases, all because the tests for doing a table
246 lookup were testing for characters less than 127 instead of 255.
248 27. If a pattern contains capturing parentheses that are not used in a match,
249 their slots in the ovector are set to -1. For those that are higher than
250 any matched groups, this happens at the end of processing. In the case when
251 there were back references that the ovector was too small to contain
252 (causing temporary malloc'd memory to be used during matching), and the
253 highest capturing number was not used, memory off the end of the ovector
254 was incorrectly being set to -1. (It was using the size of the temporary
255 memory instead of the true size.)
257 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
258 ovector size, it uses memory at the end of the block that it has got.
260 29. Check for an overlong MARK name and give an error at compile time. The
261 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
263 30. JIT compiler update.
265 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
266 Rill for the patch.
268 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
270 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
272 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
273 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
274 was enabled.
276 35. Improve JIT code generation for greedy plus quantifier.
278 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
279 1 to "aa" instead of to an empty string. The bug affected repeated groups
280 that could potentially match an empty string.
282 37. Optimizing single character iterators in JIT.
284 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
285 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
286 codepoints that are too big for the mode are faulted, and in a UTF mode,
287 disallowed codepoints are also faulted.
289 39. If PCRE was compiled with UTF support, in three places in the DFA
290 matcher there was code that should only have been obeyed in UTF mode, but
291 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
292 processing when bytes with values greater than 127 were present. In 16-bit
293 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
294 both cases the values are those that cannot be the first data item in a UTF
295 character. The three items that might have provoked this were recursions,
296 possessively repeated groups, and atomic groups.
298 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
299 and pcregrep, because some OS require shared objects to be explicitly
300 passed to ld, causing the link step to fail if they are not.
302 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
303 mode, patterns that started with \h* or \R* might be incorrectly matched.
306 Version 8.30 04-February-2012
307 -----------------------------
309 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
310 name is defined in ctype.h.
312 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
313 only in quite long subpatterns.
315 3. Removed the function pcre_info(), which has been obsolete and deprecated
316 since it was replaced by pcre_fullinfo() in February 2000.
318 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
319 match a (*MARK), and the match failed at the start of the subject, a
320 reference to memory before the start of the subject could occur. This bug
321 was introduced by fix 17 of release 8.21.
323 5. A reference to an unset group with zero minimum repetition was giving
324 totally wrong answers (in non-JavaScript-compatibility mode). For example,
325 /(another)?(\1?)test/ matched against "hello world test". This bug was
326 introduced in release 8.13.
328 6. Add support for 16-bit character strings (a large amount of work involving
329 many changes and refactorings).
331 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
332 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
333 from a file.
335 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
336 rounding is not applied in this particular case).
338 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
339 if they appear, or are escaped, in patterns.
341 10. Get rid of a number of -Wunused-but-set-variable warnings.
343 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
344 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
345 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
346 also returns the mark "x". This bug applied to capturing parentheses,
347 non-capturing parentheses, and atomic parentheses. It also applied to some
348 assertions.
350 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
351 information out of configure.ac instead of relying on pcre.h.generic, which
352 is not stored in the repository.
354 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
355 -lreadline.
357 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
359 15. Applied Graycode's patch to put the top-level frame on the stack rather
360 than the heap when not using the stack for recursion. This gives a
361 performance improvement in many cases when recursion is not deep.
363 16. Experimental code added to "pcretest -C" to output the stack frame size.
366 Version 8.21 12-Dec-2011
367 ------------------------
369 1. Updating the JIT compiler.
371 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
372 are added as well.
374 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
375 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
376 calling _pcre_jit_exec. Some extra comments are added.
378 4. (*MARK) settings inside atomic groups that do not contain any capturing
379 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
380 was introduced by change 18 for 8.20.
382 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
383 ECMA-262 standard.
385 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
386 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
387 This bug was probably introduced by change 9 of 8.13.
389 7. While fixing 6 above, I noticed that a number of other items were being
390 incorrectly rejected as "not fixed length". This arose partly because newer
391 opcodes had not been added to the fixed-length checking code. I have (a)
392 corrected the bug and added tests for these items, and (b) arranged for an
393 error to occur if an unknown opcode is encountered while checking for fixed
394 length instead of just assuming "not fixed length". The items that were
395 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
396 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
397 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
399 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
400 being incorrectly compiled and would have given unpredicatble results.
402 9. A possessively repeated subpattern with minimum repeat count greater than
403 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
404 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
405 the first (A) could occur when it should not.
407 10. Add a cast and remove a redundant test from the code.
409 11. JIT should use pcre_malloc/pcre_free for allocation.
411 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
412 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
413 is anything other than /usr, -L is still shown).
415 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
417 14. Perl does not support \N without a following name in a [] class; PCRE now
418 also gives an error.
420 15. If a forward reference was repeated with an upper limit of around 2000,
421 it caused the error "internal error: overran compiling workspace". The
422 maximum number of forward references (including repeats) was limited by the
423 internal workspace, and dependent on the LINK_SIZE. The code has been
424 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
425 the default depends on LINK_SIZE. There is a new upper limit (for safety)
426 of around 200,000 forward references. While doing this, I also speeded up
427 the filling in of repeated forward references.
429 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
430 incorrectly expecting the subject to contain another "a" after the start.
432 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
433 in the match, the SKIP should be ignored. This was not happening; instead
434 the SKIP was being treated as NOMATCH. For patterns such as
435 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
436 tested.
438 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
439 now much more compatible with Perl, in particular in cases where the result
440 is a non-match for a non-anchored pattern. For example, if
441 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
442 "m", where previously it did not return a name. A side effect of this
443 change is that for partial matches, the last encountered mark name is
444 returned, as for non matches. A number of tests that were previously not
445 Perl-compatible have been moved into the Perl-compatible test files. The
446 refactoring has had the pleasing side effect of removing one argument from
447 the match() function, thus reducing its stack requirements.
449 19. If the /S+ option was used in pcretest to study a pattern using JIT,
450 subsequent uses of /S (without +) incorrectly behaved like /S+.
452 21. Retrieve executable code size support for the JIT compiler and fixing
453 some warnings.
455 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
456 not work when the shorter character appeared right at the end of the
457 subject string.
459 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
460 systems.
462 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
463 output it when the /M option is used in pcretest.
465 25. The CheckMan script was not being included in the distribution. Also, added
466 an explicit "perl" to run Perl scripts from the PrepareRelease script
467 because this is reportedly needed in Windows.
469 26. If study data was being save in a file and studying had not found a set of
470 "starts with" bytes for the pattern, the data written to the file (though
471 never used) was taken from uninitialized memory and so caused valgrind to
472 complain.
474 27. Updated RunTest.bat as provided by Sheri Pierce.
476 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
478 29. Computation of memory usage for the table of capturing group names was
479 giving an unnecessarily large value.
482 Version 8.20 21-Oct-2011
483 ------------------------
485 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
486 a POSIX class. After further experiments with Perl, which convinced me that
487 Perl has bugs and confusions, a closing square bracket is no longer allowed
488 in a POSIX name. This bug also affected patterns with classes that started
489 with full stops.
491 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
492 captured substring, but while checking the failing first alternative,
493 substring 1 is temporarily captured. If the output vector supplied to
494 pcre_exec() was not big enough for this capture, the yield of the function
495 was still zero ("insufficient space for captured substrings"). This cannot
496 be totally fixed without adding another stack variable, which seems a lot
497 of expense for a edge case. However, I have improved the situation in cases
498 such as /(a)(b)x|abc/ matched against "abc", where the return code
499 indicates that fewer than the maximum number of slots in the ovector have
500 been set.
502 3. Related to (2) above: when there are more back references in a pattern than
503 slots in the output vector, pcre_exec() uses temporary memory during
504 matching, and copies in the captures as far as possible afterwards. It was
505 using the entire output vector, but this conflicts with the specification
506 that only 2/3 is used for passing back captured substrings. Now it uses
507 only the first 2/3, for compatibility. This is, of course, another edge
508 case.
510 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
511 main code base, and can be used by building with --enable-jit. When this is
512 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
513 runtime --no-jit option is given.
515 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
516 ovector, the return from the function was zero, implying that there were
517 other matches that did not fit. The correct "exactly full" value is now
518 returned.
520 6. If a subpattern that was called recursively or as a subroutine contained
521 (*PRUNE) or any other control that caused it to give a non-standard return,
522 invalid errors such as "Error -26 (nested recursion at the same subject
523 position)" or even infinite loops could occur.
525 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
526 computing the minimum length on reaching *ACCEPT, and so ended up with the
527 wrong value of 1 rather than 0. Further investigation indicates that
528 computing a minimum subject length in the presence of *ACCEPT is difficult
529 (think back references, subroutine calls), and so I have changed the code
530 so that no minimum is registered for a pattern that contains *ACCEPT.
532 8. If (*THEN) was present in the first (true) branch of a conditional group,
533 it was not handled as intended. [But see 16 below.]
535 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
536 Sheri Pierce.
538 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
539 the first byte in a match must be "a".
541 11. Change 17 for 8.13 increased the recursion depth for patterns like
542 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
543 pattern contains any instances of (*THEN). If it does not, the old
544 optimizations are restored. It would be nice to do this on a per-group
545 basis, but at the moment that is not feasible.
547 12. In some environments, the output of pcretest -C is CRLF terminated. This
548 broke RunTest's code that checks for the link size. A single white space
549 character after the value is now allowed for.
551 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
552 For "fr", it uses the Windows-specific input and output files.
554 14. If (*THEN) appeared in a group that was called recursively or as a
555 subroutine, it did not work as intended. [But see next item.]
557 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
558 pattern fragments (but not containing any | characters). If A and B are
559 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
560 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
561 D. In other words, Perl considers parentheses that do not contain any |
562 characters to be part of a surrounding alternative, whereas PCRE was
563 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
564 differently. PCRE now behaves in the same way as Perl, except in the case
565 of subroutine/recursion calls such as (?1) which have in any case always
566 been different (but PCRE had them first :-).
568 16. Related to 15 above: Perl does not treat the | in a conditional group as
569 creating alternatives. Such a group is treated in the same way as an
570 ordinary group without any | characters when processing (*THEN). PCRE has
571 been changed to match Perl's behaviour.
573 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
574 RunGrepTest script failed.
576 18. Change 22 for version 13 caused atomic groups to use more stack. This is
577 inevitable for groups that contain captures, but it can lead to a lot of
578 stack use in large patterns. The old behaviour has been restored for atomic
579 groups that do not contain any capturing parentheses.
581 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
582 suppress the check for a minimum subject length at run time. (If it was
583 given to pcre_exec() or pcre_dfa_exec() it did work.)
585 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
586 fail to work when decoding hex characters in data strings in EBCDIC
587 environments.
589 21. It appears that in at least one Mac OS environment, the isxdigit() function
590 is implemented as a macro that evaluates to its argument more than once,
591 contravening the C 90 Standard (I haven't checked a later standard). There
592 was an instance in pcretest which caused it to go wrong when processing
593 \x{...} escapes in subject strings. The has been rewritten to avoid using
594 things like p++ in the argument of isxdigit().
597 Version 8.13 16-Aug-2011
598 ------------------------
600 1. The Unicode data tables have been updated to Unicode 6.0.0.
602 2. Two minor typos in pcre_internal.h have been fixed.
604 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
605 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
606 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
608 4. There were a number of related bugs in the code for matching backrefences
609 caselessly in UTF-8 mode when codes for the characters concerned were
610 different numbers of bytes. For example, U+023A and U+2C65 are an upper
611 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
612 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
613 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
614 2-byte code at the end of the subject (it thought there wasn't enough data
615 left).
617 5. Comprehensive information about what went wrong is now returned by
618 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
619 as the output vector has at least 2 elements. The offset of the start of
620 the failing character and a reason code are placed in the vector.
622 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
623 now returned is for the first byte of the failing character, instead of the
624 last byte inspected. This is an incompatible change, but I hope it is small
625 enough not to be a problem. It makes the returned offset consistent with
626 pcre_exec() and pcre_dfa_exec().
628 7. pcretest now gives a text phrase as well as the error number when
629 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
630 failure, the offset and reason code are output.
632 8. When \R was used with a maximizing quantifier it failed to skip backwards
633 over a \r\n pair if the subsequent match failed. Instead, it just skipped
634 back over a single character (\n). This seems wrong (because it treated the
635 two characters as a single entity when going forwards), conflicts with the
636 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
637 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
638 has been changed.
640 9. Some internal refactoring has changed the processing so that the handling
641 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
642 time (the PCRE_DOTALL option was changed this way some time ago: version
643 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
644 which was always a bit of a fudge. It also means that there is one less
645 argument for the match() function, which reduces its stack requirements
646 slightly. This change also fixes an incompatibility with Perl: the pattern
647 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
649 10. More internal refactoring has drastically reduced the number of recursive
650 calls to match() for possessively repeated groups such as (abc)++ when
651 using pcre_exec().
653 11. While implementing 10, a number of bugs in the handling of groups were
654 discovered and fixed:
656 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
657 (a|)*(?1) gave a compile-time internal error.
658 ((a|)+)+ did not notice that the outer group could match an empty string.
659 (^a|^)+ was not marked as anchored.
660 (.*a|.*)+ was not marked as matching at start or after a newline.
662 12. Yet more internal refactoring has removed another argument from the match()
663 function. Special calls to this function are now indicated by setting a
664 value in a variable in the "match data" data block.
666 13. Be more explicit in pcre_study() instead of relying on "default" for
667 opcodes that mean there is no starting character; this means that when new
668 ones are added and accidentally left out of pcre_study(), testing should
669 pick them up.
671 14. The -s option of pcretest has been documented for ages as being an old
672 synonym of -m (show memory usage). I have changed it to mean "force study
673 for every regex", that is, assume /S for every regex. This is similar to -i
674 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
675 using it. It makes it easier to run collections of tests with and without
676 study enabled, and thereby test pcre_study() more easily. All the standard
677 tests are now run with and without -s (but some patterns can be marked as
678 "never study" - see 20 below).
680 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
681 restoration of the capturing data to the outer values was not happening
682 correctly.
684 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
685 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
686 pattern had matched an empty string, and so incorrectly returned a no
687 match.
689 17. There was optimizing code for the last branch of non-capturing parentheses,
690 and also for the obeyed branch of a conditional subexpression, which used
691 tail recursion to cut down on stack usage. Unfortunately, now that there is
692 the possibility of (*THEN) occurring in these branches, tail recursion is
693 no longer possible because the return has to be checked for (*THEN). These
694 two optimizations have therefore been removed. [But see 8.20/11 above.]
696 18. If a pattern containing \R was studied, it was assumed that \R always
697 matched two bytes, thus causing the minimum subject length to be
698 incorrectly computed because \R can also match just one byte.
700 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
701 was incorrectly computed.
703 20. If /S is present twice on a test pattern in pcretest input, it now
704 *disables* studying, thereby overriding the use of -s on the command line
705 (see 14 above). This is necessary for one or two tests to keep the output
706 identical in both cases.
708 21. When (*ACCEPT) was used in an assertion that matched an empty string and
709 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
711 22. When an atomic group that contained a capturing parenthesis was
712 successfully matched, but the branch in which it appeared failed, the
713 capturing was not being forgotten if a higher numbered group was later
714 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
715 group 1 to "a", when in fact it should be unset. This applied to multi-
716 branched capturing and non-capturing groups, repeated or not, and also to
717 positive assertions (capturing in negative assertions does not happen
718 in PCRE) and also to nested atomic groups.
720 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
721 subject after a captured substring, to make it easier to tell which of a
722 number of identical substrings has been captured.
724 24. The way atomic groups are processed by pcre_exec() has been changed so that
725 if they are repeated, backtracking one repetition now resets captured
726 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
727 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
728 "aaa". Previously, it would have been "a". As part of this code
729 refactoring, the way recursive calls are handled has also been changed.
731 25. If an assertion condition captured any substrings, they were not passed
732 back unless some other capturing happened later. For example, if
733 (?(?=(a))a) was matched against "a", no capturing was returned.
735 26. When studying a pattern that contained subroutine calls or assertions,
736 the code for finding the minimum length of a possible match was handling
737 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
738 group 1 called group 2 while simultaneously a separate group 2 called group
739 1). A stack overflow occurred in this case. I have fixed this by limiting
740 the recursion depth to 10.
742 27. Updated RunTest.bat in the distribution to the version supplied by Tom
743 Fortmann. This supports explicit test numbers on the command line, and has
744 argument validation and error reporting.
746 28. An instance of \X with an unlimited repeat could fail if at any point the
747 first character it looked at was a mark character.
749 29. Some minor code refactoring concerning Unicode properties and scripts
750 should reduce the stack requirement of match() slightly.
752 30. Added the '=' option to pcretest to check the setting of unused capturing
753 slots at the end of the pattern, which are documented as being -1, but are
754 not included in the return count.
756 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
757 compiled something random. Now it gives a compile-time error (as does
758 Perl).
760 32. A *MARK encountered during the processing of a positive assertion is now
761 recorded and passed back (compatible with Perl).
763 33. If --only-matching or --colour was set on a pcregrep call whose pattern
764 had alternative anchored branches, the search for a second match in a line
765 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
766 matched the line "0102" twice. The same bug affected patterns that started
767 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
768 twice.
770 34. Previously, PCRE did not allow quantification of assertions. However, Perl
771 does, and because of capturing effects, quantifying parenthesized
772 assertions may at times be useful. Quantifiers are now allowed for
773 parenthesized assertions.
775 35. A minor code tidy in pcre_compile() when checking options for \R usage.
777 36. \g was being checked for fancy things in a character class, when it should
778 just be a literal "g".
780 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
781 appearance of a nested POSIX class supersedes an apparent external class.
782 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
783 unescaped square brackets may also appear as part of class names. For
784 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
785 more like Perl. (But see 8.20/1 above.)
787 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
788 was because it thought it was \N{name}, which is not supported).
790 39. Add minix to OS list not supporting the -S option in pcretest.
792 40. PCRE tries to detect cases of infinite recursion at compile time, but it
793 cannot analyze patterns in sufficient detail to catch mutual recursions
794 such as ((?1))((?2)). There is now a runtime test that gives an error if a
795 subgroup is called recursively as a subpattern for a second time at the
796 same position in the subject string. In previous releases this might have
797 been caught by the recursion limit, or it might have run out of stack.
799 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
800 happen only once. PCRE was, however incorrectly giving a compile time error
801 "recursive call could loop indefinitely" because it cannot analyze the
802 pattern in sufficient detail. The compile time test no longer happens when
803 PCRE is compiling a conditional subpattern, but actual runaway loops are
804 now caught at runtime (see 40 above).
806 42. It seems that Perl allows any characters other than a closing parenthesis
807 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
808 has been changed to be the same.
810 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
811 as not to get warnings when autogen.sh is called. Also changed
812 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
814 44. To help people who use pcregrep to scan files containing exceedingly long
815 lines, the following changes have been made:
817 (a) The default value of the buffer size parameter has been increased from
818 8K to 20K. (The actual buffer used is three times this size.)
820 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
821 PCRE is built.
823 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
824 to be set at run time.
826 (d) Numerical values in pcregrep options can be followed by K or M, for
827 example --buffer-size=50K.
829 (e) If a line being scanned overflows pcregrep's buffer, an error is now
830 given and the return code is set to 2.
832 45. Add a pointer to the latest mark to the callout data block.
834 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
835 partial match of an empty string instead of no match. This was specific to
836 the use of ".".
838 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
839 complete match instead of a partial match. This bug was dependent on both
840 the PCRE_UTF8 and PCRE_DOTALL options being set.
842 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
843 starting byte set, because \b was not being ignored.
846 Version 8.12 15-Jan-2011
847 ------------------------
849 1. Fixed some typos in the markup of the man pages, and wrote a script that
850 checks for such things as part of the documentation building process.
852 2. On a big-endian 64-bit system, pcregrep did not correctly process the
853 --match-limit and --recursion-limit options (added for 8.11). In
854 particular, this made one of the standard tests fail. (The integer value
855 went into the wrong half of a long int.)
857 3. If the --colour option was given to pcregrep with -v (invert match), it
858 did strange things, either producing crazy output, or crashing. It should,
859 of course, ignore a request for colour when reporting lines that do not
860 match.
862 4. Another pcregrep bug caused similar problems if --colour was specified with
863 -M (multiline) and the pattern match finished with a line ending.
865 5. In pcregrep, when a pattern that ended with a literal newline sequence was
866 matched in multiline mode, the following line was shown as part of the
867 match. This seems wrong, so I have changed it.
869 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
870 the check for further matches in the same line (so they could be coloured)
871 to overrun the end of the current line. If another match was found, it was
872 incorrectly shown (and then shown again when found in the next line).
874 7. If pcregrep was compiled under Windows, there was a reference to the
875 function pcregrep_exit() before it was defined. I am assuming this was
876 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
877 reported by a user. I've moved the definition above the reference.
880 Version 8.11 10-Dec-2010
881 ------------------------
883 1. (*THEN) was not working properly if there were untried alternatives prior
884 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
885 backtracked to try for "b" instead of moving to the next alternative branch
886 at the same level (in this case, to look for "c"). The Perl documentation
887 is clear that when (*THEN) is backtracked onto, it goes to the "next
888 alternative in the innermost enclosing group".
890 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
891 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
892 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
893 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
894 (*THEN).
896 3. If \s appeared in a character class, it removed the VT character from
897 the class, even if it had been included by some previous item, for example
898 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
899 of \s, but is part of the POSIX "space" class.)
901 4. A partial match never returns an empty string (because you can always
902 match an empty string at the end of the subject); however the checking for
903 an empty string was starting at the "start of match" point. This has been
904 changed to the "earliest inspected character" point, because the returned
905 data for a partial match starts at this character. This means that, for
906 example, /(?<=abc)def/ gives a partial match for the subject "abc"
907 (previously it gave "no match").
909 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
910 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
911 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
912 has an implication that the given string is incomplete (because a partial
913 match is preferred over a full match). For this reason, these items now
914 give a partial match in this situation. [Aside: previously, the one case
915 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
916 match rather than a full match, which was wrong by the old rules, but is
917 now correct.]
919 6. There was a bug in the handling of #-introduced comments, recognized when
920 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
921 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
922 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
923 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
924 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
925 places in pcre_compile().
927 7. Related to (6) above, when pcre_compile() was skipping #-introduced
928 comments when looking ahead for named forward references to subpatterns,
929 the only newline sequence it recognized was NL. It now handles newlines
930 according to the set newline convention.
932 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
933 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
934 cater for a lack of strerror(). These oversights have been fixed.
936 9. Added --match-limit and --recursion-limit to pcregrep.
938 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
940 11. When the -o option was used, pcregrep was setting a return code of 1, even
941 when matches were found, and --line-buffered was not being honoured.
943 12. Added an optional parentheses number to the -o and --only-matching options
944 of pcregrep.
946 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
947 can match an empty string. The code to do it in pcretest and pcredemo
948 needed fixing:
950 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
951 only one byte after an empty string match just before CRLF (this case
952 just got forgotten; "any" and "anycrlf" were OK).
954 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
955 mode when an empty string match preceded an ASCII character followed by
956 a non-ASCII character. (The code for advancing by one character rather
957 than one byte was nonsense.)
959 (c) The pcredemo.c sample program did not have any code at all to handle
960 the cases when CRLF is a valid newline sequence.
962 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
963 as a starting offset was within the subject string. There is now a new
964 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
965 negative or greater than the length of the string. In order to test this,
966 pcretest is extended to allow the setting of negative starting offsets.
968 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
969 starting offset points to the beginning of a UTF-8 character was
970 unnecessarily clumsy. I tidied it up.
972 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
973 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
975 17. Nobody had reported that the --include_dir option, which was added in
976 release 7.7 should have been called --include-dir (hyphen, not underscore)
977 for compatibility with GNU grep. I have changed it to --include-dir, but
978 left --include_dir as an undocumented synonym, and the same for
979 --exclude-dir, though that is not available in GNU grep, at least as of
980 release 2.5.4.
982 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
983 characters from a string of bytes) have been redefined so as not to use
984 loops, in order to improve performance in some environments. At the same
985 time, I abstracted some of the common code into auxiliary macros to save
986 repetition (this should not affect the compiled code).
988 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
989 compile-time error is now given if \c is not followed by an ASCII
990 character, that is, a byte less than 128. (In EBCDIC mode, the code is
991 different, and any byte value is allowed.)
993 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
994 START_OPTIMIZE option, which is now allowed at compile time - but just
995 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
996 to pcregrep and other applications that have no direct access to PCRE
997 options. The new /Y option in pcretest sets this option when calling
998 pcre_compile().
1000 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1001 back references. Groups containing recursive back references were forced to
1002 be atomic by that change, but in the case of named groups, the amount of
1003 memory required was incorrectly computed, leading to "Failed: internal
1004 error: code overflow". This has been fixed.
1006 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1007 pcretest.c, to avoid build problems in some Borland environments.
1010 Version 8.10 25-Jun-2010
1011 ------------------------
1013 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1014 THEN.
1016 2. (*ACCEPT) was not working when inside an atomic group.
1018 3. Inside a character class, \B is treated as a literal by default, but
1019 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1020 causes the error). The code is unchanged, but I tidied the documentation.
1022 4. Inside a character class, PCRE always treated \R and \X as literals,
1023 whereas Perl faults them if its -w option is set. I have changed PCRE so
1024 that it faults them when PCRE_EXTRA is set.
1026 5. Added support for \N, which always matches any character other than
1027 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1029 6. When compiling pcregrep with newer versions of gcc which may have
1030 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1031 declared with attribute warn_unused_result" were given. Just casting the
1032 result to (void) does not stop the warnings; a more elaborate fudge is
1033 needed. I've used a macro to implement this.
1035 7. Minor change to pcretest.c to avoid a compiler warning.
1037 8. Added four artifical Unicode properties to help with an option to make
1038 \s etc use properties (see next item). The new properties are: Xan
1039 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1041 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1042 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1043 this option. Modified pcretest to add /W to test this facility. Added
1044 REG_UCP to make it available via the POSIX interface.
1046 10. Added --line-buffered to pcregrep.
1048 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1049 studied, and the match started with a letter with a code point greater than
1050 127 whose first byte was different to the first byte of the other case of
1051 the letter, the other case of this starting letter was not recognized
1052 (#976).
1054 12. If a pattern that was studied started with a repeated Unicode property
1055 test, for example, \p{Nd}+, there was the theoretical possibility of
1056 setting up an incorrect bitmap of starting bytes, but fortunately it could
1057 not have actually happened in practice until change 8 above was made (it
1058 added property types that matched character-matching opcodes).
1060 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1061 possible starting bytes for non-anchored patterns.
1063 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1064 \R, and also a number of cases that involve Unicode properties, both
1065 explicit and implicit when PCRE_UCP is set.
1067 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1068 input, it could crash or give wrong results if characters with values
1069 greater than 0xc0 were present in the subject string. (Detail: it assumed
1070 UTF-8 input when processing these items.)
1072 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1073 size_t is 64-bit (#991).
1075 17. Added a check for running out of memory when PCRE is compiled with
1076 --disable-stack-for-recursion (#990).
1078 18. If the last data line in a file for pcretest does not have a newline on
1079 the end, a newline was missing in the output.
1081 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1082 less than 128) in its various bitmaps. However, there is a facility for
1083 generating tables according to the current locale when PCRE is compiled. It
1084 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1085 characters, are recognized by isspace() and therefore were getting set in
1086 these tables, and indeed these tables seem to approximate to ISO 8859. This
1087 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1088 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1089 which of course cannot start UTF-8 characters. I have changed the code so
1090 that only real ASCII characters (less than 128) and the correct starting
1091 bytes for UTF-8 encodings are set for characters greater than 127 when in
1092 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1093 altogether.)
1095 20. Added the /T option to pcretest so as to be able to run tests with non-
1096 standard character tables, thus making it possible to include the tests
1097 used for 19 above in the standard set of tests.
1099 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1100 reference to a subpattern the other side of a comment that contains an
1101 opening parenthesis caused either an internal compiling error, or a
1102 reference to the wrong subpattern.
1105 Version 8.02 19-Mar-2010
1106 ------------------------
1108 1. The Unicode data tables have been updated to Unicode 5.2.0.
1110 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1111 configured.
1113 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1114 original author of that file, following a query about its status.
1116 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1117 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1119 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1120 quantifier applied to a forward-referencing subroutine call, could compile
1121 incorrect code or give the error "internal error: previously-checked
1122 referenced subpattern not found".
1124 6. Both MS Visual Studio and Symbian OS have problems with initializing
1125 variables to point to external functions. For these systems, therefore,
1126 pcre_malloc etc. are now initialized to local functions that call the
1127 relevant global functions.
1129 7. There were two entries missing in the vectors called coptable and poptable
1130 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1131 I've fixed the data, and added a kludgy way of testing at compile time that
1132 the lengths are correct (equal to the number of opcodes).
1134 8. Following on from 7, I added a similar kludge to check the length of the
1135 eint vector in pcreposix.c.
1137 9. Error texts for pcre_compile() are held as one long string to avoid too
1138 much relocation at load time. To find a text, the string is searched,
1139 counting zeros. There was no check for running off the end of the string,
1140 which could happen if a new error number was added without updating the
1141 string.
1143 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1145 11. \K was not working if it appeared in an atomic group or in a group that
1146 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1147 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1148 the assertion is positive, but not if it is negative.
1150 12. Change 11 fortuitously reduced the size of the stack frame used in the
1151 "match()" function of pcre_exec.c by one pointer. Forthcoming
1152 implementation of support for (*MARK) will need an extra pointer on the
1153 stack; I have reserved it now, so that the stack frame size does not
1154 decrease.
1156 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1157 item in branch that calls a recursion is a subroutine call - as in the
1158 second branch in the above example - was incorrectly given the compile-
1159 time error "recursive call could loop indefinitely" because pcre_compile()
1160 was not correctly checking the subroutine for matching a non-empty string.
1162 14. The checks for overrunning compiling workspace could trigger after an
1163 overrun had occurred. This is a "should never occur" error, but it can be
1164 triggered by pathological patterns such as hundreds of nested parentheses.
1165 The checks now trigger 100 bytes before the end of the workspace.
1167 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1170 Version 8.01 19-Jan-2010
1171 ------------------------
1173 1. If a pattern contained a conditional subpattern with only one branch (in
1174 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1175 computed the wrong minimum data length (which is of course zero for such
1176 subpatterns). This could cause incorrect "no match" results.
1178 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1179 the pattern is reset in the first branch, pcre_compile() failed with
1180 "internal error: code overflow at offset...". This happened only when
1181 the reset was to the original external option setting. (An optimization
1182 abstracts leading options settings into an external setting, which was the
1183 cause of this.)
1185 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1186 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1187 assertion pattern did not match (meaning that the assertion was true), it
1188 was incorrectly treated as false if the SKIP had been reached during the
1189 matching. This also applied to assertions used as conditions.
1191 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1192 assertion subpattern, including such a pattern used as a condition,
1193 unpredictable results occurred, instead of the error return
1196 5. The C++ GlobalReplace function was not working like Perl for the special
1197 situation when an empty string is matched. It now does the fancy magic
1198 stuff that is necessary.
1200 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1201 removed. (These were left over from very, very early versions of PCRE.)
1203 7. Some cosmetic changes to the code to make life easier when compiling it
1204 as part of something else:
1206 (a) Change DEBUG to PCRE_DEBUG.
1208 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1209 called "current" as "current_branch", to prevent a collision with the
1210 Linux macro when compiled as a kernel module.
1212 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1213 prevent a collision with the Linux macro when compiled as a kernel
1214 module.
1216 8. In pcre_compile() there are some checks for integer overflows that used to
1217 cast potentially large values to (double). This has been changed to that
1218 when building, a check for int64_t is made, and if it is found, it is used
1219 instead, thus avoiding the use of floating point arithmetic. (There is no
1220 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1221 double.
1223 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1224 2005 (difference between two addresses compared to an unsigned value).
1226 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1227 custom one, because of the following reported problem in Windows:
1229 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1230 under Win32.
1231 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1232 therefore missing the function definition.
1233 - The compiler thus generates a "C" signature for the test function.
1234 - The linker fails to find the "C" function.
1235 - PCRE fails to configure if asked to do so against libbz2.
1237 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1238 messages were output:
1240 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1241 rerunning libtoolize, to keep the correct libtool macros in-tree.
1242 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1244 I have done both of these things.
1246 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1247 most of the time, it *can* run out if it is given a pattern that contains a
1248 runaway infinite recursion. I updated the discussion in the pcrestack man
1249 page.
1251 13. Now that we have gone to the x.xx style of version numbers, the minor
1252 version may start with zero. Using 08 or 09 is a bad idea because users
1253 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1254 interpreted as invalid octal numbers. I've updated the previous comment in
1255 configure.ac, and also added a check that gives an error if 08 or 09 are
1256 used.
1258 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1259 causing partial matching to fail when the end of the subject matched \W
1260 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1262 15. There were some discrepancies between the declarations in pcre_internal.h
1263 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1264 their definitions. The declarations used "const uschar *" and the
1265 definitions used USPTR. Even though USPTR is normally defined as "const
1266 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1267 reported that: "This difference in casting confuses some C++ compilers, for
1268 example, SunCC recognizes above declarations as different functions and
1269 generates broken code for hbpcre." I have changed the declarations to use
1270 USPTR.
1272 16. GNU libtool is named differently on some systems. The autogen.sh script now
1273 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1274 (FreeBSD).
1276 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1277 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1278 comment: "Figure out how to create a longlong from a string: strtoll and
1279 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1280 instance, but it only takes 2 args instead of 3!"
1282 18. A subtle bug concerned with back references has been fixed by a change of
1283 specification, with a corresponding code fix. A pattern such as
1284 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1285 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1286 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1287 same bug. Such groups have to be quantified to be useful, or contained
1288 inside another quantified group. (If there's no repetition, the reference
1289 can never match.) The problem arises because, having left the group and
1290 moved on to the rest of the pattern, a later failure that backtracks into
1291 the group uses the captured value from the final iteration of the group
1292 rather than the correct earlier one. I have fixed this in PCRE by forcing
1293 any group that contains a reference to itself to be an atomic group; that
1294 is, there cannot be any backtracking into it once it has completed. This is
1295 similar to recursive and subroutine calls.
1298 Version 8.00 19-Oct-09
1299 ----------------------
1301 1. The table for translating pcre_compile() error codes into POSIX error codes
1302 was out-of-date, and there was no check on the pcre_compile() error code
1303 being within the table. This could lead to an OK return being given in
1304 error.
1306 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1307 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1308 in a Windows environment.
1310 3. The pcregrep --count option prints the count for each file even when it is
1311 zero, as does GNU grep. However, pcregrep was also printing all files when
1312 --files-with-matches was added. Now, when both options are given, it prints
1313 counts only for those files that have at least one match. (GNU grep just
1314 prints the file name in this circumstance, but including the count seems
1315 more useful - otherwise, why use --count?) Also ensured that the
1316 combination -clh just lists non-zero counts, with no names.
1318 4. The long form of the pcregrep -F option was incorrectly implemented as
1319 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1320 but it seems right to fix it, and I didn't think it was worth preserving
1321 the old behaviour.
1323 5. The command line items --regex=pattern and --regexp=pattern were not
1324 recognized by pcregrep, which required --regex pattern or --regexp pattern
1325 (with a space rather than an '='). The man page documented the '=' forms,
1326 which are compatible with GNU grep; these now work.
1328 6. No libpcreposix.pc file was created for pkg-config; there was just
1329 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1331 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1332 when UCP support is not needed, by modifying the Python script that
1333 generates it from Unicode data files. This should not matter if the module
1334 is correctly used as a library, but I received one complaint about 50K of
1335 unwanted data. My guess is that the person linked everything into his
1336 program rather than using a library. Anyway, it does no harm.
1338 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1339 was a minimum greater than 1 for a wide character in a possessive
1340 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1341 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1342 character. Chaos in the form of incorrect output or a compiling loop could
1343 result.
1345 9. The restrictions on what a pattern can contain when partial matching is
1346 requested for pcre_exec() have been removed. All patterns can now be
1347 partially matched by this function. In addition, if there are at least two
1348 slots in the offset vector, the offset of the earliest inspected character
1349 for the match and the offset of the end of the subject are set in them when
1350 PCRE_ERROR_PARTIAL is returned.
1352 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1353 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1354 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1355 and may be more useful for multi-segment matching.
1357 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1358 used to be given if ever the end of the subject was reached; now it is
1359 given only if matching could not proceed because another character was
1360 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1361 string "Z", which now yields "no match" instead of "partial match". In the
1362 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1363 final character ended with (*FAIL).
1365 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1366 if the pattern had a "must contain" character that was already found in the
1367 earlier partial match, unless partial matching was again requested. For
1368 example, with the pattern /dog.(body)?/, the "must contain" character is
1369 "g". If the first part-match was for the string "dog", restarting with
1370 "sbody" failed. This bug has been fixed.
1372 13. The string returned by pcre_dfa_exec() after a partial match has been
1373 changed so that it starts at the first inspected character rather than the
1374 first character of the match. This makes a difference only if the pattern
1375 starts with a lookbehind assertion or \b or \B (\K is not supported by
1376 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1377 matching functions compatible, and I think it's the right thing to do.
1379 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1380 so that the demonstration program is easily available in environments where
1381 PCRE has not been installed from source.
1383 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1384 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1385 library.
1387 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1388 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1389 is not the first non-POSIX option to be added. Clearly some people find
1390 these options useful.
1392 17. If a caller to the POSIX matching function regexec() passes a non-zero
1393 value for nmatch with a NULL value for pmatch, the value of
1394 nmatch is forced to zero.
1396 18. RunGrepTest did not have a test for the availability of the -u option of
1397 the diff command, as RunTest does. It now checks in the same way as
1398 RunTest, and also checks for the -b option.
1400 19. If an odd number of negated classes containing just a single character
1401 interposed, within parentheses, between a forward reference to a named
1402 subpattern and the definition of the subpattern, compilation crashed with
1403 an internal error, complaining that it could not find the referenced
1404 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1405 [The bug was that it was starting one character too far in when skipping
1406 over the character class, thus treating the ] as data rather than
1407 terminating the class. This meant it could skip too much.]
1409 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1410 /g option in pcretest when the pattern contains \K, which makes it possible
1411 to have an empty string match not at the start, even when the pattern is
1412 anchored. Updated pcretest and pcredemo to use this option.
1414 21. If the maximum number of capturing subpatterns in a recursion was greater
1415 than the maximum at the outer level, the higher number was returned, but
1416 with unset values at the outer level. The correct (outer level) value is
1417 now given.
1419 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1420 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1421 make it do so. The string so far is captured, making this feature
1422 compatible with Perl.
1424 23. The tests have been re-organized, adding tests 11 and 12, to make it
1425 possible to check the Perl 5.10 features against Perl 5.10.
1427 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1428 pattern matches a fixed length string. PCRE did not allow this; now it
1429 does. Neither allows recursion.
1431 25. I finally figured out how to implement a request to provide the minimum
1432 length of subject string that was needed in order to match a given pattern.
1433 (It was back references and recursion that I had previously got hung up
1434 on.) This code has now been added to pcre_study(); it finds a lower bound
1435 to the length of subject needed. It is not necessarily the greatest lower
1436 bound, but using it to avoid searching strings that are too short does give
1437 some useful speed-ups. The value is available to calling programs via
1438 pcre_fullinfo().
1440 26. While implementing 25, I discovered to my embarrassment that pcretest had
1441 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1442 study optimizations had never been tested with that matching function.
1443 Oops. What is worse, even when it was passed study data, there was a bug in
1444 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1445 were also very few tests of studied patterns with pcre_dfa_exec().
1447 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1448 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1449 on the other side of the coin, they are no longer allowed to have different
1450 names, because these cannot be distinguished in PCRE, and this has caused
1451 confusion. (This is a difference from Perl.)
1453 28. When duplicate subpattern names are present (necessarily with different
1454 numbers, as required by 27 above), and a test is made by name in a
1455 conditional pattern, either for a subpattern having been matched, or for
1456 recursion in such a pattern, all the associated numbered subpatterns are
1457 tested, and the overall condition is true if the condition is true for any
1458 one of them. This is the way Perl works, and is also more like the way
1459 testing by number works.
1462 Version 7.9 11-Apr-09
1463 ---------------------
1465 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1466 (pcretest), all targets were linked against these libraries. This included
1467 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1468 libraries. This caused unwanted dependencies to be created. This problem
1469 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1470 pcretest is linked with readline.
1472 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1473 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1474 moved outside it again, because FALSE and TRUE are already defined in AIX,
1475 but BOOL is not.
1477 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1478 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1480 4. The pcregrep documentation said spaces were inserted as well as colons (or
1481 hyphens) following file names and line numbers when outputting matching
1482 lines. This is not true; no spaces are inserted. I have also clarified the
1483 wording for the --colour (or --color) option.
1485 5. In pcregrep, when --colour was used with -o, the list of matching strings
1486 was not coloured; this is different to GNU grep, so I have changed it to be
1487 the same.
1489 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1490 each matching line was coloured. Now it goes on to look for further matches
1491 of any of the test patterns, which is the same behaviour as GNU grep.
1493 7. A pattern that could match an empty string could cause pcregrep to loop; it
1494 doesn't make sense to accept an empty string match in pcregrep, so I have
1495 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1496 seems to be how GNU grep behaves.
1498 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1499 start or after a newline", because the conditional assertion was not being
1500 correctly handled. The rule now is that both the assertion and what follows
1501 in the first alternative must satisfy the test.
1503 9. If auto-callout was enabled in a pattern with a conditional group whose
1504 condition was an assertion, PCRE could crash during matching, both with
1505 pcre_exec() and pcre_dfa_exec().
1507 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1508 used for matching.
1510 11. Unicode property support in character classes was not working for
1511 characters (bytes) greater than 127 when not in UTF-8 mode.
1513 12. Added the -M command line option to pcretest.
1515 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1517 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1519 16. Added comments and documentation about mis-use of no_arg in the C++
1520 wrapper.
1522 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1523 from Martin Jerabek that uses macro names for all relevant character and
1524 string constants.
1526 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1527 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1528 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1529 these, but not everybody uses configure.
1531 19. A conditional group that had only one branch was not being correctly
1532 recognized as an item that could match an empty string. This meant that an
1533 enclosing group might also not be so recognized, causing infinite looping
1534 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1535 with the subject "ab", where knowledge that the repeated group can match
1536 nothing is needed in order to break the loop.
1538 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1539 exec(), but without supplying a callout function, matching went wrong.
1541 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1542 leak if the size of the offset vector was greater than 30. When the vector
1543 is smaller, the saved offsets during recursion go onto a local stack
1544 vector, but for larger vectors malloc() is used. It was failing to free
1545 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1546 error, in fact).
1548 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1549 heapframe that is used only when UTF-8 support is enabled. This caused no
1550 problem, but was untidy.
1552 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1553 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1554 included within another project.
1556 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1557 slightly modified by me:
1559 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1560 not building pcregrep.
1562 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1563 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1565 25. Forward references, both numeric and by name, in patterns that made use of
1566 duplicate group numbers, could behave incorrectly or give incorrect errors,
1567 because when scanning forward to find the reference group, PCRE was not
1568 taking into account the duplicate group numbers. A pattern such as
1569 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1571 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1572 the feature of a custom pointer more persuasive (as requested by a user).
1574 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1575 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1576 pre-defined.
1578 28. Added support for (*UTF8) at the start of a pattern.
1580 29. Arrange for flags added by the "release type" setting in CMake to be shown
1581 in the configuration summary.
1584 Version 7.8 05-Sep-08
1585 ---------------------
1587 1. Replaced UCP searching code with optimized version as implemented for Ad
1588 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1589 stage table and inline lookup instead of a function, giving speed ups of 2
1590 to 5 times on some simple patterns that I tested. Permission was given to
1591 distribute the MultiStage2.py script that generates the tables (it's not in
1592 the tarball, but is in the Subversion repository).
1594 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1595 scripts.
1597 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1598 a group with a zero qualifier. The result of the study could be incorrect,
1599 or the function might crash, depending on the pattern.
1601 4. Caseless matching was not working for non-ASCII characters in back
1602 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1603 It now works when Unicode Property Support is available.
1605 5. In pcretest, an escape such as \x{de} in the data was always generating
1606 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1607 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1608 truncation.
1610 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1612 7. Added two (int) casts to pcregrep when printing the difference of two
1613 pointers, in case they are 64-bit values.
1615 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1616 test 2 if it fails.
1618 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1619 and a #define of that name to empty if it is not externally set. This is to
1620 allow users of MSVC to set it if necessary.
1622 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1623 the convenience functions in the pcre_get.c source file.
1625 11. An option change at the start of a pattern that had top-level alternatives
1626 could cause overwriting and/or a crash. This command provoked a crash in
1627 some environments:
1629 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1631 This potential security problem was recorded as CVE-2008-2371.
1633 12. For a pattern where the match had to start at the beginning or immediately
1634 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1635 pcre_dfa_exec() could read past the end of the passed subject if there was
1636 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1637 pcretest so that it places the subject at the end of its malloc-ed buffer.
1639 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1640 exec() might read past the end of the data buffer in UTF-8 mode.
1642 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1643 the data contained the byte 0x85 as part of a UTF-8 character within its
1644 first line. This applied both to normal and DFA matching.
1646 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1647 /^[^d]*?$/8 failed to match "abc".
1649 16. Added a missing copyright notice to pcrecpp_internal.h.
1651 17. Make it more clear in the documentation that values returned from
1652 pcre_exec() in ovector are byte offsets, not character counts.
1654 18. Tidied a few places to stop certain compilers from issuing warnings.
1656 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1657 supplied by Stefan Weber. I made a further small update for 7.8 because
1658 there is a change of source arrangements: the pcre_searchfuncs.c module is
1659 replaced by pcre_ucd.c.
1662 Version 7.7 07-May-08
1663 ---------------------
1665 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1666 a string to a long long, pretend we don't even have a long long." This is
1667 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1669 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1670 pre-7.6 versions, which defined a global no_arg variable instead of putting
1671 it in the RE class. (See also #8 below.)
1673 3. Remove a line of dead code, identified by coverity and reported by Nuno
1674 Lopes.
1676 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1678 (1) The include/exclude patterns were being applied to the whole pathnames
1679 of files, instead of just to the final components.
1681 (2) If there was more than one level of directory, the subdirectories were
1682 skipped unless they satisfied the include/exclude conditions. This is
1683 inconsistent with GNU grep (and could even be seen as contrary to the
1684 pcregrep specification - which I improved to make it absolutely clear).
1685 The action now is always to scan all levels of directory, and just
1686 apply the include/exclude patterns to regular files.
1688 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1689 --exclude_dir in the tests to avoid scanning .svn directories.
1691 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1692 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1693 doesn't support NULs in patterns.
1695 7. Added some missing "const"s to declarations of static tables in
1696 pcre_compile.c and pcre_dfa_exec.c.
1698 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1699 caused by fix #2 above. (Subsequently also a second patch to fix the
1700 first patch. And a third patch - this was a messy problem.)
1702 9. Applied Craig's patch to remove the use of push_back().
1704 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1705 matching function regexec().
1707 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1708 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1709 references. PCRE supports relative numbers with this syntax (I don't think
1710 Oniguruma does).
1712 12. Previously, a group with a zero repeat such as (...){0} was completely
1713 omitted from the compiled regex. However, this means that if the group
1714 was called as a subroutine from elsewhere in the pattern, things went wrong
1715 (an internal error was given). Such groups are now left in the compiled
1716 pattern, with a new opcode that causes them to be skipped at execution
1717 time.
1719 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1720 to the way PCRE behaves:
1722 (a) A lone ] character is dis-allowed (Perl treats it as data).
1724 (b) A back reference to an unmatched subpattern matches an empty string
1725 (Perl fails the current match path).
1727 (c) A data ] in a character class must be notated as \] because if the
1728 first data character in a class is ], it defines an empty class. (In
1729 Perl it is not possible to have an empty class.) The empty class []
1730 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1731 The negative empty class [^] matches any one character, independently
1732 of the DOTALL setting.
1734 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1735 non-existent subpattern following a character class starting with ']' and
1736 containing () gave an internal compiling error instead of "reference to
1737 non-existent subpattern". Fortunately, when the pattern did exist, the
1738 compiled code was correct. (When scanning forwards to check for the
1739 existencd of the subpattern, it was treating the data ']' as terminating
1740 the class, so got the count wrong. When actually compiling, the reference
1741 was subsequently set up correctly.)
1743 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1744 it was being rejected as not supported by pcre_dfa_exec(), even though
1745 other assertions are supported. I have made pcre_dfa_exec() support
1746 (*FAIL).
1748 16. The implementation of 13c above involved the invention of a new opcode,
1749 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1750 cannot be changed at match time, I realized I could make a small
1751 improvement to matching performance by compiling OP_ALLANY instead of
1752 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1753 on the OP_ANY path.
1755 17. Compiling pcretest on Windows with readline support failed without the
1756 following two fixes: (1) Make the unistd.h include conditional on
1757 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1759 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1760 ncurses library to be included for pcretest when ReadLine support is
1761 requested, but also to allow for it to be overridden. This patch came from
1762 Daniel Bergström.
1764 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1765 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1766 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1767 spotting this.
1770 Version 7.6 28-Jan-08
1771 ---------------------
1773 1. A character class containing a very large number of characters with
1774 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1775 overflow.
1777 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1778 HAVE_LONG_LONG is not defined.
1780 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1781 bring them up to date and include new features. This patch includes:
1783 - Fixed PH's badly added libz and libbz2 support.
1784 - Fixed a problem with static linking.
1785 - Added pcredemo. [But later removed - see 7 below.]
1786 - Fixed dftables problem and added an option.
1787 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1789 - Added readline support for pcretest.
1790 - Added an listing of the option settings after cmake has run.
1792 4. A user submitted a patch to Makefile that makes it easy to create
1793 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1794 Makefile.am that cause it to include this special target, without
1795 affecting anything else. Note that the same mingw target plus all
1796 the other distribution libraries and programs are now supported
1797 when configuring with CMake (see 6 below) instead of with
1798 Configure/Make.
1800 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1801 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1802 exported in the Windows port". It has not yet been confirmed that the patch
1803 solves the problem, but it does no harm.
1805 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1806 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1807 with CMake, and also correct the comment about stack recursion.
1809 7. Remove the automatic building of pcredemo from the ./configure system and
1810 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1811 of a program that users should build themselves after PCRE is installed, so
1812 building it automatically is not really right. What is more, it gave
1813 trouble in some build environments.
1815 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1818 Version 7.5 10-Jan-08
1819 ---------------------
1821 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1822 values in parens when parsing an RE using the C++ wrapper."
1824 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1825 Characters greater than 255 were excluded from the class instead of being
1826 included.
1828 3. The same bug as (2) above applied to negated POSIX classes such as
1829 [:^space:].
1831 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1832 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1833 I have changed it.
1835 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1836 first named subpattern) and a construct such as (?&a) would reference the
1837 first named subpattern whose name started with "a" (in other words, the
1838 length check was missing). Both these problems are fixed. "Subpattern name
1839 expected" is now given for (?&) (a zero-length name), and this patch also
1840 makes it give the same error for \k'' (previously it complained that that
1841 was a reference to a non-existent subpattern).
1843 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1844 this is right because (?- can be followed by option settings as well as by
1845 digits. I have, however, made the messages clearer.
1847 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1848 than the number used in the conditional) now cause a compile-time error.
1849 This is actually not compatible with Perl, which accepts such patterns, but
1850 treats the conditional as always being FALSE (as PCRE used to), but it
1851 seems to me that giving a diagnostic is better.
1853 8. Change "alphameric" to the more common word "alphanumeric" in comments
1854 and messages.
1856 9. Fix two occurrences of "backslash" in comments that should have been
1857 "backspace".
1859 10. Remove two redundant lines of code that can never be obeyed (their function
1860 was moved elsewhere).
1862 11. The program that makes PCRE's Unicode character property table had a bug
1863 which caused it to generate incorrect table entries for sequences of
1864 characters that have the same character type, but are in different scripts.
1865 It amalgamated them into a single range, with the script of the first of
1866 them. In other words, some characters were in the wrong script. There were
1867 thirteen such cases, affecting characters in the following ranges:
1869 U+002b0 - U+002c1
1870 U+0060c - U+0060d
1871 U+0061e - U+00612
1872 U+0064b - U+0065e
1873 U+0074d - U+0076d
1874 U+01800 - U+01805
1875 U+01d00 - U+01d77
1876 U+01d9b - U+01dbf
1877 U+0200b - U+0200f
1878 U+030fc - U+030fe
1879 U+03260 - U+0327f
1880 U+0fb46 - U+0fbb1
1881 U+10450 - U+1049d
1883 12. The -o option (show only the matching part of a line) for pcregrep was not
1884 compatible with GNU grep in that, if there was more than one match in a
1885 line, it showed only the first of them. It now behaves in the same way as
1886 GNU grep.
1888 13. If the -o and -v options were combined for pcregrep, it printed a blank
1889 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1890 does the same. The return code can be used to tell if there were any
1891 non-matching lines.
1893 14. Added --file-offsets and --line-offsets to pcregrep.
1895 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1896 infinitely looping recursion. The bug was that positive lookaheads were not
1897 being skipped when checking for a possible empty match (negative lookaheads
1898 and both kinds of lookbehind were skipped).
1900 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1901 inclusion of <windows.h> to before rather than after the definition of
1902 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1904 17. Specifying a possessive quantifier with a specific limit for a Unicode
1905 character property caused pcre_compile() to compile bad code, which led at
1906 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1907 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1908 caused the error; without that there was no problem.
1910 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1912 19. Added --enable-pcretest-libreadline.
1914 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1915 RE::GlobalReplace(). As a result, the number of replacements returned was
1916 double what it should be. I removed one of the increments, but Craig sent a
1917 later patch that removed the other one (the right fix) and added unit tests
1918 that check the return values (which was not done before).
1920 21. Several CMake things:
1922 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1923 the names libpcre and libpcreposix, not just pcre and pcreposix.
1925 (2) The above change means that pcretest and pcregrep are now correctly
1926 linked with the newly-built libraries, not previously installed ones.
1930 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1931 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1932 UTF-8 newline character). The key issue is that the pattern starts .*;
1933 this means that the match must be either at the beginning, or after a
1934 newline. The bug was in the code for advancing after a failed match and
1935 checking that the new position followed a newline. It was not taking
1936 account of UTF-8 characters correctly.
1938 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1939 character classes. PCRE was not treating the sequence [:...:] as a
1940 character class unless the ... were all letters. Perl, however, seems to
1941 allow any characters between [: and :], though of course it rejects as
1942 unknown any "names" that contain non-letters, because all the known class
1943 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1944 for example, whereas PCRE did not - it did not recognize a POSIX character
1945 class. This seemed a bit dangerous, so the code has been changed to be
1946 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1947 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1948 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1949 Perl does, and where it didn't before.
1951 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1952 Windows environments %n is disabled by default.
1955 Version 7.4 21-Sep-07
1956 ---------------------
1958 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1959 means that a class such as [\s] counted as "explicit reference to CR or
1960 LF". That isn't really right - the whole point of the change was to try to
1961 help when there was an actual mention of one of the two characters. So now
1962 the change happens only if \r or \n (or a literal CR or LF) character is
1963 encountered.
1965 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1966 of both had grown to the point where there were only 3 bits left.
1967 Fortunately, there was spare space in the data structure, and so I have
1968 moved the internal flags into a new 16-bit field to free up more option
1969 bits.
1971 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1972 but did not set the internal JCHANGED flag - either of these is enough to
1973 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1974 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1975 start sets both bits.
1977 4. Added options (at build time, compile time, exec time) to change \R from
1978 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1980 5. doc/pcresyntax.html was missing from the distribution.
1982 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1983 compatibility, even though it is no longer used.
1985 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1986 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1987 windows.h file is present (where different names are used). [This was
1988 reversed later after testing - see 16 below.]
1990 8. Changed all #include <config.h> to #include "config.h". There were also
1991 some further <pcre.h> cases that I changed to "pcre.h".
1993 9. When pcregrep was used with the --colour option, it missed the line ending
1994 sequence off the lines that it output.
1996 10. It was pointed out to me that arrays of string pointers cause lots of
1997 relocations when a shared library is dynamically loaded. A technique of
1998 using a single long string with a table of offsets can drastically reduce
1999 these. I have refactored PCRE in four places to do this. The result is
2000 dramatic:
2002 Originally: 290
2003 After changing UCP table: 187
2004 After changing error message table: 43
2005 After changing table of "verbs" 36
2006 After changing table of Posix names 22
2008 Thanks to the folks working on Gregex for glib for this insight.
2010 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2011 unicode-properties was also set.
2013 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2015 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2016 checked only for CRLF.
2018 14. Added casts to pcretest.c to avoid compiler warnings.
2020 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2022 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2023 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2024 entirely. This removes changes made in 7 above.
2026 17. The CMake files have been updated, and there is now more information about
2027 building with CMake in the NON-UNIX-USE document.
2030 Version 7.3 28-Aug-07
2031 ---------------------
2033 1. In the rejigging of the build system that eventually resulted in 7.1, the
2034 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2035 brackets there is not right, since it causes compilers to look for an
2036 installed pcre.h, not the version that is in the source that is being
2037 compiled (which of course may be different). I have changed it back to:
2039 #include "pcre.h"
2041 I have a vague recollection that the change was concerned with compiling in
2042 different directories, but in the new build system, that is taken care of
2043 by the VPATH setting the Makefile.
2045 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2046 when the subject happened to end in the byte 0x85 (e.g. if the last
2047 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2048 characters but of course it shouldn't be taken as a newline when it is part
2049 of another character. The bug was that, for an unlimited repeat of . in
2050 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2051 characters when looking for a newline.
2053 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2055 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2056 in debug output.
2058 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2059 long printing in the pcrecpp unittest when running under MinGW.
2061 6. ESC_K was left out of the EBCDIC table.
2063 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2064 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2065 limit also applies to "virtual nesting" when a pattern is recursive, and in
2066 this case 1000 isn't so big. I have been able to remove this limit at the
2067 expense of backing off one optimization in certain circumstances. Normally,
2068 when pcre_exec() would call its internal match() function recursively and
2069 immediately return the result unconditionally, it uses a "tail recursion"
2070 feature to save stack. However, when a subpattern that can match an empty
2071 string has an unlimited repetition quantifier, it no longer makes this
2072 optimization. That gives it a stack frame in which to save the data for
2073 checking that an empty string has been matched. Previously this was taken
2074 from the 1000-entry workspace that had been reserved. So now there is no
2075 explicit limit, but more stack is used.
2077 8. Applied Daniel's patches to solve problems with the import/export magic
2078 syntax that is required for Windows, and which was going wrong for the
2079 pcreposix and pcrecpp parts of the library. These were overlooked when this
2080 problem was solved for the main library.
2082 9. There were some crude static tests to avoid integer overflow when computing
2083 the size of patterns that contain repeated groups with explicit upper
2084 limits. As the maximum quantifier is 65535, the maximum group length was
2085 set at 30,000 so that the product of these two numbers did not overflow a
2086 32-bit integer. However, it turns out that people want to use groups that
2087 are longer than 30,000 bytes (though not repeat them that many times).
2088 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2089 made it possible to implement the integer overflow checks in a much more
2090 dynamic way, which I have now done. The artificial limitation on group
2091 length has been removed - we now have only the limit on the total length of
2092 the compiled pattern, which depends on the LINK_SIZE setting.
2094 10. Fixed a bug in the documentation for get/copy named substring when
2095 duplicate names are permitted. If none of the named substrings are set, the
2096 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2097 empty string.
2099 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2100 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2101 because the ] is interpreted as the first data character and the
2102 terminating ] is not found. PCRE has been made compatible with Perl in this
2103 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2104 cause memory overwriting.
2106 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2107 string has been matched (to stop an infinite loop). It was not recognizing
2108 a conditional subpattern that could match an empty string if that
2109 subpattern was within another subpattern. For example, it looped when
2110 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2111 condition was not nested. This bug has been fixed.
2113 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2114 past the start of the subject in the presence of bytes with the top bit
2115 set, for example "\x8aBCD".
2117 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2118 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2120 14. Optimized (?!) to (*FAIL).
2122 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2123 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2124 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2125 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2126 does: it's just the validity check that is more restrictive.
2128 16. Inserted checks for integer overflows during escape sequence (backslash)
2129 processing, and also fixed erroneous offset values for syntax errors during
2130 backslash processing.
2132 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2133 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2135 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2136 caused an overrun.
2138 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2139 something other than just ASCII characters) inside a group that had an
2140 unlimited repeat caused a loop at compile time (while checking to see
2141 whether the group could match an empty string).
2143 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2144 [\P{Any}] did so. (Error in the code for printing property names.)
2146 21. An orphan \E inside a character class could cause a crash.
2148 22. A repeated capturing bracket such as (A)? could cause a wild memory
2149 reference during compilation.
2151 23. There are several functions in pcre_compile() that scan along a compiled
2152 expression for various reasons (e.g. to see if it's fixed length for look
2153 behind). There were bugs in these functions when a repeated \p or \P was
2154 present in the pattern. These operators have additional parameters compared
2155 with \d, etc, and these were not being taken into account when moving along
2156 the compiled data. Specifically:
2158 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2159 length.
2161 (b) An item such as \pL+ within a repeated group could cause crashes or
2162 loops.
2164 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2165 "reference to non-existent subpattern" error.
2167 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2169 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2170 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2172 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2174 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2175 character were causing crashes (broken optimization).
2177 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2178 \p or \P) caused a compile-time loop.
2180 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2181 break. For example, the unstudied pattern [\r\n]A does not match the string
2182 "\r\nA" because change 7.0/46 below moves the current point on by two
2183 characters after failing to match at the start. However, the pattern \nA
2184 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2185 the same is true. There doesn't seem any very clean way out of this, but
2186 what I have chosen to do makes the common cases work: PCRE now takes note
2187 of whether there can be an explicit match for \r or \n anywhere in the
2188 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2189 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2190 pattern has explicit CR or LF references.
2192 29. Added (*CR) etc for changing newline setting at start of pattern.
2195 Version 7.2 19-Jun-07
2196 ---------------------
2198 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2199 which is apparently normally available under Windows.
2201 2. Re-jig the pcregrep tests with different newline settings in an attempt
2202 to make them independent of the local environment's newline setting.
2204 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2206 4. Some of the "internals" tests were previously cut out when the link size
2207 was not 2, because the output contained actual offsets. The recent new
2208 "Z" feature of pcretest means that these can be cut out, making the tests
2209 usable with all link sizes.
2211 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2212 stack recursion. This gives a massive performance boost under BSD, but just
2213 a small improvement under Linux. However, it saves one field in the frame
2214 in all cases.
2216 6. Added more features from the forthcoming Perl 5.10:
2218 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2219 recursion call. It refers to the nth most recently opened parentheses.
2221 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2222 to be opened parentheses.
2224 (c) Conditions that refer to capturing parentheses can be specified
2225 relatively, for example, (?(-2)... or (?(+3)...
2227 (d) \K resets the start of the current match so that everything before
2228 is not part of it.
2230 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2232 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2233 reference syntax.
2235 (g) (?| introduces a group in which the numbering of parentheses in each
2236 alternative starts with the same number.
2238 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2240 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2243 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2244 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2245 for detecting groups that can match an empty string.
2247 9. A pattern with a very large number of alternatives (more than several
2248 hundred) was running out of internal workspace during the pre-compile
2249 phase, where pcre_compile() figures out how much memory will be needed. A
2250 bit of new cunning has reduced the workspace needed for groups with
2251 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2252 workspace instead of running out of the 4096 that are available.
2254 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2256 11. Applied patch from Google to remove an optimization that didn't quite work.
2257 The report of the bug said:
2259 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2260 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2261 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2263 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2264 it matched the wrong number of bytes.
2267 Version 7.1 24-Apr-07
2268 ---------------------
2270 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2271 that is more "standard", making use of automake and other Autotools. There
2272 is some re-arrangement of the files and adjustment of comments consequent
2273 on this.
2275 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2276 for recursive directory scanning broke on some systems because the files
2277 are not scanned in any specific order and on different systems the order
2278 was different. A call to "sort" has been inserted into RunGrepTest for the
2279 approprate test as a short-term fix. In the longer term there may be an
2280 alternative.
2282 3. I had an email from Eric Raymond about problems translating some of PCRE's
2283 man pages to HTML (despite the fact that I distribute HTML pages, some
2284 people do their own conversions for various reasons). The problems
2285 concerned the use of low-level troff macros .br and .in. I have therefore
2286 removed all such uses from the man pages (some were redundant, some could
2287 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2288 HTML has been updated to handle .nf/.fi and to complain if it encounters
2289 .br or .in.
2291 4. Updated comments in configure.ac that get placed in config.h.in and also
2292 arranged for config.h to be included in the distribution, with the name
2293 config.h.generic, for the benefit of those who have to compile without
2294 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2296 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2297 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2298 makevp.bat for the current PCRE, using the additional files
2299 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2301 6. A Windows user reported a minor discrepancy with test 2, which turned out
2302 to be caused by a trailing space on an input line that had got lost in his
2303 copy. The trailing space was an accident, so I've just removed it.
2305 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2306 that is needed.
2308 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2309 as "const" (a) because they are and (b) because it helps the PHP
2310 maintainers who have recently made a script to detect big data structures
2311 in the php code that should be moved to the .rodata section. I remembered
2312 to update Builducptable as well, so it won't revert if ucptable.h is ever
2313 re-created.
2315 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2316 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2317 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2318 support is not required. This saves 1.5-2K of code, which is important in
2319 some applications.
2321 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2322 so as not to refer to the tables, even though these functions will never be
2323 called when UTF-8 support is disabled. Otherwise there are problems with a
2324 shared library.
2326 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2328 (a) It was defining its arguments as char * instead of void *.
2330 (b) It was assuming that all moves were upwards in memory; this was true
2331 a long time ago when I wrote it, but is no longer the case.
2333 The emulated memove() is provided for those environments that have neither
2334 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2335 is clearly not the case, as these two bugs were recently reported.
2337 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2338 and Detrail to create the HTML documentation, the .txt form of the man
2339 pages, and it removes trailing spaces from listed files. It also creates
2340 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2341 case, it wraps all the #defines with #ifndefs. This script should be run
2342 before "make dist".
2344 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2345 with Unicode property support.
2347 (a) For a maximizing quantifier, if the two different cases of the
2348 character were of different lengths in their UTF-8 codings (there are
2349 some cases like this - I found 11), and the matching function had to
2350 back up over a mixture of the two cases, it incorrectly assumed they
2351 were both the same length.
2353 (b) When PCRE was configured to use the heap rather than the stack for
2354 recursion during matching, it was not correctly preserving the data for
2355 the other case of a UTF-8 character when checking ahead for a match
2356 while processing a minimizing repeat. If the check also involved
2357 matching a wide character, but failed, corruption could cause an
2358 erroneous result when trying to check for a repeat of the original
2359 character.
2361 13. Some tidying changes to the testing mechanism:
2363 (a) The RunTest script now detects the internal link size and whether there
2364 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2365 values substituted by "configure". (The RunGrepTest script already did
2366 this for UTF-8.) The configure.ac script no longer substitutes the
2367 relevant variables.
2369 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2370 with length and offset values. This means that the output is different
2371 for different internal link sizes. Test 2 is skipped for link sizes
2372 other than 2 because of this, bypassing the problem. Unfortunately,
2373 there was also a test in test 3 (the locale tests) that used /B and
2374 failed for link sizes other than 2. Rather than cut the whole test out,
2375 I have added a new /Z option to pcretest that replaces the length and
2376 offset values with spaces. This is now used to make test 3 independent
2377 of link size. (Test 2 will be tidied up later.)
2379 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2380 segmentation fault instead of returning the appropriate error message.
2382 15. In multiline mode when the newline sequence was set to "any", the pattern
2383 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2384 This doesn't seem right; it now treats the CRLF combination as the line
2385 ending, and so does not match in that case. It's only a pattern such as ^$
2386 that would hit this one: something like ^ABC$ would have failed after \r
2387 and then tried again after \r\n.
2389 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2390 in an attempt to make files that differ only in their line terminators
2391 compare equal. This works on Linux.
2393 17. Under certain error circumstances pcregrep might try to free random memory
2394 as it exited. This is now fixed, thanks to valgrind.
2396 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2397 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2398 was because its rules for how to advance for /g after matching an empty
2399 string at the end of a line did not allow for this case. They now check for
2400 it specially.
2402 20. pcretest is supposed to handle patterns and data of any length, by
2403 extending its buffers when necessary. It was getting this wrong when the
2404 buffer for a data line had to be extended.
2406 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2407 CRLF as a newline sequence.
2409 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2410 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2411 I have nevertheless tidied it up.
2413 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2415 24. Added a man page for pcre-config.
2418 Version 7.0 19-Dec-06
2419 ---------------------
2421 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2422 moving to gcc 4.1.1.
2424 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2425 sys/time.h, which is documented as needed for this function. It doesn't
2426 seem to matter on Linux, but it showed up on some releases of OS X.
2428 3. It seems that there are systems where bytes whose values are greater than
2429 127 match isprint() in the "C" locale. The "C" locale should be the
2430 default when a C program starts up. In most systems, only ASCII printing
2431 characters match isprint(). This difference caused the output from pcretest
2432 to vary, making some of the tests fail. I have changed pcretest so that:
2434 (a) When it is outputting text in the compiled version of a pattern, bytes
2435 other than 32-126 are always shown as hex escapes.
2437 (b) When it is outputting text that is a matched part of a subject string,
2438 it does the same, unless a different locale has been set for the match
2439 (using the /L modifier). In this case, it uses isprint() to decide.
2441 4. Fixed a major bug that caused incorrect computation of the amount of memory
2442 required for a compiled pattern when options that changed within the
2443 pattern affected the logic of the preliminary scan that determines the
2444 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2445 that the computed length was too small. The symptoms of this bug were
2446 either the PCRE error "internal error: code overflow" from pcre_compile(),
2447 or a glibc crash with a message such as "pcretest: free(): invalid next
2448 size (fast)". Examples of patterns that provoked this bug (shown in
2449 pcretest format) are:
2451 /(?-x: )/x
2452 /(?x)(?-x: \s*#\s*)/
2453 /((?i)[\x{c0}])/8
2454 /(?i:[\x{c0}])/8
2456 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2457 is now done differently.
2459 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2460 wrapper classes; (b) implement a new function in the C++ scanner that is
2461 more efficient than the old way of doing things because it avoids levels of
2462 recursion in the regex matching; (c) add a paragraph to the documentation
2463 for the FullMatch() function.
2465 6. The escape sequence \n was being treated as whatever was defined as
2466 "newline". Not only was this contrary to the documentation, which states
2467 that \n is character 10 (hex 0A), but it also went horribly wrong when
2468 "newline" was defined as CRLF. This has been fixed.
2470 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2471 was being set to -1 for the "end of line" case (supposedly a value that no
2472 character can have). Though this value is never used (the check for end of
2473 line is "zero bytes in current character"), it caused compiler complaints.
2474 I've changed it to 0xffffffff.
2476 8. In pcre_version.c, the version string was being built by a sequence of
2477 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2478 string (as it is for production releases) called a macro with an empty
2479 argument. The C standard says the result of this is undefined. The gcc
2480 compiler treats it as an empty string (which was what was wanted) but it is
2481 reported that Visual C gives an error. The source has been hacked around to
2482 avoid this problem.
2484 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2485 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2486 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2487 of them did).
2489 10. Originally, pcretest opened its input and output without "b"; then I was
2490 told that "b" was needed in some environments, so it was added for release
2491 5.0 to both the input and output. (It makes no difference on Unix-like
2492 systems.) Later I was told that it is wrong for the input on Windows. I've
2493 now abstracted the modes into two macros, to make it easier to fiddle with
2494 them, and removed "b" from the input mode under Windows.
2496 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2498 12. Added -help and --help to pcretest as an official way of being reminded
2499 of the options.
2501 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2502 and pcrecpp.cc because they annoy compilers at high warning levels.
2504 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2506 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2507 scripts are not C programs :-) and which was not noticed because it works
2508 on Linux.
2510 16. pcretest is supposed to handle any length of pattern and data line (as one
2511 line or as a continued sequence of lines) by extending its input buffer if
2512 necessary. This feature was broken for very long pattern lines, leading to
2513 a string of junk being passed to pcre_compile() if the pattern was longer
2514 than about 50K.
2516 17. I have done a major re-factoring of the way pcre_compile() computes the
2517 amount of memory needed for a compiled pattern. Previously, there was code
2518 that made a preliminary scan of the pattern in order to do this. That was
2519 OK when PCRE was new, but as the facilities have expanded, it has become
2520 harder and harder to keep it in step with the real compile phase, and there
2521 have been a number of bugs (see for example, 4 above). I have now found a
2522 cunning way of running the real compile function in a "fake" mode that
2523 enables it to compute how much memory it would need, while actually only
2524 ever using a few hundred bytes of working memory and without too many
2525 tests of the mode. This should make future maintenance and development
2526 easier. A side effect of this work is that the limit of 200 on the nesting
2527 depth of parentheses has been removed (though this was never a serious
2528 limitation, I suspect). However, there is a downside: pcre_compile() now
2529 runs more slowly than before (30% or more, depending on the pattern). I
2530 hope this isn't a big issue. There is no effect on runtime performance.
2532 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2533 newline (only possible for the last line of a file) and it was a
2534 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2536 19. Added additional timing features to pcretest. (1) The -tm option now times
2537 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2538 separate command line item, by a number that specifies the number of
2539 repeats to use when timing. The default is 50000; this gives better
2540 precision, but takes uncomfortably long for very large patterns.
2542 20. Extended pcre_study() to be more clever in cases where a branch of a
2543 subpattern has no definite first character. For example, (a*|b*)[cd] would
2544 previously give no result from pcre_study(). Now it recognizes that the
2545 first character must be a, b, c, or d.
2547 21. There was an incorrect error "recursive call could loop indefinitely" if
2548 a subpattern (or the entire pattern) that was being tested for matching an
2549 empty string contained only one non-empty item after a nested subpattern.
2550 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2551 incorrectly, because the \d was being skipped in the check.
2553 22. The pcretest program now has a new pattern option /B and a command line
2554 option -b, which is equivalent to adding /B to every pattern. This causes
2555 it to show the compiled bytecode, without the additional information that
2556 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2557 is the same as /B/I).
2559 23. A new optimization is now able automatically to treat some sequences such
2560 as a*b as a*+b. More specifically, if something simple (such as a character
2561 or a simple class like \d) has an unlimited quantifier, and is followed by
2562 something that cannot possibly match the quantified thing, the quantifier
2563 is automatically "possessified".
2565 24. A recursive reference to a subpattern whose number was greater than 39
2566 went wrong under certain circumstances in UTF-8 mode. This bug could also
2567 have affected the operation of pcre_study().
2569 25. Realized that a little bit of performance could be had by replacing
2570 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2572 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2574 27. Possessive quantifiers such as a++ were previously implemented by turning
2575 them into atomic groups such as ($>a+). Now they have their own opcodes,
2576 which improves performance. This includes the automatically created ones
2577 from 23 above.
2579 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2580 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2581 the first matched character to be a colon. This applied both to named and
2582 numbered groups.
2584 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2586 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2587 building PCRE on Windows easier, so I have included it in the distribution.
2589 31. There is now a check in pcretest against a ridiculously large number being
2590 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2591 loop, the loop is abandoned.
2593 32. Forward references to subpatterns in conditions such as (?(2)...) where
2594 subpattern 2 is defined later cause pcre_compile() to search forwards in
2595 the pattern for the relevant set of parentheses. This search went wrong
2596 when there were unescaped parentheses in a character class, parentheses
2597 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2599 33. "Subroutine" calls and backreferences were previously restricted to
2600 referencing subpatterns earlier in the regex. This restriction has now
2601 been removed.
2603 34. Added a number of extra features that are going to be in Perl 5.10. On the
2604 whole, these are just syntactic alternatives for features that PCRE had
2605 previously implemented using the Python syntax or my own invention. The
2606 other formats are all retained for compatibility.
2608 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2609 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2610 also .NET compatible.
2612 (b) A recursion or subroutine call to a named group can now be defined as
2613 (?&name) as well as (?P>name).
2615 (c) A backreference to a named group can now be defined as \k<name> or
2616 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2617 5.10, are also .NET compatible.
2619 (d) A conditional reference to a named group can now use the syntax
2620 (?(<name>) or (?('name') as well as (?(name).
2622 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2623 groups (named and numbered) that are never evaluated inline, but can be
2624 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2625 is always false. There may be only one alternative in such a group.
2627 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2628 as the simple (?(R). The condition is true only if the most recent
2629 recursion is that of the given number or name. It does not search out
2630 through the entire recursion stack.
2632 (g) The escape \gN or \g{N} has been added, where N is a positive or
2633 negative number, specifying an absolute or relative reference.
2635 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2636 some "unreachable code" warnings.
2638 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2639 things, this adds five new scripts.
2641 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2642 There were also incompatibilities regarding the handling of \Q..\E inside
2643 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2644 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2646 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2647 matches an empty string, and forcibly breaks the loop. There were bugs in
2648 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2649 against aaaa the result was just "a" rather than "aaaa", for example. Two
2650 separate and independent bugs (that affected different cases) have been
2651 fixed.
2653 39. Refactored the code to abolish the use of different opcodes for small
2654 capturing bracket numbers. This is a tidy that I avoided doing when I
2655 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2656 The new approach is not only tidier, it makes it possible to reduce the
2657 memory needed to fix the previous bug (38).
2659 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2660 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2661 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2662 mode.
2664 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2665 report.
2667 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2668 copy construction and assignment in the C++ wrapper.
2670 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2671 couple of bugs that could have given wrong results in the "--newline=crlf"
2672 case.
2674 44. Added a number of casts and did some reorganization of signed/unsigned int
2675 variables following suggestions from Dair Grant. Also renamed the variable
2676 "this" as "item" because it is a C++ keyword.
2678 45. Arranged for dftables to add
2680 #include "pcre_internal.h"
2682 to pcre_chartables.c because without it, gcc 4.x may remove the array
2683 definition from the final binary if PCRE is built into a static library and
2684 dead code stripping is activated.
2686 46. For an unanchored pattern, if a match attempt fails at the start of a
2687 newline sequence, and the newline setting is CRLF or ANY, and the next two
2688 characters are CRLF, advance by two characters instead of one.
2691 Version 6.7 04-Jul-06
2692 ---------------------
2694 1. In order to handle tests when input lines are enormously long, pcretest has
2695 been re-factored so that it automatically extends its buffers when
2696 necessary. The code is crude, but this _is_ just a test program. The
2697 default size has been increased from 32K to 50K.
2699 2. The code in pcre_study() was using the value of the re argument before
2700 testing it for NULL. (Of course, in any sensible call of the function, it
2701 won't be NULL.)
2703 3. The memmove() emulation function in pcre_internal.h, which is used on
2704 systems that lack both memmove() and bcopy() - that is, hardly ever -
2705 was missing a "static" storage class specifier.
2707 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2708 containing an extended class (one that cannot be represented by a bitmap
2709 because it contains high-valued characters or Unicode property items, e.g.
2710 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2711 pattern, but PCRE should not loop if you do not (it no longer does).
2712 [Detail: two cases were found: (a) a repeated subpattern containing an
2713 extended class; (b) a recursive reference to a subpattern that followed a
2714 previous extended class. It wasn't skipping over the extended class
2715 correctly when UTF-8 mode was not set.]
2717 5. A negated single-character class was not being recognized as fixed-length
2718 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2719 compile error "lookbehind assertion is not fixed length".
2721 6. The RunPerlTest auxiliary script was showing an unexpected difference
2722 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2723 write a Perl script that can interpret lines of an input file either as
2724 byte characters or as UTF-8, which is what "perltest" was being required to
2725 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2726 can't do is switch easily at run time between having the "use utf8;" pragma
2727 or not. In the end, I fudged it by using the RunPerlTest script to insert
2728 "use utf8;" explicitly for the UTF-8 tests.
2730 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2731 the end of the subject string, contrary to the documentation and to what
2732 Perl does. This was true of both matching functions. Now it matches only at
2733 the start of the subject and immediately after *internal* newlines.
2735 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2736 a pointer to an int instead of a pointer to an unsigned long int. This
2737 caused problems on 64-bit systems.
2739 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2740 instance of the 'standard' template library not being so standard".
2742 10. There was no check on the number of named subpatterns nor the maximum
2743 length of a subpattern name. The product of these values is used to compute
2744 the size of the memory block for a compiled pattern. By supplying a very
2745 long subpattern name and a large number of named subpatterns, the size
2746 computation could be caused to overflow. This is now prevented by limiting
2747 the length of names to 32 characters, and the number of named subpatterns
2748 to 10,000.
2750 11. Subpatterns that are repeated with specific counts have to be replicated in
2751 the compiled pattern. The size of memory for this was computed from the
2752 length of the subpattern and the repeat count. The latter is limited to
2753 65535, but there was no limit on the former, meaning that integer overflow
2754 could in principle occur. The compiled length of a repeated subpattern is
2755 now limited to 30,000 bytes in order to prevent this.
2757 12. Added the optional facility to have named substrings with the same name.
2759 13. Added the ability to use a named substring as a condition, using the
2760 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2761 are numbers (not recommended). Forward references are permitted.
2763 14. Added forward references in named backreferences (if you see what I mean).
2765 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2766 pattern could run off the end of the subject. For example, the pattern
2767 "(?s)(.{1,5})"8 did this with the subject "ab".
2769 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2770 PCRE_CASELESS was set when matching characters that were quantified with ?
2771 or *.
2773 17. A character class other than a single negated character that had a minimum
2774 but no maximum quantifier - for example [ab]{6,} - was not handled
2775 correctly by pce_dfa_exec(). It would match only one character.
2777 18. A valid (though odd) pattern that looked like a POSIX character
2778 class but used an invalid character after [ (for example [[,abc,]]) caused
2779 pcre_compile() to give the error "Failed: internal error: code overflow" or
2780 in some cases to crash with a glibc free() error. This could even happen if
2781 the pattern terminated after [[ but there just happened to be a sequence of
2782 letters, a binary zero, and a closing ] in the memory that followed.
2784 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2785 over the years. Originally (before any Unicode support), just the bottom 8
2786 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2787 output from "man perlunicode" includes this:
2789 The regular expression compiler produces polymorphic opcodes. That
2790 is, the pattern adapts to the data and automatically switches to
2791 the Unicode character scheme when presented with Unicode data--or
2792 instead uses a traditional byte scheme when presented with byte
2793 data.
2795 Sadly, a wide octal escape does not cause a switch, and in a string with
2796 no other multibyte characters, these octal escapes are treated as before.
2797 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2798 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2799 Unicode string.
2801 I have not perpetrated such confusion in PCRE. Up till now, it took just
2802 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2803 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2804 translate to the appropriate multibyte character.
2806 29. Applied some refactoring to reduce the number of warnings from Microsoft
2807 and Borland compilers. This has included removing the fudge introduced
2808 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2809 a warning about an unused variable.
2811 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2812 characters since release 4.0, because Perl (from release 5.004) does not.
2813 [Or at least, is documented not to: some releases seem to be in conflict
2814 with the documentation.] However, when a pattern was studied with
2815 pcre_study() and all its branches started with \s, PCRE still included VT
2816 as a possible starting character. Of course, this did no harm; it just
2817 caused an unnecessary match attempt.
2819 22. Removed a now-redundant internal flag bit that recorded the fact that case
2820 dependency changed within the pattern. This was once needed for "required
2821 byte" processing, but is no longer used. This recovers a now-scarce options
2822 bit. Also moved the least significant internal flag bit to the most-
2823 significant bit of the word, which was not previously used (hangover from
2824 the days when it was an int rather than a uint) to free up another bit for
2825 the future.
2827 23. Added support for CRLF line endings as well as CR and LF. As well as the
2828 default being selectable at build time, it can now be changed at runtime
2829 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2830 specify that it is scanning data with non-default line endings.
2832 24. Changed the definition of CXXLINK to make it agree with the definition of
2833 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2835 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2836 recursions. This makes a big different to stack usage for some patterns.
2838 26. If a subpattern containing a named recursion or subroutine reference such
2839 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2840 the space required for the compiled pattern went wrong and gave too small a
2841 value. Depending on the environment, this could lead to "Failed: internal
2842 error: code overflow at offset 49" or "glibc detected double free or
2843 corruption" errors.
2845 27. Applied patches from Google (a) to support the new newline modes and (b) to
2846 advance over multibyte UTF-8 characters in GlobalReplace.
2848 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2849 difference for some implementation of PCRE in some Windows version.
2851 29. Added some extra testing facilities to pcretest:
2853 \q<number> in a data line sets the "match limit" value
2854 \Q<number> in a data line sets the "match recursion limt" value
2855 -S <number> sets the stack size, where <number> is in megabytes
2857 The -S option isn't available for Windows.
2860 Version 6.6 06-Feb-06
2861 ---------------------
2863 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2864 in pcreposix.h. I have copied the definition from pcre.h.
2866 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2867 because pcre.h is no longer a built file.
2869 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2870 not normally included in the compiled code.
2873 Version 6.5 01-Feb-06
2874 ---------------------
2876 1. When using the partial match feature with pcre_dfa_exec(), it was not
2877 anchoring the second and subsequent partial matches at the new starting
2878 point. This could lead to incorrect results. For example, with the pattern
2879 /1234/, partially matching against "123" and then "a4" gave a match.
2881 2. Changes to pcregrep:
2883 (a) All non-match returns from pcre_exec() were being treated as failures
2884 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2885 error message is output. Some extra information is given for the
2887 probably the only errors that are likely to be caused by users (by
2888 specifying a regex that has nested indefinite repeats, for instance).
2889 If there are more than 20 of these errors, pcregrep is abandoned.
2891 (b) A binary zero was treated as data while matching, but terminated the
2892 output line if it was written out. This has been fixed: binary zeroes
2893 are now no different to any other data bytes.
2895 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2896 used to set a locale for matching. The --locale=xxxx long option has
2897 been added (no short equivalent) to specify a locale explicitly on the
2898 pcregrep command, overriding the environment variables.
2900 (d) When -B was used with -n, some line numbers in the output were one less
2901 than they should have been.
2903 (e) Added the -o (--only-matching) option.
2905 (f) If -A or -C was used with -c (count only), some lines of context were
2906 accidentally printed for the final match.
2908 (g) Added the -H (--with-filename) option.
2910 (h) The combination of options -rh failed to suppress file names for files
2911 that were found from directory arguments.
2913 (i) Added the -D (--devices) and -d (--directories) options.
2915 (j) Added the -F (--fixed-strings) option.
2917 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2919 (l) Added the --colo(u)r option.
2921 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2922 is not present by default.
2924 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2925 items such as (?R) or (?1), when the recursion could match a number of
2926 alternatives. If it matched one of the alternatives, but subsequently,
2927 outside the recursion, there was a failure, the code tried to back up into
2928 the recursion. However, because of the way PCRE is implemented, this is not
2929 possible, and the result was an incorrect result from the match.
2931 In order to prevent this happening, the specification of recursion has
2932 been changed so that all such subpatterns are automatically treated as
2933 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2935 4. I had overlooked the fact that, in some locales, there are characters for
2936 which isalpha() is true but neither isupper() nor islower() are true. In
2937 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2938 and ordfeminine) are like this. This affected the treatment of \w and \W
2939 when they appeared in character classes, but not when they appeared outside
2940 a character class. The bit map for "word" characters is now created
2941 separately from the results of isalnum() instead of just taking it from the
2942 upper, lower, and digit maps. (Plus the underscore character, of course.)
2944 5. The above bug also affected the handling of POSIX character classes such as
2945 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2946 permanent tables. Instead, the bit maps for such a class were previously
2947 created as the appropriate unions of the upper, lower, and digit bitmaps.
2948 Now they are created by subtraction from the [[:word:]] class, which has
2949 its own bitmap.
2951 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2952 It is created by subtracting the vertical space characters (\x09, \x0a,
2953 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2954 subtraction was done in the overall bitmap for a character class, meaning
2955 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2956 be recognized. This bug has been fixed.
2958 7. Patches from the folks at Google:
2960 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2961 real life, but is still worth protecting against".
2963 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2964 regular expressions".
2966 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2967 have it.
2969 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2970 "configure" and the latter not, in order to fix a problem somebody had
2971 with compiling the Arg class on HP-UX.
2973 (e) Improve the error-handling of the C++ wrapper a little bit.
2975 (f) New tests for checking recursion limiting.
2977 8. The pcre_memmove() function, which is used only if the environment does not
2978 have a standard memmove() function (and is therefore rarely compiled),
2979 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2980 returning a result (though PCRE never actually uses the result).
2982 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2983 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2984 returned instead of calling malloc() with an overflowing number that would
2985 most likely cause subsequent chaos.
2987 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2989 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2990 with this option is matched, the nmatch and pmatch options of regexec() are
2991 ignored.
2993 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2994 provided in case anyone wants to the the POSIX interface with UTF-8
2995 strings.
2997 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2998 C++ linking (needed for some HP-UX environments).
3000 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3001 (unused parameter) and in the pcre_printint() function (omitted "default"
3002 switch label when the default is to do nothing).
3004 15. Added some code to make it possible, when PCRE is compiled as a C++
3005 library, to replace subject pointers for pcre_exec() with a smart pointer
3006 class, thus making it possible to process discontinuous strings.
3008 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3009 much the same function. They were added by different people who were trying
3010 to make PCRE easy to compile on non-Unix systems. It has been suggested
3011 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3012 for compiling on Windows systems. I have therefore replaced it with
3013 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3014 defaults to "extern" for C or "extern C" for C++, which works fine on
3015 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3016 SCOPE with something explicit in config.h. In addition:
3018 (a) pcreposix.h still had just "extern" instead of either of these macros;
3019 I have replaced it with PCRE_DATA_SCOPE.
3021 (b) Functions such as _pcre_xclass(), which are internal to the library,
3022 but external in the C sense, all had PCRE_EXPORT in their definitions.
3023 This is apparently wrong for the Windows case, so I have removed it.
3024 (It makes no difference on Unix-like systems.)
3026 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3027 of recursive calls to match(). This is different to MATCH_LIMIT because
3028 that limits the total number of calls to match(), not all of which increase
3029 the depth of recursion. Limiting the recursion depth limits the amount of
3030 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3031 when PCRE is compiled, and changed at run time. A patch from Google adds
3032 this functionality to the C++ interface.
3034 18. Changes to the handling of Unicode character properties:
3036 (a) Updated the table to Unicode 4.1.0.
3038 (b) Recognize characters that are not in the table as "Cn" (undefined).
3040 (c) I revised the way the table is implemented to a much improved format
3041 which includes recognition of ranges. It now supports the ranges that
3042 are defined in UnicodeData.txt, and it also amalgamates other
3043 characters into ranges. This has reduced the number of entries in the
3044 table from around 16,000 to around 3,000, thus reducing its size
3045 considerably. I realized I did not need to use a tree structure after
3046 all - a binary chop search is just as efficient. Having reduced the
3047 number of entries, I extended their size from 6 bytes to 8 bytes to
3048 allow for more data.
3050 (d) Added support for Unicode script names via properties such as \p{Han}.
3052 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3053 matching that character.
3055 20. When matching a repeated Unicode property with a minimum greater than zero,
3056 (for example \pL{2,}), PCRE could look past the end of the subject if it
3057 reached it while seeking the minimum number of characters. This could
3058 happen only if some of the characters were more than one byte long, because
3059 there is a check for at least the minimum number of bytes.
3061 21. Refactored the implementation of \p and \P so as to be more general, to
3062 allow for more different types of property in future. This has changed the
3063 compiled form incompatibly. Anybody with saved compiled patterns that use
3064 \p or \P will have to recompile them.
3066 22. Added "Any" and "L&" to the supported property types.
3068 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3069 but give a compile time error if the value is greater than 0xff.
3071 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3072 accidentally not being installed or uninstalled.
3074 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3075 made were to insert the current release number. This seemed silly, because
3076 it made things harder for people building PCRE on systems that don't run
3077 "configure". I have turned pcre.h into a distributed file, no longer built
3078 by "configure", with the version identification directly included. There is
3079 no longer a pcre.h.in file.
3081 However, this change necessitated a change to the pcre-config script as
3082 well. It is built from pcre-config.in, and one of the substitutions was the
3083 release number. I have updated configure.ac so that ./configure now finds
3084 the release number by grepping pcre.h.
3086 26. Added the ability to run the tests under valgrind.
3089 Version 6.4 05-Sep-05
3090 ---------------------
3092 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3093 "--" to be printed when multiple files were scanned, even when none of the
3094 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3095 consider it to be a bug, and have restored the previous behaviour.
3097 2. A couple of code tidies to get rid of compiler warnings.
3099 3. The pcretest program used to cheat by referring to symbols in the library
3100 whose names begin with _pcre_. These are internal symbols that are not
3101 really supposed to be visible externally, and in some environments it is
3102 possible to suppress them. The cheating is now confined to including
3103 certain files from the library's source, which is a bit cleaner.
3105 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3106 file's purpose clearer.
3108 5. Reorganized pcre_ucp_findchar().
3111 Version 6.3 15-Aug-05
3112 ---------------------
3114 1. The file libpcre.pc.in did not have general read permission in the tarball.
3116 2. There were some problems when building without C++ support:
3118 (a) If C++ support was not built, "make install" and "make test" still
3119 tried to test it.
3121 (b) There were problems when the value of CXX was explicitly set. Some
3122 changes have been made to try to fix these, and ...
3124 (c) --disable-cpp can now be used to explicitly disable C++ support.
3126 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3127 backslash in a target when C++ was disabled. This confuses some
3128 versions of "make", apparently. Using an intermediate variable solves
3129 this. (Same for CPP_LOBJ.)
3131 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3132 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3133 necessary on certain architectures.
3135 4. Added a setting of -export-symbols-regex to the link command to remove
3136 those symbols that are exported in the C sense, but actually are local
3137 within the library, and not documented. Their names all begin with
3138 "_pcre_". This is not a perfect job, because (a) we have to except some
3139 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3140 available (and never for static libraries). I have made a note to try to
3141 find a way round (a) in the future.
3144 Version 6.2 01-Aug-05
3145 ---------------------
3147 1. There was no test for integer overflow of quantifier values. A construction
3148 such as {1111111111111111} would give undefined results. What is worse, if
3149 a minimum quantifier for a parenthesized subpattern overflowed and became
3150 negative, the calculation of the memory size went wrong. This could have
3151 led to memory overwriting.
3153 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3155 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3156 operating environments where this matters.
3158 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3159 PCRE options from within the C++ wrapper.
3161 5. Named capturing subpatterns were not being correctly counted when a pattern
3162 was compiled. This caused two problems: (a) If there were more than 100
3163 such subpatterns, the calculation of the memory needed for the whole
3164 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3165 back references of the form \12, where the number was greater than 9, were
3166 not recognized as back references, even though there were sufficient
3167 previous subpatterns.
3169 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3170 versions of gcc, e.g. 2.95.4.
3173 Version 6.1 21-Jun-05
3174 ---------------------
3176 1. There was one reference to the variable "posix" in pcretest.c that was not
3177 surrounded by "#if !defined NOPOSIX".
3179 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3180 the cross-check on the old pcre_info() function, for the benefit of the
3181 cut-down version of PCRE that is currently imported into Exim.
3183 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3184 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3185 patterns that start (?i) but I don't think that matters. The use of (?i) is
3186 just an example; this all applies to the other options as well.
3188 4. Since libtool seems to echo the compile commands it is issuing, the output
3189 from "make" can be reduced a bit by putting "@" in front of each libtool
3190 compile command.
3192 5. Patch from the folks at Google for configure.in to be a bit more thorough
3193 in checking for a suitable C++ installation before trying to compile the
3194 C++ stuff. This should fix a reported problem when a compiler was present,
3195 but no suitable headers.
3197 6. The man pages all had just "PCRE" as their title. I have changed them to
3198 be the relevant file name. I have also arranged that these names are
3199 retained in the file doc/pcre.txt, which is a concatenation in text format
3200 of all the man pages except the little individual ones for each function.
3202 7. The NON-UNIX-USE file had not been updated for the different set of source
3203 files that come with release 6. I also added a few comments about the C++
3204 wrapper.
3207 Version 6.0 07-Jun-05
3208 ---------------------
3210 1. Some minor internal re-organization to help with my DFA experiments.
3212 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3213 didn't matter for the library itself when fully configured, but did matter
3214 when compiling without UCP support, or within Exim, where the ucp files are
3215 not imported.
3217 3. Refactoring of the library code to split up the various functions into
3218 different source modules. The addition of the new DFA matching code (see
3219 below) to a single monolithic source would have made it really too
3220 unwieldy, quite apart from causing all the code to be include in a
3221 statically linked application, when only some functions are used. This is
3222 relevant even without the DFA addition now that patterns can be compiled in
3223 one application and matched in another.
3225 The downside of splitting up is that there have to be some external
3226 functions and data tables that are used internally in different modules of
3227 the library but which are not part of the API. These have all had their
3228 names changed to start with "_pcre_" so that they are unlikely to clash
3229 with other external names.
3231 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3232 a different (DFA) algorithm. Although it is slower than the original
3233 function, it does have some advantages for certain types of matching
3234 problem.
3236 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3237 including restarting after a partial match.
3239 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3240 defined when compiling for Windows was sent to me. I have put it into the
3241 code, though I have no means of testing or verifying it.
3243 7. Added the pcre_refcount() auxiliary function.
3245 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3246 match before or at the first newline in the subject string. In pcretest,
3247 the /f option on a pattern can be used to set this.
3249 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3250 would behave wrongly. This has been present in PCRE since release 4.0.
3252 10. A number of changes to the pcregrep command:
3254 (a) Refactored how -x works; insert ^(...)$ instead of setting
3255 PCRE_ANCHORED and checking the length, in preparation for adding
3256 something similar for -w.
3258 (b) Added the -w (match as a word) option.
3260 (c) Refactored the way lines are read and buffered so as to have more
3261 than one at a time available.
3263 (d) Implemented a pcregrep test script.
3265 (e) Added the -M (multiline match) option. This allows patterns to match
3266 over several lines of the subject. The buffering ensures that at least
3267 8K, or the rest of the document (whichever is the shorter) is available
3268 for matching (and similarly the previous 8K for lookbehind assertions).
3270 (f) Changed the --help output so that it now says
3272 -w, --word-regex(p)
3274 instead of two lines, one with "regex" and the other with "regexp"
3275 because that confused at least one person since the short forms are the
3276 same. (This required a bit of code, as the output is generated
3277 automatically from a table. It wasn't just a text change.)
3279 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3280 option but starts with a hyphen. Could be a pattern or a path name
3281 starting with a hyphen, for instance.
3283 (h) "-" can be given as a file name to represent stdin.
3285 (i) When file names are being printed, "(standard input)" is used for
3286 the standard input, for compatibility with GNU grep. Previously
3287 "<stdin>" was used.
3289 (j) The option --label=xxx can be used to supply a name to be used for
3290 stdin when file names are being printed. There is no short form.
3292 (k) Re-factored the options decoding logic because we are going to add
3293 two more options that take data. Such options can now be given in four
3294 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3296 (l) Added the -A, -B, and -C options for requesting that lines of context
3297 around matches be printed.
3299 (m) Added the -L option to print the names of files that do not contain
3300 any matching lines, that is, the complement of -l.
3302 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3303 continue to scan other files.
3305 (o) The -s option was incorrectly implemented. For compatibility with other
3306 greps, it now suppresses the error message for a non-existent or non-
3307 accessible file (but not the return code). There is a new option called
3308 -q that suppresses the output of matching lines, which was what -s was
3309 previously doing.
3311 (p) Added --include and --exclude options to specify files for inclusion
3312 and exclusion when recursing.
3314 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3315 Hopefully, it now does.
3317 12. Missing cast in pcre_study().
3319 13. Added an "uninstall" target to the makefile.
3321 14. Replaced "extern" in the function prototypes in Makefile.in with
3322 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3323 world, but is set differently for Windows.
3325 15. Added a second compiling function called pcre_compile2(). The only
3326 difference is that it has an extra argument, which is a pointer to an
3327 integer error code. When there is a compile-time failure, this is set
3328 non-zero, in addition to the error test pointer being set to point to an
3329 error message. The new argument may be NULL if no error number is required
3330 (but then you may as well call pcre_compile(), which is now just a
3331 wrapper). This facility is provided because some applications need a
3332 numeric error indication, but it has also enabled me to tidy up the way
3333 compile-time errors are handled in the POSIX wrapper.
3335 16. Added VPATH=.libs to the makefile; this should help when building with one
3336 prefix path and installing with another. (Or so I'm told by someone who
3337 knows more about this stuff than I do.)
3339 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3340 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3341 match everything, including newlines. This is not POSIX-compatible, but
3342 somebody wanted the feature. From pcretest it can be activated by using
3343 both the P and the s flags.
3345 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3347 19. libpcre.pc was being incorrectly installed as executable.
3349 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3350 it now also looks for '\r' so that it will work unmodified on Windows.
3352 21. Added Google's contributed C++ wrapper to the distribution.
3354 22. Added some untidy missing memory free() calls in pcretest, to keep
3355 Electric Fence happy when testing.
3359 Version 5.0 13-Sep-04
3360 ---------------------
3362 1. Internal change: literal characters are no longer packed up into items
3363 containing multiple characters in a single byte-string. Each character
3364 is now matched using a separate opcode. However, there may be more than one
3365 byte in the character in UTF-8 mode.
3367 2. The pcre_callout_block structure has two new fields: pattern_position and
3368 next_item_length. These contain the offset in the pattern to the next match
3369 item, and its length, respectively.
3371 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3372 insertion of callouts before each pattern item. Added the /C option to
3373 pcretest to make use of this.
3375 4. On the advice of a Windows user, the lines
3377 #if defined(_WIN32) || defined(WIN32)
3378 _setmode( _fileno( stdout ), 0x8000 );
3379 #endif /* defined(_WIN32) || defined(WIN32) */
3381 have been added to the source of pcretest. This apparently does useful
3382 magic in relation to line terminators.
3384 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3385 for the benefit of those environments where the "b" makes a difference.
3387 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3388 to know about it. I have put a hack into configure.in that adds in code
3389 to set GCC=yes if CC=icc. This seems to end up at a point in the
3390 generated configure script that is early enough to affect the setting of
3391 compiler options, which is what is needed, but I have no means of testing
3392 whether it really works. (The user who reported this had patched the
3393 generated configure script, which of course I cannot do.)
3395 LATER: After change 22 below (new libtool files), the configure script
3396 seems to know about icc (and also ecc). Therefore, I have commented out
3397 this hack in configure.in.
3399 7. Added support for pkg-config (2 patches were sent in).
3401 8. Negated POSIX character classes that used a combination of internal tables
3402 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3403 [[:^ascii]]. Typically, they would match almost any characters. The other
3404 POSIX classes were not broken in this way.
3406 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3407 to find the match, as PCRE was deluded into thinking that the match had to
3408 start at the start point or following a newline. The same bug applied to
3409 patterns with negative forward assertions or any backward assertions
3410 preceding ".*" at the start, unless the pattern required a fixed first
3411 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3413 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3414 starting at the last subject character, bytes beyond the end of the subject
3415 string were read.
3417 11. Renamed the variable "class" as "classbits" to make life easier for C++
3418 users. (Previously there was a macro definition, but it apparently wasn't
3419 enough.)
3421 12. Added the new field "tables" to the extra data so that tables can be passed
3422 in at exec time, or the internal tables can be re-selected. This allows
3423 a compiled regex to be saved and re-used at a later time by a different
3424 program that might have everything at different addresses.
3426 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3427 -R library as well as a -L library.
3429 14. The debugging options of pcretest (-d on the command line or D on a
3430 pattern) showed incorrect output for anything following an extended class
3431 that contained multibyte characters and which was followed by a quantifier.
3433 15. Added optional support for general category Unicode character properties
3434 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3435 support. It adds about 90K to the size of the library. The meanings of the
3436 inbuilt class escapes such as \d and \s have NOT been changed.
3438 16. Updated pcredemo.c to include calls to free() to release the memory for the
3439 compiled pattern.
3441 17. The generated file chartables.c was being created in the source directory
3442 instead of in the building directory. This caused the build to fail if the
3443 source directory was different from the building directory, and was
3444 read-only.
3446 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3447 file. No doubt somebody will tell me if they don't make sense... Also added
3448 Dan Mooney's comments about building on OpenVMS.
3450 19. Added support for partial matching via the PCRE_PARTIAL option for
3451 pcre_exec() and the \P data escape in pcretest.
3453 20. Extended pcretest with 3 new pattern features:
3455 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3456 write the compiled pattern to the file whose name is "rest-of-line".
3457 This is a straight binary dump of the data, with the saved pointer to
3458 the character tables forced to be NULL. The study data, if any, is
3459 written too. After writing, pcretest reads a new pattern.
3461 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3462 compiled pattern from the given file. There must not be any
3463 occurrences of "<" in the file name (pretty unlikely); if there are,
3464 pcretest will instead treat the initial "<" as a pattern delimiter.
3465 After reading in the pattern, pcretest goes on to read data lines as
3466 usual.
3468 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3469 and 16-bit fields in a compiled pattern, to simulate a pattern that
3470 was compiled on a host of opposite endianness.
3472 21. The pcre-exec() function can now cope with patterns that were compiled on
3473 hosts of opposite endianness, with this restriction:
3475 As for any compiled expression that is saved and used later, the tables
3476 pointer field cannot be preserved; the extra_data field in the arguments
3477 to pcre_exec() should be used to pass in a tables address if a value
3478 other than the default internal tables were used at compile time.
3480 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3481 now diagnosed as an error. Previously, most of the time, a negative number
3482 would have been treated as zero, but if in addition "ovector" was passed as
3483 NULL, a crash could occur.
3485 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3486 new versions from the libtool 1.5 distribution (the last one is a copy of
3487 a file called libtool.m4). This seems to have fixed the need to patch
3488 "configure" to support Darwin 1.3 (which I used to do). However, I still
3489 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3490 workstation).
3492 24. Changed the PCRE licence to be the more standard "BSD" licence.
3495 Version 4.5 01-Dec-03
3496 ---------------------
3498 1. There has been some re-arrangement of the code for the match() function so
3499 that it can be compiled in a version that does not call itself recursively.
3500 Instead, it keeps those local variables that need separate instances for
3501 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3502 needs to "recurse". Keeping track of where control must go is done by means
3503 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3504 hide most of the details from the main code, and operates only if
3505 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3506 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3507 operating.
3509 To make it easier for callers to provide specially tailored get/free
3510 functions for this usage, two new functions, pcre_stack_malloc, and
3511 pcre_stack_free, are used. They are always called in strict stacking order,
3512 and the size of block requested is always the same.
3514 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3515 PCRE has been compiled to use the stack or the heap for recursion. The
3516 -C option of pcretest uses this to show which version is compiled.
3518 A new data escape \S, is added to pcretest; it causes the amounts of store
3519 obtained and freed by both kinds of malloc/free at match time to be added
3520 to the output.
3522 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3523 what's available on my current Linux desktop machine.
3525 3. When matching a UTF-8 string, the test for a valid string at the start has
3526 been extended. If start_offset is not zero, PCRE now checks that it points
3527 to a byte that is the start of a UTF-8 character. If not, it returns
3528 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3529 this is necessary because there may be backward assertions in the pattern.
3530 When matching the same subject several times, it may save resources to use
3531 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3533 4. The code for checking the validity of UTF-8 strings has been tightened so
3534 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3535 containing "overlong sequences".
3537 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3538 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3539 should have been "&", but it just so happened that all the cases this let
3540 through by mistake were picked up later in the function.
3542 6. I had used a variable called "isblank" - this is a C99 function, causing
3543 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3545 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3546 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3547 so the version is shown; (c) stop "make test" echoing "./RunTest".
3549 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3551 9. The prototype for memmove() for systems that don't have it was using
3552 size_t, but the inclusion of the header that defines size_t was later. I've
3553 moved the #includes for the C headers earlier to avoid this.
3555 10. Added some adjustments to the code to make it easier to compiler on certain
3556 special systems:
3558 (a) Some "const" qualifiers were missing.
3559 (b) Added the macro EXPORT before all exported functions; by default this
3560 is defined to be empty.
3561 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3562 that it reads its output file name as an argument instead of writing
3563 to the standard output and assuming this can be redirected.
3565 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3566 class containing characters with values greater than 255, PCRE compilation
3567 went into a loop.
3569 12. A recursive reference to a subpattern that was within another subpattern
3570 that had a minimum quantifier of zero caused PCRE to crash. For example,
3571 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3572 recursion. If the recursively-called subpattern itself had a zero repeat,
3573 that was OK.
3575 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3576 buffer into which it was copied (for escape processing) was still set at
3577 1024, so long lines caused crashes.
3579 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3580 "internal error: code overflow...". This applied to any character class
3581 that was followed by a possessive quantifier.
3583 15. Modified the Makefile to add libpcre.la as a prerequisite for
3584 libpcreposix.la because I was told this is needed for a parallel build to
3585 work.
3587 16. If a pattern that contained .* following optional items at the start was
3588 studied, the wrong optimizing data was generated, leading to matching
3589 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3590 matching string must start with a or b or c. The correct conclusion for
3591 this pattern is that a match can start with any character.
3594 Version 4.4 13-Aug-03
3595 ---------------------
3597 1. In UTF-8 mode, a character class containing characters with values between
3598 127 and 255 was not handled correctly if the compiled pattern was studied.
3599 In fixing this, I have also improved the studying algorithm for such
3600 classes (slightly).
3602 2. Three internal functions had redundant arguments passed to them. Removal
3603 might give a very teeny performance improvement.
3605 3. Documentation bug: the value of the capture_top field in a callout is *one
3606 more than* the number of the hightest numbered captured substring.
3608 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3609 in incorrectly linking with a previously installed version. They now link
3610 explicitly with libpcre.la.
3612 5. configure.in no longer needs to recognize Cygwin specially.
3614 6. A problem in pcre.in for Windows platforms is fixed.
3616 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3617 pcretest, it used to include the size of the study block as part of its
3618 output. Unfortunately, the structure contains a field that has a different
3619 size on different hardware architectures. This meant that the tests that
3620 showed this size failed. As the block is currently always of a fixed size,
3621 this information isn't actually particularly useful in pcretest output, so
3622 I have just removed it.
3624 8. Three pre-processor statements accidentally did not start in column 1.
3625 Sadly, there are *still* compilers around that complain, even though
3626 standard C has not required this for well over a decade. Sigh.
3628 9. In pcretest, the code for checking callouts passed small integers in the
3629 callout_data field, which is a void * field. However, some picky compilers
3630 complained about the casts involved for this on 64-bit systems. Now
3631 pcretest passes the address of the small integer instead, which should get
3632 rid of the warnings.
3634 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3635 both compile and run time, and gives an error if an invalid UTF-8 sequence
3636 is found. There is a option for disabling this check in cases where the
3637 string is known to be correct and/or the maximum performance is wanted.
3639 11. In response to a bug report, I changed one line in Makefile.in from
3641 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3642 to
3643 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3645 to look similar to other lines, but I have no way of telling whether this
3646 is the right thing to do, as I do not use Windows. No doubt I'll get told
3647 if it's wrong...
3650 Version 4.3 21-May-03
3651 ---------------------
3653 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3654 Makefile.
3656 2. Some refactoring to improve the quality of the code:
3658 (i) The utf8_table... variables are now declared "const".
3660 (ii) The code for \cx, which used the "case flipping" table to upper case
3661 lower case letters, now just substracts 32. This is ASCII-specific,
3662 but the whole concept of \cx is ASCII-specific, so it seems
3663 reasonable.
3665 (iii) PCRE was using its character types table to recognize decimal and
3666 hexadecimal digits in the pattern. This is silly, because it handles
3667 only 0-9, a-f, and A-F, but the character types table is locale-
3668 specific, which means strange things might happen. A private
3669 table is now used for this - though it costs 256 bytes, a table is
3670 much faster than multiple explicit tests. Of course, the standard
3671 character types table is still used for matching digits in subject
3672 strings against \d.
3674 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3675 ending in _t are). So I've renamed it as ESC_tee.
3677 3. The first argument for regexec() in the POSIX wrapper should have been
3678 defined as "const".
3680 4. Changed pcretest to use malloc() for its buffers so that they can be
3681 Electric Fenced for debugging.
3683 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3684 to read one or more bytes before the start of the subject string. Often this
3685 had no effect on PCRE's behaviour, but in some circumstances it could
3686 provoke a segmentation fault.
3688 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3689 to try to read one or more bytes before the start of the subject string.
3691 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3692 UTF-8 support could misbehave in various ways if the subject string
3693 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3694 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3695 back over UTF-8 characters.)
3698 Version 4.2 14-Apr-03
3699 ---------------------
3701 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3703 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3704 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3705 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3706 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3707 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3709 Note: automatic setting of the BUILD variables is not yet working
3710 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3711 compile-time but not at link-time
3712 [LINK]: use for linking executables only
3713 make different versions for Windows and non-Windows
3714 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3715 libraries
3716 [LINK_FOR_BUILD]: new variable
3717 [OBJEXT]: use throughout
3718 [EXEEXT]: use throughout
3719 <winshared>: new target
3720 <wininstall>: new target
3721 <dftables.o>: use native compiler
3722 <dftables>: use native linker
3723 <install>: handle Windows platform correctly
3724 <clean>: ditto
3725 <check>: ditto
3726 copy DLL to top builddir before testing
3728 As part of these changes, -no-undefined was removed again. This was reported
3729 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3730 in any case.
3732 3. Some tidies to get rid of compiler warnings:
3734 . In the match_data structure, match_limit was an unsigned long int, whereas
3735 match_call_count was an int. I've made them both unsigned long ints.
3737 . In pcretest the fact that a const uschar * doesn't automatically cast to
3738 a void * provoked a warning.
3740 . Turning on some more compiler warnings threw up some "shadow" variables
3741 and a few more missing casts.
3743 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3744 option, a class that contained a single character with a value between 128
3745 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3747 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3748 option, a class that contained several characters, but with at least one
3749 whose value was between 128 and 255 caused PCRE to crash.
3752 Version 4.1 12-Mar-03
3753 ---------------------
3755 1. Compiling with gcc -pedantic found a couple of places where casts were
3756 needed, and a string in dftables.c that was longer than standard compilers are
3757 required to support.
3759 2. Compiling with Sun's compiler found a few more places where the code could
3760 be tidied up in order to avoid warnings.
3762 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3763 first of these names is deprecated in the latest Autoconf in favour of the name
3764 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3765 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3766 analogy I have changed it to CFLAGS_FOR_BUILD.
3768 4. Added -no-undefined to the linking command in the Makefile, because this is
3769 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3770 linking step for the pcreposix library.
3772 5. PCRE was failing to diagnose the case of two named groups with the same
3773 name.
3775 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3776 literal character that is needed in the subject for a match, and scans along to
3777 ensure that it is present before embarking on the full matching process. This
3778 saves time in cases of nested unlimited repeats that are never going to match.
3779 Problem: the scan can take a lot of time if the subject is very long (e.g.
3780 megabytes), thus penalizing straightforward matches. It is now done only if the
3781 amount of subject to be scanned is less than 1000 bytes.
3783 7. A lesser problem with the same optimization is that it was recording the
3784 first character of an anchored pattern as "needed", thus provoking a search
3785 right along the subject, even when the first match of the pattern was going to
3786 fail. The "needed" character is now not set for anchored patterns, unless it
3787 follows something in the pattern that is of non-fixed length. Thus, it still
3788 fulfils its original purpose of finding quick non-matches in cases of nested
3789 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3792 Version 4.0 17-Feb-03
3793 ---------------------
3795 1. If a comment in an extended regex that started immediately after a meta-item
3796 extended to the end of string, PCRE compiled incorrect data. This could lead to
3797 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3799 2. Moved to autoconf 2.53 and libtool 1.4.2.
3801 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3802 the special perltest8 script is no longer needed - all the tests can be run
3803 from a single perltest script.
3805 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3806 by \s. It has now been removed in PCRE. This means it isn't recognized as
3807 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3808 class [:space:] *does* include VT, thereby creating a mess.
3810 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3811 space and tab.
3813 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3814 its new features into the main test script, reducing the number of scripts.
3816 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3817 were backward compatible, and made the (?i) apply to the whole pattern, as if
3818 /i were given. Now it behaves more logically, and applies the option setting
3819 only to what follows. PCRE has been changed to follow suit. However, if it
3820 finds options settings right at the start of the pattern, it extracts them into
3821 the global options, as before. Thus, they show up in the info data.
3823 8. Added support for the \Q...\E escape sequence. Characters in between are
3824 treated as literals. This is slightly different from Perl in that $ and @ are
3825 also handled as literals inside the quotes. In Perl, they will cause variable
3826 interpolation. Note the following examples:
3828 Pattern PCRE matches Perl matches
3830 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3831 \Qabc\$xyz\E abc\$xyz abc\$xyz
3832 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3834 For compatibility with Perl, \Q...\E sequences are recognized inside character
3835 classes as well as outside them.
3837 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3838 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3839 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3840 signed/unsigned warnings.
3842 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3843 option for pcretest, so I've replaced it by a simple function that does just
3844 that job.
3846 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3847 "pcregrep -".
3849 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3850 Java package. This provides some syntactic sugar for simple cases of what my
3851 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3852 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3853 item, you can use this simplified notation. Note that only makes sense with
3854 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3855 greediness, whatever the setting of the PCRE_UNGREEDY option.
3857 13. A change of greediness default within a pattern was not taking effect at
3858 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3859 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3860 was abstracted outside.
3862 14. PCRE now supports the \G assertion. It is true when the current matching
3863 position is at the start point of the match. This differs from \A when the
3864 starting offset is non-zero. Used with the /g option of pcretest (or similar
3865 code), it works in the same way as it does for Perl's /g option. If all
3866 alternatives of a regex begin with \G, the expression is anchored to the start
3867 match position, and the "anchored" flag is set in the compiled expression.
3869 15. Some bugs concerning the handling of certain option changes within patterns
3870 have been fixed. These applied to options other than (?ims). For example,
3871 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3872 been the other way round. Some of this was related to change 7 above.
3874 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3875 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3876 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3877 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3879 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3880 mode. Unlike ".", it always matches newline, whatever the setting of
3881 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3882 assertions. Perl allows it, but it doesn't (in general) work because it can't
3883 calculate the length of the lookbehind. At least, that's the case for Perl
3884 5.8.0 - I've been told they are going to document that it doesn't work in
3885 future.
3887 18. Added an error diagnosis for escapes that PCRE does not support: these are
3888 \L, \l, \N, \P, \p, \U, \u, and \X.
3890 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3891 reading past the end of the pattern in cases such as /[abcd/.
3893 20. PCRE was getting more memory than necessary for patterns with classes that
3894 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3896 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3897 compiling PCRE for use with Virtual Pascal.
3899 22. Small fix to the Makefile to make it work properly if the build is done
3900 outside the source tree.
3902 23. Added a new extension: a condition to go with recursion. If a conditional
3903 subpattern starts with (?(R) the "true" branch is used if recursion has
3904 happened, whereas the "false" branch is used only at the top level.
3906 24. When there was a very long string of literal characters (over 255 bytes
3907 without UTF support, over 250 bytes with UTF support), the computation of how
3908 much memory was required could be incorrect, leading to segfaults or other
3909 strange effects.
3911 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3912 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3913 there was a subsequent back reference to those brackets. This meant that, for
3914 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3915 possible to check for precisely this case. All we can do is abandon the
3916 optimization if .* occurs inside capturing brackets when there are any back
3917 references whatsoever. (See below for a better fix that came later.)
3919 26. The handling of the optimization for finding the first character of a
3920 non-anchored pattern, and for finding a character that is required later in the
3921 match were failing in some cases. This didn't break the matching; it just
3922 failed to optimize when it could. The way this is done has been re-implemented.
3924 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3926 28. Added a new feature that provides some of the functionality that Perl
3927 provides with (?{...}). The facility is termed a "callout". The way it is done
3928 in PCRE is for the caller to provide an optional function, by setting
3929 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3930 global variable. By default it is unset, which disables all calling out. To get
3931 the function called, the regex must include (?C) at appropriate points. This
3932 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3933 This provides a means of identifying different callout points. When PCRE
3934 reaches such a point in the regex, if pcre_callout has been set, the external
3935 function is called. It is provided with data in a structure called
3936 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3937 matching continues; if it returns a non-zero value, the match at the current
3938 point fails. However, backtracking will occur if possible. [This was changed
3939 later and other features added - see item 49 below.]
3941 29. pcretest is upgraded to test the callout functionality. It provides a
3942 callout function that displays information. By default, it shows the start of
3943 the match and the current position in the text. There are some new data escapes
3944 to vary what happens:
3946 \C+ in addition, show current contents of captured substrings
3947 \C- do not supply a callout function
3948 \C!n return 1 when callout number n is reached
3949 \C!n!m return 1 when callout number n is reached for the mth time
3951 30. If pcregrep was called with the -l option and just a single file name, it
3952 output "<stdin>" if a match was found, instead of the file name.
3954 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3955 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3956 pcre_exec(). This saves a malloc/free per call. The default value of
3957 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3958 when configuring.
3960 32. The default maximum size of a compiled pattern is 64K. There have been a
3961 few cases of people hitting this limit. The code now uses macros to handle the
3962 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3963 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3964 configuring. Tests 2 and 5 work only with 2-byte links because they output
3965 debugging information about compiled patterns.
3967 33. Internal code re-arrangements:
3969 (a) Moved the debugging function for printing out a compiled regex into
3970 its own source file (printint.c) and used #include to pull it into
3971 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3972 separate copies.
3974 (b) Defined the list of op-code names for debugging as a macro in
3975 internal.h so that it is next to the definition of the opcodes.
3977 (c) Defined a table of op-code lengths for simpler skipping along compiled
3978 code. This is again a macro in internal.h so that it is next to the
3979 definition of the opcodes.
3981 34. Added support for recursive calls to individual subpatterns, along the
3982 lines of Robin Houston's patch (but implemented somewhat differently).
3984 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3985 allow it to read and process whole directories in Win32. This code was
3986 contributed by Lionel Fourquaux; it has not been tested by me.
3988 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3989 used to name a group. Names consist of alphanumerics and underscores, and must
3990 be unique. Back references use the syntax (?P=name) and recursive calls use
3991 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3992 numbers. The function pcre_fullinfo() can be used after compilation to extract
3993 a name/number map. There are three relevant calls:
3995 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3996 PCRE_INFO_NAMECOUNT yields the number of entries
3997 PCRE_INFO_NAMETABLE yields a pointer to the map.
3999 The map is a vector of fixed-size entries. The size of each entry depends on
4000 the length of the longest name used. The first two bytes of each entry are the
4001 group number, most significant byte first. There follows the corresponding
4002 name, zero terminated. The names are in alphabetical order.
4004 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4005 case instead of 255. Making it the same both with and without UTF-8 support
4006 means that the same test output works with both.
4008 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4009 calling malloc() with a zero argument.
4011 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4012 optimization. I've improved things by keeping a bitmap of backreferences with
4013 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4014 fact referenced, the optimization can be applied. It is unlikely that a
4015 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4016 the match to follow \n) will appear inside brackets with a number greater than
4017 31, but if it does, any back reference > 31 suppresses the optimization.
4019 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4020 of disabling numbered capturing parentheses. Any opening parenthesis that is
4021 not followed by ? behaves as if it were followed by ?: but named parentheses
4022 can still be used for capturing (and they will acquire numbers in the usual
4023 way).
4025 41. Redesigned the return codes from the match() function into yes/no/error so
4026 that errors can be passed back from deep inside the nested calls. A malloc
4027 failure while inside a recursive subpattern call now causes the
4028 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4030 42. It is now possible to set a limit on the number of times the match()
4031 function is called in a call to pcre_exec(). This facility makes it possible to
4032 limit the amount of recursion and backtracking, though not in a directly
4033 obvious way, because the match() function is used in a number of different
4034 circumstances. The count starts from zero for each position in the subject
4035 string (for non-anchored patterns). The default limit is, for compatibility, a
4036 large number, namely 10 000 000. You can change this in two ways:
4038 (a) When configuring PCRE before making, you can use --with-match-limit=n
4039 to set a default value for the compiled library.
4041 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4042 a different value is set. See 45 below.
4044 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4046 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4047 of things that can be changed at compile time. The first argument specifies
4048 what is wanted and the second points to where the information is to be placed.
4049 The current list of available information is:
4053 The output is an integer that is set to one if UTF-8 support is available;
4054 otherwise it is set to zero.
4058 The output is an integer that it set to the value of the code that is used for
4059 newline. It is either LF (10) or CR (13).
4063 The output is an integer that contains the number of bytes used for internal
4064 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4068 The output is an integer that contains the threshold above which the POSIX
4069 interface uses malloc() for output vectors. See item 31 above.
4073 The output is an unsigned integer that contains the default limit of the number
4074 of match() calls in a pcre_exec() execution. See 42 above.
4076 44. pcretest has been upgraded by the addition of the -C option. This causes it
4077 to extract all the available output from the new pcre_config() function, and to
4078 output it. The program then exits immediately.
4080 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4081 order to support additional features. One way would have been to define
4082 pcre_exec2() (for example) with extra arguments, but this would not have been
4083 extensible, and would also have required all calls to the original function to
4084 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4085 is used for passing in "extra" data from pcre_study().
4087 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4088 contains the following fields:
4090 flags a bitmap indicating which of the following fields are set
4091 study_data opaque data from pcre_study()
4092 match_limit a way of specifying a limit on match() calls for a specific
4093 call to pcre_exec()
4094 callout_data data for callouts (see 49 below)
4096 The flag bits are also defined in pcre.h, and are
4102 The pcre_study() function now returns one of these new pcre_extra blocks, with
4103 the actual study data pointed to by the study_data field, and the
4104 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4105 before. That is, this change is entirely upwards-compatible and requires no
4106 change to existing code.
4108 If you want to pass in additional data to pcre_exec(), you can either place it
4109 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4110 block.
4112 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4113 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4114 times with different match limits, until it finds the minimum value needed for
4115 pcre_exec() to complete. The value is then output. This can be instructive; for
4116 most simple matches the number is quite small, but for pathological cases it
4117 gets very large very quickly.
4119 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4120 returns the size of the data block pointed to by the study_data field in a
4121 pcre_extra block, that is, the value that was passed as the argument to
4122 pcre_malloc() when PCRE was getting memory in which to place the information
4123 created by pcre_study(). The fourth argument should point to a size_t variable.
4124 pcretest has been extended so that this information is shown after a successful
4125 pcre_study() call when information about the compiled regex is being displayed.
4127 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4128 because what follows is always an absolute path. (Later: it turns out that this
4129 is more than cosmetic for MinGW, because it doesn't like empty path
4130 components.)
4132 49. Some changes have been made to the callout feature (see 28 above):
4134 (i) A callout function now has three choices for what it returns:
4136 0 => success, carry on matching
4137 > 0 => failure at this point, but backtrack if possible
4138 < 0 => serious error, return this value from pcre_exec()
4140 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4141 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4142 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4143 use by callout functions. It will never be used by PCRE itself.
4145 (ii) The pcre_extra structure (see 45 above) has a void * field called
4146 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4147 pcre_callout_block structure has a field of the same name. The contents of
4148 the field passed in the pcre_extra structure are passed to the callout
4149 function in the corresponding field in the callout block. This makes it
4150 easier to use the same callout-containing regex from multiple threads. For
4151 testing, the pcretest program has a new data escape
4153 \C*n pass the number n (may be negative) as callout_data
4155 If the callout function in pcretest receives a non-zero value as
4156 callout_data, it returns that value.
4158 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4159 there were some redundant $(CFLAGS) in commands that are now specified as
4160 $(LINK), which already includes $(CFLAGS).
4162 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4163 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4164 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4165 one-byte characters throughout. Note that case-insensitive matching applies
4166 only to characters whose values are less than 256. PCRE doesn't support the
4167 notion of cases for higher-valued characters.
4169 (i) A character class whose characters are all within 0-255 is handled as
4170 a bit map, and the map is inverted for negative classes. Previously, a
4171 character > 255 always failed to match such a class; however it should
4172 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4174 (ii) A negated character class with a single character < 255 is coded as
4175 "not this character" (OP_NOT). This wasn't working properly when the test
4176 character was multibyte, either singly or repeated.
4178 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4179 mode, for example: \x{100}{2,3}.
4181 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4182 singly or repeated) now correctly test multibyte characters. However,
4183 PCRE doesn't recognize any characters with values greater than 255 as
4184 digits, spaces, or word characters. Such characters always match \D, \S,
4185 and \W, and never match \d, \s, or \w.
4187 (v) Classes may now contain characters and character ranges with values
4188 greater than 255. For example: [ab\x{100}-\x{400}].
4190 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4191 PCRE in UTF-8 mode.
4193 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4194 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4195 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4196 value.)
4198 53. The single man page has become too large. I have therefore split it up into
4199 a number of separate man pages. These also give rise to individual HTML pages;
4200 these are now put in a separate directory, and there is an index.html page that
4201 lists them all. Some hyperlinking between the pages has been installed.
4203 54. Added convenience functions for handling named capturing parentheses.
4205 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4206 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4207 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4208 are faulted.
4210 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4211 calling configure. These values are used when compiling the dftables.c program
4212 which is run to generate the source of the default character tables. They
4213 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4214 you will need to set these values.
4216 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4219 Version 3.9 02-Jan-02
4220 ---------------------
4222 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4224 2. If --disable-static was given, the building process failed when trying to
4225 build pcretest and pcregrep. (For some reason it was using libtool to compile
4226 them, which is not right, as they aren't part of the library.)
4229 Version 3.8 18-Dec-01
4230 ---------------------
4232 1. The experimental UTF-8 code was completely screwed up. It was packing the
4233 bytes in the wrong order. How dumb can you get?
4236 Version 3.7 29-Oct-01
4237 ---------------------
4239 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4240 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4241 this didn't happen under Solaris 8, where I normally test things.
4243 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4244 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4245 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4246 this fix an hour or so after the initial 3.7 release.)
4249 Version 3.6 23-Oct-01
4250 ---------------------
4252 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4253 offsets passed as NULL with zero offset count.
4255 2. The config.guess and config.sub files had not been updated when I moved to
4256 the latest autoconf.
4259 Version 3.5 15-Aug-01
4260 ---------------------
4262 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4263 had been forgotten.
4265 2. By using declared but undefined structures, we can avoid using "void"
4266 definitions in pcre.h while keeping the internal definitions of the structures
4267 private.
4269 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4270 user point of view, this means that both static and shared libraries are built
4271 by default, but this can be individually controlled. More of the work of
4272 handling this static/shared cases is now inside libtool instead of PCRE's make
4273 file.
4275 4. The pcretest utility is now installed along with pcregrep because it is
4276 useful for users (to test regexs) and by doing this, it automatically gets
4277 relinked by libtool. The documentation has been turned into a man page, so
4278 there are now .1, .txt, and .html versions in /doc.
4280 5. Upgrades to pcregrep:
4281 (i) Added long-form option names like gnu grep.
4282 (ii) Added --help to list all options with an explanatory phrase.
4283 (iii) Added -r, --recursive to recurse into sub-directories.
4284 (iv) Added -f, --file to read patterns from a file.
4286 6. pcre_exec() was referring to its "code" argument before testing that
4287 argument for NULL (and giving an error if it was NULL).
4289 7. Upgraded Makefile.in to allow for compiling in a different directory from
4290 the source directory.
4292 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4293 options bits, the pointer it was passed was to an int instead of to an unsigned
4294 long int. This mattered only on 64-bit systems.
4296 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4297 generated) instead of pcre.in, which it its source. Also made the same change
4298 in several of the .c files.
4300 10. A new release of gcc defines printf() as a macro, which broke pcretest
4301 because it had an ifdef in the middle of a string argument for printf(). Fixed
4302 by using separate calls to printf().
4304 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4305 script, to force use of CR or LF instead of \n in the source. On non-Unix
4306 systems, the value can be set in config.h.
4308 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4309 absolute limit. Changed the text of the error message to make this clear, and
4310 likewise updated the man page.
4312 13. The limit of 99 on the number of capturing subpatterns has been removed.
4313 The new limit is 65535, which I hope will not be a "real" limit.
4316 Version 3.4 22-Aug-00
4317 ---------------------
4319 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4321 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4324 Version 3.3 01-Aug-00
4325 ---------------------
4327 1. If an octal character was given, but the value was greater than \377, it
4328 was not getting masked to the least significant bits, as documented. This could
4329 lead to crashes in some systems.
4331 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4332 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4334 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4335 These just pass their arguments on to (pcre_free)(), but they are provided
4336 because some uses of PCRE bind it to non-C systems that can call its functions,
4337 but cannot call free() or pcre_free() directly.
4339 4. Add "make test" as a synonym for "make check". Corrected some comments in
4340 the Makefile.
4342 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4343 Makefile.
4345 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4346 command called pgrep for grepping around the active processes.
4348 7. Added the beginnings of support for UTF-8 character strings.
4350 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4351 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4352 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4353 out for the ar command.)
4356 Version 3.2 12-May-00
4357 ---------------------
4359 This is purely a bug fixing release.
4361 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4362 of ZA. This was just one example of several cases that could provoke this bug,
4363 which was introduced by change 9 of version 2.00. The code for breaking
4364 infinite loops after an iteration that matches an empty string was't working
4365 correctly.
4367 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4368 when matched against abbab (for example). After matching an empty string, it
4369 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4370 caused it to match further down the string than it should.
4372 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4373 was there because it doesn't seem to be needed, and it causes trouble on some
4374 systems, as it is not a Standard C header. It has been removed.
4376 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4377 were reported on the Macintosh. The changes were from
4379 while ((c = *(++ptr)) != 0 && c != '\n');
4380 to
4381 while ((c = *(++ptr)) != 0 && c != '\n') ;
4383 Totally extraordinary, but if that's what it takes...
4385 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4386 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4387 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4388 assumes the way PCRE uses memmove() (always moving upwards).
4390 6. PCRE is being used in one environment where strchr() is not available. There
4391 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4392 faster code anyway.
4395 Version 3.1 09-Feb-00
4396 ---------------------
4398 The only change in this release is the fixing of some bugs in Makefile.in for
4399 the "install" target:
4401 (1) It was failing to install pcreposix.h.
4403 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4406 Version 3.0 01-Feb-00
4407 ---------------------
4409 1. Add support for the /+ modifier to perltest (to output $` like it does in
4410 pcretest).
4412 2. Add support for the /g modifier to perltest.
4414 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4415 matches null strings.
4417 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4418 pattern. Perl treats empty patterns specially - it reuses the most recent
4419 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4420 effect.
4422 5. The POSIX interface was broken in that it was just handing over the POSIX
4423 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4424 required a bigger vector, with some working space on the end. This means that
4425 the POSIX wrapper now has to get and free some memory, and copy the results.
4427 6. Added some simple autoconf support, placing the test data and the
4428 documentation in separate directories, re-organizing some of the
4429 information files, and making it build pcre-config (a GNU standard). Also added
4430 libtool support for building PCRE as a shared library, which is now the
4431 default.
4433 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4434 09 are not valid octal constants. Single digits will be used for minor values
4435 less than 10.
4437 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4438 existing programs that set these in the POSIX interface can use PCRE without
4439 modification.
4441 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4442 return all that pcre_info() returns, plus additional data. The pcre_info()
4443 function is retained for compatibility, but is considered to be obsolete.
4445 10. Added experimental recursion feature (?R) to handle one common case that
4446 Perl 5.6 will be able to do with (?p{...}).
4448 11. Added support for POSIX character classes like [:alpha:], which Perl is
4449 adopting.
4452 Version 2.08 31-Aug-99
4453 ----------------------
4455 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4456 trying to match at the startoffset position, but instead was moving forward to
4457 the next newline as if a previous match had failed.
4459 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4460 and could get into a loop if a null string was matched other than at the start
4461 of the subject.
4463 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4464 be distinguished at compile time, and for completeness also added PCRE_DATE.
4466 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4467 in GnuWin32 environments.
4470 Version 2.07 29-Jul-99
4471 ----------------------
4473 1. The documentation is now supplied in plain text form and HTML as well as in
4474 the form of man page sources.
4476 2. C++ compilers don't like assigning (void *) values to other pointer types.
4477 In particular this affects malloc(). Although there is no problem in Standard
4478 C, I've put in casts to keep C++ compilers happy.
4480 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4481 should be (const char *).
4483 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4484 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4485 However, I haven't made this a standard facility. The documentation doesn't
4486 mention it, and the Makefile doesn't support it.
4488 5. The Makefile now contains an "install" target, with editable destinations at
4489 the top of the file. The pcretest program is not installed.
4491 6. pgrep -V now gives the PCRE version number and date.
4493 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4494 causing the entire string to be ignored, instead of just the last character.
4496 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4497 non-matching string, it can take a very, very long time, even for strings of
4498 quite modest length, because of the nested recursion. PCRE now does better in
4499 some of these cases. It does this by remembering the last required literal
4500 character in the pattern, and pre-searching the subject to ensure it is present
4501 before running the real match. In other words, it applies a heuristic to detect
4502 some types of certain failure quickly, and in the above example, if presented
4503 with a string that has no trailing " it gives "no match" very quickly.
4505 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4506 other alternatives are tried instead.
4509 Version 2.06 09-Jun-99
4510 ----------------------
4512 1. Change pcretest's output for amount of store used to show just the code
4513 space, because the remainder (the data block) varies in size between 32-bit and
4514 64-bit systems.
4516 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4517 start matching at. This allows lookbehinds to work when searching for multiple
4518 occurrences in a string.
4520 3. Added additional options to pcretest for testing multiple occurrences:
4522 /+ outputs the rest of the string that follows a match
4523 /g loops for multiple occurrences, using the new startoffset argument
4524 /G loops for multiple occurrences by passing an incremented pointer
4526 4. PCRE wasn't doing the "first character" optimization for patterns starting
4527 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4528 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4529 the letter 'x'. On long subject strings, this gives a significant speed-up.
4532 Version 2.05 21-Apr-99
4533 ----------------------
4535 1. Changed the type of magic_number from int to long int so that it works
4536 properly on 16-bit systems.
4538 2. Fixed a bug which caused patterns starting with .* not to work correctly
4539 when the subject string contained newline characters. PCRE was assuming
4540 anchoring for such patterns in all cases, which is not correct because .* will
4541 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4542 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4543 must be retried after every newline in the subject.
4546 Version 2.04 18-Feb-99
4547 ----------------------
4549 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4550 computation of the store needed to hold the pattern was incorrect (too large).
4551 If such patterns were nested a few deep, this could multiply and become a real
4552 problem.
4554 2. Added /M option to pcretest to show the memory requirement of a specific
4555 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4557 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4558 compiled in such a way that the backtracking after subsequent failure was
4559 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4560 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4563 Version 2.03 02-Feb-99
4564 ----------------------
4566 1. Fixed typo and small mistake in man page.
4568 2. Added 4th condition (GPL supersedes if conflict) and created separate
4569 LICENCE file containing the conditions.
4571 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4572 Perl, that is the internal \ allows the delimiter to be included in the
4573 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4574 the final delimiter, add \ to the end of the pattern (to test the error).
4576 4. Added the convenience functions for extracting substrings after a successful
4577 match. Updated pcretest to make it able to test these functions.
4580 Version 2.02 14-Jan-99
4581 ----------------------
4583 1. Initialized the working variables associated with each extraction so that
4584 their saving and restoring doesn't refer to uninitialized store.
4586 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4587 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4588 fix the problem.
4590 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4591 calls, and wasn't printing the correct value for compiling calls. Increased the
4592 default value of LOOPREPEAT, and the number of significant figures in the
4593 times.
4595 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4597 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4598 a building problem on Windows NT with a FAT file system.
4601 Version 2.01 21-Oct-98
4602 ----------------------
4604 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4605 to character tables built by pcre_maketables() in the current locale. If NULL
4606 is passed, the default tables are used.
4609 Version 2.00 24-Sep-98
4610 ----------------------
4612 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4613 it any more.
4615 2. Allow quantification of (?>) groups, and make it work correctly.
4617 3. The first character computation wasn't working for (?>) groups.
4619 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4620 end of the subject) and add 5.005's \z, which really does match only at the
4621 very end of the subject.
4623 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4625 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4626 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4627 localized options. All options to pcre_study() were also removed.
4629 7. Add other new features from 5.005:
4631 $(?<= positive lookbehind
4632 $(?<! negative lookbehind
4633 (?imsx-imsx) added the unsetting capability
4634 such a setting is global if at outer level; local otherwise
4635 (?imsx-imsx:) non-capturing groups with option setting
4636 (?(cond)re|re) conditional pattern matching
4638 A backreference to itself in a repeated group matches the previous
4639 captured string.
4641 8. General tidying up of studying (both automatic and via "study")
4642 consequential on the addition of new assertions.
4644 9. As in 5.005, unlimited repeated groups that could match an empty substring
4645 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4646 runtime if any iteration does actually match an empty substring.
4648 10. Include the RunTest script in the distribution.
4650 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4651 discrepancies, some of which were old and were also with respect to 5.004. They
4652 have now been fixed.
4655 Version 1.09 28-Apr-98
4656 ----------------------
4658 1. A negated single character class followed by a quantifier with a minimum
4659 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4660 program crashes, or just wrong answers. This did not apply to negated classes
4661 containing more than one character, or to minima other than one.
4664 Version 1.08 27-Mar-98
4665 ----------------------
4667 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4669 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4670 latter must appear before anything that relies on it in the pattern.
4673 Version 1.07 16-Feb-98
4674 ----------------------
4676 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4677 repeat of a potentially empty string).
4680 Version 1.06 23-Jan-98
4681 ----------------------
4683 1. Added Markus Oberhumer's little patches for C++.
4685 2. Literal strings longer than 255 characters were broken.
4688 Version 1.05 23-Dec-97
4689 ----------------------
4691 1. Negated character classes containing more than one character were failing if
4692 PCRE_CASELESS was set at run time.
4695 Version 1.04 19-Dec-97
4696 ----------------------
4698 1. Corrected the man page, where some "const" qualifiers had been omitted.
4700 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4701 input syntax.
4703 3. Fixed memory leak which occurred when a regex with back references was
4704 matched with an offsets vector that wasn't big enough. The temporary memory
4705 that is used in this case wasn't being freed if the match failed.
4707 4. Tidied pcretest to ensure it frees memory that it gets.
4709 5. Temporary memory was being obtained in the case where the passed offsets
4710 vector was exactly big enough.
4712 6. Corrected definition of offsetof() from change 5 below.
4714 7. I had screwed up change 6 below and broken the rules for the use of
4715 setjmp(). Now fixed.
4718 Version 1.03 18-Dec-97
4719 ----------------------
4721 1. A erroneous regex with a missing opening parenthesis was correctly
4722 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4723 on some systems.
4725 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4726 it was reported that one broken compiler failed on the former because "code" is
4727 also an independent variable.
4729 3. The erroneous regex a[]b caused an array overrun reference.
4731 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4732 fail on data ending with that character. (It was going on too far, and checking
4733 the next character, typically a binary zero.) This was specific to the
4734 optimized code for single-character negative classes.
4736 5. Added a contributed patch from the TIN world which does the following:
4738 + Add an undef for memmove, in case the the system defines a macro for it.
4740 + Add a definition of offsetof(), in case there isn't one. (I don't know
4741 the reason behind this - offsetof() is part of the ANSI standard - but
4742 it does no harm).
4744 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4745 most of the places where whitespace preceded '#'. I have given up and
4746 allowed the remaining 2 cases to be at the margin.
4748 + Rename some variables in pcre to eliminate shadowing. This seems very
4749 pedantic, but does no harm, of course.
4751 6. Moved the call to setjmp() into its own function, to get rid of warnings
4752 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4754 7. Constructs such as \d{8,} were compiling into the equivalent of
4755 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4756 outcome, but in this particular case used more store than had been allocated,
4757 which caused the bug to be discovered because it threw up an internal error.
4759 8. The debugging code in both pcre and pcretest for outputting the compiled
4760 form of a regex was going wrong in the case of back references followed by
4761 curly-bracketed repeats.
4764 Version 1.02 12-Dec-97
4765 ----------------------
4767 1. Typos in pcre.3 and comments in the source fixed.
4769 2. Applied a contributed patch to get rid of places where it used to remove
4770 'const' from variables, and fixed some signed/unsigned and uninitialized
4771 variable warnings.
4773 3. Added the "runtest" target to Makefile.
4775 4. Set default compiler flag to -O2 rather than just -O.
4778 Version 1.01 19-Nov-97
4779 ----------------------
4781 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4782 like /([ab]*)*/, that is, for classes with more than one character in them.
4784 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4785 as /((?>a*))*/ (a PCRE_EXTRA facility).
4788 Version 1.00 18-Nov-97
4789 ----------------------
4791 1. Added compile-time macros to support systems such as SunOS4 which don't have
4792 memmove() or strerror() but have other things that can be used instead.
4794 2. Arranged that "make clean" removes the executables.
4797 Version 0.99 27-Oct-97
4798 ----------------------
4800 1. Fixed bug in code for optimizing classes with only one character. It was
4801 initializing a 32-byte map regardless, which could cause it to run off the end
4802 of the memory it had got.
4804 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4807 Version 0.98 22-Oct-97
4808 ----------------------
4810 1. Fixed bug in code for handling temporary memory usage when there are more
4811 back references than supplied space in the ovector. This could cause segfaults.
4814 Version 0.97 21-Oct-97
4815 ----------------------
4817 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4819 2. Optimized negated single characters not to use a bit map.
4821 3. Brought error texts together as macro definitions; clarified some of them;
4822 fixed one that was wrong - it said "range out of order" when it meant "invalid
4823 escape sequence".
4825 4. Changed some char * arguments to const char *.
4827 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4829 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4830 pcretest.
4833 Version 0.96 16-Oct-97
4834 ----------------------
4836 1. Added a simple "pgrep" utility to the distribution.
4838 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4839 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4840 where "ddd" means "one or more decimal digits".
4842 3. Fixed serious bug. If a pattern had a back reference, but the call to
4843 pcre_exec() didn't supply a large enough ovector to record the related
4844 identifying subpattern, the match always failed. PCRE now remembers the number
4845 of the largest back reference, and gets some temporary memory in which to save
4846 the offsets during matching if necessary, in order to ensure that
4847 backreferences always work.
4849 4. Increased the compatibility with Perl in a number of ways:
4851 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4852 to request this handling. The option can be set at compile or exec time.
4854 (b) $ matches before a terminating newline by default; an option
4855 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4856 mode). The option can be set at compile or exec time.
4858 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4859 the same as Perl's. If the decimal number it represents is less than 10
4860 or there aren't that many previous left capturing parentheses, an octal
4861 escape is read. Inside a character class, it's always an octal escape,
4862 even if it is a single digit.
4864 (d) An escaped but undefined alphabetic character is taken as a literal,
4865 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4866 escapes.
4868 (e) {0} is now permitted. (The previous item is removed from the compiled
4869 pattern).
4871 5. Changed all the names of code files so that the basic parts are no longer
4872 than 10 characters, and abolished the teeny "globals.c" file.
4874 6. Changed the handling of character classes; they are now done with a 32-byte
4875 bit map always.
4877 7. Added the -d and /D options to pcretest to make it possible to look at the
4878 internals of compilation without having to recompile pcre.
4881 Version 0.95 23-Sep-97
4882 ----------------------
4884 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4885 \x20 at the start of a run of normal characters. These were being treated as
4886 real characters, instead of the source characters being re-checked.
4889 Version 0.94 18-Sep-97
4890 ----------------------
4892 1. The functions are now thread-safe, with the caveat that the global variables
4893 containing pointers to malloc() and free() or alternative functions are the
4894 same for all threads.
4896 2. Get pcre_study() to generate a bitmap of initial characters for non-
4897 anchored patterns when this is possible, and use it if passed to pcre_exec().
4900 Version 0.93 15-Sep-97
4901 ----------------------
4903 1. /(b)|(:+)/ was computing an incorrect first character.
4905 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4906 but not actually doing anything yet.
4908 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4909 as Perl does (e.g. [-az] or [az-]).
4911 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4912 all possible positions.
4914 5. Split up into different modules to avoid including unneeded functions in a
4915 compiled binary. However, compile and exec are still in one module. The "study"
4916 function is split off.
4918 6. The character tables are now in a separate module whose source is generated
4919 by an auxiliary program - but can then be edited by hand if required. There are
4920 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4921 toupper() in the code.
4923 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4924 make them global. Abolish the function for setting them, as the caller can now
4925 set them directly.
4928 Version 0.92 11-Sep-97
4929 ----------------------
4931 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4932 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4934 2. Caseless matching was not working in character classes if the characters in
4935 the pattern were in upper case.
4937 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4939 4. Make PCRE_ANCHORED public and accept as a compile option.
4941 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4942 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4943 pass them.
4945 6. Give an error if bad option bits passed at compile or run time.
4947 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4948 pcretest to cause it to pass that flag.
4950 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4951 options, and the first character, if set.
4953 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4956 Version 0.91 10-Sep-97
4957 ----------------------
4959 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4960 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4962 2. PCRE was looping on encountering an indefinitely repeated back reference to
4963 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4964 Perl does - treats the match as successful.
4966 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5