ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1050 - (show annotations)
Sun Sep 30 18:20:10 2012 UTC (7 years, 2 months ago) by chpe
File size: 237787 byte(s)
Error occurred while calculating annotation data.
unicode: Update to Unicode 6.2
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 Version 8.31 06-July-2012
122 -------------------------
124 1. Fixing a wrong JIT test case and some compiler warnings.
126 2. Removed a bashism from the RunTest script.
128 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
129 to unsigned type, result still unsigned" that was given by an MS compiler
130 on encountering the code "-sizeof(xxx)".
132 4. Partial matching support is added to the JIT compiler.
134 5. Fixed several bugs concerned with partial matching of items that consist
135 of more than one character:
137 (a) /^(..)\1/ did not partially match "aba" because checking references was
138 done on an "all or nothing" basis. This also applied to repeated
139 references.
141 (b) \R did not give a hard partial match if \r was found at the end of the
142 subject.
144 (c) \X did not give a hard partial match after matching one or more
145 characters at the end of the subject.
147 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
148 a partial match for the string "\r".
150 (e) When newline was set to CRLF, the metacharacter "." did not recognize
151 a partial match for a CR character at the end of the subject string.
153 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
154 running pcretest, the text "(JIT)" added to the output whenever JIT is
155 actually used to run the match.
157 7. Individual JIT compile options can be set in pcretest by following -s+[+]
158 or /S+[+] with a digit between 1 and 7.
160 8. OP_NOT now supports any UTF character not just single-byte ones.
162 9. (*MARK) control verb is now supported by the JIT compiler.
164 10. The command "./RunTest list" lists the available tests without actually
165 running any of them. (Because I keep forgetting what they all are.)
169 12. Applied a (slightly modified) user-supplied patch that improves performance
170 when the heap is used for recursion (compiled with --disable-stack-for-
171 recursion). Instead of malloc and free for each heap frame each time a
172 logical recursion happens, frames are retained on a chain and re-used where
173 possible. This sometimes gives as much as 30% improvement.
175 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
176 call.
178 14. As documented, (*COMMIT) is now confined to within a positive assertion.
180 15. It is now possible to link pcretest with libedit as an alternative to
181 libreadline.
183 16. (*COMMIT) control verb is now supported by the JIT compiler.
185 17. The Unicode data tables have been updated to Unicode 6.1.0.
187 18. Added --file-list option to pcregrep.
189 19. Added binary file support to pcregrep, including the -a, --binary-files,
190 -I, and --text options.
192 20. The madvise function is renamed for posix_madvise for QNX compatibility
193 reasons. Fixed by Giuseppe D'Angelo.
195 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
197 22. Changed the option for creating symbolic links for 16-bit man pages from
198 -s to -sf so that re-installing does not cause issues.
200 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
202 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
203 matches in certain environments (the workspace was not being correctly
204 retained). Also added to pcre_dfa_exec() a simple plausibility check on
205 some of the workspace data at the beginning of a restart.
207 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
208 was not doing so when it should - probably a typo introduced by SVN 528
209 (change 8.10/14).
211 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
212 \w+ when the character tables indicated that \x{c4} was a word character.
213 There were several related cases, all because the tests for doing a table
214 lookup were testing for characters less than 127 instead of 255.
216 27. If a pattern contains capturing parentheses that are not used in a match,
217 their slots in the ovector are set to -1. For those that are higher than
218 any matched groups, this happens at the end of processing. In the case when
219 there were back references that the ovector was too small to contain
220 (causing temporary malloc'd memory to be used during matching), and the
221 highest capturing number was not used, memory off the end of the ovector
222 was incorrectly being set to -1. (It was using the size of the temporary
223 memory instead of the true size.)
225 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
226 ovector size, it uses memory at the end of the block that it has got.
228 29. Check for an overlong MARK name and give an error at compile time. The
229 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
231 30. JIT compiler update.
233 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
234 Rill for the patch.
236 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
238 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
240 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
241 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
242 was enabled.
244 35. Improve JIT code generation for greedy plus quantifier.
246 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
247 1 to "aa" instead of to an empty string. The bug affected repeated groups
248 that could potentially match an empty string.
250 37. Optimizing single character iterators in JIT.
252 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
253 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
254 codepoints that are too big for the mode are faulted, and in a UTF mode,
255 disallowed codepoints are also faulted.
257 39. If PCRE was compiled with UTF support, in three places in the DFA
258 matcher there was code that should only have been obeyed in UTF mode, but
259 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
260 processing when bytes with values greater than 127 were present. In 16-bit
261 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
262 both cases the values are those that cannot be the first data item in a UTF
263 character. The three items that might have provoked this were recursions,
264 possessively repeated groups, and atomic groups.
266 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
267 and pcregrep, because some OS require shared objects to be explicitly
268 passed to ld, causing the link step to fail if they are not.
270 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
271 mode, patterns that started with \h* or \R* might be incorrectly matched.
274 Version 8.30 04-February-2012
275 -----------------------------
277 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
278 name is defined in ctype.h.
280 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
281 only in quite long subpatterns.
283 3. Removed the function pcre_info(), which has been obsolete and deprecated
284 since it was replaced by pcre_fullinfo() in February 2000.
286 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
287 match a (*MARK), and the match failed at the start of the subject, a
288 reference to memory before the start of the subject could occur. This bug
289 was introduced by fix 17 of release 8.21.
291 5. A reference to an unset group with zero minimum repetition was giving
292 totally wrong answers (in non-JavaScript-compatibility mode). For example,
293 /(another)?(\1?)test/ matched against "hello world test". This bug was
294 introduced in release 8.13.
296 6. Add support for 16-bit character strings (a large amount of work involving
297 many changes and refactorings).
299 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
300 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
301 from a file.
303 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
304 rounding is not applied in this particular case).
306 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
307 if they appear, or are escaped, in patterns.
309 10. Get rid of a number of -Wunused-but-set-variable warnings.
311 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
312 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
313 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
314 also returns the mark "x". This bug applied to capturing parentheses,
315 non-capturing parentheses, and atomic parentheses. It also applied to some
316 assertions.
318 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
319 information out of configure.ac instead of relying on pcre.h.generic, which
320 is not stored in the repository.
322 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
323 -lreadline.
325 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
327 15. Applied Graycode's patch to put the top-level frame on the stack rather
328 than the heap when not using the stack for recursion. This gives a
329 performance improvement in many cases when recursion is not deep.
331 16. Experimental code added to "pcretest -C" to output the stack frame size.
334 Version 8.21 12-Dec-2011
335 ------------------------
337 1. Updating the JIT compiler.
339 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
340 are added as well.
342 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
343 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
344 calling _pcre_jit_exec. Some extra comments are added.
346 4. (*MARK) settings inside atomic groups that do not contain any capturing
347 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
348 was introduced by change 18 for 8.20.
350 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
351 ECMA-262 standard.
353 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
354 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
355 This bug was probably introduced by change 9 of 8.13.
357 7. While fixing 6 above, I noticed that a number of other items were being
358 incorrectly rejected as "not fixed length". This arose partly because newer
359 opcodes had not been added to the fixed-length checking code. I have (a)
360 corrected the bug and added tests for these items, and (b) arranged for an
361 error to occur if an unknown opcode is encountered while checking for fixed
362 length instead of just assuming "not fixed length". The items that were
363 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
364 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
365 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
367 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
368 being incorrectly compiled and would have given unpredicatble results.
370 9. A possessively repeated subpattern with minimum repeat count greater than
371 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
372 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
373 the first (A) could occur when it should not.
375 10. Add a cast and remove a redundant test from the code.
377 11. JIT should use pcre_malloc/pcre_free for allocation.
379 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
380 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
381 is anything other than /usr, -L is still shown).
383 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
385 14. Perl does not support \N without a following name in a [] class; PCRE now
386 also gives an error.
388 15. If a forward reference was repeated with an upper limit of around 2000,
389 it caused the error "internal error: overran compiling workspace". The
390 maximum number of forward references (including repeats) was limited by the
391 internal workspace, and dependent on the LINK_SIZE. The code has been
392 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
393 the default depends on LINK_SIZE. There is a new upper limit (for safety)
394 of around 200,000 forward references. While doing this, I also speeded up
395 the filling in of repeated forward references.
397 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
398 incorrectly expecting the subject to contain another "a" after the start.
400 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
401 in the match, the SKIP should be ignored. This was not happening; instead
402 the SKIP was being treated as NOMATCH. For patterns such as
403 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
404 tested.
406 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
407 now much more compatible with Perl, in particular in cases where the result
408 is a non-match for a non-anchored pattern. For example, if
409 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
410 "m", where previously it did not return a name. A side effect of this
411 change is that for partial matches, the last encountered mark name is
412 returned, as for non matches. A number of tests that were previously not
413 Perl-compatible have been moved into the Perl-compatible test files. The
414 refactoring has had the pleasing side effect of removing one argument from
415 the match() function, thus reducing its stack requirements.
417 19. If the /S+ option was used in pcretest to study a pattern using JIT,
418 subsequent uses of /S (without +) incorrectly behaved like /S+.
420 21. Retrieve executable code size support for the JIT compiler and fixing
421 some warnings.
423 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
424 not work when the shorter character appeared right at the end of the
425 subject string.
427 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
428 systems.
430 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
431 output it when the /M option is used in pcretest.
433 25. The CheckMan script was not being included in the distribution. Also, added
434 an explicit "perl" to run Perl scripts from the PrepareRelease script
435 because this is reportedly needed in Windows.
437 26. If study data was being save in a file and studying had not found a set of
438 "starts with" bytes for the pattern, the data written to the file (though
439 never used) was taken from uninitialized memory and so caused valgrind to
440 complain.
442 27. Updated RunTest.bat as provided by Sheri Pierce.
444 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
446 29. Computation of memory usage for the table of capturing group names was
447 giving an unnecessarily large value.
450 Version 8.20 21-Oct-2011
451 ------------------------
453 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
454 a POSIX class. After further experiments with Perl, which convinced me that
455 Perl has bugs and confusions, a closing square bracket is no longer allowed
456 in a POSIX name. This bug also affected patterns with classes that started
457 with full stops.
459 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
460 captured substring, but while checking the failing first alternative,
461 substring 1 is temporarily captured. If the output vector supplied to
462 pcre_exec() was not big enough for this capture, the yield of the function
463 was still zero ("insufficient space for captured substrings"). This cannot
464 be totally fixed without adding another stack variable, which seems a lot
465 of expense for a edge case. However, I have improved the situation in cases
466 such as /(a)(b)x|abc/ matched against "abc", where the return code
467 indicates that fewer than the maximum number of slots in the ovector have
468 been set.
470 3. Related to (2) above: when there are more back references in a pattern than
471 slots in the output vector, pcre_exec() uses temporary memory during
472 matching, and copies in the captures as far as possible afterwards. It was
473 using the entire output vector, but this conflicts with the specification
474 that only 2/3 is used for passing back captured substrings. Now it uses
475 only the first 2/3, for compatibility. This is, of course, another edge
476 case.
478 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
479 main code base, and can be used by building with --enable-jit. When this is
480 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
481 runtime --no-jit option is given.
483 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
484 ovector, the return from the function was zero, implying that there were
485 other matches that did not fit. The correct "exactly full" value is now
486 returned.
488 6. If a subpattern that was called recursively or as a subroutine contained
489 (*PRUNE) or any other control that caused it to give a non-standard return,
490 invalid errors such as "Error -26 (nested recursion at the same subject
491 position)" or even infinite loops could occur.
493 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
494 computing the minimum length on reaching *ACCEPT, and so ended up with the
495 wrong value of 1 rather than 0. Further investigation indicates that
496 computing a minimum subject length in the presence of *ACCEPT is difficult
497 (think back references, subroutine calls), and so I have changed the code
498 so that no minimum is registered for a pattern that contains *ACCEPT.
500 8. If (*THEN) was present in the first (true) branch of a conditional group,
501 it was not handled as intended. [But see 16 below.]
503 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
504 Sheri Pierce.
506 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
507 the first byte in a match must be "a".
509 11. Change 17 for 8.13 increased the recursion depth for patterns like
510 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
511 pattern contains any instances of (*THEN). If it does not, the old
512 optimizations are restored. It would be nice to do this on a per-group
513 basis, but at the moment that is not feasible.
515 12. In some environments, the output of pcretest -C is CRLF terminated. This
516 broke RunTest's code that checks for the link size. A single white space
517 character after the value is now allowed for.
519 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
520 For "fr", it uses the Windows-specific input and output files.
522 14. If (*THEN) appeared in a group that was called recursively or as a
523 subroutine, it did not work as intended. [But see next item.]
525 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
526 pattern fragments (but not containing any | characters). If A and B are
527 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
528 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
529 D. In other words, Perl considers parentheses that do not contain any |
530 characters to be part of a surrounding alternative, whereas PCRE was
531 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
532 differently. PCRE now behaves in the same way as Perl, except in the case
533 of subroutine/recursion calls such as (?1) which have in any case always
534 been different (but PCRE had them first :-).
536 16. Related to 15 above: Perl does not treat the | in a conditional group as
537 creating alternatives. Such a group is treated in the same way as an
538 ordinary group without any | characters when processing (*THEN). PCRE has
539 been changed to match Perl's behaviour.
541 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
542 RunGrepTest script failed.
544 18. Change 22 for version 13 caused atomic groups to use more stack. This is
545 inevitable for groups that contain captures, but it can lead to a lot of
546 stack use in large patterns. The old behaviour has been restored for atomic
547 groups that do not contain any capturing parentheses.
549 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
550 suppress the check for a minimum subject length at run time. (If it was
551 given to pcre_exec() or pcre_dfa_exec() it did work.)
553 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
554 fail to work when decoding hex characters in data strings in EBCDIC
555 environments.
557 21. It appears that in at least one Mac OS environment, the isxdigit() function
558 is implemented as a macro that evaluates to its argument more than once,
559 contravening the C 90 Standard (I haven't checked a later standard). There
560 was an instance in pcretest which caused it to go wrong when processing
561 \x{...} escapes in subject strings. The has been rewritten to avoid using
562 things like p++ in the argument of isxdigit().
565 Version 8.13 16-Aug-2011
566 ------------------------
568 1. The Unicode data tables have been updated to Unicode 6.0.0.
570 2. Two minor typos in pcre_internal.h have been fixed.
572 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
573 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
574 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
576 4. There were a number of related bugs in the code for matching backrefences
577 caselessly in UTF-8 mode when codes for the characters concerned were
578 different numbers of bytes. For example, U+023A and U+2C65 are an upper
579 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
580 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
581 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
582 2-byte code at the end of the subject (it thought there wasn't enough data
583 left).
585 5. Comprehensive information about what went wrong is now returned by
586 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
587 as the output vector has at least 2 elements. The offset of the start of
588 the failing character and a reason code are placed in the vector.
590 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
591 now returned is for the first byte of the failing character, instead of the
592 last byte inspected. This is an incompatible change, but I hope it is small
593 enough not to be a problem. It makes the returned offset consistent with
594 pcre_exec() and pcre_dfa_exec().
596 7. pcretest now gives a text phrase as well as the error number when
597 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
598 failure, the offset and reason code are output.
600 8. When \R was used with a maximizing quantifier it failed to skip backwards
601 over a \r\n pair if the subsequent match failed. Instead, it just skipped
602 back over a single character (\n). This seems wrong (because it treated the
603 two characters as a single entity when going forwards), conflicts with the
604 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
605 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
606 has been changed.
608 9. Some internal refactoring has changed the processing so that the handling
609 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
610 time (the PCRE_DOTALL option was changed this way some time ago: version
611 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
612 which was always a bit of a fudge. It also means that there is one less
613 argument for the match() function, which reduces its stack requirements
614 slightly. This change also fixes an incompatibility with Perl: the pattern
615 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
617 10. More internal refactoring has drastically reduced the number of recursive
618 calls to match() for possessively repeated groups such as (abc)++ when
619 using pcre_exec().
621 11. While implementing 10, a number of bugs in the handling of groups were
622 discovered and fixed:
624 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
625 (a|)*(?1) gave a compile-time internal error.
626 ((a|)+)+ did not notice that the outer group could match an empty string.
627 (^a|^)+ was not marked as anchored.
628 (.*a|.*)+ was not marked as matching at start or after a newline.
630 12. Yet more internal refactoring has removed another argument from the match()
631 function. Special calls to this function are now indicated by setting a
632 value in a variable in the "match data" data block.
634 13. Be more explicit in pcre_study() instead of relying on "default" for
635 opcodes that mean there is no starting character; this means that when new
636 ones are added and accidentally left out of pcre_study(), testing should
637 pick them up.
639 14. The -s option of pcretest has been documented for ages as being an old
640 synonym of -m (show memory usage). I have changed it to mean "force study
641 for every regex", that is, assume /S for every regex. This is similar to -i
642 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
643 using it. It makes it easier to run collections of tests with and without
644 study enabled, and thereby test pcre_study() more easily. All the standard
645 tests are now run with and without -s (but some patterns can be marked as
646 "never study" - see 20 below).
648 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
649 restoration of the capturing data to the outer values was not happening
650 correctly.
652 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
653 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
654 pattern had matched an empty string, and so incorrectly returned a no
655 match.
657 17. There was optimizing code for the last branch of non-capturing parentheses,
658 and also for the obeyed branch of a conditional subexpression, which used
659 tail recursion to cut down on stack usage. Unfortunately, now that there is
660 the possibility of (*THEN) occurring in these branches, tail recursion is
661 no longer possible because the return has to be checked for (*THEN). These
662 two optimizations have therefore been removed. [But see 8.20/11 above.]
664 18. If a pattern containing \R was studied, it was assumed that \R always
665 matched two bytes, thus causing the minimum subject length to be
666 incorrectly computed because \R can also match just one byte.
668 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
669 was incorrectly computed.
671 20. If /S is present twice on a test pattern in pcretest input, it now
672 *disables* studying, thereby overriding the use of -s on the command line
673 (see 14 above). This is necessary for one or two tests to keep the output
674 identical in both cases.
676 21. When (*ACCEPT) was used in an assertion that matched an empty string and
677 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
679 22. When an atomic group that contained a capturing parenthesis was
680 successfully matched, but the branch in which it appeared failed, the
681 capturing was not being forgotten if a higher numbered group was later
682 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
683 group 1 to "a", when in fact it should be unset. This applied to multi-
684 branched capturing and non-capturing groups, repeated or not, and also to
685 positive assertions (capturing in negative assertions does not happen
686 in PCRE) and also to nested atomic groups.
688 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
689 subject after a captured substring, to make it easier to tell which of a
690 number of identical substrings has been captured.
692 24. The way atomic groups are processed by pcre_exec() has been changed so that
693 if they are repeated, backtracking one repetition now resets captured
694 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
695 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
696 "aaa". Previously, it would have been "a". As part of this code
697 refactoring, the way recursive calls are handled has also been changed.
699 25. If an assertion condition captured any substrings, they were not passed
700 back unless some other capturing happened later. For example, if
701 (?(?=(a))a) was matched against "a", no capturing was returned.
703 26. When studying a pattern that contained subroutine calls or assertions,
704 the code for finding the minimum length of a possible match was handling
705 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
706 group 1 called group 2 while simultaneously a separate group 2 called group
707 1). A stack overflow occurred in this case. I have fixed this by limiting
708 the recursion depth to 10.
710 27. Updated RunTest.bat in the distribution to the version supplied by Tom
711 Fortmann. This supports explicit test numbers on the command line, and has
712 argument validation and error reporting.
714 28. An instance of \X with an unlimited repeat could fail if at any point the
715 first character it looked at was a mark character.
717 29. Some minor code refactoring concerning Unicode properties and scripts
718 should reduce the stack requirement of match() slightly.
720 30. Added the '=' option to pcretest to check the setting of unused capturing
721 slots at the end of the pattern, which are documented as being -1, but are
722 not included in the return count.
724 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
725 compiled something random. Now it gives a compile-time error (as does
726 Perl).
728 32. A *MARK encountered during the processing of a positive assertion is now
729 recorded and passed back (compatible with Perl).
731 33. If --only-matching or --colour was set on a pcregrep call whose pattern
732 had alternative anchored branches, the search for a second match in a line
733 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
734 matched the line "0102" twice. The same bug affected patterns that started
735 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
736 twice.
738 34. Previously, PCRE did not allow quantification of assertions. However, Perl
739 does, and because of capturing effects, quantifying parenthesized
740 assertions may at times be useful. Quantifiers are now allowed for
741 parenthesized assertions.
743 35. A minor code tidy in pcre_compile() when checking options for \R usage.
745 36. \g was being checked for fancy things in a character class, when it should
746 just be a literal "g".
748 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
749 appearance of a nested POSIX class supersedes an apparent external class.
750 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
751 unescaped square brackets may also appear as part of class names. For
752 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
753 more like Perl. (But see 8.20/1 above.)
755 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
756 was because it thought it was \N{name}, which is not supported).
758 39. Add minix to OS list not supporting the -S option in pcretest.
760 40. PCRE tries to detect cases of infinite recursion at compile time, but it
761 cannot analyze patterns in sufficient detail to catch mutual recursions
762 such as ((?1))((?2)). There is now a runtime test that gives an error if a
763 subgroup is called recursively as a subpattern for a second time at the
764 same position in the subject string. In previous releases this might have
765 been caught by the recursion limit, or it might have run out of stack.
767 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
768 happen only once. PCRE was, however incorrectly giving a compile time error
769 "recursive call could loop indefinitely" because it cannot analyze the
770 pattern in sufficient detail. The compile time test no longer happens when
771 PCRE is compiling a conditional subpattern, but actual runaway loops are
772 now caught at runtime (see 40 above).
774 42. It seems that Perl allows any characters other than a closing parenthesis
775 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
776 has been changed to be the same.
778 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
779 as not to get warnings when autogen.sh is called. Also changed
780 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
782 44. To help people who use pcregrep to scan files containing exceedingly long
783 lines, the following changes have been made:
785 (a) The default value of the buffer size parameter has been increased from
786 8K to 20K. (The actual buffer used is three times this size.)
788 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
789 PCRE is built.
791 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
792 to be set at run time.
794 (d) Numerical values in pcregrep options can be followed by K or M, for
795 example --buffer-size=50K.
797 (e) If a line being scanned overflows pcregrep's buffer, an error is now
798 given and the return code is set to 2.
800 45. Add a pointer to the latest mark to the callout data block.
802 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
803 partial match of an empty string instead of no match. This was specific to
804 the use of ".".
806 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
807 complete match instead of a partial match. This bug was dependent on both
808 the PCRE_UTF8 and PCRE_DOTALL options being set.
810 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
811 starting byte set, because \b was not being ignored.
814 Version 8.12 15-Jan-2011
815 ------------------------
817 1. Fixed some typos in the markup of the man pages, and wrote a script that
818 checks for such things as part of the documentation building process.
820 2. On a big-endian 64-bit system, pcregrep did not correctly process the
821 --match-limit and --recursion-limit options (added for 8.11). In
822 particular, this made one of the standard tests fail. (The integer value
823 went into the wrong half of a long int.)
825 3. If the --colour option was given to pcregrep with -v (invert match), it
826 did strange things, either producing crazy output, or crashing. It should,
827 of course, ignore a request for colour when reporting lines that do not
828 match.
830 4. Another pcregrep bug caused similar problems if --colour was specified with
831 -M (multiline) and the pattern match finished with a line ending.
833 5. In pcregrep, when a pattern that ended with a literal newline sequence was
834 matched in multiline mode, the following line was shown as part of the
835 match. This seems wrong, so I have changed it.
837 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
838 the check for further matches in the same line (so they could be coloured)
839 to overrun the end of the current line. If another match was found, it was
840 incorrectly shown (and then shown again when found in the next line).
842 7. If pcregrep was compiled under Windows, there was a reference to the
843 function pcregrep_exit() before it was defined. I am assuming this was
844 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
845 reported by a user. I've moved the definition above the reference.
848 Version 8.11 10-Dec-2010
849 ------------------------
851 1. (*THEN) was not working properly if there were untried alternatives prior
852 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
853 backtracked to try for "b" instead of moving to the next alternative branch
854 at the same level (in this case, to look for "c"). The Perl documentation
855 is clear that when (*THEN) is backtracked onto, it goes to the "next
856 alternative in the innermost enclosing group".
858 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
859 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
860 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
861 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
862 (*THEN).
864 3. If \s appeared in a character class, it removed the VT character from
865 the class, even if it had been included by some previous item, for example
866 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
867 of \s, but is part of the POSIX "space" class.)
869 4. A partial match never returns an empty string (because you can always
870 match an empty string at the end of the subject); however the checking for
871 an empty string was starting at the "start of match" point. This has been
872 changed to the "earliest inspected character" point, because the returned
873 data for a partial match starts at this character. This means that, for
874 example, /(?<=abc)def/ gives a partial match for the subject "abc"
875 (previously it gave "no match").
877 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
878 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
879 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
880 has an implication that the given string is incomplete (because a partial
881 match is preferred over a full match). For this reason, these items now
882 give a partial match in this situation. [Aside: previously, the one case
883 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
884 match rather than a full match, which was wrong by the old rules, but is
885 now correct.]
887 6. There was a bug in the handling of #-introduced comments, recognized when
888 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
889 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
890 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
891 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
892 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
893 places in pcre_compile().
895 7. Related to (6) above, when pcre_compile() was skipping #-introduced
896 comments when looking ahead for named forward references to subpatterns,
897 the only newline sequence it recognized was NL. It now handles newlines
898 according to the set newline convention.
900 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
901 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
902 cater for a lack of strerror(). These oversights have been fixed.
904 9. Added --match-limit and --recursion-limit to pcregrep.
906 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
908 11. When the -o option was used, pcregrep was setting a return code of 1, even
909 when matches were found, and --line-buffered was not being honoured.
911 12. Added an optional parentheses number to the -o and --only-matching options
912 of pcregrep.
914 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
915 can match an empty string. The code to do it in pcretest and pcredemo
916 needed fixing:
918 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
919 only one byte after an empty string match just before CRLF (this case
920 just got forgotten; "any" and "anycrlf" were OK).
922 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
923 mode when an empty string match preceded an ASCII character followed by
924 a non-ASCII character. (The code for advancing by one character rather
925 than one byte was nonsense.)
927 (c) The pcredemo.c sample program did not have any code at all to handle
928 the cases when CRLF is a valid newline sequence.
930 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
931 as a starting offset was within the subject string. There is now a new
932 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
933 negative or greater than the length of the string. In order to test this,
934 pcretest is extended to allow the setting of negative starting offsets.
936 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
937 starting offset points to the beginning of a UTF-8 character was
938 unnecessarily clumsy. I tidied it up.
940 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
941 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
943 17. Nobody had reported that the --include_dir option, which was added in
944 release 7.7 should have been called --include-dir (hyphen, not underscore)
945 for compatibility with GNU grep. I have changed it to --include-dir, but
946 left --include_dir as an undocumented synonym, and the same for
947 --exclude-dir, though that is not available in GNU grep, at least as of
948 release 2.5.4.
950 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
951 characters from a string of bytes) have been redefined so as not to use
952 loops, in order to improve performance in some environments. At the same
953 time, I abstracted some of the common code into auxiliary macros to save
954 repetition (this should not affect the compiled code).
956 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
957 compile-time error is now given if \c is not followed by an ASCII
958 character, that is, a byte less than 128. (In EBCDIC mode, the code is
959 different, and any byte value is allowed.)
961 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
962 START_OPTIMIZE option, which is now allowed at compile time - but just
963 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
964 to pcregrep and other applications that have no direct access to PCRE
965 options. The new /Y option in pcretest sets this option when calling
966 pcre_compile().
968 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
969 back references. Groups containing recursive back references were forced to
970 be atomic by that change, but in the case of named groups, the amount of
971 memory required was incorrectly computed, leading to "Failed: internal
972 error: code overflow". This has been fixed.
974 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
975 pcretest.c, to avoid build problems in some Borland environments.
978 Version 8.10 25-Jun-2010
979 ------------------------
981 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
982 THEN.
984 2. (*ACCEPT) was not working when inside an atomic group.
986 3. Inside a character class, \B is treated as a literal by default, but
987 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
988 causes the error). The code is unchanged, but I tidied the documentation.
990 4. Inside a character class, PCRE always treated \R and \X as literals,
991 whereas Perl faults them if its -w option is set. I have changed PCRE so
992 that it faults them when PCRE_EXTRA is set.
994 5. Added support for \N, which always matches any character other than
995 newline. (It is the same as "." when PCRE_DOTALL is not set.)
997 6. When compiling pcregrep with newer versions of gcc which may have
998 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
999 declared with attribute warn_unused_result" were given. Just casting the
1000 result to (void) does not stop the warnings; a more elaborate fudge is
1001 needed. I've used a macro to implement this.
1003 7. Minor change to pcretest.c to avoid a compiler warning.
1005 8. Added four artifical Unicode properties to help with an option to make
1006 \s etc use properties (see next item). The new properties are: Xan
1007 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1009 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1010 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1011 this option. Modified pcretest to add /W to test this facility. Added
1012 REG_UCP to make it available via the POSIX interface.
1014 10. Added --line-buffered to pcregrep.
1016 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1017 studied, and the match started with a letter with a code point greater than
1018 127 whose first byte was different to the first byte of the other case of
1019 the letter, the other case of this starting letter was not recognized
1020 (#976).
1022 12. If a pattern that was studied started with a repeated Unicode property
1023 test, for example, \p{Nd}+, there was the theoretical possibility of
1024 setting up an incorrect bitmap of starting bytes, but fortunately it could
1025 not have actually happened in practice until change 8 above was made (it
1026 added property types that matched character-matching opcodes).
1028 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1029 possible starting bytes for non-anchored patterns.
1031 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1032 \R, and also a number of cases that involve Unicode properties, both
1033 explicit and implicit when PCRE_UCP is set.
1035 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1036 input, it could crash or give wrong results if characters with values
1037 greater than 0xc0 were present in the subject string. (Detail: it assumed
1038 UTF-8 input when processing these items.)
1040 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1041 size_t is 64-bit (#991).
1043 17. Added a check for running out of memory when PCRE is compiled with
1044 --disable-stack-for-recursion (#990).
1046 18. If the last data line in a file for pcretest does not have a newline on
1047 the end, a newline was missing in the output.
1049 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1050 less than 128) in its various bitmaps. However, there is a facility for
1051 generating tables according to the current locale when PCRE is compiled. It
1052 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1053 characters, are recognized by isspace() and therefore were getting set in
1054 these tables, and indeed these tables seem to approximate to ISO 8859. This
1055 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1056 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1057 which of course cannot start UTF-8 characters. I have changed the code so
1058 that only real ASCII characters (less than 128) and the correct starting
1059 bytes for UTF-8 encodings are set for characters greater than 127 when in
1060 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1061 altogether.)
1063 20. Added the /T option to pcretest so as to be able to run tests with non-
1064 standard character tables, thus making it possible to include the tests
1065 used for 19 above in the standard set of tests.
1067 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1068 reference to a subpattern the other side of a comment that contains an
1069 opening parenthesis caused either an internal compiling error, or a
1070 reference to the wrong subpattern.
1073 Version 8.02 19-Mar-2010
1074 ------------------------
1076 1. The Unicode data tables have been updated to Unicode 5.2.0.
1078 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1079 configured.
1081 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1082 original author of that file, following a query about its status.
1084 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1085 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1087 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1088 quantifier applied to a forward-referencing subroutine call, could compile
1089 incorrect code or give the error "internal error: previously-checked
1090 referenced subpattern not found".
1092 6. Both MS Visual Studio and Symbian OS have problems with initializing
1093 variables to point to external functions. For these systems, therefore,
1094 pcre_malloc etc. are now initialized to local functions that call the
1095 relevant global functions.
1097 7. There were two entries missing in the vectors called coptable and poptable
1098 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1099 I've fixed the data, and added a kludgy way of testing at compile time that
1100 the lengths are correct (equal to the number of opcodes).
1102 8. Following on from 7, I added a similar kludge to check the length of the
1103 eint vector in pcreposix.c.
1105 9. Error texts for pcre_compile() are held as one long string to avoid too
1106 much relocation at load time. To find a text, the string is searched,
1107 counting zeros. There was no check for running off the end of the string,
1108 which could happen if a new error number was added without updating the
1109 string.
1111 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1113 11. \K was not working if it appeared in an atomic group or in a group that
1114 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1115 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1116 the assertion is positive, but not if it is negative.
1118 12. Change 11 fortuitously reduced the size of the stack frame used in the
1119 "match()" function of pcre_exec.c by one pointer. Forthcoming
1120 implementation of support for (*MARK) will need an extra pointer on the
1121 stack; I have reserved it now, so that the stack frame size does not
1122 decrease.
1124 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1125 item in branch that calls a recursion is a subroutine call - as in the
1126 second branch in the above example - was incorrectly given the compile-
1127 time error "recursive call could loop indefinitely" because pcre_compile()
1128 was not correctly checking the subroutine for matching a non-empty string.
1130 14. The checks for overrunning compiling workspace could trigger after an
1131 overrun had occurred. This is a "should never occur" error, but it can be
1132 triggered by pathological patterns such as hundreds of nested parentheses.
1133 The checks now trigger 100 bytes before the end of the workspace.
1135 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1138 Version 8.01 19-Jan-2010
1139 ------------------------
1141 1. If a pattern contained a conditional subpattern with only one branch (in
1142 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1143 computed the wrong minimum data length (which is of course zero for such
1144 subpatterns). This could cause incorrect "no match" results.
1146 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1147 the pattern is reset in the first branch, pcre_compile() failed with
1148 "internal error: code overflow at offset...". This happened only when
1149 the reset was to the original external option setting. (An optimization
1150 abstracts leading options settings into an external setting, which was the
1151 cause of this.)
1153 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1154 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1155 assertion pattern did not match (meaning that the assertion was true), it
1156 was incorrectly treated as false if the SKIP had been reached during the
1157 matching. This also applied to assertions used as conditions.
1159 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1160 assertion subpattern, including such a pattern used as a condition,
1161 unpredictable results occurred, instead of the error return
1164 5. The C++ GlobalReplace function was not working like Perl for the special
1165 situation when an empty string is matched. It now does the fancy magic
1166 stuff that is necessary.
1168 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1169 removed. (These were left over from very, very early versions of PCRE.)
1171 7. Some cosmetic changes to the code to make life easier when compiling it
1172 as part of something else:
1174 (a) Change DEBUG to PCRE_DEBUG.
1176 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1177 called "current" as "current_branch", to prevent a collision with the
1178 Linux macro when compiled as a kernel module.
1180 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1181 prevent a collision with the Linux macro when compiled as a kernel
1182 module.
1184 8. In pcre_compile() there are some checks for integer overflows that used to
1185 cast potentially large values to (double). This has been changed to that
1186 when building, a check for int64_t is made, and if it is found, it is used
1187 instead, thus avoiding the use of floating point arithmetic. (There is no
1188 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1189 double.
1191 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1192 2005 (difference between two addresses compared to an unsigned value).
1194 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1195 custom one, because of the following reported problem in Windows:
1197 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1198 under Win32.
1199 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1200 therefore missing the function definition.
1201 - The compiler thus generates a "C" signature for the test function.
1202 - The linker fails to find the "C" function.
1203 - PCRE fails to configure if asked to do so against libbz2.
1205 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1206 messages were output:
1208 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1209 rerunning libtoolize, to keep the correct libtool macros in-tree.
1210 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1212 I have done both of these things.
1214 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1215 most of the time, it *can* run out if it is given a pattern that contains a
1216 runaway infinite recursion. I updated the discussion in the pcrestack man
1217 page.
1219 13. Now that we have gone to the x.xx style of version numbers, the minor
1220 version may start with zero. Using 08 or 09 is a bad idea because users
1221 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1222 interpreted as invalid octal numbers. I've updated the previous comment in
1223 configure.ac, and also added a check that gives an error if 08 or 09 are
1224 used.
1226 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1227 causing partial matching to fail when the end of the subject matched \W
1228 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1230 15. There were some discrepancies between the declarations in pcre_internal.h
1231 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1232 their definitions. The declarations used "const uschar *" and the
1233 definitions used USPTR. Even though USPTR is normally defined as "const
1234 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1235 reported that: "This difference in casting confuses some C++ compilers, for
1236 example, SunCC recognizes above declarations as different functions and
1237 generates broken code for hbpcre." I have changed the declarations to use
1238 USPTR.
1240 16. GNU libtool is named differently on some systems. The autogen.sh script now
1241 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1242 (FreeBSD).
1244 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1245 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1246 comment: "Figure out how to create a longlong from a string: strtoll and
1247 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1248 instance, but it only takes 2 args instead of 3!"
1250 18. A subtle bug concerned with back references has been fixed by a change of
1251 specification, with a corresponding code fix. A pattern such as
1252 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1253 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1254 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1255 same bug. Such groups have to be quantified to be useful, or contained
1256 inside another quantified group. (If there's no repetition, the reference
1257 can never match.) The problem arises because, having left the group and
1258 moved on to the rest of the pattern, a later failure that backtracks into
1259 the group uses the captured value from the final iteration of the group
1260 rather than the correct earlier one. I have fixed this in PCRE by forcing
1261 any group that contains a reference to itself to be an atomic group; that
1262 is, there cannot be any backtracking into it once it has completed. This is
1263 similar to recursive and subroutine calls.
1266 Version 8.00 19-Oct-09
1267 ----------------------
1269 1. The table for translating pcre_compile() error codes into POSIX error codes
1270 was out-of-date, and there was no check on the pcre_compile() error code
1271 being within the table. This could lead to an OK return being given in
1272 error.
1274 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1275 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1276 in a Windows environment.
1278 3. The pcregrep --count option prints the count for each file even when it is
1279 zero, as does GNU grep. However, pcregrep was also printing all files when
1280 --files-with-matches was added. Now, when both options are given, it prints
1281 counts only for those files that have at least one match. (GNU grep just
1282 prints the file name in this circumstance, but including the count seems
1283 more useful - otherwise, why use --count?) Also ensured that the
1284 combination -clh just lists non-zero counts, with no names.
1286 4. The long form of the pcregrep -F option was incorrectly implemented as
1287 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1288 but it seems right to fix it, and I didn't think it was worth preserving
1289 the old behaviour.
1291 5. The command line items --regex=pattern and --regexp=pattern were not
1292 recognized by pcregrep, which required --regex pattern or --regexp pattern
1293 (with a space rather than an '='). The man page documented the '=' forms,
1294 which are compatible with GNU grep; these now work.
1296 6. No libpcreposix.pc file was created for pkg-config; there was just
1297 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1299 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1300 when UCP support is not needed, by modifying the Python script that
1301 generates it from Unicode data files. This should not matter if the module
1302 is correctly used as a library, but I received one complaint about 50K of
1303 unwanted data. My guess is that the person linked everything into his
1304 program rather than using a library. Anyway, it does no harm.
1306 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1307 was a minimum greater than 1 for a wide character in a possessive
1308 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1309 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1310 character. Chaos in the form of incorrect output or a compiling loop could
1311 result.
1313 9. The restrictions on what a pattern can contain when partial matching is
1314 requested for pcre_exec() have been removed. All patterns can now be
1315 partially matched by this function. In addition, if there are at least two
1316 slots in the offset vector, the offset of the earliest inspected character
1317 for the match and the offset of the end of the subject are set in them when
1318 PCRE_ERROR_PARTIAL is returned.
1320 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1321 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1322 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1323 and may be more useful for multi-segment matching.
1325 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1326 used to be given if ever the end of the subject was reached; now it is
1327 given only if matching could not proceed because another character was
1328 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1329 string "Z", which now yields "no match" instead of "partial match". In the
1330 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1331 final character ended with (*FAIL).
1333 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1334 if the pattern had a "must contain" character that was already found in the
1335 earlier partial match, unless partial matching was again requested. For
1336 example, with the pattern /dog.(body)?/, the "must contain" character is
1337 "g". If the first part-match was for the string "dog", restarting with
1338 "sbody" failed. This bug has been fixed.
1340 13. The string returned by pcre_dfa_exec() after a partial match has been
1341 changed so that it starts at the first inspected character rather than the
1342 first character of the match. This makes a difference only if the pattern
1343 starts with a lookbehind assertion or \b or \B (\K is not supported by
1344 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1345 matching functions compatible, and I think it's the right thing to do.
1347 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1348 so that the demonstration program is easily available in environments where
1349 PCRE has not been installed from source.
1351 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1352 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1353 library.
1355 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1356 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1357 is not the first non-POSIX option to be added. Clearly some people find
1358 these options useful.
1360 17. If a caller to the POSIX matching function regexec() passes a non-zero
1361 value for nmatch with a NULL value for pmatch, the value of
1362 nmatch is forced to zero.
1364 18. RunGrepTest did not have a test for the availability of the -u option of
1365 the diff command, as RunTest does. It now checks in the same way as
1366 RunTest, and also checks for the -b option.
1368 19. If an odd number of negated classes containing just a single character
1369 interposed, within parentheses, between a forward reference to a named
1370 subpattern and the definition of the subpattern, compilation crashed with
1371 an internal error, complaining that it could not find the referenced
1372 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1373 [The bug was that it was starting one character too far in when skipping
1374 over the character class, thus treating the ] as data rather than
1375 terminating the class. This meant it could skip too much.]
1377 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1378 /g option in pcretest when the pattern contains \K, which makes it possible
1379 to have an empty string match not at the start, even when the pattern is
1380 anchored. Updated pcretest and pcredemo to use this option.
1382 21. If the maximum number of capturing subpatterns in a recursion was greater
1383 than the maximum at the outer level, the higher number was returned, but
1384 with unset values at the outer level. The correct (outer level) value is
1385 now given.
1387 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1388 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1389 make it do so. The string so far is captured, making this feature
1390 compatible with Perl.
1392 23. The tests have been re-organized, adding tests 11 and 12, to make it
1393 possible to check the Perl 5.10 features against Perl 5.10.
1395 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1396 pattern matches a fixed length string. PCRE did not allow this; now it
1397 does. Neither allows recursion.
1399 25. I finally figured out how to implement a request to provide the minimum
1400 length of subject string that was needed in order to match a given pattern.
1401 (It was back references and recursion that I had previously got hung up
1402 on.) This code has now been added to pcre_study(); it finds a lower bound
1403 to the length of subject needed. It is not necessarily the greatest lower
1404 bound, but using it to avoid searching strings that are too short does give
1405 some useful speed-ups. The value is available to calling programs via
1406 pcre_fullinfo().
1408 26. While implementing 25, I discovered to my embarrassment that pcretest had
1409 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1410 study optimizations had never been tested with that matching function.
1411 Oops. What is worse, even when it was passed study data, there was a bug in
1412 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1413 were also very few tests of studied patterns with pcre_dfa_exec().
1415 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1416 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1417 on the other side of the coin, they are no longer allowed to have different
1418 names, because these cannot be distinguished in PCRE, and this has caused
1419 confusion. (This is a difference from Perl.)
1421 28. When duplicate subpattern names are present (necessarily with different
1422 numbers, as required by 27 above), and a test is made by name in a
1423 conditional pattern, either for a subpattern having been matched, or for
1424 recursion in such a pattern, all the associated numbered subpatterns are
1425 tested, and the overall condition is true if the condition is true for any
1426 one of them. This is the way Perl works, and is also more like the way
1427 testing by number works.
1430 Version 7.9 11-Apr-09
1431 ---------------------
1433 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1434 (pcretest), all targets were linked against these libraries. This included
1435 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1436 libraries. This caused unwanted dependencies to be created. This problem
1437 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1438 pcretest is linked with readline.
1440 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1441 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1442 moved outside it again, because FALSE and TRUE are already defined in AIX,
1443 but BOOL is not.
1445 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1446 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1448 4. The pcregrep documentation said spaces were inserted as well as colons (or
1449 hyphens) following file names and line numbers when outputting matching
1450 lines. This is not true; no spaces are inserted. I have also clarified the
1451 wording for the --colour (or --color) option.
1453 5. In pcregrep, when --colour was used with -o, the list of matching strings
1454 was not coloured; this is different to GNU grep, so I have changed it to be
1455 the same.
1457 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1458 each matching line was coloured. Now it goes on to look for further matches
1459 of any of the test patterns, which is the same behaviour as GNU grep.
1461 7. A pattern that could match an empty string could cause pcregrep to loop; it
1462 doesn't make sense to accept an empty string match in pcregrep, so I have
1463 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1464 seems to be how GNU grep behaves.
1466 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1467 start or after a newline", because the conditional assertion was not being
1468 correctly handled. The rule now is that both the assertion and what follows
1469 in the first alternative must satisfy the test.
1471 9. If auto-callout was enabled in a pattern with a conditional group whose
1472 condition was an assertion, PCRE could crash during matching, both with
1473 pcre_exec() and pcre_dfa_exec().
1475 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1476 used for matching.
1478 11. Unicode property support in character classes was not working for
1479 characters (bytes) greater than 127 when not in UTF-8 mode.
1481 12. Added the -M command line option to pcretest.
1483 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1485 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1487 16. Added comments and documentation about mis-use of no_arg in the C++
1488 wrapper.
1490 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1491 from Martin Jerabek that uses macro names for all relevant character and
1492 string constants.
1494 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1495 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1496 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1497 these, but not everybody uses configure.
1499 19. A conditional group that had only one branch was not being correctly
1500 recognized as an item that could match an empty string. This meant that an
1501 enclosing group might also not be so recognized, causing infinite looping
1502 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1503 with the subject "ab", where knowledge that the repeated group can match
1504 nothing is needed in order to break the loop.
1506 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1507 exec(), but without supplying a callout function, matching went wrong.
1509 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1510 leak if the size of the offset vector was greater than 30. When the vector
1511 is smaller, the saved offsets during recursion go onto a local stack
1512 vector, but for larger vectors malloc() is used. It was failing to free
1513 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1514 error, in fact).
1516 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1517 heapframe that is used only when UTF-8 support is enabled. This caused no
1518 problem, but was untidy.
1520 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1521 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1522 included within another project.
1524 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1525 slightly modified by me:
1527 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1528 not building pcregrep.
1530 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1531 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1533 25. Forward references, both numeric and by name, in patterns that made use of
1534 duplicate group numbers, could behave incorrectly or give incorrect errors,
1535 because when scanning forward to find the reference group, PCRE was not
1536 taking into account the duplicate group numbers. A pattern such as
1537 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1539 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1540 the feature of a custom pointer more persuasive (as requested by a user).
1542 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1543 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1544 pre-defined.
1546 28. Added support for (*UTF8) at the start of a pattern.
1548 29. Arrange for flags added by the "release type" setting in CMake to be shown
1549 in the configuration summary.
1552 Version 7.8 05-Sep-08
1553 ---------------------
1555 1. Replaced UCP searching code with optimized version as implemented for Ad
1556 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1557 stage table and inline lookup instead of a function, giving speed ups of 2
1558 to 5 times on some simple patterns that I tested. Permission was given to
1559 distribute the MultiStage2.py script that generates the tables (it's not in
1560 the tarball, but is in the Subversion repository).
1562 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1563 scripts.
1565 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1566 a group with a zero qualifier. The result of the study could be incorrect,
1567 or the function might crash, depending on the pattern.
1569 4. Caseless matching was not working for non-ASCII characters in back
1570 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1571 It now works when Unicode Property Support is available.
1573 5. In pcretest, an escape such as \x{de} in the data was always generating
1574 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1575 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1576 truncation.
1578 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1580 7. Added two (int) casts to pcregrep when printing the difference of two
1581 pointers, in case they are 64-bit values.
1583 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1584 test 2 if it fails.
1586 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1587 and a #define of that name to empty if it is not externally set. This is to
1588 allow users of MSVC to set it if necessary.
1590 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1591 the convenience functions in the pcre_get.c source file.
1593 11. An option change at the start of a pattern that had top-level alternatives
1594 could cause overwriting and/or a crash. This command provoked a crash in
1595 some environments:
1597 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1599 This potential security problem was recorded as CVE-2008-2371.
1601 12. For a pattern where the match had to start at the beginning or immediately
1602 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1603 pcre_dfa_exec() could read past the end of the passed subject if there was
1604 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1605 pcretest so that it places the subject at the end of its malloc-ed buffer.
1607 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1608 exec() might read past the end of the data buffer in UTF-8 mode.
1610 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1611 the data contained the byte 0x85 as part of a UTF-8 character within its
1612 first line. This applied both to normal and DFA matching.
1614 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1615 /^[^d]*?$/8 failed to match "abc".
1617 16. Added a missing copyright notice to pcrecpp_internal.h.
1619 17. Make it more clear in the documentation that values returned from
1620 pcre_exec() in ovector are byte offsets, not character counts.
1622 18. Tidied a few places to stop certain compilers from issuing warnings.
1624 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1625 supplied by Stefan Weber. I made a further small update for 7.8 because
1626 there is a change of source arrangements: the pcre_searchfuncs.c module is
1627 replaced by pcre_ucd.c.
1630 Version 7.7 07-May-08
1631 ---------------------
1633 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1634 a string to a long long, pretend we don't even have a long long." This is
1635 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1637 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1638 pre-7.6 versions, which defined a global no_arg variable instead of putting
1639 it in the RE class. (See also #8 below.)
1641 3. Remove a line of dead code, identified by coverity and reported by Nuno
1642 Lopes.
1644 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1646 (1) The include/exclude patterns were being applied to the whole pathnames
1647 of files, instead of just to the final components.
1649 (2) If there was more than one level of directory, the subdirectories were
1650 skipped unless they satisfied the include/exclude conditions. This is
1651 inconsistent with GNU grep (and could even be seen as contrary to the
1652 pcregrep specification - which I improved to make it absolutely clear).
1653 The action now is always to scan all levels of directory, and just
1654 apply the include/exclude patterns to regular files.
1656 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1657 --exclude_dir in the tests to avoid scanning .svn directories.
1659 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1660 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1661 doesn't support NULs in patterns.
1663 7. Added some missing "const"s to declarations of static tables in
1664 pcre_compile.c and pcre_dfa_exec.c.
1666 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1667 caused by fix #2 above. (Subsequently also a second patch to fix the
1668 first patch. And a third patch - this was a messy problem.)
1670 9. Applied Craig's patch to remove the use of push_back().
1672 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1673 matching function regexec().
1675 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1676 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1677 references. PCRE supports relative numbers with this syntax (I don't think
1678 Oniguruma does).
1680 12. Previously, a group with a zero repeat such as (...){0} was completely
1681 omitted from the compiled regex. However, this means that if the group
1682 was called as a subroutine from elsewhere in the pattern, things went wrong
1683 (an internal error was given). Such groups are now left in the compiled
1684 pattern, with a new opcode that causes them to be skipped at execution
1685 time.
1687 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1688 to the way PCRE behaves:
1690 (a) A lone ] character is dis-allowed (Perl treats it as data).
1692 (b) A back reference to an unmatched subpattern matches an empty string
1693 (Perl fails the current match path).
1695 (c) A data ] in a character class must be notated as \] because if the
1696 first data character in a class is ], it defines an empty class. (In
1697 Perl it is not possible to have an empty class.) The empty class []
1698 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1699 The negative empty class [^] matches any one character, independently
1700 of the DOTALL setting.
1702 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1703 non-existent subpattern following a character class starting with ']' and
1704 containing () gave an internal compiling error instead of "reference to
1705 non-existent subpattern". Fortunately, when the pattern did exist, the
1706 compiled code was correct. (When scanning forwards to check for the
1707 existencd of the subpattern, it was treating the data ']' as terminating
1708 the class, so got the count wrong. When actually compiling, the reference
1709 was subsequently set up correctly.)
1711 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1712 it was being rejected as not supported by pcre_dfa_exec(), even though
1713 other assertions are supported. I have made pcre_dfa_exec() support
1714 (*FAIL).
1716 16. The implementation of 13c above involved the invention of a new opcode,
1717 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1718 cannot be changed at match time, I realized I could make a small
1719 improvement to matching performance by compiling OP_ALLANY instead of
1720 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1721 on the OP_ANY path.
1723 17. Compiling pcretest on Windows with readline support failed without the
1724 following two fixes: (1) Make the unistd.h include conditional on
1725 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1727 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1728 ncurses library to be included for pcretest when ReadLine support is
1729 requested, but also to allow for it to be overridden. This patch came from
1730 Daniel Bergström.
1732 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1733 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1734 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1735 spotting this.
1738 Version 7.6 28-Jan-08
1739 ---------------------
1741 1. A character class containing a very large number of characters with
1742 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1743 overflow.
1745 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1746 HAVE_LONG_LONG is not defined.
1748 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1749 bring them up to date and include new features. This patch includes:
1751 - Fixed PH's badly added libz and libbz2 support.
1752 - Fixed a problem with static linking.
1753 - Added pcredemo. [But later removed - see 7 below.]
1754 - Fixed dftables problem and added an option.
1755 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1757 - Added readline support for pcretest.
1758 - Added an listing of the option settings after cmake has run.
1760 4. A user submitted a patch to Makefile that makes it easy to create
1761 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1762 Makefile.am that cause it to include this special target, without
1763 affecting anything else. Note that the same mingw target plus all
1764 the other distribution libraries and programs are now supported
1765 when configuring with CMake (see 6 below) instead of with
1766 Configure/Make.
1768 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1769 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1770 exported in the Windows port". It has not yet been confirmed that the patch
1771 solves the problem, but it does no harm.
1773 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1774 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1775 with CMake, and also correct the comment about stack recursion.
1777 7. Remove the automatic building of pcredemo from the ./configure system and
1778 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1779 of a program that users should build themselves after PCRE is installed, so
1780 building it automatically is not really right. What is more, it gave
1781 trouble in some build environments.
1783 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1786 Version 7.5 10-Jan-08
1787 ---------------------
1789 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1790 values in parens when parsing an RE using the C++ wrapper."
1792 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1793 Characters greater than 255 were excluded from the class instead of being
1794 included.
1796 3. The same bug as (2) above applied to negated POSIX classes such as
1797 [:^space:].
1799 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1800 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1801 I have changed it.
1803 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1804 first named subpattern) and a construct such as (?&a) would reference the
1805 first named subpattern whose name started with "a" (in other words, the
1806 length check was missing). Both these problems are fixed. "Subpattern name
1807 expected" is now given for (?&) (a zero-length name), and this patch also
1808 makes it give the same error for \k'' (previously it complained that that
1809 was a reference to a non-existent subpattern).
1811 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1812 this is right because (?- can be followed by option settings as well as by
1813 digits. I have, however, made the messages clearer.
1815 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1816 than the number used in the conditional) now cause a compile-time error.
1817 This is actually not compatible with Perl, which accepts such patterns, but
1818 treats the conditional as always being FALSE (as PCRE used to), but it
1819 seems to me that giving a diagnostic is better.
1821 8. Change "alphameric" to the more common word "alphanumeric" in comments
1822 and messages.
1824 9. Fix two occurrences of "backslash" in comments that should have been
1825 "backspace".
1827 10. Remove two redundant lines of code that can never be obeyed (their function
1828 was moved elsewhere).
1830 11. The program that makes PCRE's Unicode character property table had a bug
1831 which caused it to generate incorrect table entries for sequences of
1832 characters that have the same character type, but are in different scripts.
1833 It amalgamated them into a single range, with the script of the first of
1834 them. In other words, some characters were in the wrong script. There were
1835 thirteen such cases, affecting characters in the following ranges:
1837 U+002b0 - U+002c1
1838 U+0060c - U+0060d
1839 U+0061e - U+00612
1840 U+0064b - U+0065e
1841 U+0074d - U+0076d
1842 U+01800 - U+01805
1843 U+01d00 - U+01d77
1844 U+01d9b - U+01dbf
1845 U+0200b - U+0200f
1846 U+030fc - U+030fe
1847 U+03260 - U+0327f
1848 U+0fb46 - U+0fbb1
1849 U+10450 - U+1049d
1851 12. The -o option (show only the matching part of a line) for pcregrep was not
1852 compatible with GNU grep in that, if there was more than one match in a
1853 line, it showed only the first of them. It now behaves in the same way as
1854 GNU grep.
1856 13. If the -o and -v options were combined for pcregrep, it printed a blank
1857 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1858 does the same. The return code can be used to tell if there were any
1859 non-matching lines.
1861 14. Added --file-offsets and --line-offsets to pcregrep.
1863 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1864 infinitely looping recursion. The bug was that positive lookaheads were not
1865 being skipped when checking for a possible empty match (negative lookaheads
1866 and both kinds of lookbehind were skipped).
1868 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1869 inclusion of <windows.h> to before rather than after the definition of
1870 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1872 17. Specifying a possessive quantifier with a specific limit for a Unicode
1873 character property caused pcre_compile() to compile bad code, which led at
1874 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1875 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1876 caused the error; without that there was no problem.
1878 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1880 19. Added --enable-pcretest-libreadline.
1882 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1883 RE::GlobalReplace(). As a result, the number of replacements returned was
1884 double what it should be. I removed one of the increments, but Craig sent a
1885 later patch that removed the other one (the right fix) and added unit tests
1886 that check the return values (which was not done before).
1888 21. Several CMake things:
1890 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1891 the names libpcre and libpcreposix, not just pcre and pcreposix.
1893 (2) The above change means that pcretest and pcregrep are now correctly
1894 linked with the newly-built libraries, not previously installed ones.
1898 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1899 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1900 UTF-8 newline character). The key issue is that the pattern starts .*;
1901 this means that the match must be either at the beginning, or after a
1902 newline. The bug was in the code for advancing after a failed match and
1903 checking that the new position followed a newline. It was not taking
1904 account of UTF-8 characters correctly.
1906 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1907 character classes. PCRE was not treating the sequence [:...:] as a
1908 character class unless the ... were all letters. Perl, however, seems to
1909 allow any characters between [: and :], though of course it rejects as
1910 unknown any "names" that contain non-letters, because all the known class
1911 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1912 for example, whereas PCRE did not - it did not recognize a POSIX character
1913 class. This seemed a bit dangerous, so the code has been changed to be
1914 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1915 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1916 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1917 Perl does, and where it didn't before.
1919 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1920 Windows environments %n is disabled by default.
1923 Version 7.4 21-Sep-07
1924 ---------------------
1926 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1927 means that a class such as [\s] counted as "explicit reference to CR or
1928 LF". That isn't really right - the whole point of the change was to try to
1929 help when there was an actual mention of one of the two characters. So now
1930 the change happens only if \r or \n (or a literal CR or LF) character is
1931 encountered.
1933 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1934 of both had grown to the point where there were only 3 bits left.
1935 Fortunately, there was spare space in the data structure, and so I have
1936 moved the internal flags into a new 16-bit field to free up more option
1937 bits.
1939 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1940 but did not set the internal JCHANGED flag - either of these is enough to
1941 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1942 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1943 start sets both bits.
1945 4. Added options (at build time, compile time, exec time) to change \R from
1946 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1948 5. doc/pcresyntax.html was missing from the distribution.
1950 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1951 compatibility, even though it is no longer used.
1953 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1954 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1955 windows.h file is present (where different names are used). [This was
1956 reversed later after testing - see 16 below.]
1958 8. Changed all #include <config.h> to #include "config.h". There were also
1959 some further <pcre.h> cases that I changed to "pcre.h".
1961 9. When pcregrep was used with the --colour option, it missed the line ending
1962 sequence off the lines that it output.
1964 10. It was pointed out to me that arrays of string pointers cause lots of
1965 relocations when a shared library is dynamically loaded. A technique of
1966 using a single long string with a table of offsets can drastically reduce
1967 these. I have refactored PCRE in four places to do this. The result is
1968 dramatic:
1970 Originally: 290
1971 After changing UCP table: 187
1972 After changing error message table: 43
1973 After changing table of "verbs" 36
1974 After changing table of Posix names 22
1976 Thanks to the folks working on Gregex for glib for this insight.
1978 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1979 unicode-properties was also set.
1981 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1983 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1984 checked only for CRLF.
1986 14. Added casts to pcretest.c to avoid compiler warnings.
1988 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1990 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1991 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1992 entirely. This removes changes made in 7 above.
1994 17. The CMake files have been updated, and there is now more information about
1995 building with CMake in the NON-UNIX-USE document.
1998 Version 7.3 28-Aug-07
1999 ---------------------
2001 1. In the rejigging of the build system that eventually resulted in 7.1, the
2002 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2003 brackets there is not right, since it causes compilers to look for an
2004 installed pcre.h, not the version that is in the source that is being
2005 compiled (which of course may be different). I have changed it back to:
2007 #include "pcre.h"
2009 I have a vague recollection that the change was concerned with compiling in
2010 different directories, but in the new build system, that is taken care of
2011 by the VPATH setting the Makefile.
2013 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2014 when the subject happened to end in the byte 0x85 (e.g. if the last
2015 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2016 characters but of course it shouldn't be taken as a newline when it is part
2017 of another character. The bug was that, for an unlimited repeat of . in
2018 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2019 characters when looking for a newline.
2021 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2023 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2024 in debug output.
2026 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2027 long printing in the pcrecpp unittest when running under MinGW.
2029 6. ESC_K was left out of the EBCDIC table.
2031 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2032 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2033 limit also applies to "virtual nesting" when a pattern is recursive, and in
2034 this case 1000 isn't so big. I have been able to remove this limit at the
2035 expense of backing off one optimization in certain circumstances. Normally,
2036 when pcre_exec() would call its internal match() function recursively and
2037 immediately return the result unconditionally, it uses a "tail recursion"
2038 feature to save stack. However, when a subpattern that can match an empty
2039 string has an unlimited repetition quantifier, it no longer makes this
2040 optimization. That gives it a stack frame in which to save the data for
2041 checking that an empty string has been matched. Previously this was taken
2042 from the 1000-entry workspace that had been reserved. So now there is no
2043 explicit limit, but more stack is used.
2045 8. Applied Daniel's patches to solve problems with the import/export magic
2046 syntax that is required for Windows, and which was going wrong for the
2047 pcreposix and pcrecpp parts of the library. These were overlooked when this
2048 problem was solved for the main library.
2050 9. There were some crude static tests to avoid integer overflow when computing
2051 the size of patterns that contain repeated groups with explicit upper
2052 limits. As the maximum quantifier is 65535, the maximum group length was
2053 set at 30,000 so that the product of these two numbers did not overflow a
2054 32-bit integer. However, it turns out that people want to use groups that
2055 are longer than 30,000 bytes (though not repeat them that many times).
2056 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2057 made it possible to implement the integer overflow checks in a much more
2058 dynamic way, which I have now done. The artificial limitation on group
2059 length has been removed - we now have only the limit on the total length of
2060 the compiled pattern, which depends on the LINK_SIZE setting.
2062 10. Fixed a bug in the documentation for get/copy named substring when
2063 duplicate names are permitted. If none of the named substrings are set, the
2064 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2065 empty string.
2067 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2068 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2069 because the ] is interpreted as the first data character and the
2070 terminating ] is not found. PCRE has been made compatible with Perl in this
2071 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2072 cause memory overwriting.
2074 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2075 string has been matched (to stop an infinite loop). It was not recognizing
2076 a conditional subpattern that could match an empty string if that
2077 subpattern was within another subpattern. For example, it looped when
2078 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2079 condition was not nested. This bug has been fixed.
2081 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2082 past the start of the subject in the presence of bytes with the top bit
2083 set, for example "\x8aBCD".
2085 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2086 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2088 14. Optimized (?!) to (*FAIL).
2090 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2091 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2092 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2093 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2094 does: it's just the validity check that is more restrictive.
2096 16. Inserted checks for integer overflows during escape sequence (backslash)
2097 processing, and also fixed erroneous offset values for syntax errors during
2098 backslash processing.
2100 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2101 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2103 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2104 caused an overrun.
2106 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2107 something other than just ASCII characters) inside a group that had an
2108 unlimited repeat caused a loop at compile time (while checking to see
2109 whether the group could match an empty string).
2111 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2112 [\P{Any}] did so. (Error in the code for printing property names.)
2114 21. An orphan \E inside a character class could cause a crash.
2116 22. A repeated capturing bracket such as (A)? could cause a wild memory
2117 reference during compilation.
2119 23. There are several functions in pcre_compile() that scan along a compiled
2120 expression for various reasons (e.g. to see if it's fixed length for look
2121 behind). There were bugs in these functions when a repeated \p or \P was
2122 present in the pattern. These operators have additional parameters compared
2123 with \d, etc, and these were not being taken into account when moving along
2124 the compiled data. Specifically:
2126 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2127 length.
2129 (b) An item such as \pL+ within a repeated group could cause crashes or
2130 loops.
2132 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2133 "reference to non-existent subpattern" error.
2135 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2137 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2138 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2140 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2142 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2143 character were causing crashes (broken optimization).
2145 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2146 \p or \P) caused a compile-time loop.
2148 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2149 break. For example, the unstudied pattern [\r\n]A does not match the string
2150 "\r\nA" because change 7.0/46 below moves the current point on by two
2151 characters after failing to match at the start. However, the pattern \nA
2152 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2153 the same is true. There doesn't seem any very clean way out of this, but
2154 what I have chosen to do makes the common cases work: PCRE now takes note
2155 of whether there can be an explicit match for \r or \n anywhere in the
2156 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2157 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2158 pattern has explicit CR or LF references.
2160 29. Added (*CR) etc for changing newline setting at start of pattern.
2163 Version 7.2 19-Jun-07
2164 ---------------------
2166 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2167 which is apparently normally available under Windows.
2169 2. Re-jig the pcregrep tests with different newline settings in an attempt
2170 to make them independent of the local environment's newline setting.
2172 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2174 4. Some of the "internals" tests were previously cut out when the link size
2175 was not 2, because the output contained actual offsets. The recent new
2176 "Z" feature of pcretest means that these can be cut out, making the tests
2177 usable with all link sizes.
2179 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2180 stack recursion. This gives a massive performance boost under BSD, but just
2181 a small improvement under Linux. However, it saves one field in the frame
2182 in all cases.
2184 6. Added more features from the forthcoming Perl 5.10:
2186 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2187 recursion call. It refers to the nth most recently opened parentheses.
2189 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2190 to be opened parentheses.
2192 (c) Conditions that refer to capturing parentheses can be specified
2193 relatively, for example, (?(-2)... or (?(+3)...
2195 (d) \K resets the start of the current match so that everything before
2196 is not part of it.
2198 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2200 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2201 reference syntax.
2203 (g) (?| introduces a group in which the numbering of parentheses in each
2204 alternative starts with the same number.
2206 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2208 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2211 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2212 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2213 for detecting groups that can match an empty string.
2215 9. A pattern with a very large number of alternatives (more than several
2216 hundred) was running out of internal workspace during the pre-compile
2217 phase, where pcre_compile() figures out how much memory will be needed. A
2218 bit of new cunning has reduced the workspace needed for groups with
2219 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2220 workspace instead of running out of the 4096 that are available.
2222 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2224 11. Applied patch from Google to remove an optimization that didn't quite work.
2225 The report of the bug said:
2227 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2228 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2229 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2231 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2232 it matched the wrong number of bytes.
2235 Version 7.1 24-Apr-07
2236 ---------------------
2238 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2239 that is more "standard", making use of automake and other Autotools. There
2240 is some re-arrangement of the files and adjustment of comments consequent
2241 on this.
2243 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2244 for recursive directory scanning broke on some systems because the files
2245 are not scanned in any specific order and on different systems the order
2246 was different. A call to "sort" has been inserted into RunGrepTest for the
2247 approprate test as a short-term fix. In the longer term there may be an
2248 alternative.
2250 3. I had an email from Eric Raymond about problems translating some of PCRE's
2251 man pages to HTML (despite the fact that I distribute HTML pages, some
2252 people do their own conversions for various reasons). The problems
2253 concerned the use of low-level troff macros .br and .in. I have therefore
2254 removed all such uses from the man pages (some were redundant, some could
2255 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2256 HTML has been updated to handle .nf/.fi and to complain if it encounters
2257 .br or .in.
2259 4. Updated comments in configure.ac that get placed in config.h.in and also
2260 arranged for config.h to be included in the distribution, with the name
2261 config.h.generic, for the benefit of those who have to compile without
2262 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2264 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2265 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2266 makevp.bat for the current PCRE, using the additional files
2267 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2269 6. A Windows user reported a minor discrepancy with test 2, which turned out
2270 to be caused by a trailing space on an input line that had got lost in his
2271 copy. The trailing space was an accident, so I've just removed it.
2273 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2274 that is needed.
2276 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2277 as "const" (a) because they are and (b) because it helps the PHP
2278 maintainers who have recently made a script to detect big data structures
2279 in the php code that should be moved to the .rodata section. I remembered
2280 to update Builducptable as well, so it won't revert if ucptable.h is ever
2281 re-created.
2283 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2284 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2285 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2286 support is not required. This saves 1.5-2K of code, which is important in
2287 some applications.
2289 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2290 so as not to refer to the tables, even though these functions will never be
2291 called when UTF-8 support is disabled. Otherwise there are problems with a
2292 shared library.
2294 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2296 (a) It was defining its arguments as char * instead of void *.
2298 (b) It was assuming that all moves were upwards in memory; this was true
2299 a long time ago when I wrote it, but is no longer the case.
2301 The emulated memove() is provided for those environments that have neither
2302 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2303 is clearly not the case, as these two bugs were recently reported.
2305 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2306 and Detrail to create the HTML documentation, the .txt form of the man
2307 pages, and it removes trailing spaces from listed files. It also creates
2308 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2309 case, it wraps all the #defines with #ifndefs. This script should be run
2310 before "make dist".
2312 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2313 with Unicode property support.
2315 (a) For a maximizing quantifier, if the two different cases of the
2316 character were of different lengths in their UTF-8 codings (there are
2317 some cases like this - I found 11), and the matching function had to
2318 back up over a mixture of the two cases, it incorrectly assumed they
2319 were both the same length.
2321 (b) When PCRE was configured to use the heap rather than the stack for
2322 recursion during matching, it was not correctly preserving the data for
2323 the other case of a UTF-8 character when checking ahead for a match
2324 while processing a minimizing repeat. If the check also involved
2325 matching a wide character, but failed, corruption could cause an
2326 erroneous result when trying to check for a repeat of the original
2327 character.
2329 13. Some tidying changes to the testing mechanism:
2331 (a) The RunTest script now detects the internal link size and whether there
2332 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2333 values substituted by "configure". (The RunGrepTest script already did
2334 this for UTF-8.) The configure.ac script no longer substitutes the
2335 relevant variables.
2337 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2338 with length and offset values. This means that the output is different
2339 for different internal link sizes. Test 2 is skipped for link sizes
2340 other than 2 because of this, bypassing the problem. Unfortunately,
2341 there was also a test in test 3 (the locale tests) that used /B and
2342 failed for link sizes other than 2. Rather than cut the whole test out,
2343 I have added a new /Z option to pcretest that replaces the length and
2344 offset values with spaces. This is now used to make test 3 independent
2345 of link size. (Test 2 will be tidied up later.)
2347 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2348 segmentation fault instead of returning the appropriate error message.
2350 15. In multiline mode when the newline sequence was set to "any", the pattern
2351 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2352 This doesn't seem right; it now treats the CRLF combination as the line
2353 ending, and so does not match in that case. It's only a pattern such as ^$
2354 that would hit this one: something like ^ABC$ would have failed after \r
2355 and then tried again after \r\n.
2357 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2358 in an attempt to make files that differ only in their line terminators
2359 compare equal. This works on Linux.
2361 17. Under certain error circumstances pcregrep might try to free random memory
2362 as it exited. This is now fixed, thanks to valgrind.
2364 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2365 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2366 was because its rules for how to advance for /g after matching an empty
2367 string at the end of a line did not allow for this case. They now check for
2368 it specially.
2370 20. pcretest is supposed to handle patterns and data of any length, by
2371 extending its buffers when necessary. It was getting this wrong when the
2372 buffer for a data line had to be extended.
2374 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2375 CRLF as a newline sequence.
2377 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2378 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2379 I have nevertheless tidied it up.
2381 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2383 24. Added a man page for pcre-config.
2386 Version 7.0 19-Dec-06
2387 ---------------------
2389 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2390 moving to gcc 4.1.1.
2392 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2393 sys/time.h, which is documented as needed for this function. It doesn't
2394 seem to matter on Linux, but it showed up on some releases of OS X.
2396 3. It seems that there are systems where bytes whose values are greater than
2397 127 match isprint() in the "C" locale. The "C" locale should be the
2398 default when a C program starts up. In most systems, only ASCII printing
2399 characters match isprint(). This difference caused the output from pcretest
2400 to vary, making some of the tests fail. I have changed pcretest so that:
2402 (a) When it is outputting text in the compiled version of a pattern, bytes
2403 other than 32-126 are always shown as hex escapes.
2405 (b) When it is outputting text that is a matched part of a subject string,
2406 it does the same, unless a different locale has been set for the match
2407 (using the /L modifier). In this case, it uses isprint() to decide.
2409 4. Fixed a major bug that caused incorrect computation of the amount of memory
2410 required for a compiled pattern when options that changed within the
2411 pattern affected the logic of the preliminary scan that determines the
2412 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2413 that the computed length was too small. The symptoms of this bug were
2414 either the PCRE error "internal error: code overflow" from pcre_compile(),
2415 or a glibc crash with a message such as "pcretest: free(): invalid next
2416 size (fast)". Examples of patterns that provoked this bug (shown in
2417 pcretest format) are:
2419 /(?-x: )/x
2420 /(?x)(?-x: \s*#\s*)/
2421 /((?i)[\x{c0}])/8
2422 /(?i:[\x{c0}])/8
2424 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2425 is now done differently.
2427 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2428 wrapper classes; (b) implement a new function in the C++ scanner that is
2429 more efficient than the old way of doing things because it avoids levels of
2430 recursion in the regex matching; (c) add a paragraph to the documentation
2431 for the FullMatch() function.
2433 6. The escape sequence \n was being treated as whatever was defined as
2434 "newline". Not only was this contrary to the documentation, which states
2435 that \n is character 10 (hex 0A), but it also went horribly wrong when
2436 "newline" was defined as CRLF. This has been fixed.
2438 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2439 was being set to -1 for the "end of line" case (supposedly a value that no
2440 character can have). Though this value is never used (the check for end of
2441 line is "zero bytes in current character"), it caused compiler complaints.
2442 I've changed it to 0xffffffff.
2444 8. In pcre_version.c, the version string was being built by a sequence of
2445 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2446 string (as it is for production releases) called a macro with an empty
2447 argument. The C standard says the result of this is undefined. The gcc
2448 compiler treats it as an empty string (which was what was wanted) but it is
2449 reported that Visual C gives an error. The source has been hacked around to
2450 avoid this problem.
2452 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2453 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2454 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2455 of them did).
2457 10. Originally, pcretest opened its input and output without "b"; then I was
2458 told that "b" was needed in some environments, so it was added for release
2459 5.0 to both the input and output. (It makes no difference on Unix-like
2460 systems.) Later I was told that it is wrong for the input on Windows. I've
2461 now abstracted the modes into two macros, to make it easier to fiddle with
2462 them, and removed "b" from the input mode under Windows.
2464 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2466 12. Added -help and --help to pcretest as an official way of being reminded
2467 of the options.
2469 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2470 and pcrecpp.cc because they annoy compilers at high warning levels.
2472 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2474 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2475 scripts are not C programs :-) and which was not noticed because it works
2476 on Linux.
2478 16. pcretest is supposed to handle any length of pattern and data line (as one
2479 line or as a continued sequence of lines) by extending its input buffer if
2480 necessary. This feature was broken for very long pattern lines, leading to
2481 a string of junk being passed to pcre_compile() if the pattern was longer
2482 than about 50K.
2484 17. I have done a major re-factoring of the way pcre_compile() computes the
2485 amount of memory needed for a compiled pattern. Previously, there was code
2486 that made a preliminary scan of the pattern in order to do this. That was
2487 OK when PCRE was new, but as the facilities have expanded, it has become
2488 harder and harder to keep it in step with the real compile phase, and there
2489 have been a number of bugs (see for example, 4 above). I have now found a
2490 cunning way of running the real compile function in a "fake" mode that
2491 enables it to compute how much memory it would need, while actually only
2492 ever using a few hundred bytes of working memory and without too many
2493 tests of the mode. This should make future maintenance and development
2494 easier. A side effect of this work is that the limit of 200 on the nesting
2495 depth of parentheses has been removed (though this was never a serious
2496 limitation, I suspect). However, there is a downside: pcre_compile() now
2497 runs more slowly than before (30% or more, depending on the pattern). I
2498 hope this isn't a big issue. There is no effect on runtime performance.
2500 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2501 newline (only possible for the last line of a file) and it was a
2502 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2504 19. Added additional timing features to pcretest. (1) The -tm option now times
2505 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2506 separate command line item, by a number that specifies the number of
2507 repeats to use when timing. The default is 50000; this gives better
2508 precision, but takes uncomfortably long for very large patterns.
2510 20. Extended pcre_study() to be more clever in cases where a branch of a
2511 subpattern has no definite first character. For example, (a*|b*)[cd] would
2512 previously give no result from pcre_study(). Now it recognizes that the
2513 first character must be a, b, c, or d.
2515 21. There was an incorrect error "recursive call could loop indefinitely" if
2516 a subpattern (or the entire pattern) that was being tested for matching an
2517 empty string contained only one non-empty item after a nested subpattern.
2518 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2519 incorrectly, because the \d was being skipped in the check.
2521 22. The pcretest program now has a new pattern option /B and a command line
2522 option -b, which is equivalent to adding /B to every pattern. This causes
2523 it to show the compiled bytecode, without the additional information that
2524 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2525 is the same as /B/I).
2527 23. A new optimization is now able automatically to treat some sequences such
2528 as a*b as a*+b. More specifically, if something simple (such as a character
2529 or a simple class like \d) has an unlimited quantifier, and is followed by
2530 something that cannot possibly match the quantified thing, the quantifier
2531 is automatically "possessified".
2533 24. A recursive reference to a subpattern whose number was greater than 39
2534 went wrong under certain circumstances in UTF-8 mode. This bug could also
2535 have affected the operation of pcre_study().
2537 25. Realized that a little bit of performance could be had by replacing
2538 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2540 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2542 27. Possessive quantifiers such as a++ were previously implemented by turning
2543 them into atomic groups such as ($>a+). Now they have their own opcodes,
2544 which improves performance. This includes the automatically created ones
2545 from 23 above.
2547 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2548 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2549 the first matched character to be a colon. This applied both to named and
2550 numbered groups.
2552 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2554 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2555 building PCRE on Windows easier, so I have included it in the distribution.
2557 31. There is now a check in pcretest against a ridiculously large number being
2558 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2559 loop, the loop is abandoned.
2561 32. Forward references to subpatterns in conditions such as (?(2)...) where
2562 subpattern 2 is defined later cause pcre_compile() to search forwards in
2563 the pattern for the relevant set of parentheses. This search went wrong
2564 when there were unescaped parentheses in a character class, parentheses
2565 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2567 33. "Subroutine" calls and backreferences were previously restricted to
2568 referencing subpatterns earlier in the regex. This restriction has now
2569 been removed.
2571 34. Added a number of extra features that are going to be in Perl 5.10. On the
2572 whole, these are just syntactic alternatives for features that PCRE had
2573 previously implemented using the Python syntax or my own invention. The
2574 other formats are all retained for compatibility.
2576 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2577 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2578 also .NET compatible.
2580 (b) A recursion or subroutine call to a named group can now be defined as
2581 (?&name) as well as (?P>name).
2583 (c) A backreference to a named group can now be defined as \k<name> or
2584 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2585 5.10, are also .NET compatible.
2587 (d) A conditional reference to a named group can now use the syntax
2588 (?(<name>) or (?('name') as well as (?(name).
2590 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2591 groups (named and numbered) that are never evaluated inline, but can be
2592 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2593 is always false. There may be only one alternative in such a group.
2595 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2596 as the simple (?(R). The condition is true only if the most recent
2597 recursion is that of the given number or name. It does not search out
2598 through the entire recursion stack.
2600 (g) The escape \gN or \g{N} has been added, where N is a positive or
2601 negative number, specifying an absolute or relative reference.
2603 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2604 some "unreachable code" warnings.
2606 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2607 things, this adds five new scripts.
2609 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2610 There were also incompatibilities regarding the handling of \Q..\E inside
2611 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2612 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2614 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2615 matches an empty string, and forcibly breaks the loop. There were bugs in
2616 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2617 against aaaa the result was just "a" rather than "aaaa", for example. Two
2618 separate and independent bugs (that affected different cases) have been
2619 fixed.
2621 39. Refactored the code to abolish the use of different opcodes for small
2622 capturing bracket numbers. This is a tidy that I avoided doing when I
2623 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2624 The new approach is not only tidier, it makes it possible to reduce the
2625 memory needed to fix the previous bug (38).
2627 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2628 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2629 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2630 mode.
2632 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2633 report.
2635 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2636 copy construction and assignment in the C++ wrapper.
2638 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2639 couple of bugs that could have given wrong results in the "--newline=crlf"
2640 case.
2642 44. Added a number of casts and did some reorganization of signed/unsigned int
2643 variables following suggestions from Dair Grant. Also renamed the variable
2644 "this" as "item" because it is a C++ keyword.
2646 45. Arranged for dftables to add
2648 #include "pcre_internal.h"
2650 to pcre_chartables.c because without it, gcc 4.x may remove the array
2651 definition from the final binary if PCRE is built into a static library and
2652 dead code stripping is activated.
2654 46. For an unanchored pattern, if a match attempt fails at the start of a
2655 newline sequence, and the newline setting is CRLF or ANY, and the next two
2656 characters are CRLF, advance by two characters instead of one.
2659 Version 6.7 04-Jul-06
2660 ---------------------
2662 1. In order to handle tests when input lines are enormously long, pcretest has
2663 been re-factored so that it automatically extends its buffers when
2664 necessary. The code is crude, but this _is_ just a test program. The
2665 default size has been increased from 32K to 50K.
2667 2. The code in pcre_study() was using the value of the re argument before
2668 testing it for NULL. (Of course, in any sensible call of the function, it
2669 won't be NULL.)
2671 3. The memmove() emulation function in pcre_internal.h, which is used on
2672 systems that lack both memmove() and bcopy() - that is, hardly ever -
2673 was missing a "static" storage class specifier.
2675 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2676 containing an extended class (one that cannot be represented by a bitmap
2677 because it contains high-valued characters or Unicode property items, e.g.
2678 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2679 pattern, but PCRE should not loop if you do not (it no longer does).
2680 [Detail: two cases were found: (a) a repeated subpattern containing an
2681 extended class; (b) a recursive reference to a subpattern that followed a
2682 previous extended class. It wasn't skipping over the extended class
2683 correctly when UTF-8 mode was not set.]
2685 5. A negated single-character class was not being recognized as fixed-length
2686 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2687 compile error "lookbehind assertion is not fixed length".
2689 6. The RunPerlTest auxiliary script was showing an unexpected difference
2690 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2691 write a Perl script that can interpret lines of an input file either as
2692 byte characters or as UTF-8, which is what "perltest" was being required to
2693 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2694 can't do is switch easily at run time between having the "use utf8;" pragma
2695 or not. In the end, I fudged it by using the RunPerlTest script to insert
2696 "use utf8;" explicitly for the UTF-8 tests.
2698 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2699 the end of the subject string, contrary to the documentation and to what
2700 Perl does. This was true of both matching functions. Now it matches only at
2701 the start of the subject and immediately after *internal* newlines.
2703 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2704 a pointer to an int instead of a pointer to an unsigned long int. This
2705 caused problems on 64-bit systems.
2707 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2708 instance of the 'standard' template library not being so standard".
2710 10. There was no check on the number of named subpatterns nor the maximum
2711 length of a subpattern name. The product of these values is used to compute
2712 the size of the memory block for a compiled pattern. By supplying a very
2713 long subpattern name and a large number of named subpatterns, the size
2714 computation could be caused to overflow. This is now prevented by limiting
2715 the length of names to 32 characters, and the number of named subpatterns
2716 to 10,000.
2718 11. Subpatterns that are repeated with specific counts have to be replicated in
2719 the compiled pattern. The size of memory for this was computed from the
2720 length of the subpattern and the repeat count. The latter is limited to
2721 65535, but there was no limit on the former, meaning that integer overflow
2722 could in principle occur. The compiled length of a repeated subpattern is
2723 now limited to 30,000 bytes in order to prevent this.
2725 12. Added the optional facility to have named substrings with the same name.
2727 13. Added the ability to use a named substring as a condition, using the
2728 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2729 are numbers (not recommended). Forward references are permitted.
2731 14. Added forward references in named backreferences (if you see what I mean).
2733 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2734 pattern could run off the end of the subject. For example, the pattern
2735 "(?s)(.{1,5})"8 did this with the subject "ab".
2737 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2738 PCRE_CASELESS was set when matching characters that were quantified with ?
2739 or *.
2741 17. A character class other than a single negated character that had a minimum
2742 but no maximum quantifier - for example [ab]{6,} - was not handled
2743 correctly by pce_dfa_exec(). It would match only one character.
2745 18. A valid (though odd) pattern that looked like a POSIX character
2746 class but used an invalid character after [ (for example [[,abc,]]) caused
2747 pcre_compile() to give the error "Failed: internal error: code overflow" or
2748 in some cases to crash with a glibc free() error. This could even happen if
2749 the pattern terminated after [[ but there just happened to be a sequence of
2750 letters, a binary zero, and a closing ] in the memory that followed.
2752 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2753 over the years. Originally (before any Unicode support), just the bottom 8
2754 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2755 output from "man perlunicode" includes this:
2757 The regular expression compiler produces polymorphic opcodes. That
2758 is, the pattern adapts to the data and automatically switches to
2759 the Unicode character scheme when presented with Unicode data--or
2760 instead uses a traditional byte scheme when presented with byte
2761 data.
2763 Sadly, a wide octal escape does not cause a switch, and in a string with
2764 no other multibyte characters, these octal escapes are treated as before.
2765 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2766 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2767 Unicode string.
2769 I have not perpetrated such confusion in PCRE. Up till now, it took just
2770 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2771 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2772 translate to the appropriate multibyte character.
2774 29. Applied some refactoring to reduce the number of warnings from Microsoft
2775 and Borland compilers. This has included removing the fudge introduced
2776 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2777 a warning about an unused variable.
2779 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2780 characters since release 4.0, because Perl (from release 5.004) does not.
2781 [Or at least, is documented not to: some releases seem to be in conflict
2782 with the documentation.] However, when a pattern was studied with
2783 pcre_study() and all its branches started with \s, PCRE still included VT
2784 as a possible starting character. Of course, this did no harm; it just
2785 caused an unnecessary match attempt.
2787 22. Removed a now-redundant internal flag bit that recorded the fact that case
2788 dependency changed within the pattern. This was once needed for "required
2789 byte" processing, but is no longer used. This recovers a now-scarce options
2790 bit. Also moved the least significant internal flag bit to the most-
2791 significant bit of the word, which was not previously used (hangover from
2792 the days when it was an int rather than a uint) to free up another bit for
2793 the future.
2795 23. Added support for CRLF line endings as well as CR and LF. As well as the
2796 default being selectable at build time, it can now be changed at runtime
2797 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2798 specify that it is scanning data with non-default line endings.
2800 24. Changed the definition of CXXLINK to make it agree with the definition of
2801 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2803 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2804 recursions. This makes a big different to stack usage for some patterns.
2806 26. If a subpattern containing a named recursion or subroutine reference such
2807 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2808 the space required for the compiled pattern went wrong and gave too small a
2809 value. Depending on the environment, this could lead to "Failed: internal
2810 error: code overflow at offset 49" or "glibc detected double free or
2811 corruption" errors.
2813 27. Applied patches from Google (a) to support the new newline modes and (b) to
2814 advance over multibyte UTF-8 characters in GlobalReplace.
2816 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2817 difference for some implementation of PCRE in some Windows version.
2819 29. Added some extra testing facilities to pcretest:
2821 \q<number> in a data line sets the "match limit" value
2822 \Q<number> in a data line sets the "match recursion limt" value
2823 -S <number> sets the stack size, where <number> is in megabytes
2825 The -S option isn't available for Windows.
2828 Version 6.6 06-Feb-06
2829 ---------------------
2831 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2832 in pcreposix.h. I have copied the definition from pcre.h.
2834 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2835 because pcre.h is no longer a built file.
2837 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2838 not normally included in the compiled code.
2841 Version 6.5 01-Feb-06
2842 ---------------------
2844 1. When using the partial match feature with pcre_dfa_exec(), it was not
2845 anchoring the second and subsequent partial matches at the new starting
2846 point. This could lead to incorrect results. For example, with the pattern
2847 /1234/, partially matching against "123" and then "a4" gave a match.
2849 2. Changes to pcregrep:
2851 (a) All non-match returns from pcre_exec() were being treated as failures
2852 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2853 error message is output. Some extra information is given for the
2855 probably the only errors that are likely to be caused by users (by
2856 specifying a regex that has nested indefinite repeats, for instance).
2857 If there are more than 20 of these errors, pcregrep is abandoned.
2859 (b) A binary zero was treated as data while matching, but terminated the
2860 output line if it was written out. This has been fixed: binary zeroes
2861 are now no different to any other data bytes.
2863 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2864 used to set a locale for matching. The --locale=xxxx long option has
2865 been added (no short equivalent) to specify a locale explicitly on the
2866 pcregrep command, overriding the environment variables.
2868 (d) When -B was used with -n, some line numbers in the output were one less
2869 than they should have been.
2871 (e) Added the -o (--only-matching) option.
2873 (f) If -A or -C was used with -c (count only), some lines of context were
2874 accidentally printed for the final match.
2876 (g) Added the -H (--with-filename) option.
2878 (h) The combination of options -rh failed to suppress file names for files
2879 that were found from directory arguments.
2881 (i) Added the -D (--devices) and -d (--directories) options.
2883 (j) Added the -F (--fixed-strings) option.
2885 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2887 (l) Added the --colo(u)r option.
2889 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2890 is not present by default.
2892 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2893 items such as (?R) or (?1), when the recursion could match a number of
2894 alternatives. If it matched one of the alternatives, but subsequently,
2895 outside the recursion, there was a failure, the code tried to back up into
2896 the recursion. However, because of the way PCRE is implemented, this is not
2897 possible, and the result was an incorrect result from the match.
2899 In order to prevent this happening, the specification of recursion has
2900 been changed so that all such subpatterns are automatically treated as
2901 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2903 4. I had overlooked the fact that, in some locales, there are characters for
2904 which isalpha() is true but neither isupper() nor islower() are true. In
2905 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2906 and ordfeminine) are like this. This affected the treatment of \w and \W
2907 when they appeared in character classes, but not when they appeared outside
2908 a character class. The bit map for "word" characters is now created
2909 separately from the results of isalnum() instead of just taking it from the
2910 upper, lower, and digit maps. (Plus the underscore character, of course.)
2912 5. The above bug also affected the handling of POSIX character classes such as
2913 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2914 permanent tables. Instead, the bit maps for such a class were previously
2915 created as the appropriate unions of the upper, lower, and digit bitmaps.
2916 Now they are created by subtraction from the [[:word:]] class, which has
2917 its own bitmap.
2919 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2920 It is created by subtracting the vertical space characters (\x09, \x0a,
2921 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2922 subtraction was done in the overall bitmap for a character class, meaning
2923 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2924 be recognized. This bug has been fixed.
2926 7. Patches from the folks at Google:
2928 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2929 real life, but is still worth protecting against".
2931 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2932 regular expressions".
2934 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2935 have it.
2937 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2938 "configure" and the latter not, in order to fix a problem somebody had
2939 with compiling the Arg class on HP-UX.
2941 (e) Improve the error-handling of the C++ wrapper a little bit.
2943 (f) New tests for checking recursion limiting.
2945 8. The pcre_memmove() function, which is used only if the environment does not
2946 have a standard memmove() function (and is therefore rarely compiled),
2947 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2948 returning a result (though PCRE never actually uses the result).
2950 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2951 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2952 returned instead of calling malloc() with an overflowing number that would
2953 most likely cause subsequent chaos.
2955 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2957 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2958 with this option is matched, the nmatch and pmatch options of regexec() are
2959 ignored.
2961 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2962 provided in case anyone wants to the the POSIX interface with UTF-8
2963 strings.
2965 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2966 C++ linking (needed for some HP-UX environments).
2968 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2969 (unused parameter) and in the pcre_printint() function (omitted "default"
2970 switch label when the default is to do nothing).
2972 15. Added some code to make it possible, when PCRE is compiled as a C++
2973 library, to replace subject pointers for pcre_exec() with a smart pointer
2974 class, thus making it possible to process discontinuous strings.
2976 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2977 much the same function. They were added by different people who were trying
2978 to make PCRE easy to compile on non-Unix systems. It has been suggested
2979 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2980 for compiling on Windows systems. I have therefore replaced it with
2981 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2982 defaults to "extern" for C or "extern C" for C++, which works fine on
2983 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2984 SCOPE with something explicit in config.h. In addition:
2986 (a) pcreposix.h still had just "extern" instead of either of these macros;
2987 I have replaced it with PCRE_DATA_SCOPE.
2989 (b) Functions such as _pcre_xclass(), which are internal to the library,
2990 but external in the C sense, all had PCRE_EXPORT in their definitions.
2991 This is apparently wrong for the Windows case, so I have removed it.
2992 (It makes no difference on Unix-like systems.)
2994 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2995 of recursive calls to match(). This is different to MATCH_LIMIT because
2996 that limits the total number of calls to match(), not all of which increase
2997 the depth of recursion. Limiting the recursion depth limits the amount of
2998 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2999 when PCRE is compiled, and changed at run time. A patch from Google adds
3000 this functionality to the C++ interface.
3002 18. Changes to the handling of Unicode character properties:
3004 (a) Updated the table to Unicode 4.1.0.
3006 (b) Recognize characters that are not in the table as "Cn" (undefined).
3008 (c) I revised the way the table is implemented to a much improved format
3009 which includes recognition of ranges. It now supports the ranges that
3010 are defined in UnicodeData.txt, and it also amalgamates other
3011 characters into ranges. This has reduced the number of entries in the
3012 table from around 16,000 to around 3,000, thus reducing its size
3013 considerably. I realized I did not need to use a tree structure after
3014 all - a binary chop search is just as efficient. Having reduced the
3015 number of entries, I extended their size from 6 bytes to 8 bytes to
3016 allow for more data.
3018 (d) Added support for Unicode script names via properties such as \p{Han}.
3020 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3021 matching that character.
3023 20. When matching a repeated Unicode property with a minimum greater than zero,
3024 (for example \pL{2,}), PCRE could look past the end of the subject if it
3025 reached it while seeking the minimum number of characters. This could
3026 happen only if some of the characters were more than one byte long, because
3027 there is a check for at least the minimum number of bytes.
3029 21. Refactored the implementation of \p and \P so as to be more general, to
3030 allow for more different types of property in future. This has changed the
3031 compiled form incompatibly. Anybody with saved compiled patterns that use
3032 \p or \P will have to recompile them.
3034 22. Added "Any" and "L&" to the supported property types.
3036 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3037 but give a compile time error if the value is greater than 0xff.
3039 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3040 accidentally not being installed or uninstalled.
3042 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3043 made were to insert the current release number. This seemed silly, because
3044 it made things harder for people building PCRE on systems that don't run
3045 "configure". I have turned pcre.h into a distributed file, no longer built
3046 by "configure", with the version identification directly included. There is
3047 no longer a pcre.h.in file.
3049 However, this change necessitated a change to the pcre-config script as
3050 well. It is built from pcre-config.in, and one of the substitutions was the
3051 release number. I have updated configure.ac so that ./configure now finds
3052 the release number by grepping pcre.h.
3054 26. Added the ability to run the tests under valgrind.
3057 Version 6.4 05-Sep-05
3058 ---------------------
3060 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3061 "--" to be printed when multiple files were scanned, even when none of the
3062 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3063 consider it to be a bug, and have restored the previous behaviour.
3065 2. A couple of code tidies to get rid of compiler warnings.
3067 3. The pcretest program used to cheat by referring to symbols in the library
3068 whose names begin with _pcre_. These are internal symbols that are not
3069 really supposed to be visible externally, and in some environments it is
3070 possible to suppress them. The cheating is now confined to including
3071 certain files from the library's source, which is a bit cleaner.
3073 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3074 file's purpose clearer.
3076 5. Reorganized pcre_ucp_findchar().
3079 Version 6.3 15-Aug-05
3080 ---------------------
3082 1. The file libpcre.pc.in did not have general read permission in the tarball.
3084 2. There were some problems when building without C++ support:
3086 (a) If C++ support was not built, "make install" and "make test" still
3087 tried to test it.
3089 (b) There were problems when the value of CXX was explicitly set. Some
3090 changes have been made to try to fix these, and ...
3092 (c) --disable-cpp can now be used to explicitly disable C++ support.
3094 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3095 backslash in a target when C++ was disabled. This confuses some
3096 versions of "make", apparently. Using an intermediate variable solves
3097 this. (Same for CPP_LOBJ.)
3099 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3100 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3101 necessary on certain architectures.
3103 4. Added a setting of -export-symbols-regex to the link command to remove
3104 those symbols that are exported in the C sense, but actually are local
3105 within the library, and not documented. Their names all begin with
3106 "_pcre_". This is not a perfect job, because (a) we have to except some
3107 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3108 available (and never for static libraries). I have made a note to try to
3109 find a way round (a) in the future.
3112 Version 6.2 01-Aug-05
3113 ---------------------
3115 1. There was no test for integer overflow of quantifier values. A construction
3116 such as {1111111111111111} would give undefined results. What is worse, if
3117 a minimum quantifier for a parenthesized subpattern overflowed and became
3118 negative, the calculation of the memory size went wrong. This could have
3119 led to memory overwriting.
3121 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3123 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3124 operating environments where this matters.
3126 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3127 PCRE options from within the C++ wrapper.
3129 5. Named capturing subpatterns were not being correctly counted when a pattern
3130 was compiled. This caused two problems: (a) If there were more than 100
3131 such subpatterns, the calculation of the memory needed for the whole
3132 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3133 back references of the form \12, where the number was greater than 9, were
3134 not recognized as back references, even though there were sufficient
3135 previous subpatterns.
3137 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3138 versions of gcc, e.g. 2.95.4.
3141 Version 6.1 21-Jun-05
3142 ---------------------
3144 1. There was one reference to the variable "posix" in pcretest.c that was not
3145 surrounded by "#if !defined NOPOSIX".
3147 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3148 the cross-check on the old pcre_info() function, for the benefit of the
3149 cut-down version of PCRE that is currently imported into Exim.
3151 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3152 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3153 patterns that start (?i) but I don't think that matters. The use of (?i) is
3154 just an example; this all applies to the other options as well.
3156 4. Since libtool seems to echo the compile commands it is issuing, the output
3157 from "make" can be reduced a bit by putting "@" in front of each libtool
3158 compile command.
3160 5. Patch from the folks at Google for configure.in to be a bit more thorough
3161 in checking for a suitable C++ installation before trying to compile the
3162 C++ stuff. This should fix a reported problem when a compiler was present,
3163 but no suitable headers.
3165 6. The man pages all had just "PCRE" as their title. I have changed them to
3166 be the relevant file name. I have also arranged that these names are
3167 retained in the file doc/pcre.txt, which is a concatenation in text format
3168 of all the man pages except the little individual ones for each function.
3170 7. The NON-UNIX-USE file had not been updated for the different set of source
3171 files that come with release 6. I also added a few comments about the C++
3172 wrapper.
3175 Version 6.0 07-Jun-05
3176 ---------------------
3178 1. Some minor internal re-organization to help with my DFA experiments.
3180 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3181 didn't matter for the library itself when fully configured, but did matter
3182 when compiling without UCP support, or within Exim, where the ucp files are
3183 not imported.
3185 3. Refactoring of the library code to split up the various functions into
3186 different source modules. The addition of the new DFA matching code (see
3187 below) to a single monolithic source would have made it really too
3188 unwieldy, quite apart from causing all the code to be include in a
3189 statically linked application, when only some functions are used. This is
3190 relevant even without the DFA addition now that patterns can be compiled in
3191 one application and matched in another.
3193 The downside of splitting up is that there have to be some external
3194 functions and data tables that are used internally in different modules of
3195 the library but which are not part of the API. These have all had their
3196 names changed to start with "_pcre_" so that they are unlikely to clash
3197 with other external names.
3199 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3200 a different (DFA) algorithm. Although it is slower than the original
3201 function, it does have some advantages for certain types of matching
3202 problem.
3204 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3205 including restarting after a partial match.
3207 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3208 defined when compiling for Windows was sent to me. I have put it into the
3209 code, though I have no means of testing or verifying it.
3211 7. Added the pcre_refcount() auxiliary function.
3213 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3214 match before or at the first newline in the subject string. In pcretest,
3215 the /f option on a pattern can be used to set this.
3217 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3218 would behave wrongly. This has been present in PCRE since release 4.0.
3220 10. A number of changes to the pcregrep command:
3222 (a) Refactored how -x works; insert ^(...)$ instead of setting
3223 PCRE_ANCHORED and checking the length, in preparation for adding
3224 something similar for -w.
3226 (b) Added the -w (match as a word) option.
3228 (c) Refactored the way lines are read and buffered so as to have more
3229 than one at a time available.
3231 (d) Implemented a pcregrep test script.
3233 (e) Added the -M (multiline match) option. This allows patterns to match
3234 over several lines of the subject. The buffering ensures that at least
3235 8K, or the rest of the document (whichever is the shorter) is available
3236 for matching (and similarly the previous 8K for lookbehind assertions).
3238 (f) Changed the --help output so that it now says
3240 -w, --word-regex(p)
3242 instead of two lines, one with "regex" and the other with "regexp"
3243 because that confused at least one person since the short forms are the
3244 same. (This required a bit of code, as the output is generated
3245 automatically from a table. It wasn't just a text change.)
3247 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3248 option but starts with a hyphen. Could be a pattern or a path name
3249 starting with a hyphen, for instance.
3251 (h) "-" can be given as a file name to represent stdin.
3253 (i) When file names are being printed, "(standard input)" is used for
3254 the standard input, for compatibility with GNU grep. Previously
3255 "<stdin>" was used.
3257 (j) The option --label=xxx can be used to supply a name to be used for
3258 stdin when file names are being printed. There is no short form.
3260 (k) Re-factored the options decoding logic because we are going to add
3261 two more options that take data. Such options can now be given in four
3262 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3264 (l) Added the -A, -B, and -C options for requesting that lines of context
3265 around matches be printed.
3267 (m) Added the -L option to print the names of files that do not contain
3268 any matching lines, that is, the complement of -l.
3270 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3271 continue to scan other files.
3273 (o) The -s option was incorrectly implemented. For compatibility with other
3274 greps, it now suppresses the error message for a non-existent or non-
3275 accessible file (but not the return code). There is a new option called
3276 -q that suppresses the output of matching lines, which was what -s was
3277 previously doing.
3279 (p) Added --include and --exclude options to specify files for inclusion
3280 and exclusion when recursing.
3282 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3283 Hopefully, it now does.
3285 12. Missing cast in pcre_study().
3287 13. Added an "uninstall" target to the makefile.
3289 14. Replaced "extern" in the function prototypes in Makefile.in with
3290 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3291 world, but is set differently for Windows.
3293 15. Added a second compiling function called pcre_compile2(). The only
3294 difference is that it has an extra argument, which is a pointer to an
3295 integer error code. When there is a compile-time failure, this is set
3296 non-zero, in addition to the error test pointer being set to point to an
3297 error message. The new argument may be NULL if no error number is required
3298 (but then you may as well call pcre_compile(), which is now just a
3299 wrapper). This facility is provided because some applications need a
3300 numeric error indication, but it has also enabled me to tidy up the way
3301 compile-time errors are handled in the POSIX wrapper.
3303 16. Added VPATH=.libs to the makefile; this should help when building with one
3304 prefix path and installing with another. (Or so I'm told by someone who
3305 knows more about this stuff than I do.)
3307 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3308 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3309 match everything, including newlines. This is not POSIX-compatible, but
3310 somebody wanted the feature. From pcretest it can be activated by using
3311 both the P and the s flags.
3313 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3315 19. libpcre.pc was being incorrectly installed as executable.
3317 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3318 it now also looks for '\r' so that it will work unmodified on Windows.
3320 21. Added Google's contributed C++ wrapper to the distribution.
3322 22. Added some untidy missing memory free() calls in pcretest, to keep
3323 Electric Fence happy when testing.
3327 Version 5.0 13-Sep-04
3328 ---------------------
3330 1. Internal change: literal characters are no longer packed up into items
3331 containing multiple characters in a single byte-string. Each character
3332 is now matched using a separate opcode. However, there may be more than one
3333 byte in the character in UTF-8 mode.
3335 2. The pcre_callout_block structure has two new fields: pattern_position and
3336 next_item_length. These contain the offset in the pattern to the next match
3337 item, and its length, respectively.
3339 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3340 insertion of callouts before each pattern item. Added the /C option to
3341 pcretest to make use of this.
3343 4. On the advice of a Windows user, the lines
3345 #if defined(_WIN32) || defined(WIN32)
3346 _setmode( _fileno( stdout ), 0x8000 );
3347 #endif /* defined(_WIN32) || defined(WIN32) */
3349 have been added to the source of pcretest. This apparently does useful
3350 magic in relation to line terminators.
3352 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3353 for the benefit of those environments where the "b" makes a difference.
3355 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3356 to know about it. I have put a hack into configure.in that adds in code
3357 to set GCC=yes if CC=icc. This seems to end up at a point in the
3358 generated configure script that is early enough to affect the setting of
3359 compiler options, which is what is needed, but I have no means of testing
3360 whether it really works. (The user who reported this had patched the
3361 generated configure script, which of course I cannot do.)
3363 LATER: After change 22 below (new libtool files), the configure script
3364 seems to know about icc (and also ecc). Therefore, I have commented out
3365 this hack in configure.in.
3367 7. Added support for pkg-config (2 patches were sent in).
3369 8. Negated POSIX character classes that used a combination of internal tables
3370 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3371 [[:^ascii]]. Typically, they would match almost any characters. The other
3372 POSIX classes were not broken in this way.
3374 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3375 to find the match, as PCRE was deluded into thinking that the match had to
3376 start at the start point or following a newline. The same bug applied to
3377 patterns with negative forward assertions or any backward assertions
3378 preceding ".*" at the start, unless the pattern required a fixed first
3379 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3381 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3382 starting at the last subject character, bytes beyond the end of the subject
3383 string were read.
3385 11. Renamed the variable "class" as "classbits" to make life easier for C++
3386 users. (Previously there was a macro definition, but it apparently wasn't
3387 enough.)
3389 12. Added the new field "tables" to the extra data so that tables can be passed
3390 in at exec time, or the internal tables can be re-selected. This allows
3391 a compiled regex to be saved and re-used at a later time by a different
3392 program that might have everything at different addresses.
3394 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3395 -R library as well as a -L library.
3397 14. The debugging options of pcretest (-d on the command line or D on a
3398 pattern) showed incorrect output for anything following an extended class
3399 that contained multibyte characters and which was followed by a quantifier.
3401 15. Added optional support for general category Unicode character properties
3402 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3403 support. It adds about 90K to the size of the library. The meanings of the
3404 inbuilt class escapes such as \d and \s have NOT been changed.
3406 16. Updated pcredemo.c to include calls to free() to release the memory for the
3407 compiled pattern.
3409 17. The generated file chartables.c was being created in the source directory
3410 instead of in the building directory. This caused the build to fail if the
3411 source directory was different from the building directory, and was
3412 read-only.
3414 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3415 file. No doubt somebody will tell me if they don't make sense... Also added
3416 Dan Mooney's comments about building on OpenVMS.
3418 19. Added support for partial matching via the PCRE_PARTIAL option for
3419 pcre_exec() and the \P data escape in pcretest.
3421 20. Extended pcretest with 3 new pattern features:
3423 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3424 write the compiled pattern to the file whose name is "rest-of-line".
3425 This is a straight binary dump of the data, with the saved pointer to
3426 the character tables forced to be NULL. The study data, if any, is
3427 written too. After writing, pcretest reads a new pattern.
3429 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3430 compiled pattern from the given file. There must not be any
3431 occurrences of "<" in the file name (pretty unlikely); if there are,
3432 pcretest will instead treat the initial "<" as a pattern delimiter.
3433 After reading in the pattern, pcretest goes on to read data lines as
3434 usual.
3436 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3437 and 16-bit fields in a compiled pattern, to simulate a pattern that
3438 was compiled on a host of opposite endianness.
3440 21. The pcre-exec() function can now cope with patterns that were compiled on
3441 hosts of opposite endianness, with this restriction:
3443 As for any compiled expression that is saved and used later, the tables
3444 pointer field cannot be preserved; the extra_data field in the arguments
3445 to pcre_exec() should be used to pass in a tables address if a value
3446 other than the default internal tables were used at compile time.
3448 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3449 now diagnosed as an error. Previously, most of the time, a negative number
3450 would have been treated as zero, but if in addition "ovector" was passed as
3451 NULL, a crash could occur.
3453 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3454 new versions from the libtool 1.5 distribution (the last one is a copy of
3455 a file called libtool.m4). This seems to have fixed the need to patch
3456 "configure" to support Darwin 1.3 (which I used to do). However, I still
3457 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3458 workstation).
3460 24. Changed the PCRE licence to be the more standard "BSD" licence.
3463 Version 4.5 01-Dec-03
3464 ---------------------
3466 1. There has been some re-arrangement of the code for the match() function so
3467 that it can be compiled in a version that does not call itself recursively.
3468 Instead, it keeps those local variables that need separate instances for
3469 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3470 needs to "recurse". Keeping track of where control must go is done by means
3471 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3472 hide most of the details from the main code, and operates only if
3473 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3474 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3475 operating.
3477 To make it easier for callers to provide specially tailored get/free
3478 functions for this usage, two new functions, pcre_stack_malloc, and
3479 pcre_stack_free, are used. They are always called in strict stacking order,
3480 and the size of block requested is always the same.
3482 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3483 PCRE has been compiled to use the stack or the heap for recursion. The
3484 -C option of pcretest uses this to show which version is compiled.
3486 A new data escape \S, is added to pcretest; it causes the amounts of store
3487 obtained and freed by both kinds of malloc/free at match time to be added
3488 to the output.
3490 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3491 what's available on my current Linux desktop machine.
3493 3. When matching a UTF-8 string, the test for a valid string at the start has
3494 been extended. If start_offset is not zero, PCRE now checks that it points
3495 to a byte that is the start of a UTF-8 character. If not, it returns
3496 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3497 this is necessary because there may be backward assertions in the pattern.
3498 When matching the same subject several times, it may save resources to use
3499 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3501 4. The code for checking the validity of UTF-8 strings has been tightened so
3502 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3503 containing "overlong sequences".
3505 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3506 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3507 should have been "&", but it just so happened that all the cases this let
3508 through by mistake were picked up later in the function.
3510 6. I had used a variable called "isblank" - this is a C99 function, causing
3511 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3513 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3514 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3515 so the version is shown; (c) stop "make test" echoing "./RunTest".
3517 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3519 9. The prototype for memmove() for systems that don't have it was using
3520 size_t, but the inclusion of the header that defines size_t was later. I've
3521 moved the #includes for the C headers earlier to avoid this.
3523 10. Added some adjustments to the code to make it easier to compiler on certain
3524 special systems:
3526 (a) Some "const" qualifiers were missing.
3527 (b) Added the macro EXPORT before all exported functions; by default this
3528 is defined to be empty.
3529 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3530 that it reads its output file name as an argument instead of writing
3531 to the standard output and assuming this can be redirected.
3533 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3534 class containing characters with values greater than 255, PCRE compilation
3535 went into a loop.
3537 12. A recursive reference to a subpattern that was within another subpattern
3538 that had a minimum quantifier of zero caused PCRE to crash. For example,
3539 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3540 recursion. If the recursively-called subpattern itself had a zero repeat,
3541 that was OK.
3543 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3544 buffer into which it was copied (for escape processing) was still set at
3545 1024, so long lines caused crashes.
3547 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3548 "internal error: code overflow...". This applied to any character class
3549 that was followed by a possessive quantifier.
3551 15. Modified the Makefile to add libpcre.la as a prerequisite for
3552 libpcreposix.la because I was told this is needed for a parallel build to
3553 work.
3555 16. If a pattern that contained .* following optional items at the start was
3556 studied, the wrong optimizing data was generated, leading to matching
3557 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3558 matching string must start with a or b or c. The correct conclusion for
3559 this pattern is that a match can start with any character.
3562 Version 4.4 13-Aug-03
3563 ---------------------
3565 1. In UTF-8 mode, a character class containing characters with values between
3566 127 and 255 was not handled correctly if the compiled pattern was studied.
3567 In fixing this, I have also improved the studying algorithm for such
3568 classes (slightly).
3570 2. Three internal functions had redundant arguments passed to them. Removal
3571 might give a very teeny performance improvement.
3573 3. Documentation bug: the value of the capture_top field in a callout is *one
3574 more than* the number of the hightest numbered captured substring.
3576 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3577 in incorrectly linking with a previously installed version. They now link
3578 explicitly with libpcre.la.
3580 5. configure.in no longer needs to recognize Cygwin specially.
3582 6. A problem in pcre.in for Windows platforms is fixed.
3584 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3585 pcretest, it used to include the size of the study block as part of its
3586 output. Unfortunately, the structure contains a field that has a different
3587 size on different hardware architectures. This meant that the tests that
3588 showed this size failed. As the block is currently always of a fixed size,
3589 this information isn't actually particularly useful in pcretest output, so
3590 I have just removed it.
3592 8. Three pre-processor statements accidentally did not start in column 1.
3593 Sadly, there are *still* compilers around that complain, even though
3594 standard C has not required this for well over a decade. Sigh.
3596 9. In pcretest, the code for checking callouts passed small integers in the
3597 callout_data field, which is a void * field. However, some picky compilers
3598 complained about the casts involved for this on 64-bit systems. Now
3599 pcretest passes the address of the small integer instead, which should get
3600 rid of the warnings.
3602 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3603 both compile and run time, and gives an error if an invalid UTF-8 sequence
3604 is found. There is a option for disabling this check in cases where the
3605 string is known to be correct and/or the maximum performance is wanted.
3607 11. In response to a bug report, I changed one line in Makefile.in from
3609 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3610 to
3611 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3613 to look similar to other lines, but I have no way of telling whether this
3614 is the right thing to do, as I do not use Windows. No doubt I'll get told
3615 if it's wrong...
3618 Version 4.3 21-May-03
3619 ---------------------
3621 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3622 Makefile.
3624 2. Some refactoring to improve the quality of the code:
3626 (i) The utf8_table... variables are now declared "const".
3628 (ii) The code for \cx, which used the "case flipping" table to upper case
3629 lower case letters, now just substracts 32. This is ASCII-specific,
3630 but the whole concept of \cx is ASCII-specific, so it seems
3631 reasonable.
3633 (iii) PCRE was using its character types table to recognize decimal and
3634 hexadecimal digits in the pattern. This is silly, because it handles
3635 only 0-9, a-f, and A-F, but the character types table is locale-
3636 specific, which means strange things might happen. A private
3637 table is now used for this - though it costs 256 bytes, a table is
3638 much faster than multiple explicit tests. Of course, the standard
3639 character types table is still used for matching digits in subject
3640 strings against \d.
3642 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3643 ending in _t are). So I've renamed it as ESC_tee.
3645 3. The first argument for regexec() in the POSIX wrapper should have been
3646 defined as "const".
3648 4. Changed pcretest to use malloc() for its buffers so that they can be
3649 Electric Fenced for debugging.
3651 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3652 to read one or more bytes before the start of the subject string. Often this
3653 had no effect on PCRE's behaviour, but in some circumstances it could
3654 provoke a segmentation fault.
3656 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3657 to try to read one or more bytes before the start of the subject string.
3659 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3660 UTF-8 support could misbehave in various ways if the subject string
3661 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3662 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3663 back over UTF-8 characters.)
3666 Version 4.2 14-Apr-03
3667 ---------------------
3669 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3671 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3672 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3673 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3674 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3675 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3677 Note: automatic setting of the BUILD variables is not yet working
3678 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3679 compile-time but not at link-time
3680 [LINK]: use for linking executables only
3681 make different versions for Windows and non-Windows
3682 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3683 libraries
3684 [LINK_FOR_BUILD]: new variable
3685 [OBJEXT]: use throughout
3686 [EXEEXT]: use throughout
3687 <winshared>: new target
3688 <wininstall>: new target
3689 <dftables.o>: use native compiler
3690 <dftables>: use native linker
3691 <install>: handle Windows platform correctly
3692 <clean>: ditto
3693 <check>: ditto
3694 copy DLL to top builddir before testing
3696 As part of these changes, -no-undefined was removed again. This was reported
3697 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3698 in any case.
3700 3. Some tidies to get rid of compiler warnings:
3702 . In the match_data structure, match_limit was an unsigned long int, whereas
3703 match_call_count was an int. I've made them both unsigned long ints.
3705 . In pcretest the fact that a const uschar * doesn't automatically cast to
3706 a void * provoked a warning.
3708 . Turning on some more compiler warnings threw up some "shadow" variables
3709 and a few more missing casts.
3711 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3712 option, a class that contained a single character with a value between 128
3713 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3715 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3716 option, a class that contained several characters, but with at least one
3717 whose value was between 128 and 255 caused PCRE to crash.
3720 Version 4.1 12-Mar-03
3721 ---------------------
3723 1. Compiling with gcc -pedantic found a couple of places where casts were
3724 needed, and a string in dftables.c that was longer than standard compilers are
3725 required to support.
3727 2. Compiling with Sun's compiler found a few more places where the code could
3728 be tidied up in order to avoid warnings.
3730 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3731 first of these names is deprecated in the latest Autoconf in favour of the name
3732 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3733 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3734 analogy I have changed it to CFLAGS_FOR_BUILD.
3736 4. Added -no-undefined to the linking command in the Makefile, because this is
3737 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3738 linking step for the pcreposix library.
3740 5. PCRE was failing to diagnose the case of two named groups with the same
3741 name.
3743 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3744 literal character that is needed in the subject for a match, and scans along to
3745 ensure that it is present before embarking on the full matching process. This
3746 saves time in cases of nested unlimited repeats that are never going to match.
3747 Problem: the scan can take a lot of time if the subject is very long (e.g.
3748 megabytes), thus penalizing straightforward matches. It is now done only if the
3749 amount of subject to be scanned is less than 1000 bytes.
3751 7. A lesser problem with the same optimization is that it was recording the
3752 first character of an anchored pattern as "needed", thus provoking a search
3753 right along the subject, even when the first match of the pattern was going to
3754 fail. The "needed" character is now not set for anchored patterns, unless it
3755 follows something in the pattern that is of non-fixed length. Thus, it still
3756 fulfils its original purpose of finding quick non-matches in cases of nested
3757 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3760 Version 4.0 17-Feb-03
3761 ---------------------
3763 1. If a comment in an extended regex that started immediately after a meta-item
3764 extended to the end of string, PCRE compiled incorrect data. This could lead to
3765 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3767 2. Moved to autoconf 2.53 and libtool 1.4.2.
3769 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3770 the special perltest8 script is no longer needed - all the tests can be run
3771 from a single perltest script.
3773 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3774 by \s. It has now been removed in PCRE. This means it isn't recognized as
3775 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3776 class [:space:] *does* include VT, thereby creating a mess.
3778 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3779 space and tab.
3781 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3782 its new features into the main test script, reducing the number of scripts.
3784 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3785 were backward compatible, and made the (?i) apply to the whole pattern, as if
3786 /i were given. Now it behaves more logically, and applies the option setting
3787 only to what follows. PCRE has been changed to follow suit. However, if it
3788 finds options settings right at the start of the pattern, it extracts them into
3789 the global options, as before. Thus, they show up in the info data.
3791 8. Added support for the \Q...\E escape sequence. Characters in between are
3792 treated as literals. This is slightly different from Perl in that $ and @ are
3793 also handled as literals inside the quotes. In Perl, they will cause variable
3794 interpolation. Note the following examples:
3796 Pattern PCRE matches Perl matches
3798 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3799 \Qabc\$xyz\E abc\$xyz abc\$xyz
3800 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3802 For compatibility with Perl, \Q...\E sequences are recognized inside character
3803 classes as well as outside them.
3805 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3806 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3807 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3808 signed/unsigned warnings.
3810 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3811 option for pcretest, so I've replaced it by a simple function that does just
3812 that job.
3814 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3815 "pcregrep -".
3817 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3818 Java package. This provides some syntactic sugar for simple cases of what my
3819 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3820 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3821 item, you can use this simplified notation. Note that only makes sense with
3822 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3823 greediness, whatever the setting of the PCRE_UNGREEDY option.
3825 13. A change of greediness default within a pattern was not taking effect at
3826 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3827 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3828 was abstracted outside.
3830 14. PCRE now supports the \G assertion. It is true when the current matching
3831 position is at the start point of the match. This differs from \A when the
3832 starting offset is non-zero. Used with the /g option of pcretest (or similar
3833 code), it works in the same way as it does for Perl's /g option. If all
3834 alternatives of a regex begin with \G, the expression is anchored to the start
3835 match position, and the "anchored" flag is set in the compiled expression.
3837 15. Some bugs concerning the handling of certain option changes within patterns
3838 have been fixed. These applied to options other than (?ims). For example,
3839 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3840 been the other way round. Some of this was related to change 7 above.
3842 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3843 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3844 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3845 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3847 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3848 mode. Unlike ".", it always matches newline, whatever the setting of
3849 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3850 assertions. Perl allows it, but it doesn't (in general) work because it can't
3851 calculate the length of the lookbehind. At least, that's the case for Perl
3852 5.8.0 - I've been told they are going to document that it doesn't work in
3853 future.
3855 18. Added an error diagnosis for escapes that PCRE does not support: these are
3856 \L, \l, \N, \P, \p, \U, \u, and \X.
3858 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3859 reading past the end of the pattern in cases such as /[abcd/.
3861 20. PCRE was getting more memory than necessary for patterns with classes that
3862 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3864 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3865 compiling PCRE for use with Virtual Pascal.
3867 22. Small fix to the Makefile to make it work properly if the build is done
3868 outside the source tree.
3870 23. Added a new extension: a condition to go with recursion. If a conditional
3871 subpattern starts with (?(R) the "true" branch is used if recursion has
3872 happened, whereas the "false" branch is used only at the top level.
3874 24. When there was a very long string of literal characters (over 255 bytes
3875 without UTF support, over 250 bytes with UTF support), the computation of how
3876 much memory was required could be incorrect, leading to segfaults or other
3877 strange effects.
3879 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3880 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3881 there was a subsequent back reference to those brackets. This meant that, for
3882 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3883 possible to check for precisely this case. All we can do is abandon the
3884 optimization if .* occurs inside capturing brackets when there are any back
3885 references whatsoever. (See below for a better fix that came later.)
3887 26. The handling of the optimization for finding the first character of a
3888 non-anchored pattern, and for finding a character that is required later in the
3889 match were failing in some cases. This didn't break the matching; it just
3890 failed to optimize when it could. The way this is done has been re-implemented.
3892 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3894 28. Added a new feature that provides some of the functionality that Perl
3895 provides with (?{...}). The facility is termed a "callout". The way it is done
3896 in PCRE is for the caller to provide an optional function, by setting
3897 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3898 global variable. By default it is unset, which disables all calling out. To get
3899 the function called, the regex must include (?C) at appropriate points. This
3900 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3901 This provides a means of identifying different callout points. When PCRE
3902 reaches such a point in the regex, if pcre_callout has been set, the external
3903 function is called. It is provided with data in a structure called
3904 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3905 matching continues; if it returns a non-zero value, the match at the current
3906 point fails. However, backtracking will occur if possible. [This was changed
3907 later and other features added - see item 49 below.]
3909 29. pcretest is upgraded to test the callout functionality. It provides a
3910 callout function that displays information. By default, it shows the start of
3911 the match and the current position in the text. There are some new data escapes
3912 to vary what happens:
3914 \C+ in addition, show current contents of captured substrings
3915 \C- do not supply a callout function
3916 \C!n return 1 when callout number n is reached
3917 \C!n!m return 1 when callout number n is reached for the mth time
3919 30. If pcregrep was called with the -l option and just a single file name, it
3920 output "<stdin>" if a match was found, instead of the file name.
3922 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3923 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3924 pcre_exec(). This saves a malloc/free per call. The default value of
3925 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3926 when configuring.
3928 32. The default maximum size of a compiled pattern is 64K. There have been a
3929 few cases of people hitting this limit. The code now uses macros to handle the
3930 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3931 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3932 configuring. Tests 2 and 5 work only with 2-byte links because they output
3933 debugging information about compiled patterns.
3935 33. Internal code re-arrangements:
3937 (a) Moved the debugging function for printing out a compiled regex into
3938 its own source file (printint.c) and used #include to pull it into
3939 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3940 separate copies.
3942 (b) Defined the list of op-code names for debugging as a macro in
3943 internal.h so that it is next to the definition of the opcodes.
3945 (c) Defined a table of op-code lengths for simpler skipping along compiled
3946 code. This is again a macro in internal.h so that it is next to the
3947 definition of the opcodes.
3949 34. Added support for recursive calls to individual subpatterns, along the
3950 lines of Robin Houston's patch (but implemented somewhat differently).
3952 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3953 allow it to read and process whole directories in Win32. This code was
3954 contributed by Lionel Fourquaux; it has not been tested by me.
3956 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3957 used to name a group. Names consist of alphanumerics and underscores, and must
3958 be unique. Back references use the syntax (?P=name) and recursive calls use
3959 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3960 numbers. The function pcre_fullinfo() can be used after compilation to extract
3961 a name/number map. There are three relevant calls:
3963 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3964 PCRE_INFO_NAMECOUNT yields the number of entries
3965 PCRE_INFO_NAMETABLE yields a pointer to the map.
3967 The map is a vector of fixed-size entries. The size of each entry depends on
3968 the length of the longest name used. The first two bytes of each entry are the
3969 group number, most significant byte first. There follows the corresponding
3970 name, zero terminated. The names are in alphabetical order.
3972 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3973 case instead of 255. Making it the same both with and without UTF-8 support
3974 means that the same test output works with both.
3976 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3977 calling malloc() with a zero argument.
3979 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3980 optimization. I've improved things by keeping a bitmap of backreferences with
3981 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3982 fact referenced, the optimization can be applied. It is unlikely that a
3983 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3984 the match to follow \n) will appear inside brackets with a number greater than
3985 31, but if it does, any back reference > 31 suppresses the optimization.
3987 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3988 of disabling numbered capturing parentheses. Any opening parenthesis that is
3989 not followed by ? behaves as if it were followed by ?: but named parentheses
3990 can still be used for capturing (and they will acquire numbers in the usual
3991 way).
3993 41. Redesigned the return codes from the match() function into yes/no/error so
3994 that errors can be passed back from deep inside the nested calls. A malloc
3995 failure while inside a recursive subpattern call now causes the
3996 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3998 42. It is now possible to set a limit on the number of times the match()
3999 function is called in a call to pcre_exec(). This facility makes it possible to
4000 limit the amount of recursion and backtracking, though not in a directly
4001 obvious way, because the match() function is used in a number of different
4002 circumstances. The count starts from zero for each position in the subject
4003 string (for non-anchored patterns). The default limit is, for compatibility, a
4004 large number, namely 10 000 000. You can change this in two ways:
4006 (a) When configuring PCRE before making, you can use --with-match-limit=n
4007 to set a default value for the compiled library.
4009 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4010 a different value is set. See 45 below.
4012 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4014 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4015 of things that can be changed at compile time. The first argument specifies
4016 what is wanted and the second points to where the information is to be placed.
4017 The current list of available information is:
4021 The output is an integer that is set to one if UTF-8 support is available;
4022 otherwise it is set to zero.
4026 The output is an integer that it set to the value of the code that is used for
4027 newline. It is either LF (10) or CR (13).
4031 The output is an integer that contains the number of bytes used for internal
4032 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4036 The output is an integer that contains the threshold above which the POSIX
4037 interface uses malloc() for output vectors. See item 31 above.
4041 The output is an unsigned integer that contains the default limit of the number
4042 of match() calls in a pcre_exec() execution. See 42 above.
4044 44. pcretest has been upgraded by the addition of the -C option. This causes it
4045 to extract all the available output from the new pcre_config() function, and to
4046 output it. The program then exits immediately.
4048 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4049 order to support additional features. One way would have been to define
4050 pcre_exec2() (for example) with extra arguments, but this would not have been
4051 extensible, and would also have required all calls to the original function to
4052 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4053 is used for passing in "extra" data from pcre_study().
4055 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4056 contains the following fields:
4058 flags a bitmap indicating which of the following fields are set
4059 study_data opaque data from pcre_study()
4060 match_limit a way of specifying a limit on match() calls for a specific
4061 call to pcre_exec()
4062 callout_data data for callouts (see 49 below)
4064 The flag bits are also defined in pcre.h, and are
4070 The pcre_study() function now returns one of these new pcre_extra blocks, with
4071 the actual study data pointed to by the study_data field, and the
4072 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4073 before. That is, this change is entirely upwards-compatible and requires no
4074 change to existing code.
4076 If you want to pass in additional data to pcre_exec(), you can either place it
4077 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4078 block.
4080 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4081 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4082 times with different match limits, until it finds the minimum value needed for
4083 pcre_exec() to complete. The value is then output. This can be instructive; for
4084 most simple matches the number is quite small, but for pathological cases it
4085 gets very large very quickly.
4087 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4088 returns the size of the data block pointed to by the study_data field in a
4089 pcre_extra block, that is, the value that was passed as the argument to
4090 pcre_malloc() when PCRE was getting memory in which to place the information
4091 created by pcre_study(). The fourth argument should point to a size_t variable.
4092 pcretest has been extended so that this information is shown after a successful
4093 pcre_study() call when information about the compiled regex is being displayed.
4095 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4096 because what follows is always an absolute path. (Later: it turns out that this
4097 is more than cosmetic for MinGW, because it doesn't like empty path
4098 components.)
4100 49. Some changes have been made to the callout feature (see 28 above):
4102 (i) A callout function now has three choices for what it returns:
4104 0 => success, carry on matching
4105 > 0 => failure at this point, but backtrack if possible
4106 < 0 => serious error, return this value from pcre_exec()
4108 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4109 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4110 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4111 use by callout functions. It will never be used by PCRE itself.
4113 (ii) The pcre_extra structure (see 45 above) has a void * field called
4114 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4115 pcre_callout_block structure has a field of the same name. The contents of
4116 the field passed in the pcre_extra structure are passed to the callout
4117 function in the corresponding field in the callout block. This makes it
4118 easier to use the same callout-containing regex from multiple threads. For
4119 testing, the pcretest program has a new data escape
4121 \C*n pass the number n (may be negative) as callout_data
4123 If the callout function in pcretest receives a non-zero value as
4124 callout_data, it returns that value.
4126 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4127 there were some redundant $(CFLAGS) in commands that are now specified as
4128 $(LINK), which already includes $(CFLAGS).
4130 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4131 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4132 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4133 one-byte characters throughout. Note that case-insensitive matching applies
4134 only to characters whose values are less than 256. PCRE doesn't support the
4135 notion of cases for higher-valued characters.
4137 (i) A character class whose characters are all within 0-255 is handled as
4138 a bit map, and the map is inverted for negative classes. Previously, a
4139 character > 255 always failed to match such a class; however it should
4140 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4142 (ii) A negated character class with a single character < 255 is coded as
4143 "not this character" (OP_NOT). This wasn't working properly when the test
4144 character was multibyte, either singly or repeated.
4146 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4147 mode, for example: \x{100}{2,3}.
4149 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4150 singly or repeated) now correctly test multibyte characters. However,
4151 PCRE doesn't recognize any characters with values greater than 255 as
4152 digits, spaces, or word characters. Such characters always match \D, \S,
4153 and \W, and never match \d, \s, or \w.
4155 (v) Classes may now contain characters and character ranges with values
4156 greater than 255. For example: [ab\x{100}-\x{400}].
4158 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4159 PCRE in UTF-8 mode.
4161 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4162 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4163 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4164 value.)
4166 53. The single man page has become too large. I have therefore split it up into
4167 a number of separate man pages. These also give rise to individual HTML pages;
4168 these are now put in a separate directory, and there is an index.html page that
4169 lists them all. Some hyperlinking between the pages has been installed.
4171 54. Added convenience functions for handling named capturing parentheses.
4173 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4174 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4175 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4176 are faulted.
4178 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4179 calling configure. These values are used when compiling the dftables.c program
4180 which is run to generate the source of the default character tables. They
4181 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4182 you will need to set these values.
4184 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4187 Version 3.9 02-Jan-02
4188 ---------------------
4190 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4192 2. If --disable-static was given, the building process failed when trying to
4193 build pcretest and pcregrep. (For some reason it was using libtool to compile
4194 them, which is not right, as they aren't part of the library.)
4197 Version 3.8 18-Dec-01
4198 ---------------------
4200 1. The experimental UTF-8 code was completely screwed up. It was packing the
4201 bytes in the wrong order. How dumb can you get?
4204 Version 3.7 29-Oct-01
4205 ---------------------
4207 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4208 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4209 this didn't happen under Solaris 8, where I normally test things.
4211 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4212 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4213 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4214 this fix an hour or so after the initial 3.7 release.)
4217 Version 3.6 23-Oct-01
4218 ---------------------
4220 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4221 offsets passed as NULL with zero offset count.
4223 2. The config.guess and config.sub files had not been updated when I moved to
4224 the latest autoconf.
4227 Version 3.5 15-Aug-01
4228 ---------------------
4230 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4231 had been forgotten.
4233 2. By using declared but undefined structures, we can avoid using "void"
4234 definitions in pcre.h while keeping the internal definitions of the structures
4235 private.
4237 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4238 user point of view, this means that both static and shared libraries are built
4239 by default, but this can be individually controlled. More of the work of
4240 handling this static/shared cases is now inside libtool instead of PCRE's make
4241 file.
4243 4. The pcretest utility is now installed along with pcregrep because it is
4244 useful for users (to test regexs) and by doing this, it automatically gets
4245 relinked by libtool. The documentation has been turned into a man page, so
4246 there are now .1, .txt, and .html versions in /doc.
4248 5. Upgrades to pcregrep:
4249 (i) Added long-form option names like gnu grep.
4250 (ii) Added --help to list all options with an explanatory phrase.
4251 (iii) Added -r, --recursive to recurse into sub-directories.
4252 (iv) Added -f, --file to read patterns from a file.
4254 6. pcre_exec() was referring to its "code" argument before testing that
4255 argument for NULL (and giving an error if it was NULL).
4257 7. Upgraded Makefile.in to allow for compiling in a different directory from
4258 the source directory.
4260 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4261 options bits, the pointer it was passed was to an int instead of to an unsigned
4262 long int. This mattered only on 64-bit systems.
4264 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4265 generated) instead of pcre.in, which it its source. Also made the same change
4266 in several of the .c files.
4268 10. A new release of gcc defines printf() as a macro, which broke pcretest
4269 because it had an ifdef in the middle of a string argument for printf(). Fixed
4270 by using separate calls to printf().
4272 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4273 script, to force use of CR or LF instead of \n in the source. On non-Unix
4274 systems, the value can be set in config.h.
4276 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4277 absolute limit. Changed the text of the error message to make this clear, and
4278 likewise updated the man page.
4280 13. The limit of 99 on the number of capturing subpatterns has been removed.
4281 The new limit is 65535, which I hope will not be a "real" limit.
4284 Version 3.4 22-Aug-00
4285 ---------------------
4287 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4289 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4292 Version 3.3 01-Aug-00
4293 ---------------------
4295 1. If an octal character was given, but the value was greater than \377, it
4296 was not getting masked to the least significant bits, as documented. This could
4297 lead to crashes in some systems.
4299 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4300 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4302 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4303 These just pass their arguments on to (pcre_free)(), but they are provided
4304 because some uses of PCRE bind it to non-C systems that can call its functions,
4305 but cannot call free() or pcre_free() directly.
4307 4. Add "make test" as a synonym for "make check". Corrected some comments in
4308 the Makefile.
4310 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4311 Makefile.
4313 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4314 command called pgrep for grepping around the active processes.
4316 7. Added the beginnings of support for UTF-8 character strings.
4318 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4319 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4320 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4321 out for the ar command.)
4324 Version 3.2 12-May-00
4325 ---------------------
4327 This is purely a bug fixing release.
4329 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4330 of ZA. This was just one example of several cases that could provoke this bug,
4331 which was introduced by change 9 of version 2.00. The code for breaking
4332 infinite loops after an iteration that matches an empty string was't working
4333 correctly.
4335 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4336 when matched against abbab (for example). After matching an empty string, it
4337 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4338 caused it to match further down the string than it should.
4340 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4341 was there because it doesn't seem to be needed, and it causes trouble on some
4342 systems, as it is not a Standard C header. It has been removed.
4344 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4345 were reported on the Macintosh. The changes were from
4347 while ((c = *(++ptr)) != 0 && c != '\n');
4348 to
4349 while ((c = *(++ptr)) != 0 && c != '\n') ;
4351 Totally extraordinary, but if that's what it takes...
4353 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4354 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4355 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4356 assumes the way PCRE uses memmove() (always moving upwards).
4358 6. PCRE is being used in one environment where strchr() is not available. There
4359 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4360 faster code anyway.
4363 Version 3.1 09-Feb-00
4364 ---------------------
4366 The only change in this release is the fixing of some bugs in Makefile.in for
4367 the "install" target:
4369 (1) It was failing to install pcreposix.h.
4371 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4374 Version 3.0 01-Feb-00
4375 ---------------------
4377 1. Add support for the /+ modifier to perltest (to output $` like it does in
4378 pcretest).
4380 2. Add support for the /g modifier to perltest.
4382 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4383 matches null strings.
4385 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4386 pattern. Perl treats empty patterns specially - it reuses the most recent
4387 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4388 effect.
4390 5. The POSIX interface was broken in that it was just handing over the POSIX
4391 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4392 required a bigger vector, with some working space on the end. This means that
4393 the POSIX wrapper now has to get and free some memory, and copy the results.
4395 6. Added some simple autoconf support, placing the test data and the
4396 documentation in separate directories, re-organizing some of the
4397 information files, and making it build pcre-config (a GNU standard). Also added
4398 libtool support for building PCRE as a shared library, which is now the
4399 default.
4401 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4402 09 are not valid octal constants. Single digits will be used for minor values
4403 less than 10.
4405 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4406 existing programs that set these in the POSIX interface can use PCRE without
4407 modification.
4409 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4410 return all that pcre_info() returns, plus additional data. The pcre_info()
4411 function is retained for compatibility, but is considered to be obsolete.
4413 10. Added experimental recursion feature (?R) to handle one common case that
4414 Perl 5.6 will be able to do with (?p{...}).
4416 11. Added support for POSIX character classes like [:alpha:], which Perl is
4417 adopting.
4420 Version 2.08 31-Aug-99
4421 ----------------------
4423 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4424 trying to match at the startoffset position, but instead was moving forward to
4425 the next newline as if a previous match had failed.
4427 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4428 and could get into a loop if a null string was matched other than at the start
4429 of the subject.
4431 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4432 be distinguished at compile time, and for completeness also added PCRE_DATE.
4434 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4435 in GnuWin32 environments.
4438 Version 2.07 29-Jul-99
4439 ----------------------
4441 1. The documentation is now supplied in plain text form and HTML as well as in
4442 the form of man page sources.
4444 2. C++ compilers don't like assigning (void *) values to other pointer types.
4445 In particular this affects malloc(). Although there is no problem in Standard
4446 C, I've put in casts to keep C++ compilers happy.
4448 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4449 should be (const char *).
4451 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4452 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4453 However, I haven't made this a standard facility. The documentation doesn't
4454 mention it, and the Makefile doesn't support it.
4456 5. The Makefile now contains an "install" target, with editable destinations at
4457 the top of the file. The pcretest program is not installed.
4459 6. pgrep -V now gives the PCRE version number and date.
4461 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4462 causing the entire string to be ignored, instead of just the last character.
4464 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4465 non-matching string, it can take a very, very long time, even for strings of
4466 quite modest length, because of the nested recursion. PCRE now does better in
4467 some of these cases. It does this by remembering the last required literal
4468 character in the pattern, and pre-searching the subject to ensure it is present
4469 before running the real match. In other words, it applies a heuristic to detect
4470 some types of certain failure quickly, and in the above example, if presented
4471 with a string that has no trailing " it gives "no match" very quickly.
4473 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4474 other alternatives are tried instead.
4477 Version 2.06 09-Jun-99
4478 ----------------------
4480 1. Change pcretest's output for amount of store used to show just the code
4481 space, because the remainder (the data block) varies in size between 32-bit and
4482 64-bit systems.
4484 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4485 start matching at. This allows lookbehinds to work when searching for multiple
4486 occurrences in a string.
4488 3. Added additional options to pcretest for testing multiple occurrences:
4490 /+ outputs the rest of the string that follows a match
4491 /g loops for multiple occurrences, using the new startoffset argument
4492 /G loops for multiple occurrences by passing an incremented pointer
4494 4. PCRE wasn't doing the "first character" optimization for patterns starting
4495 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4496 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4497 the letter 'x'. On long subject strings, this gives a significant speed-up.
4500 Version 2.05 21-Apr-99
4501 ----------------------
4503 1. Changed the type of magic_number from int to long int so that it works
4504 properly on 16-bit systems.
4506 2. Fixed a bug which caused patterns starting with .* not to work correctly
4507 when the subject string contained newline characters. PCRE was assuming
4508 anchoring for such patterns in all cases, which is not correct because .* will
4509 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4510 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4511 must be retried after every newline in the subject.
4514 Version 2.04 18-Feb-99
4515 ----------------------
4517 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4518 computation of the store needed to hold the pattern was incorrect (too large).
4519 If such patterns were nested a few deep, this could multiply and become a real
4520 problem.
4522 2. Added /M option to pcretest to show the memory requirement of a specific
4523 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4525 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4526 compiled in such a way that the backtracking after subsequent failure was
4527 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4528 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4531 Version 2.03 02-Feb-99
4532 ----------------------
4534 1. Fixed typo and small mistake in man page.
4536 2. Added 4th condition (GPL supersedes if conflict) and created separate
4537 LICENCE file containing the conditions.
4539 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4540 Perl, that is the internal \ allows the delimiter to be included in the
4541 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4542 the final delimiter, add \ to the end of the pattern (to test the error).
4544 4. Added the convenience functions for extracting substrings after a successful
4545 match. Updated pcretest to make it able to test these functions.
4548 Version 2.02 14-Jan-99
4549 ----------------------
4551 1. Initialized the working variables associated with each extraction so that
4552 their saving and restoring doesn't refer to uninitialized store.
4554 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4555 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4556 fix the problem.
4558 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4559 calls, and wasn't printing the correct value for compiling calls. Increased the
4560 default value of LOOPREPEAT, and the number of significant figures in the
4561 times.
4563 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4565 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4566 a building problem on Windows NT with a FAT file system.
4569 Version 2.01 21-Oct-98
4570 ----------------------
4572 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4573 to character tables built by pcre_maketables() in the current locale. If NULL
4574 is passed, the default tables are used.
4577 Version 2.00 24-Sep-98
4578 ----------------------
4580 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4581 it any more.
4583 2. Allow quantification of (?>) groups, and make it work correctly.
4585 3. The first character computation wasn't working for (?>) groups.
4587 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4588 end of the subject) and add 5.005's \z, which really does match only at the
4589 very end of the subject.
4591 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4593 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4594 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4595 localized options. All options to pcre_study() were also removed.
4597 7. Add other new features from 5.005:
4599 $(?<= positive lookbehind
4600 $(?<! negative lookbehind
4601 (?imsx-imsx) added the unsetting capability
4602 such a setting is global if at outer level; local otherwise
4603 (?imsx-imsx:) non-capturing groups with option setting
4604 (?(cond)re|re) conditional pattern matching
4606 A backreference to itself in a repeated group matches the previous
4607 captured string.
4609 8. General tidying up of studying (both automatic and via "study")
4610 consequential on the addition of new assertions.
4612 9. As in 5.005, unlimited repeated groups that could match an empty substring
4613 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4614 runtime if any iteration does actually match an empty substring.
4616 10. Include the RunTest script in the distribution.
4618 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4619 discrepancies, some of which were old and were also with respect to 5.004. They
4620 have now been fixed.
4623 Version 1.09 28-Apr-98
4624 ----------------------
4626 1. A negated single character class followed by a quantifier with a minimum
4627 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4628 program crashes, or just wrong answers. This did not apply to negated classes
4629 containing more than one character, or to minima other than one.
4632 Version 1.08 27-Mar-98
4633 ----------------------
4635 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4637 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4638 latter must appear before anything that relies on it in the pattern.
4641 Version 1.07 16-Feb-98
4642 ----------------------
4644 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4645 repeat of a potentially empty string).
4648 Version 1.06 23-Jan-98
4649 ----------------------
4651 1. Added Markus Oberhumer's little patches for C++.
4653 2. Literal strings longer than 255 characters were broken.
4656 Version 1.05 23-Dec-97
4657 ----------------------
4659 1. Negated character classes containing more than one character were failing if
4660 PCRE_CASELESS was set at run time.
4663 Version 1.04 19-Dec-97
4664 ----------------------
4666 1. Corrected the man page, where some "const" qualifiers had been omitted.
4668 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4669 input syntax.
4671 3. Fixed memory leak which occurred when a regex with back references was
4672 matched with an offsets vector that wasn't big enough. The temporary memory
4673 that is used in this case wasn't being freed if the match failed.
4675 4. Tidied pcretest to ensure it frees memory that it gets.
4677 5. Temporary memory was being obtained in the case where the passed offsets
4678 vector was exactly big enough.
4680 6. Corrected definition of offsetof() from change 5 below.
4682 7. I had screwed up change 6 below and broken the rules for the use of
4683 setjmp(). Now fixed.
4686 Version 1.03 18-Dec-97
4687 ----------------------
4689 1. A erroneous regex with a missing opening parenthesis was correctly
4690 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4691 on some systems.
4693 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4694 it was reported that one broken compiler failed on the former because "code" is
4695 also an independent variable.
4697 3. The erroneous regex a[]b caused an array overrun reference.
4699 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4700 fail on data ending with that character. (It was going on too far, and checking
4701 the next character, typically a binary zero.) This was specific to the
4702 optimized code for single-character negative classes.
4704 5. Added a contributed patch from the TIN world which does the following:
4706 + Add an undef for memmove, in case the the system defines a macro for it.
4708 + Add a definition of offsetof(), in case there isn't one. (I don't know
4709 the reason behind this - offsetof() is part of the ANSI standard - but
4710 it does no harm).
4712 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4713 most of the places where whitespace preceded '#'. I have given up and
4714 allowed the remaining 2 cases to be at the margin.
4716 + Rename some variables in pcre to eliminate shadowing. This seems very
4717 pedantic, but does no harm, of course.
4719 6. Moved the call to setjmp() into its own function, to get rid of warnings
4720 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4722 7. Constructs such as \d{8,} were compiling into the equivalent of
4723 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4724 outcome, but in this particular case used more store than had been allocated,
4725 which caused the bug to be discovered because it threw up an internal error.
4727 8. The debugging code in both pcre and pcretest for outputting the compiled
4728 form of a regex was going wrong in the case of back references followed by
4729 curly-bracketed repeats.
4732 Version 1.02 12-Dec-97
4733 ----------------------
4735 1. Typos in pcre.3 and comments in the source fixed.
4737 2. Applied a contributed patch to get rid of places where it used to remove
4738 'const' from variables, and fixed some signed/unsigned and uninitialized
4739 variable warnings.
4741 3. Added the "runtest" target to Makefile.
4743 4. Set default compiler flag to -O2 rather than just -O.
4746 Version 1.01 19-Nov-97
4747 ----------------------
4749 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4750 like /([ab]*)*/, that is, for classes with more than one character in them.
4752 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4753 as /((?>a*))*/ (a PCRE_EXTRA facility).
4756 Version 1.00 18-Nov-97
4757 ----------------------
4759 1. Added compile-time macros to support systems such as SunOS4 which don't have
4760 memmove() or strerror() but have other things that can be used instead.
4762 2. Arranged that "make clean" removes the executables.
4765 Version 0.99 27-Oct-97
4766 ----------------------
4768 1. Fixed bug in code for optimizing classes with only one character. It was
4769 initializing a 32-byte map regardless, which could cause it to run off the end
4770 of the memory it had got.
4772 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4775 Version 0.98 22-Oct-97
4776 ----------------------
4778 1. Fixed bug in code for handling temporary memory usage when there are more
4779 back references than supplied space in the ovector. This could cause segfaults.
4782 Version 0.97 21-Oct-97
4783 ----------------------
4785 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4787 2. Optimized negated single characters not to use a bit map.
4789 3. Brought error texts together as macro definitions; clarified some of them;
4790 fixed one that was wrong - it said "range out of order" when it meant "invalid
4791 escape sequence".
4793 4. Changed some char * arguments to const char *.
4795 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4797 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4798 pcretest.
4801 Version 0.96 16-Oct-97
4802 ----------------------
4804 1. Added a simple "pgrep" utility to the distribution.
4806 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4807 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4808 where "ddd" means "one or more decimal digits".
4810 3. Fixed serious bug. If a pattern had a back reference, but the call to
4811 pcre_exec() didn't supply a large enough ovector to record the related
4812 identifying subpattern, the match always failed. PCRE now remembers the number
4813 of the largest back reference, and gets some temporary memory in which to save
4814 the offsets during matching if necessary, in order to ensure that
4815 backreferences always work.
4817 4. Increased the compatibility with Perl in a number of ways:
4819 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4820 to request this handling. The option can be set at compile or exec time.
4822 (b) $ matches before a terminating newline by default; an option
4823 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4824 mode). The option can be set at compile or exec time.
4826 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4827 the same as Perl's. If the decimal number it represents is less than 10
4828 or there aren't that many previous left capturing parentheses, an octal
4829 escape is read. Inside a character class, it's always an octal escape,
4830 even if it is a single digit.
4832 (d) An escaped but undefined alphabetic character is taken as a literal,
4833 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4834 escapes.
4836 (e) {0} is now permitted. (The previous item is removed from the compiled
4837 pattern).
4839 5. Changed all the names of code files so that the basic parts are no longer
4840 than 10 characters, and abolished the teeny "globals.c" file.
4842 6. Changed the handling of character classes; they are now done with a 32-byte
4843 bit map always.
4845 7. Added the -d and /D options to pcretest to make it possible to look at the
4846 internals of compilation without having to recompile pcre.
4849 Version 0.95 23-Sep-97
4850 ----------------------
4852 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4853 \x20 at the start of a run of normal characters. These were being treated as
4854 real characters, instead of the source characters being re-checked.
4857 Version 0.94 18-Sep-97
4858 ----------------------
4860 1. The functions are now thread-safe, with the caveat that the global variables
4861 containing pointers to malloc() and free() or alternative functions are the
4862 same for all threads.
4864 2. Get pcre_study() to generate a bitmap of initial characters for non-
4865 anchored patterns when this is possible, and use it if passed to pcre_exec().
4868 Version 0.93 15-Sep-97
4869 ----------------------
4871 1. /(b)|(:+)/ was computing an incorrect first character.
4873 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4874 but not actually doing anything yet.
4876 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4877 as Perl does (e.g. [-az] or [az-]).
4879 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4880 all possible positions.
4882 5. Split up into different modules to avoid including unneeded functions in a
4883 compiled binary. However, compile and exec are still in one module. The "study"
4884 function is split off.
4886 6. The character tables are now in a separate module whose source is generated
4887 by an auxiliary program - but can then be edited by hand if required. There are
4888 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4889 toupper() in the code.
4891 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4892 make them global. Abolish the function for setting them, as the caller can now
4893 set them directly.
4896 Version 0.92 11-Sep-97
4897 ----------------------
4899 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4900 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4902 2. Caseless matching was not working in character classes if the characters in
4903 the pattern were in upper case.
4905 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4907 4. Make PCRE_ANCHORED public and accept as a compile option.
4909 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4910 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4911 pass them.
4913 6. Give an error if bad option bits passed at compile or run time.
4915 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4916 pcretest to cause it to pass that flag.
4918 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4919 options, and the first character, if set.
4921 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4924 Version 0.91 10-Sep-97
4925 ----------------------
4927 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4928 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4930 2. PCRE was looping on encountering an indefinitely repeated back reference to
4931 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4932 Perl does - treats the match as successful.
4934 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5