ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1045 - (show annotations)
Sun Sep 23 16:50:00 2012 UTC (7 years, 5 months ago) by ph10
File size: 237465 byte(s)
Error occurred while calculating annotation data.
Update character class handling to use new character case information; rework 
\h, \H, \v, and \V to use the same apparatus with centrally defined lists.
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.
92 (a) The Unicode property table now has offsets into a new table of sets of
93 three or more characters that are case-equivalent. The MultiStage2.py
94 script that generates these tables (the pcre_ucd.c file) now scans
95 CaseFolding.txt instead of UnicodeData.txt for character case
96 information.
98 (b) The code for adding characters or ranges of characters to a character
99 class has been abstracted into a generalized function that also handles
100 case-independence. In UTF-mode with UCP support, this uses the new data
101 to handle characters with more than one other case.
103 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
104 whose other case is greater than 256 are now correctly matched
105 caselessly. Previously, the high codepoint matched the low one, but not
106 vice versa.
108 (d) The processing of \h, \H, \v, and \ in character classes now makes use
109 of the new class addition function, using character lists defined as
110 macros alongside the case definitions of 20 above.
113 Version 8.31 06-July-2012
114 -------------------------
116 1. Fixing a wrong JIT test case and some compiler warnings.
118 2. Removed a bashism from the RunTest script.
120 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
121 to unsigned type, result still unsigned" that was given by an MS compiler
122 on encountering the code "-sizeof(xxx)".
124 4. Partial matching support is added to the JIT compiler.
126 5. Fixed several bugs concerned with partial matching of items that consist
127 of more than one character:
129 (a) /^(..)\1/ did not partially match "aba" because checking references was
130 done on an "all or nothing" basis. This also applied to repeated
131 references.
133 (b) \R did not give a hard partial match if \r was found at the end of the
134 subject.
136 (c) \X did not give a hard partial match after matching one or more
137 characters at the end of the subject.
139 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
140 a partial match for the string "\r".
142 (e) When newline was set to CRLF, the metacharacter "." did not recognize
143 a partial match for a CR character at the end of the subject string.
145 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
146 running pcretest, the text "(JIT)" added to the output whenever JIT is
147 actually used to run the match.
149 7. Individual JIT compile options can be set in pcretest by following -s+[+]
150 or /S+[+] with a digit between 1 and 7.
152 8. OP_NOT now supports any UTF character not just single-byte ones.
154 9. (*MARK) control verb is now supported by the JIT compiler.
156 10. The command "./RunTest list" lists the available tests without actually
157 running any of them. (Because I keep forgetting what they all are.)
161 12. Applied a (slightly modified) user-supplied patch that improves performance
162 when the heap is used for recursion (compiled with --disable-stack-for-
163 recursion). Instead of malloc and free for each heap frame each time a
164 logical recursion happens, frames are retained on a chain and re-used where
165 possible. This sometimes gives as much as 30% improvement.
167 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
168 call.
170 14. As documented, (*COMMIT) is now confined to within a positive assertion.
172 15. It is now possible to link pcretest with libedit as an alternative to
173 libreadline.
175 16. (*COMMIT) control verb is now supported by the JIT compiler.
177 17. The Unicode data tables have been updated to Unicode 6.1.0.
179 18. Added --file-list option to pcregrep.
181 19. Added binary file support to pcregrep, including the -a, --binary-files,
182 -I, and --text options.
184 20. The madvise function is renamed for posix_madvise for QNX compatibility
185 reasons. Fixed by Giuseppe D'Angelo.
187 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
189 22. Changed the option for creating symbolic links for 16-bit man pages from
190 -s to -sf so that re-installing does not cause issues.
192 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
194 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
195 matches in certain environments (the workspace was not being correctly
196 retained). Also added to pcre_dfa_exec() a simple plausibility check on
197 some of the workspace data at the beginning of a restart.
199 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
200 was not doing so when it should - probably a typo introduced by SVN 528
201 (change 8.10/14).
203 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
204 \w+ when the character tables indicated that \x{c4} was a word character.
205 There were several related cases, all because the tests for doing a table
206 lookup were testing for characters less than 127 instead of 255.
208 27. If a pattern contains capturing parentheses that are not used in a match,
209 their slots in the ovector are set to -1. For those that are higher than
210 any matched groups, this happens at the end of processing. In the case when
211 there were back references that the ovector was too small to contain
212 (causing temporary malloc'd memory to be used during matching), and the
213 highest capturing number was not used, memory off the end of the ovector
214 was incorrectly being set to -1. (It was using the size of the temporary
215 memory instead of the true size.)
217 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
218 ovector size, it uses memory at the end of the block that it has got.
220 29. Check for an overlong MARK name and give an error at compile time. The
221 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
223 30. JIT compiler update.
225 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
226 Rill for the patch.
228 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
230 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
232 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
233 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
234 was enabled.
236 35. Improve JIT code generation for greedy plus quantifier.
238 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
239 1 to "aa" instead of to an empty string. The bug affected repeated groups
240 that could potentially match an empty string.
242 37. Optimizing single character iterators in JIT.
244 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
245 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
246 codepoints that are too big for the mode are faulted, and in a UTF mode,
247 disallowed codepoints are also faulted.
249 39. If PCRE was compiled with UTF support, in three places in the DFA
250 matcher there was code that should only have been obeyed in UTF mode, but
251 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
252 processing when bytes with values greater than 127 were present. In 16-bit
253 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
254 both cases the values are those that cannot be the first data item in a UTF
255 character. The three items that might have provoked this were recursions,
256 possessively repeated groups, and atomic groups.
258 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
259 and pcregrep, because some OS require shared objects to be explicitly
260 passed to ld, causing the link step to fail if they are not.
262 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
263 mode, patterns that started with \h* or \R* might be incorrectly matched.
266 Version 8.30 04-February-2012
267 -----------------------------
269 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
270 name is defined in ctype.h.
272 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
273 only in quite long subpatterns.
275 3. Removed the function pcre_info(), which has been obsolete and deprecated
276 since it was replaced by pcre_fullinfo() in February 2000.
278 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
279 match a (*MARK), and the match failed at the start of the subject, a
280 reference to memory before the start of the subject could occur. This bug
281 was introduced by fix 17 of release 8.21.
283 5. A reference to an unset group with zero minimum repetition was giving
284 totally wrong answers (in non-JavaScript-compatibility mode). For example,
285 /(another)?(\1?)test/ matched against "hello world test". This bug was
286 introduced in release 8.13.
288 6. Add support for 16-bit character strings (a large amount of work involving
289 many changes and refactorings).
291 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
292 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
293 from a file.
295 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
296 rounding is not applied in this particular case).
298 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
299 if they appear, or are escaped, in patterns.
301 10. Get rid of a number of -Wunused-but-set-variable warnings.
303 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
304 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
305 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
306 also returns the mark "x". This bug applied to capturing parentheses,
307 non-capturing parentheses, and atomic parentheses. It also applied to some
308 assertions.
310 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
311 information out of configure.ac instead of relying on pcre.h.generic, which
312 is not stored in the repository.
314 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
315 -lreadline.
317 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
319 15. Applied Graycode's patch to put the top-level frame on the stack rather
320 than the heap when not using the stack for recursion. This gives a
321 performance improvement in many cases when recursion is not deep.
323 16. Experimental code added to "pcretest -C" to output the stack frame size.
326 Version 8.21 12-Dec-2011
327 ------------------------
329 1. Updating the JIT compiler.
331 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
332 are added as well.
334 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
335 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
336 calling _pcre_jit_exec. Some extra comments are added.
338 4. (*MARK) settings inside atomic groups that do not contain any capturing
339 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
340 was introduced by change 18 for 8.20.
342 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
343 ECMA-262 standard.
345 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
346 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
347 This bug was probably introduced by change 9 of 8.13.
349 7. While fixing 6 above, I noticed that a number of other items were being
350 incorrectly rejected as "not fixed length". This arose partly because newer
351 opcodes had not been added to the fixed-length checking code. I have (a)
352 corrected the bug and added tests for these items, and (b) arranged for an
353 error to occur if an unknown opcode is encountered while checking for fixed
354 length instead of just assuming "not fixed length". The items that were
355 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
356 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
357 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
359 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
360 being incorrectly compiled and would have given unpredicatble results.
362 9. A possessively repeated subpattern with minimum repeat count greater than
363 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
364 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
365 the first (A) could occur when it should not.
367 10. Add a cast and remove a redundant test from the code.
369 11. JIT should use pcre_malloc/pcre_free for allocation.
371 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
372 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
373 is anything other than /usr, -L is still shown).
375 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
377 14. Perl does not support \N without a following name in a [] class; PCRE now
378 also gives an error.
380 15. If a forward reference was repeated with an upper limit of around 2000,
381 it caused the error "internal error: overran compiling workspace". The
382 maximum number of forward references (including repeats) was limited by the
383 internal workspace, and dependent on the LINK_SIZE. The code has been
384 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
385 the default depends on LINK_SIZE. There is a new upper limit (for safety)
386 of around 200,000 forward references. While doing this, I also speeded up
387 the filling in of repeated forward references.
389 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
390 incorrectly expecting the subject to contain another "a" after the start.
392 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
393 in the match, the SKIP should be ignored. This was not happening; instead
394 the SKIP was being treated as NOMATCH. For patterns such as
395 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
396 tested.
398 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
399 now much more compatible with Perl, in particular in cases where the result
400 is a non-match for a non-anchored pattern. For example, if
401 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
402 "m", where previously it did not return a name. A side effect of this
403 change is that for partial matches, the last encountered mark name is
404 returned, as for non matches. A number of tests that were previously not
405 Perl-compatible have been moved into the Perl-compatible test files. The
406 refactoring has had the pleasing side effect of removing one argument from
407 the match() function, thus reducing its stack requirements.
409 19. If the /S+ option was used in pcretest to study a pattern using JIT,
410 subsequent uses of /S (without +) incorrectly behaved like /S+.
412 21. Retrieve executable code size support for the JIT compiler and fixing
413 some warnings.
415 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
416 not work when the shorter character appeared right at the end of the
417 subject string.
419 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
420 systems.
422 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
423 output it when the /M option is used in pcretest.
425 25. The CheckMan script was not being included in the distribution. Also, added
426 an explicit "perl" to run Perl scripts from the PrepareRelease script
427 because this is reportedly needed in Windows.
429 26. If study data was being save in a file and studying had not found a set of
430 "starts with" bytes for the pattern, the data written to the file (though
431 never used) was taken from uninitialized memory and so caused valgrind to
432 complain.
434 27. Updated RunTest.bat as provided by Sheri Pierce.
436 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
438 29. Computation of memory usage for the table of capturing group names was
439 giving an unnecessarily large value.
442 Version 8.20 21-Oct-2011
443 ------------------------
445 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
446 a POSIX class. After further experiments with Perl, which convinced me that
447 Perl has bugs and confusions, a closing square bracket is no longer allowed
448 in a POSIX name. This bug also affected patterns with classes that started
449 with full stops.
451 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
452 captured substring, but while checking the failing first alternative,
453 substring 1 is temporarily captured. If the output vector supplied to
454 pcre_exec() was not big enough for this capture, the yield of the function
455 was still zero ("insufficient space for captured substrings"). This cannot
456 be totally fixed without adding another stack variable, which seems a lot
457 of expense for a edge case. However, I have improved the situation in cases
458 such as /(a)(b)x|abc/ matched against "abc", where the return code
459 indicates that fewer than the maximum number of slots in the ovector have
460 been set.
462 3. Related to (2) above: when there are more back references in a pattern than
463 slots in the output vector, pcre_exec() uses temporary memory during
464 matching, and copies in the captures as far as possible afterwards. It was
465 using the entire output vector, but this conflicts with the specification
466 that only 2/3 is used for passing back captured substrings. Now it uses
467 only the first 2/3, for compatibility. This is, of course, another edge
468 case.
470 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
471 main code base, and can be used by building with --enable-jit. When this is
472 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
473 runtime --no-jit option is given.
475 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
476 ovector, the return from the function was zero, implying that there were
477 other matches that did not fit. The correct "exactly full" value is now
478 returned.
480 6. If a subpattern that was called recursively or as a subroutine contained
481 (*PRUNE) or any other control that caused it to give a non-standard return,
482 invalid errors such as "Error -26 (nested recursion at the same subject
483 position)" or even infinite loops could occur.
485 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
486 computing the minimum length on reaching *ACCEPT, and so ended up with the
487 wrong value of 1 rather than 0. Further investigation indicates that
488 computing a minimum subject length in the presence of *ACCEPT is difficult
489 (think back references, subroutine calls), and so I have changed the code
490 so that no minimum is registered for a pattern that contains *ACCEPT.
492 8. If (*THEN) was present in the first (true) branch of a conditional group,
493 it was not handled as intended. [But see 16 below.]
495 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
496 Sheri Pierce.
498 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
499 the first byte in a match must be "a".
501 11. Change 17 for 8.13 increased the recursion depth for patterns like
502 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
503 pattern contains any instances of (*THEN). If it does not, the old
504 optimizations are restored. It would be nice to do this on a per-group
505 basis, but at the moment that is not feasible.
507 12. In some environments, the output of pcretest -C is CRLF terminated. This
508 broke RunTest's code that checks for the link size. A single white space
509 character after the value is now allowed for.
511 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
512 For "fr", it uses the Windows-specific input and output files.
514 14. If (*THEN) appeared in a group that was called recursively or as a
515 subroutine, it did not work as intended. [But see next item.]
517 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
518 pattern fragments (but not containing any | characters). If A and B are
519 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
520 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
521 D. In other words, Perl considers parentheses that do not contain any |
522 characters to be part of a surrounding alternative, whereas PCRE was
523 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
524 differently. PCRE now behaves in the same way as Perl, except in the case
525 of subroutine/recursion calls such as (?1) which have in any case always
526 been different (but PCRE had them first :-).
528 16. Related to 15 above: Perl does not treat the | in a conditional group as
529 creating alternatives. Such a group is treated in the same way as an
530 ordinary group without any | characters when processing (*THEN). PCRE has
531 been changed to match Perl's behaviour.
533 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
534 RunGrepTest script failed.
536 18. Change 22 for version 13 caused atomic groups to use more stack. This is
537 inevitable for groups that contain captures, but it can lead to a lot of
538 stack use in large patterns. The old behaviour has been restored for atomic
539 groups that do not contain any capturing parentheses.
541 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
542 suppress the check for a minimum subject length at run time. (If it was
543 given to pcre_exec() or pcre_dfa_exec() it did work.)
545 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
546 fail to work when decoding hex characters in data strings in EBCDIC
547 environments.
549 21. It appears that in at least one Mac OS environment, the isxdigit() function
550 is implemented as a macro that evaluates to its argument more than once,
551 contravening the C 90 Standard (I haven't checked a later standard). There
552 was an instance in pcretest which caused it to go wrong when processing
553 \x{...} escapes in subject strings. The has been rewritten to avoid using
554 things like p++ in the argument of isxdigit().
557 Version 8.13 16-Aug-2011
558 ------------------------
560 1. The Unicode data tables have been updated to Unicode 6.0.0.
562 2. Two minor typos in pcre_internal.h have been fixed.
564 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
565 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
566 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
568 4. There were a number of related bugs in the code for matching backrefences
569 caselessly in UTF-8 mode when codes for the characters concerned were
570 different numbers of bytes. For example, U+023A and U+2C65 are an upper
571 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
572 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
573 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
574 2-byte code at the end of the subject (it thought there wasn't enough data
575 left).
577 5. Comprehensive information about what went wrong is now returned by
578 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
579 as the output vector has at least 2 elements. The offset of the start of
580 the failing character and a reason code are placed in the vector.
582 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
583 now returned is for the first byte of the failing character, instead of the
584 last byte inspected. This is an incompatible change, but I hope it is small
585 enough not to be a problem. It makes the returned offset consistent with
586 pcre_exec() and pcre_dfa_exec().
588 7. pcretest now gives a text phrase as well as the error number when
589 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
590 failure, the offset and reason code are output.
592 8. When \R was used with a maximizing quantifier it failed to skip backwards
593 over a \r\n pair if the subsequent match failed. Instead, it just skipped
594 back over a single character (\n). This seems wrong (because it treated the
595 two characters as a single entity when going forwards), conflicts with the
596 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
597 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
598 has been changed.
600 9. Some internal refactoring has changed the processing so that the handling
601 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
602 time (the PCRE_DOTALL option was changed this way some time ago: version
603 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
604 which was always a bit of a fudge. It also means that there is one less
605 argument for the match() function, which reduces its stack requirements
606 slightly. This change also fixes an incompatibility with Perl: the pattern
607 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
609 10. More internal refactoring has drastically reduced the number of recursive
610 calls to match() for possessively repeated groups such as (abc)++ when
611 using pcre_exec().
613 11. While implementing 10, a number of bugs in the handling of groups were
614 discovered and fixed:
616 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
617 (a|)*(?1) gave a compile-time internal error.
618 ((a|)+)+ did not notice that the outer group could match an empty string.
619 (^a|^)+ was not marked as anchored.
620 (.*a|.*)+ was not marked as matching at start or after a newline.
622 12. Yet more internal refactoring has removed another argument from the match()
623 function. Special calls to this function are now indicated by setting a
624 value in a variable in the "match data" data block.
626 13. Be more explicit in pcre_study() instead of relying on "default" for
627 opcodes that mean there is no starting character; this means that when new
628 ones are added and accidentally left out of pcre_study(), testing should
629 pick them up.
631 14. The -s option of pcretest has been documented for ages as being an old
632 synonym of -m (show memory usage). I have changed it to mean "force study
633 for every regex", that is, assume /S for every regex. This is similar to -i
634 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
635 using it. It makes it easier to run collections of tests with and without
636 study enabled, and thereby test pcre_study() more easily. All the standard
637 tests are now run with and without -s (but some patterns can be marked as
638 "never study" - see 20 below).
640 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
641 restoration of the capturing data to the outer values was not happening
642 correctly.
644 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
645 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
646 pattern had matched an empty string, and so incorrectly returned a no
647 match.
649 17. There was optimizing code for the last branch of non-capturing parentheses,
650 and also for the obeyed branch of a conditional subexpression, which used
651 tail recursion to cut down on stack usage. Unfortunately, now that there is
652 the possibility of (*THEN) occurring in these branches, tail recursion is
653 no longer possible because the return has to be checked for (*THEN). These
654 two optimizations have therefore been removed. [But see 8.20/11 above.]
656 18. If a pattern containing \R was studied, it was assumed that \R always
657 matched two bytes, thus causing the minimum subject length to be
658 incorrectly computed because \R can also match just one byte.
660 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
661 was incorrectly computed.
663 20. If /S is present twice on a test pattern in pcretest input, it now
664 *disables* studying, thereby overriding the use of -s on the command line
665 (see 14 above). This is necessary for one or two tests to keep the output
666 identical in both cases.
668 21. When (*ACCEPT) was used in an assertion that matched an empty string and
669 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
671 22. When an atomic group that contained a capturing parenthesis was
672 successfully matched, but the branch in which it appeared failed, the
673 capturing was not being forgotten if a higher numbered group was later
674 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
675 group 1 to "a", when in fact it should be unset. This applied to multi-
676 branched capturing and non-capturing groups, repeated or not, and also to
677 positive assertions (capturing in negative assertions does not happen
678 in PCRE) and also to nested atomic groups.
680 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
681 subject after a captured substring, to make it easier to tell which of a
682 number of identical substrings has been captured.
684 24. The way atomic groups are processed by pcre_exec() has been changed so that
685 if they are repeated, backtracking one repetition now resets captured
686 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
687 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
688 "aaa". Previously, it would have been "a". As part of this code
689 refactoring, the way recursive calls are handled has also been changed.
691 25. If an assertion condition captured any substrings, they were not passed
692 back unless some other capturing happened later. For example, if
693 (?(?=(a))a) was matched against "a", no capturing was returned.
695 26. When studying a pattern that contained subroutine calls or assertions,
696 the code for finding the minimum length of a possible match was handling
697 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
698 group 1 called group 2 while simultaneously a separate group 2 called group
699 1). A stack overflow occurred in this case. I have fixed this by limiting
700 the recursion depth to 10.
702 27. Updated RunTest.bat in the distribution to the version supplied by Tom
703 Fortmann. This supports explicit test numbers on the command line, and has
704 argument validation and error reporting.
706 28. An instance of \X with an unlimited repeat could fail if at any point the
707 first character it looked at was a mark character.
709 29. Some minor code refactoring concerning Unicode properties and scripts
710 should reduce the stack requirement of match() slightly.
712 30. Added the '=' option to pcretest to check the setting of unused capturing
713 slots at the end of the pattern, which are documented as being -1, but are
714 not included in the return count.
716 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
717 compiled something random. Now it gives a compile-time error (as does
718 Perl).
720 32. A *MARK encountered during the processing of a positive assertion is now
721 recorded and passed back (compatible with Perl).
723 33. If --only-matching or --colour was set on a pcregrep call whose pattern
724 had alternative anchored branches, the search for a second match in a line
725 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
726 matched the line "0102" twice. The same bug affected patterns that started
727 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
728 twice.
730 34. Previously, PCRE did not allow quantification of assertions. However, Perl
731 does, and because of capturing effects, quantifying parenthesized
732 assertions may at times be useful. Quantifiers are now allowed for
733 parenthesized assertions.
735 35. A minor code tidy in pcre_compile() when checking options for \R usage.
737 36. \g was being checked for fancy things in a character class, when it should
738 just be a literal "g".
740 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
741 appearance of a nested POSIX class supersedes an apparent external class.
742 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
743 unescaped square brackets may also appear as part of class names. For
744 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
745 more like Perl. (But see 8.20/1 above.)
747 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
748 was because it thought it was \N{name}, which is not supported).
750 39. Add minix to OS list not supporting the -S option in pcretest.
752 40. PCRE tries to detect cases of infinite recursion at compile time, but it
753 cannot analyze patterns in sufficient detail to catch mutual recursions
754 such as ((?1))((?2)). There is now a runtime test that gives an error if a
755 subgroup is called recursively as a subpattern for a second time at the
756 same position in the subject string. In previous releases this might have
757 been caught by the recursion limit, or it might have run out of stack.
759 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
760 happen only once. PCRE was, however incorrectly giving a compile time error
761 "recursive call could loop indefinitely" because it cannot analyze the
762 pattern in sufficient detail. The compile time test no longer happens when
763 PCRE is compiling a conditional subpattern, but actual runaway loops are
764 now caught at runtime (see 40 above).
766 42. It seems that Perl allows any characters other than a closing parenthesis
767 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
768 has been changed to be the same.
770 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
771 as not to get warnings when autogen.sh is called. Also changed
772 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
774 44. To help people who use pcregrep to scan files containing exceedingly long
775 lines, the following changes have been made:
777 (a) The default value of the buffer size parameter has been increased from
778 8K to 20K. (The actual buffer used is three times this size.)
780 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
781 PCRE is built.
783 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
784 to be set at run time.
786 (d) Numerical values in pcregrep options can be followed by K or M, for
787 example --buffer-size=50K.
789 (e) If a line being scanned overflows pcregrep's buffer, an error is now
790 given and the return code is set to 2.
792 45. Add a pointer to the latest mark to the callout data block.
794 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
795 partial match of an empty string instead of no match. This was specific to
796 the use of ".".
798 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
799 complete match instead of a partial match. This bug was dependent on both
800 the PCRE_UTF8 and PCRE_DOTALL options being set.
802 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
803 starting byte set, because \b was not being ignored.
806 Version 8.12 15-Jan-2011
807 ------------------------
809 1. Fixed some typos in the markup of the man pages, and wrote a script that
810 checks for such things as part of the documentation building process.
812 2. On a big-endian 64-bit system, pcregrep did not correctly process the
813 --match-limit and --recursion-limit options (added for 8.11). In
814 particular, this made one of the standard tests fail. (The integer value
815 went into the wrong half of a long int.)
817 3. If the --colour option was given to pcregrep with -v (invert match), it
818 did strange things, either producing crazy output, or crashing. It should,
819 of course, ignore a request for colour when reporting lines that do not
820 match.
822 4. Another pcregrep bug caused similar problems if --colour was specified with
823 -M (multiline) and the pattern match finished with a line ending.
825 5. In pcregrep, when a pattern that ended with a literal newline sequence was
826 matched in multiline mode, the following line was shown as part of the
827 match. This seems wrong, so I have changed it.
829 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
830 the check for further matches in the same line (so they could be coloured)
831 to overrun the end of the current line. If another match was found, it was
832 incorrectly shown (and then shown again when found in the next line).
834 7. If pcregrep was compiled under Windows, there was a reference to the
835 function pcregrep_exit() before it was defined. I am assuming this was
836 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
837 reported by a user. I've moved the definition above the reference.
840 Version 8.11 10-Dec-2010
841 ------------------------
843 1. (*THEN) was not working properly if there were untried alternatives prior
844 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
845 backtracked to try for "b" instead of moving to the next alternative branch
846 at the same level (in this case, to look for "c"). The Perl documentation
847 is clear that when (*THEN) is backtracked onto, it goes to the "next
848 alternative in the innermost enclosing group".
850 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
851 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
852 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
853 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
854 (*THEN).
856 3. If \s appeared in a character class, it removed the VT character from
857 the class, even if it had been included by some previous item, for example
858 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
859 of \s, but is part of the POSIX "space" class.)
861 4. A partial match never returns an empty string (because you can always
862 match an empty string at the end of the subject); however the checking for
863 an empty string was starting at the "start of match" point. This has been
864 changed to the "earliest inspected character" point, because the returned
865 data for a partial match starts at this character. This means that, for
866 example, /(?<=abc)def/ gives a partial match for the subject "abc"
867 (previously it gave "no match").
869 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
870 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
871 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
872 has an implication that the given string is incomplete (because a partial
873 match is preferred over a full match). For this reason, these items now
874 give a partial match in this situation. [Aside: previously, the one case
875 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
876 match rather than a full match, which was wrong by the old rules, but is
877 now correct.]
879 6. There was a bug in the handling of #-introduced comments, recognized when
880 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
881 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
882 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
883 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
884 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
885 places in pcre_compile().
887 7. Related to (6) above, when pcre_compile() was skipping #-introduced
888 comments when looking ahead for named forward references to subpatterns,
889 the only newline sequence it recognized was NL. It now handles newlines
890 according to the set newline convention.
892 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
893 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
894 cater for a lack of strerror(). These oversights have been fixed.
896 9. Added --match-limit and --recursion-limit to pcregrep.
898 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
900 11. When the -o option was used, pcregrep was setting a return code of 1, even
901 when matches were found, and --line-buffered was not being honoured.
903 12. Added an optional parentheses number to the -o and --only-matching options
904 of pcregrep.
906 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
907 can match an empty string. The code to do it in pcretest and pcredemo
908 needed fixing:
910 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
911 only one byte after an empty string match just before CRLF (this case
912 just got forgotten; "any" and "anycrlf" were OK).
914 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
915 mode when an empty string match preceded an ASCII character followed by
916 a non-ASCII character. (The code for advancing by one character rather
917 than one byte was nonsense.)
919 (c) The pcredemo.c sample program did not have any code at all to handle
920 the cases when CRLF is a valid newline sequence.
922 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
923 as a starting offset was within the subject string. There is now a new
924 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
925 negative or greater than the length of the string. In order to test this,
926 pcretest is extended to allow the setting of negative starting offsets.
928 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
929 starting offset points to the beginning of a UTF-8 character was
930 unnecessarily clumsy. I tidied it up.
932 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
933 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
935 17. Nobody had reported that the --include_dir option, which was added in
936 release 7.7 should have been called --include-dir (hyphen, not underscore)
937 for compatibility with GNU grep. I have changed it to --include-dir, but
938 left --include_dir as an undocumented synonym, and the same for
939 --exclude-dir, though that is not available in GNU grep, at least as of
940 release 2.5.4.
942 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
943 characters from a string of bytes) have been redefined so as not to use
944 loops, in order to improve performance in some environments. At the same
945 time, I abstracted some of the common code into auxiliary macros to save
946 repetition (this should not affect the compiled code).
948 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
949 compile-time error is now given if \c is not followed by an ASCII
950 character, that is, a byte less than 128. (In EBCDIC mode, the code is
951 different, and any byte value is allowed.)
953 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
954 START_OPTIMIZE option, which is now allowed at compile time - but just
955 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
956 to pcregrep and other applications that have no direct access to PCRE
957 options. The new /Y option in pcretest sets this option when calling
958 pcre_compile().
960 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
961 back references. Groups containing recursive back references were forced to
962 be atomic by that change, but in the case of named groups, the amount of
963 memory required was incorrectly computed, leading to "Failed: internal
964 error: code overflow". This has been fixed.
966 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
967 pcretest.c, to avoid build problems in some Borland environments.
970 Version 8.10 25-Jun-2010
971 ------------------------
973 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
974 THEN.
976 2. (*ACCEPT) was not working when inside an atomic group.
978 3. Inside a character class, \B is treated as a literal by default, but
979 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
980 causes the error). The code is unchanged, but I tidied the documentation.
982 4. Inside a character class, PCRE always treated \R and \X as literals,
983 whereas Perl faults them if its -w option is set. I have changed PCRE so
984 that it faults them when PCRE_EXTRA is set.
986 5. Added support for \N, which always matches any character other than
987 newline. (It is the same as "." when PCRE_DOTALL is not set.)
989 6. When compiling pcregrep with newer versions of gcc which may have
990 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
991 declared with attribute warn_unused_result" were given. Just casting the
992 result to (void) does not stop the warnings; a more elaborate fudge is
993 needed. I've used a macro to implement this.
995 7. Minor change to pcretest.c to avoid a compiler warning.
997 8. Added four artifical Unicode properties to help with an option to make
998 \s etc use properties (see next item). The new properties are: Xan
999 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1001 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1002 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1003 this option. Modified pcretest to add /W to test this facility. Added
1004 REG_UCP to make it available via the POSIX interface.
1006 10. Added --line-buffered to pcregrep.
1008 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1009 studied, and the match started with a letter with a code point greater than
1010 127 whose first byte was different to the first byte of the other case of
1011 the letter, the other case of this starting letter was not recognized
1012 (#976).
1014 12. If a pattern that was studied started with a repeated Unicode property
1015 test, for example, \p{Nd}+, there was the theoretical possibility of
1016 setting up an incorrect bitmap of starting bytes, but fortunately it could
1017 not have actually happened in practice until change 8 above was made (it
1018 added property types that matched character-matching opcodes).
1020 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1021 possible starting bytes for non-anchored patterns.
1023 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1024 \R, and also a number of cases that involve Unicode properties, both
1025 explicit and implicit when PCRE_UCP is set.
1027 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1028 input, it could crash or give wrong results if characters with values
1029 greater than 0xc0 were present in the subject string. (Detail: it assumed
1030 UTF-8 input when processing these items.)
1032 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1033 size_t is 64-bit (#991).
1035 17. Added a check for running out of memory when PCRE is compiled with
1036 --disable-stack-for-recursion (#990).
1038 18. If the last data line in a file for pcretest does not have a newline on
1039 the end, a newline was missing in the output.
1041 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1042 less than 128) in its various bitmaps. However, there is a facility for
1043 generating tables according to the current locale when PCRE is compiled. It
1044 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1045 characters, are recognized by isspace() and therefore were getting set in
1046 these tables, and indeed these tables seem to approximate to ISO 8859. This
1047 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1048 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1049 which of course cannot start UTF-8 characters. I have changed the code so
1050 that only real ASCII characters (less than 128) and the correct starting
1051 bytes for UTF-8 encodings are set for characters greater than 127 when in
1052 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1053 altogether.)
1055 20. Added the /T option to pcretest so as to be able to run tests with non-
1056 standard character tables, thus making it possible to include the tests
1057 used for 19 above in the standard set of tests.
1059 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1060 reference to a subpattern the other side of a comment that contains an
1061 opening parenthesis caused either an internal compiling error, or a
1062 reference to the wrong subpattern.
1065 Version 8.02 19-Mar-2010
1066 ------------------------
1068 1. The Unicode data tables have been updated to Unicode 5.2.0.
1070 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1071 configured.
1073 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1074 original author of that file, following a query about its status.
1076 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1077 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1079 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1080 quantifier applied to a forward-referencing subroutine call, could compile
1081 incorrect code or give the error "internal error: previously-checked
1082 referenced subpattern not found".
1084 6. Both MS Visual Studio and Symbian OS have problems with initializing
1085 variables to point to external functions. For these systems, therefore,
1086 pcre_malloc etc. are now initialized to local functions that call the
1087 relevant global functions.
1089 7. There were two entries missing in the vectors called coptable and poptable
1090 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1091 I've fixed the data, and added a kludgy way of testing at compile time that
1092 the lengths are correct (equal to the number of opcodes).
1094 8. Following on from 7, I added a similar kludge to check the length of the
1095 eint vector in pcreposix.c.
1097 9. Error texts for pcre_compile() are held as one long string to avoid too
1098 much relocation at load time. To find a text, the string is searched,
1099 counting zeros. There was no check for running off the end of the string,
1100 which could happen if a new error number was added without updating the
1101 string.
1103 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1105 11. \K was not working if it appeared in an atomic group or in a group that
1106 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1107 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1108 the assertion is positive, but not if it is negative.
1110 12. Change 11 fortuitously reduced the size of the stack frame used in the
1111 "match()" function of pcre_exec.c by one pointer. Forthcoming
1112 implementation of support for (*MARK) will need an extra pointer on the
1113 stack; I have reserved it now, so that the stack frame size does not
1114 decrease.
1116 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1117 item in branch that calls a recursion is a subroutine call - as in the
1118 second branch in the above example - was incorrectly given the compile-
1119 time error "recursive call could loop indefinitely" because pcre_compile()
1120 was not correctly checking the subroutine for matching a non-empty string.
1122 14. The checks for overrunning compiling workspace could trigger after an
1123 overrun had occurred. This is a "should never occur" error, but it can be
1124 triggered by pathological patterns such as hundreds of nested parentheses.
1125 The checks now trigger 100 bytes before the end of the workspace.
1127 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1130 Version 8.01 19-Jan-2010
1131 ------------------------
1133 1. If a pattern contained a conditional subpattern with only one branch (in
1134 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1135 computed the wrong minimum data length (which is of course zero for such
1136 subpatterns). This could cause incorrect "no match" results.
1138 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1139 the pattern is reset in the first branch, pcre_compile() failed with
1140 "internal error: code overflow at offset...". This happened only when
1141 the reset was to the original external option setting. (An optimization
1142 abstracts leading options settings into an external setting, which was the
1143 cause of this.)
1145 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1146 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1147 assertion pattern did not match (meaning that the assertion was true), it
1148 was incorrectly treated as false if the SKIP had been reached during the
1149 matching. This also applied to assertions used as conditions.
1151 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1152 assertion subpattern, including such a pattern used as a condition,
1153 unpredictable results occurred, instead of the error return
1156 5. The C++ GlobalReplace function was not working like Perl for the special
1157 situation when an empty string is matched. It now does the fancy magic
1158 stuff that is necessary.
1160 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1161 removed. (These were left over from very, very early versions of PCRE.)
1163 7. Some cosmetic changes to the code to make life easier when compiling it
1164 as part of something else:
1166 (a) Change DEBUG to PCRE_DEBUG.
1168 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1169 called "current" as "current_branch", to prevent a collision with the
1170 Linux macro when compiled as a kernel module.
1172 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1173 prevent a collision with the Linux macro when compiled as a kernel
1174 module.
1176 8. In pcre_compile() there are some checks for integer overflows that used to
1177 cast potentially large values to (double). This has been changed to that
1178 when building, a check for int64_t is made, and if it is found, it is used
1179 instead, thus avoiding the use of floating point arithmetic. (There is no
1180 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1181 double.
1183 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1184 2005 (difference between two addresses compared to an unsigned value).
1186 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1187 custom one, because of the following reported problem in Windows:
1189 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1190 under Win32.
1191 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1192 therefore missing the function definition.
1193 - The compiler thus generates a "C" signature for the test function.
1194 - The linker fails to find the "C" function.
1195 - PCRE fails to configure if asked to do so against libbz2.
1197 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1198 messages were output:
1200 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1201 rerunning libtoolize, to keep the correct libtool macros in-tree.
1202 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1204 I have done both of these things.
1206 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1207 most of the time, it *can* run out if it is given a pattern that contains a
1208 runaway infinite recursion. I updated the discussion in the pcrestack man
1209 page.
1211 13. Now that we have gone to the x.xx style of version numbers, the minor
1212 version may start with zero. Using 08 or 09 is a bad idea because users
1213 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1214 interpreted as invalid octal numbers. I've updated the previous comment in
1215 configure.ac, and also added a check that gives an error if 08 or 09 are
1216 used.
1218 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1219 causing partial matching to fail when the end of the subject matched \W
1220 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1222 15. There were some discrepancies between the declarations in pcre_internal.h
1223 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1224 their definitions. The declarations used "const uschar *" and the
1225 definitions used USPTR. Even though USPTR is normally defined as "const
1226 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1227 reported that: "This difference in casting confuses some C++ compilers, for
1228 example, SunCC recognizes above declarations as different functions and
1229 generates broken code for hbpcre." I have changed the declarations to use
1230 USPTR.
1232 16. GNU libtool is named differently on some systems. The autogen.sh script now
1233 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1234 (FreeBSD).
1236 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1237 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1238 comment: "Figure out how to create a longlong from a string: strtoll and
1239 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1240 instance, but it only takes 2 args instead of 3!"
1242 18. A subtle bug concerned with back references has been fixed by a change of
1243 specification, with a corresponding code fix. A pattern such as
1244 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1245 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1246 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1247 same bug. Such groups have to be quantified to be useful, or contained
1248 inside another quantified group. (If there's no repetition, the reference
1249 can never match.) The problem arises because, having left the group and
1250 moved on to the rest of the pattern, a later failure that backtracks into
1251 the group uses the captured value from the final iteration of the group
1252 rather than the correct earlier one. I have fixed this in PCRE by forcing
1253 any group that contains a reference to itself to be an atomic group; that
1254 is, there cannot be any backtracking into it once it has completed. This is
1255 similar to recursive and subroutine calls.
1258 Version 8.00 19-Oct-09
1259 ----------------------
1261 1. The table for translating pcre_compile() error codes into POSIX error codes
1262 was out-of-date, and there was no check on the pcre_compile() error code
1263 being within the table. This could lead to an OK return being given in
1264 error.
1266 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1267 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1268 in a Windows environment.
1270 3. The pcregrep --count option prints the count for each file even when it is
1271 zero, as does GNU grep. However, pcregrep was also printing all files when
1272 --files-with-matches was added. Now, when both options are given, it prints
1273 counts only for those files that have at least one match. (GNU grep just
1274 prints the file name in this circumstance, but including the count seems
1275 more useful - otherwise, why use --count?) Also ensured that the
1276 combination -clh just lists non-zero counts, with no names.
1278 4. The long form of the pcregrep -F option was incorrectly implemented as
1279 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1280 but it seems right to fix it, and I didn't think it was worth preserving
1281 the old behaviour.
1283 5. The command line items --regex=pattern and --regexp=pattern were not
1284 recognized by pcregrep, which required --regex pattern or --regexp pattern
1285 (with a space rather than an '='). The man page documented the '=' forms,
1286 which are compatible with GNU grep; these now work.
1288 6. No libpcreposix.pc file was created for pkg-config; there was just
1289 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1291 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1292 when UCP support is not needed, by modifying the Python script that
1293 generates it from Unicode data files. This should not matter if the module
1294 is correctly used as a library, but I received one complaint about 50K of
1295 unwanted data. My guess is that the person linked everything into his
1296 program rather than using a library. Anyway, it does no harm.
1298 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1299 was a minimum greater than 1 for a wide character in a possessive
1300 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1301 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1302 character. Chaos in the form of incorrect output or a compiling loop could
1303 result.
1305 9. The restrictions on what a pattern can contain when partial matching is
1306 requested for pcre_exec() have been removed. All patterns can now be
1307 partially matched by this function. In addition, if there are at least two
1308 slots in the offset vector, the offset of the earliest inspected character
1309 for the match and the offset of the end of the subject are set in them when
1310 PCRE_ERROR_PARTIAL is returned.
1312 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1313 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1314 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1315 and may be more useful for multi-segment matching.
1317 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1318 used to be given if ever the end of the subject was reached; now it is
1319 given only if matching could not proceed because another character was
1320 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1321 string "Z", which now yields "no match" instead of "partial match". In the
1322 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1323 final character ended with (*FAIL).
1325 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1326 if the pattern had a "must contain" character that was already found in the
1327 earlier partial match, unless partial matching was again requested. For
1328 example, with the pattern /dog.(body)?/, the "must contain" character is
1329 "g". If the first part-match was for the string "dog", restarting with
1330 "sbody" failed. This bug has been fixed.
1332 13. The string returned by pcre_dfa_exec() after a partial match has been
1333 changed so that it starts at the first inspected character rather than the
1334 first character of the match. This makes a difference only if the pattern
1335 starts with a lookbehind assertion or \b or \B (\K is not supported by
1336 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1337 matching functions compatible, and I think it's the right thing to do.
1339 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1340 so that the demonstration program is easily available in environments where
1341 PCRE has not been installed from source.
1343 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1344 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1345 library.
1347 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1348 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1349 is not the first non-POSIX option to be added. Clearly some people find
1350 these options useful.
1352 17. If a caller to the POSIX matching function regexec() passes a non-zero
1353 value for nmatch with a NULL value for pmatch, the value of
1354 nmatch is forced to zero.
1356 18. RunGrepTest did not have a test for the availability of the -u option of
1357 the diff command, as RunTest does. It now checks in the same way as
1358 RunTest, and also checks for the -b option.
1360 19. If an odd number of negated classes containing just a single character
1361 interposed, within parentheses, between a forward reference to a named
1362 subpattern and the definition of the subpattern, compilation crashed with
1363 an internal error, complaining that it could not find the referenced
1364 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1365 [The bug was that it was starting one character too far in when skipping
1366 over the character class, thus treating the ] as data rather than
1367 terminating the class. This meant it could skip too much.]
1369 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1370 /g option in pcretest when the pattern contains \K, which makes it possible
1371 to have an empty string match not at the start, even when the pattern is
1372 anchored. Updated pcretest and pcredemo to use this option.
1374 21. If the maximum number of capturing subpatterns in a recursion was greater
1375 than the maximum at the outer level, the higher number was returned, but
1376 with unset values at the outer level. The correct (outer level) value is
1377 now given.
1379 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1380 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1381 make it do so. The string so far is captured, making this feature
1382 compatible with Perl.
1384 23. The tests have been re-organized, adding tests 11 and 12, to make it
1385 possible to check the Perl 5.10 features against Perl 5.10.
1387 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1388 pattern matches a fixed length string. PCRE did not allow this; now it
1389 does. Neither allows recursion.
1391 25. I finally figured out how to implement a request to provide the minimum
1392 length of subject string that was needed in order to match a given pattern.
1393 (It was back references and recursion that I had previously got hung up
1394 on.) This code has now been added to pcre_study(); it finds a lower bound
1395 to the length of subject needed. It is not necessarily the greatest lower
1396 bound, but using it to avoid searching strings that are too short does give
1397 some useful speed-ups. The value is available to calling programs via
1398 pcre_fullinfo().
1400 26. While implementing 25, I discovered to my embarrassment that pcretest had
1401 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1402 study optimizations had never been tested with that matching function.
1403 Oops. What is worse, even when it was passed study data, there was a bug in
1404 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1405 were also very few tests of studied patterns with pcre_dfa_exec().
1407 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1408 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1409 on the other side of the coin, they are no longer allowed to have different
1410 names, because these cannot be distinguished in PCRE, and this has caused
1411 confusion. (This is a difference from Perl.)
1413 28. When duplicate subpattern names are present (necessarily with different
1414 numbers, as required by 27 above), and a test is made by name in a
1415 conditional pattern, either for a subpattern having been matched, or for
1416 recursion in such a pattern, all the associated numbered subpatterns are
1417 tested, and the overall condition is true if the condition is true for any
1418 one of them. This is the way Perl works, and is also more like the way
1419 testing by number works.
1422 Version 7.9 11-Apr-09
1423 ---------------------
1425 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1426 (pcretest), all targets were linked against these libraries. This included
1427 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1428 libraries. This caused unwanted dependencies to be created. This problem
1429 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1430 pcretest is linked with readline.
1432 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1433 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1434 moved outside it again, because FALSE and TRUE are already defined in AIX,
1435 but BOOL is not.
1437 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1438 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1440 4. The pcregrep documentation said spaces were inserted as well as colons (or
1441 hyphens) following file names and line numbers when outputting matching
1442 lines. This is not true; no spaces are inserted. I have also clarified the
1443 wording for the --colour (or --color) option.
1445 5. In pcregrep, when --colour was used with -o, the list of matching strings
1446 was not coloured; this is different to GNU grep, so I have changed it to be
1447 the same.
1449 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1450 each matching line was coloured. Now it goes on to look for further matches
1451 of any of the test patterns, which is the same behaviour as GNU grep.
1453 7. A pattern that could match an empty string could cause pcregrep to loop; it
1454 doesn't make sense to accept an empty string match in pcregrep, so I have
1455 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1456 seems to be how GNU grep behaves.
1458 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1459 start or after a newline", because the conditional assertion was not being
1460 correctly handled. The rule now is that both the assertion and what follows
1461 in the first alternative must satisfy the test.
1463 9. If auto-callout was enabled in a pattern with a conditional group whose
1464 condition was an assertion, PCRE could crash during matching, both with
1465 pcre_exec() and pcre_dfa_exec().
1467 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1468 used for matching.
1470 11. Unicode property support in character classes was not working for
1471 characters (bytes) greater than 127 when not in UTF-8 mode.
1473 12. Added the -M command line option to pcretest.
1475 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1477 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1479 16. Added comments and documentation about mis-use of no_arg in the C++
1480 wrapper.
1482 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1483 from Martin Jerabek that uses macro names for all relevant character and
1484 string constants.
1486 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1487 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1488 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1489 these, but not everybody uses configure.
1491 19. A conditional group that had only one branch was not being correctly
1492 recognized as an item that could match an empty string. This meant that an
1493 enclosing group might also not be so recognized, causing infinite looping
1494 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1495 with the subject "ab", where knowledge that the repeated group can match
1496 nothing is needed in order to break the loop.
1498 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1499 exec(), but without supplying a callout function, matching went wrong.
1501 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1502 leak if the size of the offset vector was greater than 30. When the vector
1503 is smaller, the saved offsets during recursion go onto a local stack
1504 vector, but for larger vectors malloc() is used. It was failing to free
1505 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1506 error, in fact).
1508 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1509 heapframe that is used only when UTF-8 support is enabled. This caused no
1510 problem, but was untidy.
1512 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1513 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1514 included within another project.
1516 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1517 slightly modified by me:
1519 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1520 not building pcregrep.
1522 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1523 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1525 25. Forward references, both numeric and by name, in patterns that made use of
1526 duplicate group numbers, could behave incorrectly or give incorrect errors,
1527 because when scanning forward to find the reference group, PCRE was not
1528 taking into account the duplicate group numbers. A pattern such as
1529 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1531 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1532 the feature of a custom pointer more persuasive (as requested by a user).
1534 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1535 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1536 pre-defined.
1538 28. Added support for (*UTF8) at the start of a pattern.
1540 29. Arrange for flags added by the "release type" setting in CMake to be shown
1541 in the configuration summary.
1544 Version 7.8 05-Sep-08
1545 ---------------------
1547 1. Replaced UCP searching code with optimized version as implemented for Ad
1548 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1549 stage table and inline lookup instead of a function, giving speed ups of 2
1550 to 5 times on some simple patterns that I tested. Permission was given to
1551 distribute the MultiStage2.py script that generates the tables (it's not in
1552 the tarball, but is in the Subversion repository).
1554 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1555 scripts.
1557 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1558 a group with a zero qualifier. The result of the study could be incorrect,
1559 or the function might crash, depending on the pattern.
1561 4. Caseless matching was not working for non-ASCII characters in back
1562 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1563 It now works when Unicode Property Support is available.
1565 5. In pcretest, an escape such as \x{de} in the data was always generating
1566 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1567 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1568 truncation.
1570 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1572 7. Added two (int) casts to pcregrep when printing the difference of two
1573 pointers, in case they are 64-bit values.
1575 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1576 test 2 if it fails.
1578 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1579 and a #define of that name to empty if it is not externally set. This is to
1580 allow users of MSVC to set it if necessary.
1582 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1583 the convenience functions in the pcre_get.c source file.
1585 11. An option change at the start of a pattern that had top-level alternatives
1586 could cause overwriting and/or a crash. This command provoked a crash in
1587 some environments:
1589 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1591 This potential security problem was recorded as CVE-2008-2371.
1593 12. For a pattern where the match had to start at the beginning or immediately
1594 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1595 pcre_dfa_exec() could read past the end of the passed subject if there was
1596 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1597 pcretest so that it places the subject at the end of its malloc-ed buffer.
1599 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1600 exec() might read past the end of the data buffer in UTF-8 mode.
1602 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1603 the data contained the byte 0x85 as part of a UTF-8 character within its
1604 first line. This applied both to normal and DFA matching.
1606 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1607 /^[^d]*?$/8 failed to match "abc".
1609 16. Added a missing copyright notice to pcrecpp_internal.h.
1611 17. Make it more clear in the documentation that values returned from
1612 pcre_exec() in ovector are byte offsets, not character counts.
1614 18. Tidied a few places to stop certain compilers from issuing warnings.
1616 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1617 supplied by Stefan Weber. I made a further small update for 7.8 because
1618 there is a change of source arrangements: the pcre_searchfuncs.c module is
1619 replaced by pcre_ucd.c.
1622 Version 7.7 07-May-08
1623 ---------------------
1625 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1626 a string to a long long, pretend we don't even have a long long." This is
1627 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1629 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1630 pre-7.6 versions, which defined a global no_arg variable instead of putting
1631 it in the RE class. (See also #8 below.)
1633 3. Remove a line of dead code, identified by coverity and reported by Nuno
1634 Lopes.
1636 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1638 (1) The include/exclude patterns were being applied to the whole pathnames
1639 of files, instead of just to the final components.
1641 (2) If there was more than one level of directory, the subdirectories were
1642 skipped unless they satisfied the include/exclude conditions. This is
1643 inconsistent with GNU grep (and could even be seen as contrary to the
1644 pcregrep specification - which I improved to make it absolutely clear).
1645 The action now is always to scan all levels of directory, and just
1646 apply the include/exclude patterns to regular files.
1648 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1649 --exclude_dir in the tests to avoid scanning .svn directories.
1651 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1652 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1653 doesn't support NULs in patterns.
1655 7. Added some missing "const"s to declarations of static tables in
1656 pcre_compile.c and pcre_dfa_exec.c.
1658 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1659 caused by fix #2 above. (Subsequently also a second patch to fix the
1660 first patch. And a third patch - this was a messy problem.)
1662 9. Applied Craig's patch to remove the use of push_back().
1664 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1665 matching function regexec().
1667 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1668 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1669 references. PCRE supports relative numbers with this syntax (I don't think
1670 Oniguruma does).
1672 12. Previously, a group with a zero repeat such as (...){0} was completely
1673 omitted from the compiled regex. However, this means that if the group
1674 was called as a subroutine from elsewhere in the pattern, things went wrong
1675 (an internal error was given). Such groups are now left in the compiled
1676 pattern, with a new opcode that causes them to be skipped at execution
1677 time.
1679 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1680 to the way PCRE behaves:
1682 (a) A lone ] character is dis-allowed (Perl treats it as data).
1684 (b) A back reference to an unmatched subpattern matches an empty string
1685 (Perl fails the current match path).
1687 (c) A data ] in a character class must be notated as \] because if the
1688 first data character in a class is ], it defines an empty class. (In
1689 Perl it is not possible to have an empty class.) The empty class []
1690 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1691 The negative empty class [^] matches any one character, independently
1692 of the DOTALL setting.
1694 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1695 non-existent subpattern following a character class starting with ']' and
1696 containing () gave an internal compiling error instead of "reference to
1697 non-existent subpattern". Fortunately, when the pattern did exist, the
1698 compiled code was correct. (When scanning forwards to check for the
1699 existencd of the subpattern, it was treating the data ']' as terminating
1700 the class, so got the count wrong. When actually compiling, the reference
1701 was subsequently set up correctly.)
1703 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1704 it was being rejected as not supported by pcre_dfa_exec(), even though
1705 other assertions are supported. I have made pcre_dfa_exec() support
1706 (*FAIL).
1708 16. The implementation of 13c above involved the invention of a new opcode,
1709 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1710 cannot be changed at match time, I realized I could make a small
1711 improvement to matching performance by compiling OP_ALLANY instead of
1712 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1713 on the OP_ANY path.
1715 17. Compiling pcretest on Windows with readline support failed without the
1716 following two fixes: (1) Make the unistd.h include conditional on
1717 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1719 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1720 ncurses library to be included for pcretest when ReadLine support is
1721 requested, but also to allow for it to be overridden. This patch came from
1722 Daniel Bergström.
1724 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1725 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1726 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1727 spotting this.
1730 Version 7.6 28-Jan-08
1731 ---------------------
1733 1. A character class containing a very large number of characters with
1734 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1735 overflow.
1737 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1738 HAVE_LONG_LONG is not defined.
1740 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1741 bring them up to date and include new features. This patch includes:
1743 - Fixed PH's badly added libz and libbz2 support.
1744 - Fixed a problem with static linking.
1745 - Added pcredemo. [But later removed - see 7 below.]
1746 - Fixed dftables problem and added an option.
1747 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1749 - Added readline support for pcretest.
1750 - Added an listing of the option settings after cmake has run.
1752 4. A user submitted a patch to Makefile that makes it easy to create
1753 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1754 Makefile.am that cause it to include this special target, without
1755 affecting anything else. Note that the same mingw target plus all
1756 the other distribution libraries and programs are now supported
1757 when configuring with CMake (see 6 below) instead of with
1758 Configure/Make.
1760 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1761 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1762 exported in the Windows port". It has not yet been confirmed that the patch
1763 solves the problem, but it does no harm.
1765 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1766 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1767 with CMake, and also correct the comment about stack recursion.
1769 7. Remove the automatic building of pcredemo from the ./configure system and
1770 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1771 of a program that users should build themselves after PCRE is installed, so
1772 building it automatically is not really right. What is more, it gave
1773 trouble in some build environments.
1775 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1778 Version 7.5 10-Jan-08
1779 ---------------------
1781 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1782 values in parens when parsing an RE using the C++ wrapper."
1784 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1785 Characters greater than 255 were excluded from the class instead of being
1786 included.
1788 3. The same bug as (2) above applied to negated POSIX classes such as
1789 [:^space:].
1791 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1792 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1793 I have changed it.
1795 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1796 first named subpattern) and a construct such as (?&a) would reference the
1797 first named subpattern whose name started with "a" (in other words, the
1798 length check was missing). Both these problems are fixed. "Subpattern name
1799 expected" is now given for (?&) (a zero-length name), and this patch also
1800 makes it give the same error for \k'' (previously it complained that that
1801 was a reference to a non-existent subpattern).
1803 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1804 this is right because (?- can be followed by option settings as well as by
1805 digits. I have, however, made the messages clearer.
1807 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1808 than the number used in the conditional) now cause a compile-time error.
1809 This is actually not compatible with Perl, which accepts such patterns, but
1810 treats the conditional as always being FALSE (as PCRE used to), but it
1811 seems to me that giving a diagnostic is better.
1813 8. Change "alphameric" to the more common word "alphanumeric" in comments
1814 and messages.
1816 9. Fix two occurrences of "backslash" in comments that should have been
1817 "backspace".
1819 10. Remove two redundant lines of code that can never be obeyed (their function
1820 was moved elsewhere).
1822 11. The program that makes PCRE's Unicode character property table had a bug
1823 which caused it to generate incorrect table entries for sequences of
1824 characters that have the same character type, but are in different scripts.
1825 It amalgamated them into a single range, with the script of the first of
1826 them. In other words, some characters were in the wrong script. There were
1827 thirteen such cases, affecting characters in the following ranges:
1829 U+002b0 - U+002c1
1830 U+0060c - U+0060d
1831 U+0061e - U+00612
1832 U+0064b - U+0065e
1833 U+0074d - U+0076d
1834 U+01800 - U+01805
1835 U+01d00 - U+01d77
1836 U+01d9b - U+01dbf
1837 U+0200b - U+0200f
1838 U+030fc - U+030fe
1839 U+03260 - U+0327f
1840 U+0fb46 - U+0fbb1
1841 U+10450 - U+1049d
1843 12. The -o option (show only the matching part of a line) for pcregrep was not
1844 compatible with GNU grep in that, if there was more than one match in a
1845 line, it showed only the first of them. It now behaves in the same way as
1846 GNU grep.
1848 13. If the -o and -v options were combined for pcregrep, it printed a blank
1849 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1850 does the same. The return code can be used to tell if there were any
1851 non-matching lines.
1853 14. Added --file-offsets and --line-offsets to pcregrep.
1855 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1856 infinitely looping recursion. The bug was that positive lookaheads were not
1857 being skipped when checking for a possible empty match (negative lookaheads
1858 and both kinds of lookbehind were skipped).
1860 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1861 inclusion of <windows.h> to before rather than after the definition of
1862 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1864 17. Specifying a possessive quantifier with a specific limit for a Unicode
1865 character property caused pcre_compile() to compile bad code, which led at
1866 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1867 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1868 caused the error; without that there was no problem.
1870 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1872 19. Added --enable-pcretest-libreadline.
1874 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1875 RE::GlobalReplace(). As a result, the number of replacements returned was
1876 double what it should be. I removed one of the increments, but Craig sent a
1877 later patch that removed the other one (the right fix) and added unit tests
1878 that check the return values (which was not done before).
1880 21. Several CMake things:
1882 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1883 the names libpcre and libpcreposix, not just pcre and pcreposix.
1885 (2) The above change means that pcretest and pcregrep are now correctly
1886 linked with the newly-built libraries, not previously installed ones.
1890 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1891 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1892 UTF-8 newline character). The key issue is that the pattern starts .*;
1893 this means that the match must be either at the beginning, or after a
1894 newline. The bug was in the code for advancing after a failed match and
1895 checking that the new position followed a newline. It was not taking
1896 account of UTF-8 characters correctly.
1898 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1899 character classes. PCRE was not treating the sequence [:...:] as a
1900 character class unless the ... were all letters. Perl, however, seems to
1901 allow any characters between [: and :], though of course it rejects as
1902 unknown any "names" that contain non-letters, because all the known class
1903 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1904 for example, whereas PCRE did not - it did not recognize a POSIX character
1905 class. This seemed a bit dangerous, so the code has been changed to be
1906 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1907 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1908 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1909 Perl does, and where it didn't before.
1911 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1912 Windows environments %n is disabled by default.
1915 Version 7.4 21-Sep-07
1916 ---------------------
1918 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1919 means that a class such as [\s] counted as "explicit reference to CR or
1920 LF". That isn't really right - the whole point of the change was to try to
1921 help when there was an actual mention of one of the two characters. So now
1922 the change happens only if \r or \n (or a literal CR or LF) character is
1923 encountered.
1925 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1926 of both had grown to the point where there were only 3 bits left.
1927 Fortunately, there was spare space in the data structure, and so I have
1928 moved the internal flags into a new 16-bit field to free up more option
1929 bits.
1931 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1932 but did not set the internal JCHANGED flag - either of these is enough to
1933 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1934 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1935 start sets both bits.
1937 4. Added options (at build time, compile time, exec time) to change \R from
1938 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1940 5. doc/pcresyntax.html was missing from the distribution.
1942 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1943 compatibility, even though it is no longer used.
1945 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1946 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1947 windows.h file is present (where different names are used). [This was
1948 reversed later after testing - see 16 below.]
1950 8. Changed all #include <config.h> to #include "config.h". There were also
1951 some further <pcre.h> cases that I changed to "pcre.h".
1953 9. When pcregrep was used with the --colour option, it missed the line ending
1954 sequence off the lines that it output.
1956 10. It was pointed out to me that arrays of string pointers cause lots of
1957 relocations when a shared library is dynamically loaded. A technique of
1958 using a single long string with a table of offsets can drastically reduce
1959 these. I have refactored PCRE in four places to do this. The result is
1960 dramatic:
1962 Originally: 290
1963 After changing UCP table: 187
1964 After changing error message table: 43
1965 After changing table of "verbs" 36
1966 After changing table of Posix names 22
1968 Thanks to the folks working on Gregex for glib for this insight.
1970 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1971 unicode-properties was also set.
1973 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1975 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1976 checked only for CRLF.
1978 14. Added casts to pcretest.c to avoid compiler warnings.
1980 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1982 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1983 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1984 entirely. This removes changes made in 7 above.
1986 17. The CMake files have been updated, and there is now more information about
1987 building with CMake in the NON-UNIX-USE document.
1990 Version 7.3 28-Aug-07
1991 ---------------------
1993 1. In the rejigging of the build system that eventually resulted in 7.1, the
1994 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1995 brackets there is not right, since it causes compilers to look for an
1996 installed pcre.h, not the version that is in the source that is being
1997 compiled (which of course may be different). I have changed it back to:
1999 #include "pcre.h"
2001 I have a vague recollection that the change was concerned with compiling in
2002 different directories, but in the new build system, that is taken care of
2003 by the VPATH setting the Makefile.
2005 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2006 when the subject happened to end in the byte 0x85 (e.g. if the last
2007 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2008 characters but of course it shouldn't be taken as a newline when it is part
2009 of another character. The bug was that, for an unlimited repeat of . in
2010 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2011 characters when looking for a newline.
2013 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2015 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2016 in debug output.
2018 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2019 long printing in the pcrecpp unittest when running under MinGW.
2021 6. ESC_K was left out of the EBCDIC table.
2023 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2024 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2025 limit also applies to "virtual nesting" when a pattern is recursive, and in
2026 this case 1000 isn't so big. I have been able to remove this limit at the
2027 expense of backing off one optimization in certain circumstances. Normally,
2028 when pcre_exec() would call its internal match() function recursively and
2029 immediately return the result unconditionally, it uses a "tail recursion"
2030 feature to save stack. However, when a subpattern that can match an empty
2031 string has an unlimited repetition quantifier, it no longer makes this
2032 optimization. That gives it a stack frame in which to save the data for
2033 checking that an empty string has been matched. Previously this was taken
2034 from the 1000-entry workspace that had been reserved. So now there is no
2035 explicit limit, but more stack is used.
2037 8. Applied Daniel's patches to solve problems with the import/export magic
2038 syntax that is required for Windows, and which was going wrong for the
2039 pcreposix and pcrecpp parts of the library. These were overlooked when this
2040 problem was solved for the main library.
2042 9. There were some crude static tests to avoid integer overflow when computing
2043 the size of patterns that contain repeated groups with explicit upper
2044 limits. As the maximum quantifier is 65535, the maximum group length was
2045 set at 30,000 so that the product of these two numbers did not overflow a
2046 32-bit integer. However, it turns out that people want to use groups that
2047 are longer than 30,000 bytes (though not repeat them that many times).
2048 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2049 made it possible to implement the integer overflow checks in a much more
2050 dynamic way, which I have now done. The artificial limitation on group
2051 length has been removed - we now have only the limit on the total length of
2052 the compiled pattern, which depends on the LINK_SIZE setting.
2054 10. Fixed a bug in the documentation for get/copy named substring when
2055 duplicate names are permitted. If none of the named substrings are set, the
2056 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2057 empty string.
2059 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2060 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2061 because the ] is interpreted as the first data character and the
2062 terminating ] is not found. PCRE has been made compatible with Perl in this
2063 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2064 cause memory overwriting.
2066 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2067 string has been matched (to stop an infinite loop). It was not recognizing
2068 a conditional subpattern that could match an empty string if that
2069 subpattern was within another subpattern. For example, it looped when
2070 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2071 condition was not nested. This bug has been fixed.
2073 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2074 past the start of the subject in the presence of bytes with the top bit
2075 set, for example "\x8aBCD".
2077 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2078 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2080 14. Optimized (?!) to (*FAIL).
2082 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2083 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2084 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2085 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2086 does: it's just the validity check that is more restrictive.
2088 16. Inserted checks for integer overflows during escape sequence (backslash)
2089 processing, and also fixed erroneous offset values for syntax errors during
2090 backslash processing.
2092 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2093 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2095 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2096 caused an overrun.
2098 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2099 something other than just ASCII characters) inside a group that had an
2100 unlimited repeat caused a loop at compile time (while checking to see
2101 whether the group could match an empty string).
2103 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2104 [\P{Any}] did so. (Error in the code for printing property names.)
2106 21. An orphan \E inside a character class could cause a crash.
2108 22. A repeated capturing bracket such as (A)? could cause a wild memory
2109 reference during compilation.
2111 23. There are several functions in pcre_compile() that scan along a compiled
2112 expression for various reasons (e.g. to see if it's fixed length for look
2113 behind). There were bugs in these functions when a repeated \p or \P was
2114 present in the pattern. These operators have additional parameters compared
2115 with \d, etc, and these were not being taken into account when moving along
2116 the compiled data. Specifically:
2118 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2119 length.
2121 (b) An item such as \pL+ within a repeated group could cause crashes or
2122 loops.
2124 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2125 "reference to non-existent subpattern" error.
2127 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2129 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2130 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2132 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2134 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2135 character were causing crashes (broken optimization).
2137 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2138 \p or \P) caused a compile-time loop.
2140 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2141 break. For example, the unstudied pattern [\r\n]A does not match the string
2142 "\r\nA" because change 7.0/46 below moves the current point on by two
2143 characters after failing to match at the start. However, the pattern \nA
2144 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2145 the same is true. There doesn't seem any very clean way out of this, but
2146 what I have chosen to do makes the common cases work: PCRE now takes note
2147 of whether there can be an explicit match for \r or \n anywhere in the
2148 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2149 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2150 pattern has explicit CR or LF references.
2152 29. Added (*CR) etc for changing newline setting at start of pattern.
2155 Version 7.2 19-Jun-07
2156 ---------------------
2158 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2159 which is apparently normally available under Windows.
2161 2. Re-jig the pcregrep tests with different newline settings in an attempt
2162 to make them independent of the local environment's newline setting.
2164 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2166 4. Some of the "internals" tests were previously cut out when the link size
2167 was not 2, because the output contained actual offsets. The recent new
2168 "Z" feature of pcretest means that these can be cut out, making the tests
2169 usable with all link sizes.
2171 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2172 stack recursion. This gives a massive performance boost under BSD, but just
2173 a small improvement under Linux. However, it saves one field in the frame
2174 in all cases.
2176 6. Added more features from the forthcoming Perl 5.10:
2178 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2179 recursion call. It refers to the nth most recently opened parentheses.
2181 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2182 to be opened parentheses.
2184 (c) Conditions that refer to capturing parentheses can be specified
2185 relatively, for example, (?(-2)... or (?(+3)...
2187 (d) \K resets the start of the current match so that everything before
2188 is not part of it.
2190 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2192 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2193 reference syntax.
2195 (g) (?| introduces a group in which the numbering of parentheses in each
2196 alternative starts with the same number.
2198 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2200 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2203 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2204 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2205 for detecting groups that can match an empty string.
2207 9. A pattern with a very large number of alternatives (more than several
2208 hundred) was running out of internal workspace during the pre-compile
2209 phase, where pcre_compile() figures out how much memory will be needed. A
2210 bit of new cunning has reduced the workspace needed for groups with
2211 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2212 workspace instead of running out of the 4096 that are available.
2214 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2216 11. Applied patch from Google to remove an optimization that didn't quite work.
2217 The report of the bug said:
2219 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2220 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2221 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2223 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2224 it matched the wrong number of bytes.
2227 Version 7.1 24-Apr-07
2228 ---------------------
2230 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2231 that is more "standard", making use of automake and other Autotools. There
2232 is some re-arrangement of the files and adjustment of comments consequent
2233 on this.
2235 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2236 for recursive directory scanning broke on some systems because the files
2237 are not scanned in any specific order and on different systems the order
2238 was different. A call to "sort" has been inserted into RunGrepTest for the
2239 approprate test as a short-term fix. In the longer term there may be an
2240 alternative.
2242 3. I had an email from Eric Raymond about problems translating some of PCRE's
2243 man pages to HTML (despite the fact that I distribute HTML pages, some
2244 people do their own conversions for various reasons). The problems
2245 concerned the use of low-level troff macros .br and .in. I have therefore
2246 removed all such uses from the man pages (some were redundant, some could
2247 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2248 HTML has been updated to handle .nf/.fi and to complain if it encounters
2249 .br or .in.
2251 4. Updated comments in configure.ac that get placed in config.h.in and also
2252 arranged for config.h to be included in the distribution, with the name
2253 config.h.generic, for the benefit of those who have to compile without
2254 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2256 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2257 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2258 makevp.bat for the current PCRE, using the additional files
2259 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2261 6. A Windows user reported a minor discrepancy with test 2, which turned out
2262 to be caused by a trailing space on an input line that had got lost in his
2263 copy. The trailing space was an accident, so I've just removed it.
2265 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2266 that is needed.
2268 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2269 as "const" (a) because they are and (b) because it helps the PHP
2270 maintainers who have recently made a script to detect big data structures
2271 in the php code that should be moved to the .rodata section. I remembered
2272 to update Builducptable as well, so it won't revert if ucptable.h is ever
2273 re-created.
2275 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2276 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2277 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2278 support is not required. This saves 1.5-2K of code, which is important in
2279 some applications.
2281 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2282 so as not to refer to the tables, even though these functions will never be
2283 called when UTF-8 support is disabled. Otherwise there are problems with a
2284 shared library.
2286 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2288 (a) It was defining its arguments as char * instead of void *.
2290 (b) It was assuming that all moves were upwards in memory; this was true
2291 a long time ago when I wrote it, but is no longer the case.
2293 The emulated memove() is provided for those environments that have neither
2294 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2295 is clearly not the case, as these two bugs were recently reported.
2297 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2298 and Detrail to create the HTML documentation, the .txt form of the man
2299 pages, and it removes trailing spaces from listed files. It also creates
2300 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2301 case, it wraps all the #defines with #ifndefs. This script should be run
2302 before "make dist".
2304 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2305 with Unicode property support.
2307 (a) For a maximizing quantifier, if the two different cases of the
2308 character were of different lengths in their UTF-8 codings (there are
2309 some cases like this - I found 11), and the matching function had to
2310 back up over a mixture of the two cases, it incorrectly assumed they
2311 were both the same length.
2313 (b) When PCRE was configured to use the heap rather than the stack for
2314 recursion during matching, it was not correctly preserving the data for
2315 the other case of a UTF-8 character when checking ahead for a match
2316 while processing a minimizing repeat. If the check also involved
2317 matching a wide character, but failed, corruption could cause an
2318 erroneous result when trying to check for a repeat of the original
2319 character.
2321 13. Some tidying changes to the testing mechanism:
2323 (a) The RunTest script now detects the internal link size and whether there
2324 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2325 values substituted by "configure". (The RunGrepTest script already did
2326 this for UTF-8.) The configure.ac script no longer substitutes the
2327 relevant variables.
2329 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2330 with length and offset values. This means that the output is different
2331 for different internal link sizes. Test 2 is skipped for link sizes
2332 other than 2 because of this, bypassing the problem. Unfortunately,
2333 there was also a test in test 3 (the locale tests) that used /B and
2334 failed for link sizes other than 2. Rather than cut the whole test out,
2335 I have added a new /Z option to pcretest that replaces the length and
2336 offset values with spaces. This is now used to make test 3 independent
2337 of link size. (Test 2 will be tidied up later.)
2339 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2340 segmentation fault instead of returning the appropriate error message.
2342 15. In multiline mode when the newline sequence was set to "any", the pattern
2343 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2344 This doesn't seem right; it now treats the CRLF combination as the line
2345 ending, and so does not match in that case. It's only a pattern such as ^$
2346 that would hit this one: something like ^ABC$ would have failed after \r
2347 and then tried again after \r\n.
2349 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2350 in an attempt to make files that differ only in their line terminators
2351 compare equal. This works on Linux.
2353 17. Under certain error circumstances pcregrep might try to free random memory
2354 as it exited. This is now fixed, thanks to valgrind.
2356 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2357 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2358 was because its rules for how to advance for /g after matching an empty
2359 string at the end of a line did not allow for this case. They now check for
2360 it specially.
2362 20. pcretest is supposed to handle patterns and data of any length, by
2363 extending its buffers when necessary. It was getting this wrong when the
2364 buffer for a data line had to be extended.
2366 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2367 CRLF as a newline sequence.
2369 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2370 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2371 I have nevertheless tidied it up.
2373 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2375 24. Added a man page for pcre-config.
2378 Version 7.0 19-Dec-06
2379 ---------------------
2381 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2382 moving to gcc 4.1.1.
2384 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2385 sys/time.h, which is documented as needed for this function. It doesn't
2386 seem to matter on Linux, but it showed up on some releases of OS X.
2388 3. It seems that there are systems where bytes whose values are greater than
2389 127 match isprint() in the "C" locale. The "C" locale should be the
2390 default when a C program starts up. In most systems, only ASCII printing
2391 characters match isprint(). This difference caused the output from pcretest
2392 to vary, making some of the tests fail. I have changed pcretest so that:
2394 (a) When it is outputting text in the compiled version of a pattern, bytes
2395 other than 32-126 are always shown as hex escapes.
2397 (b) When it is outputting text that is a matched part of a subject string,
2398 it does the same, unless a different locale has been set for the match
2399 (using the /L modifier). In this case, it uses isprint() to decide.
2401 4. Fixed a major bug that caused incorrect computation of the amount of memory
2402 required for a compiled pattern when options that changed within the
2403 pattern affected the logic of the preliminary scan that determines the
2404 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2405 that the computed length was too small. The symptoms of this bug were
2406 either the PCRE error "internal error: code overflow" from pcre_compile(),
2407 or a glibc crash with a message such as "pcretest: free(): invalid next
2408 size (fast)". Examples of patterns that provoked this bug (shown in
2409 pcretest format) are:
2411 /(?-x: )/x
2412 /(?x)(?-x: \s*#\s*)/
2413 /((?i)[\x{c0}])/8
2414 /(?i:[\x{c0}])/8
2416 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2417 is now done differently.
2419 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2420 wrapper classes; (b) implement a new function in the C++ scanner that is
2421 more efficient than the old way of doing things because it avoids levels of
2422 recursion in the regex matching; (c) add a paragraph to the documentation
2423 for the FullMatch() function.
2425 6. The escape sequence \n was being treated as whatever was defined as
2426 "newline". Not only was this contrary to the documentation, which states
2427 that \n is character 10 (hex 0A), but it also went horribly wrong when
2428 "newline" was defined as CRLF. This has been fixed.
2430 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2431 was being set to -1 for the "end of line" case (supposedly a value that no
2432 character can have). Though this value is never used (the check for end of
2433 line is "zero bytes in current character"), it caused compiler complaints.
2434 I've changed it to 0xffffffff.
2436 8. In pcre_version.c, the version string was being built by a sequence of
2437 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2438 string (as it is for production releases) called a macro with an empty
2439 argument. The C standard says the result of this is undefined. The gcc
2440 compiler treats it as an empty string (which was what was wanted) but it is
2441 reported that Visual C gives an error. The source has been hacked around to
2442 avoid this problem.
2444 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2445 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2446 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2447 of them did).
2449 10. Originally, pcretest opened its input and output without "b"; then I was
2450 told that "b" was needed in some environments, so it was added for release
2451 5.0 to both the input and output. (It makes no difference on Unix-like
2452 systems.) Later I was told that it is wrong for the input on Windows. I've
2453 now abstracted the modes into two macros, to make it easier to fiddle with
2454 them, and removed "b" from the input mode under Windows.
2456 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2458 12. Added -help and --help to pcretest as an official way of being reminded
2459 of the options.
2461 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2462 and pcrecpp.cc because they annoy compilers at high warning levels.
2464 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2466 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2467 scripts are not C programs :-) and which was not noticed because it works
2468 on Linux.
2470 16. pcretest is supposed to handle any length of pattern and data line (as one
2471 line or as a continued sequence of lines) by extending its input buffer if
2472 necessary. This feature was broken for very long pattern lines, leading to
2473 a string of junk being passed to pcre_compile() if the pattern was longer
2474 than about 50K.
2476 17. I have done a major re-factoring of the way pcre_compile() computes the
2477 amount of memory needed for a compiled pattern. Previously, there was code
2478 that made a preliminary scan of the pattern in order to do this. That was
2479 OK when PCRE was new, but as the facilities have expanded, it has become
2480 harder and harder to keep it in step with the real compile phase, and there
2481 have been a number of bugs (see for example, 4 above). I have now found a
2482 cunning way of running the real compile function in a "fake" mode that
2483 enables it to compute how much memory it would need, while actually only
2484 ever using a few hundred bytes of working memory and without too many
2485 tests of the mode. This should make future maintenance and development
2486 easier. A side effect of this work is that the limit of 200 on the nesting
2487 depth of parentheses has been removed (though this was never a serious
2488 limitation, I suspect). However, there is a downside: pcre_compile() now
2489 runs more slowly than before (30% or more, depending on the pattern). I
2490 hope this isn't a big issue. There is no effect on runtime performance.
2492 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2493 newline (only possible for the last line of a file) and it was a
2494 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2496 19. Added additional timing features to pcretest. (1) The -tm option now times
2497 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2498 separate command line item, by a number that specifies the number of
2499 repeats to use when timing. The default is 50000; this gives better
2500 precision, but takes uncomfortably long for very large patterns.
2502 20. Extended pcre_study() to be more clever in cases where a branch of a
2503 subpattern has no definite first character. For example, (a*|b*)[cd] would
2504 previously give no result from pcre_study(). Now it recognizes that the
2505 first character must be a, b, c, or d.
2507 21. There was an incorrect error "recursive call could loop indefinitely" if
2508 a subpattern (or the entire pattern) that was being tested for matching an
2509 empty string contained only one non-empty item after a nested subpattern.
2510 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2511 incorrectly, because the \d was being skipped in the check.
2513 22. The pcretest program now has a new pattern option /B and a command line
2514 option -b, which is equivalent to adding /B to every pattern. This causes
2515 it to show the compiled bytecode, without the additional information that
2516 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2517 is the same as /B/I).
2519 23. A new optimization is now able automatically to treat some sequences such
2520 as a*b as a*+b. More specifically, if something simple (such as a character
2521 or a simple class like \d) has an unlimited quantifier, and is followed by
2522 something that cannot possibly match the quantified thing, the quantifier
2523 is automatically "possessified".
2525 24. A recursive reference to a subpattern whose number was greater than 39
2526 went wrong under certain circumstances in UTF-8 mode. This bug could also
2527 have affected the operation of pcre_study().
2529 25. Realized that a little bit of performance could be had by replacing
2530 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2532 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2534 27. Possessive quantifiers such as a++ were previously implemented by turning
2535 them into atomic groups such as ($>a+). Now they have their own opcodes,
2536 which improves performance. This includes the automatically created ones
2537 from 23 above.
2539 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2540 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2541 the first matched character to be a colon. This applied both to named and
2542 numbered groups.
2544 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2546 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2547 building PCRE on Windows easier, so I have included it in the distribution.
2549 31. There is now a check in pcretest against a ridiculously large number being
2550 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2551 loop, the loop is abandoned.
2553 32. Forward references to subpatterns in conditions such as (?(2)...) where
2554 subpattern 2 is defined later cause pcre_compile() to search forwards in
2555 the pattern for the relevant set of parentheses. This search went wrong
2556 when there were unescaped parentheses in a character class, parentheses
2557 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2559 33. "Subroutine" calls and backreferences were previously restricted to
2560 referencing subpatterns earlier in the regex. This restriction has now
2561 been removed.
2563 34. Added a number of extra features that are going to be in Perl 5.10. On the
2564 whole, these are just syntactic alternatives for features that PCRE had
2565 previously implemented using the Python syntax or my own invention. The
2566 other formats are all retained for compatibility.
2568 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2569 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2570 also .NET compatible.
2572 (b) A recursion or subroutine call to a named group can now be defined as
2573 (?&name) as well as (?P>name).
2575 (c) A backreference to a named group can now be defined as \k<name> or
2576 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2577 5.10, are also .NET compatible.
2579 (d) A conditional reference to a named group can now use the syntax
2580 (?(<name>) or (?('name') as well as (?(name).
2582 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2583 groups (named and numbered) that are never evaluated inline, but can be
2584 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2585 is always false. There may be only one alternative in such a group.
2587 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2588 as the simple (?(R). The condition is true only if the most recent
2589 recursion is that of the given number or name. It does not search out
2590 through the entire recursion stack.
2592 (g) The escape \gN or \g{N} has been added, where N is a positive or
2593 negative number, specifying an absolute or relative reference.
2595 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2596 some "unreachable code" warnings.
2598 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2599 things, this adds five new scripts.
2601 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2602 There were also incompatibilities regarding the handling of \Q..\E inside
2603 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2604 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2606 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2607 matches an empty string, and forcibly breaks the loop. There were bugs in
2608 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2609 against aaaa the result was just "a" rather than "aaaa", for example. Two
2610 separate and independent bugs (that affected different cases) have been
2611 fixed.
2613 39. Refactored the code to abolish the use of different opcodes for small
2614 capturing bracket numbers. This is a tidy that I avoided doing when I
2615 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2616 The new approach is not only tidier, it makes it possible to reduce the
2617 memory needed to fix the previous bug (38).
2619 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2620 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2621 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2622 mode.
2624 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2625 report.
2627 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2628 copy construction and assignment in the C++ wrapper.
2630 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2631 couple of bugs that could have given wrong results in the "--newline=crlf"
2632 case.
2634 44. Added a number of casts and did some reorganization of signed/unsigned int
2635 variables following suggestions from Dair Grant. Also renamed the variable
2636 "this" as "item" because it is a C++ keyword.
2638 45. Arranged for dftables to add
2640 #include "pcre_internal.h"
2642 to pcre_chartables.c because without it, gcc 4.x may remove the array
2643 definition from the final binary if PCRE is built into a static library and
2644 dead code stripping is activated.
2646 46. For an unanchored pattern, if a match attempt fails at the start of a
2647 newline sequence, and the newline setting is CRLF or ANY, and the next two
2648 characters are CRLF, advance by two characters instead of one.
2651 Version 6.7 04-Jul-06
2652 ---------------------
2654 1. In order to handle tests when input lines are enormously long, pcretest has
2655 been re-factored so that it automatically extends its buffers when
2656 necessary. The code is crude, but this _is_ just a test program. The
2657 default size has been increased from 32K to 50K.
2659 2. The code in pcre_study() was using the value of the re argument before
2660 testing it for NULL. (Of course, in any sensible call of the function, it
2661 won't be NULL.)
2663 3. The memmove() emulation function in pcre_internal.h, which is used on
2664 systems that lack both memmove() and bcopy() - that is, hardly ever -
2665 was missing a "static" storage class specifier.
2667 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2668 containing an extended class (one that cannot be represented by a bitmap
2669 because it contains high-valued characters or Unicode property items, e.g.
2670 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2671 pattern, but PCRE should not loop if you do not (it no longer does).
2672 [Detail: two cases were found: (a) a repeated subpattern containing an
2673 extended class; (b) a recursive reference to a subpattern that followed a
2674 previous extended class. It wasn't skipping over the extended class
2675 correctly when UTF-8 mode was not set.]
2677 5. A negated single-character class was not being recognized as fixed-length
2678 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2679 compile error "lookbehind assertion is not fixed length".
2681 6. The RunPerlTest auxiliary script was showing an unexpected difference
2682 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2683 write a Perl script that can interpret lines of an input file either as
2684 byte characters or as UTF-8, which is what "perltest" was being required to
2685 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2686 can't do is switch easily at run time between having the "use utf8;" pragma
2687 or not. In the end, I fudged it by using the RunPerlTest script to insert
2688 "use utf8;" explicitly for the UTF-8 tests.
2690 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2691 the end of the subject string, contrary to the documentation and to what
2692 Perl does. This was true of both matching functions. Now it matches only at
2693 the start of the subject and immediately after *internal* newlines.
2695 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2696 a pointer to an int instead of a pointer to an unsigned long int. This
2697 caused problems on 64-bit systems.
2699 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2700 instance of the 'standard' template library not being so standard".
2702 10. There was no check on the number of named subpatterns nor the maximum
2703 length of a subpattern name. The product of these values is used to compute
2704 the size of the memory block for a compiled pattern. By supplying a very
2705 long subpattern name and a large number of named subpatterns, the size
2706 computation could be caused to overflow. This is now prevented by limiting
2707 the length of names to 32 characters, and the number of named subpatterns
2708 to 10,000.
2710 11. Subpatterns that are repeated with specific counts have to be replicated in
2711 the compiled pattern. The size of memory for this was computed from the
2712 length of the subpattern and the repeat count. The latter is limited to
2713 65535, but there was no limit on the former, meaning that integer overflow
2714 could in principle occur. The compiled length of a repeated subpattern is
2715 now limited to 30,000 bytes in order to prevent this.
2717 12. Added the optional facility to have named substrings with the same name.
2719 13. Added the ability to use a named substring as a condition, using the
2720 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2721 are numbers (not recommended). Forward references are permitted.
2723 14. Added forward references in named backreferences (if you see what I mean).
2725 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2726 pattern could run off the end of the subject. For example, the pattern
2727 "(?s)(.{1,5})"8 did this with the subject "ab".
2729 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2730 PCRE_CASELESS was set when matching characters that were quantified with ?
2731 or *.
2733 17. A character class other than a single negated character that had a minimum
2734 but no maximum quantifier - for example [ab]{6,} - was not handled
2735 correctly by pce_dfa_exec(). It would match only one character.
2737 18. A valid (though odd) pattern that looked like a POSIX character
2738 class but used an invalid character after [ (for example [[,abc,]]) caused
2739 pcre_compile() to give the error "Failed: internal error: code overflow" or
2740 in some cases to crash with a glibc free() error. This could even happen if
2741 the pattern terminated after [[ but there just happened to be a sequence of
2742 letters, a binary zero, and a closing ] in the memory that followed.
2744 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2745 over the years. Originally (before any Unicode support), just the bottom 8
2746 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2747 output from "man perlunicode" includes this:
2749 The regular expression compiler produces polymorphic opcodes. That
2750 is, the pattern adapts to the data and automatically switches to
2751 the Unicode character scheme when presented with Unicode data--or
2752 instead uses a traditional byte scheme when presented with byte
2753 data.
2755 Sadly, a wide octal escape does not cause a switch, and in a string with
2756 no other multibyte characters, these octal escapes are treated as before.
2757 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2758 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2759 Unicode string.
2761 I have not perpetrated such confusion in PCRE. Up till now, it took just
2762 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2763 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2764 translate to the appropriate multibyte character.
2766 29. Applied some refactoring to reduce the number of warnings from Microsoft
2767 and Borland compilers. This has included removing the fudge introduced
2768 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2769 a warning about an unused variable.
2771 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2772 characters since release 4.0, because Perl (from release 5.004) does not.
2773 [Or at least, is documented not to: some releases seem to be in conflict
2774 with the documentation.] However, when a pattern was studied with
2775 pcre_study() and all its branches started with \s, PCRE still included VT
2776 as a possible starting character. Of course, this did no harm; it just
2777 caused an unnecessary match attempt.
2779 22. Removed a now-redundant internal flag bit that recorded the fact that case
2780 dependency changed within the pattern. This was once needed for "required
2781 byte" processing, but is no longer used. This recovers a now-scarce options
2782 bit. Also moved the least significant internal flag bit to the most-
2783 significant bit of the word, which was not previously used (hangover from
2784 the days when it was an int rather than a uint) to free up another bit for
2785 the future.
2787 23. Added support for CRLF line endings as well as CR and LF. As well as the
2788 default being selectable at build time, it can now be changed at runtime
2789 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2790 specify that it is scanning data with non-default line endings.
2792 24. Changed the definition of CXXLINK to make it agree with the definition of
2793 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2795 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2796 recursions. This makes a big different to stack usage for some patterns.
2798 26. If a subpattern containing a named recursion or subroutine reference such
2799 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2800 the space required for the compiled pattern went wrong and gave too small a
2801 value. Depending on the environment, this could lead to "Failed: internal
2802 error: code overflow at offset 49" or "glibc detected double free or
2803 corruption" errors.
2805 27. Applied patches from Google (a) to support the new newline modes and (b) to
2806 advance over multibyte UTF-8 characters in GlobalReplace.
2808 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2809 difference for some implementation of PCRE in some Windows version.
2811 29. Added some extra testing facilities to pcretest:
2813 \q<number> in a data line sets the "match limit" value
2814 \Q<number> in a data line sets the "match recursion limt" value
2815 -S <number> sets the stack size, where <number> is in megabytes
2817 The -S option isn't available for Windows.
2820 Version 6.6 06-Feb-06
2821 ---------------------
2823 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2824 in pcreposix.h. I have copied the definition from pcre.h.
2826 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2827 because pcre.h is no longer a built file.
2829 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2830 not normally included in the compiled code.
2833 Version 6.5 01-Feb-06
2834 ---------------------
2836 1. When using the partial match feature with pcre_dfa_exec(), it was not
2837 anchoring the second and subsequent partial matches at the new starting
2838 point. This could lead to incorrect results. For example, with the pattern
2839 /1234/, partially matching against "123" and then "a4" gave a match.
2841 2. Changes to pcregrep:
2843 (a) All non-match returns from pcre_exec() were being treated as failures
2844 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2845 error message is output. Some extra information is given for the
2847 probably the only errors that are likely to be caused by users (by
2848 specifying a regex that has nested indefinite repeats, for instance).
2849 If there are more than 20 of these errors, pcregrep is abandoned.
2851 (b) A binary zero was treated as data while matching, but terminated the
2852 output line if it was written out. This has been fixed: binary zeroes
2853 are now no different to any other data bytes.
2855 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2856 used to set a locale for matching. The --locale=xxxx long option has
2857 been added (no short equivalent) to specify a locale explicitly on the
2858 pcregrep command, overriding the environment variables.
2860 (d) When -B was used with -n, some line numbers in the output were one less
2861 than they should have been.
2863 (e) Added the -o (--only-matching) option.
2865 (f) If -A or -C was used with -c (count only), some lines of context were
2866 accidentally printed for the final match.
2868 (g) Added the -H (--with-filename) option.
2870 (h) The combination of options -rh failed to suppress file names for files
2871 that were found from directory arguments.
2873 (i) Added the -D (--devices) and -d (--directories) options.
2875 (j) Added the -F (--fixed-strings) option.
2877 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2879 (l) Added the --colo(u)r option.
2881 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2882 is not present by default.
2884 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2885 items such as (?R) or (?1), when the recursion could match a number of
2886 alternatives. If it matched one of the alternatives, but subsequently,
2887 outside the recursion, there was a failure, the code tried to back up into
2888 the recursion. However, because of the way PCRE is implemented, this is not
2889 possible, and the result was an incorrect result from the match.
2891 In order to prevent this happening, the specification of recursion has
2892 been changed so that all such subpatterns are automatically treated as
2893 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2895 4. I had overlooked the fact that, in some locales, there are characters for
2896 which isalpha() is true but neither isupper() nor islower() are true. In
2897 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2898 and ordfeminine) are like this. This affected the treatment of \w and \W
2899 when they appeared in character classes, but not when they appeared outside
2900 a character class. The bit map for "word" characters is now created
2901 separately from the results of isalnum() instead of just taking it from the
2902 upper, lower, and digit maps. (Plus the underscore character, of course.)
2904 5. The above bug also affected the handling of POSIX character classes such as
2905 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2906 permanent tables. Instead, the bit maps for such a class were previously
2907 created as the appropriate unions of the upper, lower, and digit bitmaps.
2908 Now they are created by subtraction from the [[:word:]] class, which has
2909 its own bitmap.
2911 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2912 It is created by subtracting the vertical space characters (\x09, \x0a,
2913 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2914 subtraction was done in the overall bitmap for a character class, meaning
2915 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2916 be recognized. This bug has been fixed.
2918 7. Patches from the folks at Google:
2920 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2921 real life, but is still worth protecting against".
2923 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2924 regular expressions".
2926 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2927 have it.
2929 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2930 "configure" and the latter not, in order to fix a problem somebody had
2931 with compiling the Arg class on HP-UX.
2933 (e) Improve the error-handling of the C++ wrapper a little bit.
2935 (f) New tests for checking recursion limiting.
2937 8. The pcre_memmove() function, which is used only if the environment does not
2938 have a standard memmove() function (and is therefore rarely compiled),
2939 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2940 returning a result (though PCRE never actually uses the result).
2942 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2943 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2944 returned instead of calling malloc() with an overflowing number that would
2945 most likely cause subsequent chaos.
2947 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2949 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2950 with this option is matched, the nmatch and pmatch options of regexec() are
2951 ignored.
2953 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2954 provided in case anyone wants to the the POSIX interface with UTF-8
2955 strings.
2957 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2958 C++ linking (needed for some HP-UX environments).
2960 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2961 (unused parameter) and in the pcre_printint() function (omitted "default"
2962 switch label when the default is to do nothing).
2964 15. Added some code to make it possible, when PCRE is compiled as a C++
2965 library, to replace subject pointers for pcre_exec() with a smart pointer
2966 class, thus making it possible to process discontinuous strings.
2968 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2969 much the same function. They were added by different people who were trying
2970 to make PCRE easy to compile on non-Unix systems. It has been suggested
2971 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2972 for compiling on Windows systems. I have therefore replaced it with
2973 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2974 defaults to "extern" for C or "extern C" for C++, which works fine on
2975 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2976 SCOPE with something explicit in config.h. In addition:
2978 (a) pcreposix.h still had just "extern" instead of either of these macros;
2979 I have replaced it with PCRE_DATA_SCOPE.
2981 (b) Functions such as _pcre_xclass(), which are internal to the library,
2982 but external in the C sense, all had PCRE_EXPORT in their definitions.
2983 This is apparently wrong for the Windows case, so I have removed it.
2984 (It makes no difference on Unix-like systems.)
2986 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2987 of recursive calls to match(). This is different to MATCH_LIMIT because
2988 that limits the total number of calls to match(), not all of which increase
2989 the depth of recursion. Limiting the recursion depth limits the amount of
2990 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2991 when PCRE is compiled, and changed at run time. A patch from Google adds
2992 this functionality to the C++ interface.
2994 18. Changes to the handling of Unicode character properties:
2996 (a) Updated the table to Unicode 4.1.0.
2998 (b) Recognize characters that are not in the table as "Cn" (undefined).
3000 (c) I revised the way the table is implemented to a much improved format
3001 which includes recognition of ranges. It now supports the ranges that
3002 are defined in UnicodeData.txt, and it also amalgamates other
3003 characters into ranges. This has reduced the number of entries in the
3004 table from around 16,000 to around 3,000, thus reducing its size
3005 considerably. I realized I did not need to use a tree structure after
3006 all - a binary chop search is just as efficient. Having reduced the
3007 number of entries, I extended their size from 6 bytes to 8 bytes to
3008 allow for more data.
3010 (d) Added support for Unicode script names via properties such as \p{Han}.
3012 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3013 matching that character.
3015 20. When matching a repeated Unicode property with a minimum greater than zero,
3016 (for example \pL{2,}), PCRE could look past the end of the subject if it
3017 reached it while seeking the minimum number of characters. This could
3018 happen only if some of the characters were more than one byte long, because
3019 there is a check for at least the minimum number of bytes.
3021 21. Refactored the implementation of \p and \P so as to be more general, to
3022 allow for more different types of property in future. This has changed the
3023 compiled form incompatibly. Anybody with saved compiled patterns that use
3024 \p or \P will have to recompile them.
3026 22. Added "Any" and "L&" to the supported property types.
3028 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3029 but give a compile time error if the value is greater than 0xff.
3031 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3032 accidentally not being installed or uninstalled.
3034 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3035 made were to insert the current release number. This seemed silly, because
3036 it made things harder for people building PCRE on systems that don't run
3037 "configure". I have turned pcre.h into a distributed file, no longer built
3038 by "configure", with the version identification directly included. There is
3039 no longer a pcre.h.in file.
3041 However, this change necessitated a change to the pcre-config script as
3042 well. It is built from pcre-config.in, and one of the substitutions was the
3043 release number. I have updated configure.ac so that ./configure now finds
3044 the release number by grepping pcre.h.
3046 26. Added the ability to run the tests under valgrind.
3049 Version 6.4 05-Sep-05
3050 ---------------------
3052 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3053 "--" to be printed when multiple files were scanned, even when none of the
3054 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3055 consider it to be a bug, and have restored the previous behaviour.
3057 2. A couple of code tidies to get rid of compiler warnings.
3059 3. The pcretest program used to cheat by referring to symbols in the library
3060 whose names begin with _pcre_. These are internal symbols that are not
3061 really supposed to be visible externally, and in some environments it is
3062 possible to suppress them. The cheating is now confined to including
3063 certain files from the library's source, which is a bit cleaner.
3065 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3066 file's purpose clearer.
3068 5. Reorganized pcre_ucp_findchar().
3071 Version 6.3 15-Aug-05
3072 ---------------------
3074 1. The file libpcre.pc.in did not have general read permission in the tarball.
3076 2. There were some problems when building without C++ support:
3078 (a) If C++ support was not built, "make install" and "make test" still
3079 tried to test it.
3081 (b) There were problems when the value of CXX was explicitly set. Some
3082 changes have been made to try to fix these, and ...
3084 (c) --disable-cpp can now be used to explicitly disable C++ support.
3086 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3087 backslash in a target when C++ was disabled. This confuses some
3088 versions of "make", apparently. Using an intermediate variable solves
3089 this. (Same for CPP_LOBJ.)
3091 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3092 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3093 necessary on certain architectures.
3095 4. Added a setting of -export-symbols-regex to the link command to remove
3096 those symbols that are exported in the C sense, but actually are local
3097 within the library, and not documented. Their names all begin with
3098 "_pcre_". This is not a perfect job, because (a) we have to except some
3099 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3100 available (and never for static libraries). I have made a note to try to
3101 find a way round (a) in the future.
3104 Version 6.2 01-Aug-05
3105 ---------------------
3107 1. There was no test for integer overflow of quantifier values. A construction
3108 such as {1111111111111111} would give undefined results. What is worse, if
3109 a minimum quantifier for a parenthesized subpattern overflowed and became
3110 negative, the calculation of the memory size went wrong. This could have
3111 led to memory overwriting.
3113 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3115 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3116 operating environments where this matters.
3118 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3119 PCRE options from within the C++ wrapper.
3121 5. Named capturing subpatterns were not being correctly counted when a pattern
3122 was compiled. This caused two problems: (a) If there were more than 100
3123 such subpatterns, the calculation of the memory needed for the whole
3124 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3125 back references of the form \12, where the number was greater than 9, were
3126 not recognized as back references, even though there were sufficient
3127 previous subpatterns.
3129 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3130 versions of gcc, e.g. 2.95.4.
3133 Version 6.1 21-Jun-05
3134 ---------------------
3136 1. There was one reference to the variable "posix" in pcretest.c that was not
3137 surrounded by "#if !defined NOPOSIX".
3139 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3140 the cross-check on the old pcre_info() function, for the benefit of the
3141 cut-down version of PCRE that is currently imported into Exim.
3143 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3144 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3145 patterns that start (?i) but I don't think that matters. The use of (?i) is
3146 just an example; this all applies to the other options as well.
3148 4. Since libtool seems to echo the compile commands it is issuing, the output
3149 from "make" can be reduced a bit by putting "@" in front of each libtool
3150 compile command.
3152 5. Patch from the folks at Google for configure.in to be a bit more thorough
3153 in checking for a suitable C++ installation before trying to compile the
3154 C++ stuff. This should fix a reported problem when a compiler was present,
3155 but no suitable headers.
3157 6. The man pages all had just "PCRE" as their title. I have changed them to
3158 be the relevant file name. I have also arranged that these names are
3159 retained in the file doc/pcre.txt, which is a concatenation in text format
3160 of all the man pages except the little individual ones for each function.
3162 7. The NON-UNIX-USE file had not been updated for the different set of source
3163 files that come with release 6. I also added a few comments about the C++
3164 wrapper.
3167 Version 6.0 07-Jun-05
3168 ---------------------
3170 1. Some minor internal re-organization to help with my DFA experiments.
3172 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3173 didn't matter for the library itself when fully configured, but did matter
3174 when compiling without UCP support, or within Exim, where the ucp files are
3175 not imported.
3177 3. Refactoring of the library code to split up the various functions into
3178 different source modules. The addition of the new DFA matching code (see
3179 below) to a single monolithic source would have made it really too
3180 unwieldy, quite apart from causing all the code to be include in a
3181 statically linked application, when only some functions are used. This is
3182 relevant even without the DFA addition now that patterns can be compiled in
3183 one application and matched in another.
3185 The downside of splitting up is that there have to be some external
3186 functions and data tables that are used internally in different modules of
3187 the library but which are not part of the API. These have all had their
3188 names changed to start with "_pcre_" so that they are unlikely to clash
3189 with other external names.
3191 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3192 a different (DFA) algorithm. Although it is slower than the original
3193 function, it does have some advantages for certain types of matching
3194 problem.
3196 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3197 including restarting after a partial match.
3199 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3200 defined when compiling for Windows was sent to me. I have put it into the
3201 code, though I have no means of testing or verifying it.
3203 7. Added the pcre_refcount() auxiliary function.
3205 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3206 match before or at the first newline in the subject string. In pcretest,
3207 the /f option on a pattern can be used to set this.
3209 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3210 would behave wrongly. This has been present in PCRE since release 4.0.
3212 10. A number of changes to the pcregrep command:
3214 (a) Refactored how -x works; insert ^(...)$ instead of setting
3215 PCRE_ANCHORED and checking the length, in preparation for adding
3216 something similar for -w.
3218 (b) Added the -w (match as a word) option.
3220 (c) Refactored the way lines are read and buffered so as to have more
3221 than one at a time available.
3223 (d) Implemented a pcregrep test script.
3225 (e) Added the -M (multiline match) option. This allows patterns to match
3226 over several lines of the subject. The buffering ensures that at least
3227 8K, or the rest of the document (whichever is the shorter) is available
3228 for matching (and similarly the previous 8K for lookbehind assertions).
3230 (f) Changed the --help output so that it now says
3232 -w, --word-regex(p)
3234 instead of two lines, one with "regex" and the other with "regexp"
3235 because that confused at least one person since the short forms are the
3236 same. (This required a bit of code, as the output is generated
3237 automatically from a table. It wasn't just a text change.)
3239 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3240 option but starts with a hyphen. Could be a pattern or a path name
3241 starting with a hyphen, for instance.
3243 (h) "-" can be given as a file name to represent stdin.
3245 (i) When file names are being printed, "(standard input)" is used for
3246 the standard input, for compatibility with GNU grep. Previously
3247 "<stdin>" was used.
3249 (j) The option --label=xxx can be used to supply a name to be used for
3250 stdin when file names are being printed. There is no short form.
3252 (k) Re-factored the options decoding logic because we are going to add
3253 two more options that take data. Such options can now be given in four
3254 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3256 (l) Added the -A, -B, and -C options for requesting that lines of context
3257 around matches be printed.
3259 (m) Added the -L option to print the names of files that do not contain
3260 any matching lines, that is, the complement of -l.
3262 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3263 continue to scan other files.
3265 (o) The -s option was incorrectly implemented. For compatibility with other
3266 greps, it now suppresses the error message for a non-existent or non-
3267 accessible file (but not the return code). There is a new option called
3268 -q that suppresses the output of matching lines, which was what -s was
3269 previously doing.
3271 (p) Added --include and --exclude options to specify files for inclusion
3272 and exclusion when recursing.
3274 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3275 Hopefully, it now does.
3277 12. Missing cast in pcre_study().
3279 13. Added an "uninstall" target to the makefile.
3281 14. Replaced "extern" in the function prototypes in Makefile.in with
3282 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3283 world, but is set differently for Windows.
3285 15. Added a second compiling function called pcre_compile2(). The only
3286 difference is that it has an extra argument, which is a pointer to an
3287 integer error code. When there is a compile-time failure, this is set
3288 non-zero, in addition to the error test pointer being set to point to an
3289 error message. The new argument may be NULL if no error number is required
3290 (but then you may as well call pcre_compile(), which is now just a
3291 wrapper). This facility is provided because some applications need a
3292 numeric error indication, but it has also enabled me to tidy up the way
3293 compile-time errors are handled in the POSIX wrapper.
3295 16. Added VPATH=.libs to the makefile; this should help when building with one
3296 prefix path and installing with another. (Or so I'm told by someone who
3297 knows more about this stuff than I do.)
3299 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3300 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3301 match everything, including newlines. This is not POSIX-compatible, but
3302 somebody wanted the feature. From pcretest it can be activated by using
3303 both the P and the s flags.
3305 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3307 19. libpcre.pc was being incorrectly installed as executable.
3309 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3310 it now also looks for '\r' so that it will work unmodified on Windows.
3312 21. Added Google's contributed C++ wrapper to the distribution.
3314 22. Added some untidy missing memory free() calls in pcretest, to keep
3315 Electric Fence happy when testing.
3319 Version 5.0 13-Sep-04
3320 ---------------------
3322 1. Internal change: literal characters are no longer packed up into items
3323 containing multiple characters in a single byte-string. Each character
3324 is now matched using a separate opcode. However, there may be more than one
3325 byte in the character in UTF-8 mode.
3327 2. The pcre_callout_block structure has two new fields: pattern_position and
3328 next_item_length. These contain the offset in the pattern to the next match
3329 item, and its length, respectively.
3331 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3332 insertion of callouts before each pattern item. Added the /C option to
3333 pcretest to make use of this.
3335 4. On the advice of a Windows user, the lines
3337 #if defined(_WIN32) || defined(WIN32)
3338 _setmode( _fileno( stdout ), 0x8000 );
3339 #endif /* defined(_WIN32) || defined(WIN32) */
3341 have been added to the source of pcretest. This apparently does useful
3342 magic in relation to line terminators.
3344 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3345 for the benefit of those environments where the "b" makes a difference.
3347 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3348 to know about it. I have put a hack into configure.in that adds in code
3349 to set GCC=yes if CC=icc. This seems to end up at a point in the
3350 generated configure script that is early enough to affect the setting of
3351 compiler options, which is what is needed, but I have no means of testing
3352 whether it really works. (The user who reported this had patched the
3353 generated configure script, which of course I cannot do.)
3355 LATER: After change 22 below (new libtool files), the configure script
3356 seems to know about icc (and also ecc). Therefore, I have commented out
3357 this hack in configure.in.
3359 7. Added support for pkg-config (2 patches were sent in).
3361 8. Negated POSIX character classes that used a combination of internal tables
3362 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3363 [[:^ascii]]. Typically, they would match almost any characters. The other
3364 POSIX classes were not broken in this way.
3366 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3367 to find the match, as PCRE was deluded into thinking that the match had to
3368 start at the start point or following a newline. The same bug applied to
3369 patterns with negative forward assertions or any backward assertions
3370 preceding ".*" at the start, unless the pattern required a fixed first
3371 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3373 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3374 starting at the last subject character, bytes beyond the end of the subject
3375 string were read.
3377 11. Renamed the variable "class" as "classbits" to make life easier for C++
3378 users. (Previously there was a macro definition, but it apparently wasn't
3379 enough.)
3381 12. Added the new field "tables" to the extra data so that tables can be passed
3382 in at exec time, or the internal tables can be re-selected. This allows
3383 a compiled regex to be saved and re-used at a later time by a different
3384 program that might have everything at different addresses.
3386 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3387 -R library as well as a -L library.
3389 14. The debugging options of pcretest (-d on the command line or D on a
3390 pattern) showed incorrect output for anything following an extended class
3391 that contained multibyte characters and which was followed by a quantifier.
3393 15. Added optional support for general category Unicode character properties
3394 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3395 support. It adds about 90K to the size of the library. The meanings of the
3396 inbuilt class escapes such as \d and \s have NOT been changed.
3398 16. Updated pcredemo.c to include calls to free() to release the memory for the
3399 compiled pattern.
3401 17. The generated file chartables.c was being created in the source directory
3402 instead of in the building directory. This caused the build to fail if the
3403 source directory was different from the building directory, and was
3404 read-only.
3406 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3407 file. No doubt somebody will tell me if they don't make sense... Also added
3408 Dan Mooney's comments about building on OpenVMS.
3410 19. Added support for partial matching via the PCRE_PARTIAL option for
3411 pcre_exec() and the \P data escape in pcretest.
3413 20. Extended pcretest with 3 new pattern features:
3415 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3416 write the compiled pattern to the file whose name is "rest-of-line".
3417 This is a straight binary dump of the data, with the saved pointer to
3418 the character tables forced to be NULL. The study data, if any, is
3419 written too. After writing, pcretest reads a new pattern.
3421 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3422 compiled pattern from the given file. There must not be any
3423 occurrences of "<" in the file name (pretty unlikely); if there are,
3424 pcretest will instead treat the initial "<" as a pattern delimiter.
3425 After reading in the pattern, pcretest goes on to read data lines as
3426 usual.
3428 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3429 and 16-bit fields in a compiled pattern, to simulate a pattern that
3430 was compiled on a host of opposite endianness.
3432 21. The pcre-exec() function can now cope with patterns that were compiled on
3433 hosts of opposite endianness, with this restriction:
3435 As for any compiled expression that is saved and used later, the tables
3436 pointer field cannot be preserved; the extra_data field in the arguments
3437 to pcre_exec() should be used to pass in a tables address if a value
3438 other than the default internal tables were used at compile time.
3440 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3441 now diagnosed as an error. Previously, most of the time, a negative number
3442 would have been treated as zero, but if in addition "ovector" was passed as
3443 NULL, a crash could occur.
3445 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3446 new versions from the libtool 1.5 distribution (the last one is a copy of
3447 a file called libtool.m4). This seems to have fixed the need to patch
3448 "configure" to support Darwin 1.3 (which I used to do). However, I still
3449 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3450 workstation).
3452 24. Changed the PCRE licence to be the more standard "BSD" licence.
3455 Version 4.5 01-Dec-03
3456 ---------------------
3458 1. There has been some re-arrangement of the code for the match() function so
3459 that it can be compiled in a version that does not call itself recursively.
3460 Instead, it keeps those local variables that need separate instances for
3461 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3462 needs to "recurse". Keeping track of where control must go is done by means
3463 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3464 hide most of the details from the main code, and operates only if
3465 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3466 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3467 operating.
3469 To make it easier for callers to provide specially tailored get/free
3470 functions for this usage, two new functions, pcre_stack_malloc, and
3471 pcre_stack_free, are used. They are always called in strict stacking order,
3472 and the size of block requested is always the same.
3474 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3475 PCRE has been compiled to use the stack or the heap for recursion. The
3476 -C option of pcretest uses this to show which version is compiled.
3478 A new data escape \S, is added to pcretest; it causes the amounts of store
3479 obtained and freed by both kinds of malloc/free at match time to be added
3480 to the output.
3482 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3483 what's available on my current Linux desktop machine.
3485 3. When matching a UTF-8 string, the test for a valid string at the start has
3486 been extended. If start_offset is not zero, PCRE now checks that it points
3487 to a byte that is the start of a UTF-8 character. If not, it returns
3488 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3489 this is necessary because there may be backward assertions in the pattern.
3490 When matching the same subject several times, it may save resources to use
3491 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3493 4. The code for checking the validity of UTF-8 strings has been tightened so
3494 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3495 containing "overlong sequences".
3497 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3498 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3499 should have been "&", but it just so happened that all the cases this let
3500 through by mistake were picked up later in the function.
3502 6. I had used a variable called "isblank" - this is a C99 function, causing
3503 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3505 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3506 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3507 so the version is shown; (c) stop "make test" echoing "./RunTest".
3509 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3511 9. The prototype for memmove() for systems that don't have it was using
3512 size_t, but the inclusion of the header that defines size_t was later. I've
3513 moved the #includes for the C headers earlier to avoid this.
3515 10. Added some adjustments to the code to make it easier to compiler on certain
3516 special systems:
3518 (a) Some "const" qualifiers were missing.
3519 (b) Added the macro EXPORT before all exported functions; by default this
3520 is defined to be empty.
3521 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3522 that it reads its output file name as an argument instead of writing
3523 to the standard output and assuming this can be redirected.
3525 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3526 class containing characters with values greater than 255, PCRE compilation
3527 went into a loop.
3529 12. A recursive reference to a subpattern that was within another subpattern
3530 that had a minimum quantifier of zero caused PCRE to crash. For example,
3531 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3532 recursion. If the recursively-called subpattern itself had a zero repeat,
3533 that was OK.
3535 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3536 buffer into which it was copied (for escape processing) was still set at
3537 1024, so long lines caused crashes.
3539 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3540 "internal error: code overflow...". This applied to any character class
3541 that was followed by a possessive quantifier.
3543 15. Modified the Makefile to add libpcre.la as a prerequisite for
3544 libpcreposix.la because I was told this is needed for a parallel build to
3545 work.
3547 16. If a pattern that contained .* following optional items at the start was
3548 studied, the wrong optimizing data was generated, leading to matching
3549 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3550 matching string must start with a or b or c. The correct conclusion for
3551 this pattern is that a match can start with any character.
3554 Version 4.4 13-Aug-03
3555 ---------------------
3557 1. In UTF-8 mode, a character class containing characters with values between
3558 127 and 255 was not handled correctly if the compiled pattern was studied.
3559 In fixing this, I have also improved the studying algorithm for such
3560 classes (slightly).
3562 2. Three internal functions had redundant arguments passed to them. Removal
3563 might give a very teeny performance improvement.
3565 3. Documentation bug: the value of the capture_top field in a callout is *one
3566 more than* the number of the hightest numbered captured substring.
3568 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3569 in incorrectly linking with a previously installed version. They now link
3570 explicitly with libpcre.la.
3572 5. configure.in no longer needs to recognize Cygwin specially.
3574 6. A problem in pcre.in for Windows platforms is fixed.
3576 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3577 pcretest, it used to include the size of the study block as part of its
3578 output. Unfortunately, the structure contains a field that has a different
3579 size on different hardware architectures. This meant that the tests that
3580 showed this size failed. As the block is currently always of a fixed size,
3581 this information isn't actually particularly useful in pcretest output, so
3582 I have just removed it.
3584 8. Three pre-processor statements accidentally did not start in column 1.
3585 Sadly, there are *still* compilers around that complain, even though
3586 standard C has not required this for well over a decade. Sigh.
3588 9. In pcretest, the code for checking callouts passed small integers in the
3589 callout_data field, which is a void * field. However, some picky compilers
3590 complained about the casts involved for this on 64-bit systems. Now
3591 pcretest passes the address of the small integer instead, which should get
3592 rid of the warnings.
3594 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3595 both compile and run time, and gives an error if an invalid UTF-8 sequence
3596 is found. There is a option for disabling this check in cases where the
3597 string is known to be correct and/or the maximum performance is wanted.
3599 11. In response to a bug report, I changed one line in Makefile.in from
3601 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3602 to
3603 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3605 to look similar to other lines, but I have no way of telling whether this
3606 is the right thing to do, as I do not use Windows. No doubt I'll get told
3607 if it's wrong...
3610 Version 4.3 21-May-03
3611 ---------------------
3613 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3614 Makefile.
3616 2. Some refactoring to improve the quality of the code:
3618 (i) The utf8_table... variables are now declared "const".
3620 (ii) The code for \cx, which used the "case flipping" table to upper case
3621 lower case letters, now just substracts 32. This is ASCII-specific,
3622 but the whole concept of \cx is ASCII-specific, so it seems
3623 reasonable.
3625 (iii) PCRE was using its character types table to recognize decimal and
3626 hexadecimal digits in the pattern. This is silly, because it handles
3627 only 0-9, a-f, and A-F, but the character types table is locale-
3628 specific, which means strange things might happen. A private
3629 table is now used for this - though it costs 256 bytes, a table is
3630 much faster than multiple explicit tests. Of course, the standard
3631 character types table is still used for matching digits in subject
3632 strings against \d.
3634 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3635 ending in _t are). So I've renamed it as ESC_tee.
3637 3. The first argument for regexec() in the POSIX wrapper should have been
3638 defined as "const".
3640 4. Changed pcretest to use malloc() for its buffers so that they can be
3641 Electric Fenced for debugging.
3643 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3644 to read one or more bytes before the start of the subject string. Often this
3645 had no effect on PCRE's behaviour, but in some circumstances it could
3646 provoke a segmentation fault.
3648 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3649 to try to read one or more bytes before the start of the subject string.
3651 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3652 UTF-8 support could misbehave in various ways if the subject string
3653 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3654 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3655 back over UTF-8 characters.)
3658 Version 4.2 14-Apr-03
3659 ---------------------
3661 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3663 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3664 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3665 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3666 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3667 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3669 Note: automatic setting of the BUILD variables is not yet working
3670 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3671 compile-time but not at link-time
3672 [LINK]: use for linking executables only
3673 make different versions for Windows and non-Windows
3674 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3675 libraries
3676 [LINK_FOR_BUILD]: new variable
3677 [OBJEXT]: use throughout
3678 [EXEEXT]: use throughout
3679 <winshared>: new target
3680 <wininstall>: new target
3681 <dftables.o>: use native compiler
3682 <dftables>: use native linker
3683 <install>: handle Windows platform correctly
3684 <clean>: ditto
3685 <check>: ditto
3686 copy DLL to top builddir before testing
3688 As part of these changes, -no-undefined was removed again. This was reported
3689 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3690 in any case.
3692 3. Some tidies to get rid of compiler warnings:
3694 . In the match_data structure, match_limit was an unsigned long int, whereas
3695 match_call_count was an int. I've made them both unsigned long ints.
3697 . In pcretest the fact that a const uschar * doesn't automatically cast to
3698 a void * provoked a warning.
3700 . Turning on some more compiler warnings threw up some "shadow" variables
3701 and a few more missing casts.
3703 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3704 option, a class that contained a single character with a value between 128
3705 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3707 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3708 option, a class that contained several characters, but with at least one
3709 whose value was between 128 and 255 caused PCRE to crash.
3712 Version 4.1 12-Mar-03
3713 ---------------------
3715 1. Compiling with gcc -pedantic found a couple of places where casts were
3716 needed, and a string in dftables.c that was longer than standard compilers are
3717 required to support.
3719 2. Compiling with Sun's compiler found a few more places where the code could
3720 be tidied up in order to avoid warnings.
3722 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3723 first of these names is deprecated in the latest Autoconf in favour of the name
3724 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3725 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3726 analogy I have changed it to CFLAGS_FOR_BUILD.
3728 4. Added -no-undefined to the linking command in the Makefile, because this is
3729 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3730 linking step for the pcreposix library.
3732 5. PCRE was failing to diagnose the case of two named groups with the same
3733 name.
3735 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3736 literal character that is needed in the subject for a match, and scans along to
3737 ensure that it is present before embarking on the full matching process. This
3738 saves time in cases of nested unlimited repeats that are never going to match.
3739 Problem: the scan can take a lot of time if the subject is very long (e.g.
3740 megabytes), thus penalizing straightforward matches. It is now done only if the
3741 amount of subject to be scanned is less than 1000 bytes.
3743 7. A lesser problem with the same optimization is that it was recording the
3744 first character of an anchored pattern as "needed", thus provoking a search
3745 right along the subject, even when the first match of the pattern was going to
3746 fail. The "needed" character is now not set for anchored patterns, unless it
3747 follows something in the pattern that is of non-fixed length. Thus, it still
3748 fulfils its original purpose of finding quick non-matches in cases of nested
3749 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3752 Version 4.0 17-Feb-03
3753 ---------------------
3755 1. If a comment in an extended regex that started immediately after a meta-item
3756 extended to the end of string, PCRE compiled incorrect data. This could lead to
3757 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3759 2. Moved to autoconf 2.53 and libtool 1.4.2.
3761 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3762 the special perltest8 script is no longer needed - all the tests can be run
3763 from a single perltest script.
3765 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3766 by \s. It has now been removed in PCRE. This means it isn't recognized as
3767 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3768 class [:space:] *does* include VT, thereby creating a mess.
3770 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3771 space and tab.
3773 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3774 its new features into the main test script, reducing the number of scripts.
3776 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3777 were backward compatible, and made the (?i) apply to the whole pattern, as if
3778 /i were given. Now it behaves more logically, and applies the option setting
3779 only to what follows. PCRE has been changed to follow suit. However, if it
3780 finds options settings right at the start of the pattern, it extracts them into
3781 the global options, as before. Thus, they show up in the info data.
3783 8. Added support for the \Q...\E escape sequence. Characters in between are
3784 treated as literals. This is slightly different from Perl in that $ and @ are
3785 also handled as literals inside the quotes. In Perl, they will cause variable
3786 interpolation. Note the following examples:
3788 Pattern PCRE matches Perl matches
3790 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3791 \Qabc\$xyz\E abc\$xyz abc\$xyz
3792 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3794 For compatibility with Perl, \Q...\E sequences are recognized inside character
3795 classes as well as outside them.
3797 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3798 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3799 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3800 signed/unsigned warnings.
3802 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3803 option for pcretest, so I've replaced it by a simple function that does just
3804 that job.
3806 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3807 "pcregrep -".
3809 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3810 Java package. This provides some syntactic sugar for simple cases of what my
3811 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3812 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3813 item, you can use this simplified notation. Note that only makes sense with
3814 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3815 greediness, whatever the setting of the PCRE_UNGREEDY option.
3817 13. A change of greediness default within a pattern was not taking effect at
3818 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3819 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3820 was abstracted outside.
3822 14. PCRE now supports the \G assertion. It is true when the current matching
3823 position is at the start point of the match. This differs from \A when the
3824 starting offset is non-zero. Used with the /g option of pcretest (or similar
3825 code), it works in the same way as it does for Perl's /g option. If all
3826 alternatives of a regex begin with \G, the expression is anchored to the start
3827 match position, and the "anchored" flag is set in the compiled expression.
3829 15. Some bugs concerning the handling of certain option changes within patterns
3830 have been fixed. These applied to options other than (?ims). For example,
3831 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3832 been the other way round. Some of this was related to change 7 above.
3834 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3835 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3836 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3837 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3839 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3840 mode. Unlike ".", it always matches newline, whatever the setting of
3841 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3842 assertions. Perl allows it, but it doesn't (in general) work because it can't
3843 calculate the length of the lookbehind. At least, that's the case for Perl
3844 5.8.0 - I've been told they are going to document that it doesn't work in
3845 future.
3847 18. Added an error diagnosis for escapes that PCRE does not support: these are
3848 \L, \l, \N, \P, \p, \U, \u, and \X.
3850 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3851 reading past the end of the pattern in cases such as /[abcd/.
3853 20. PCRE was getting more memory than necessary for patterns with classes that
3854 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3856 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3857 compiling PCRE for use with Virtual Pascal.
3859 22. Small fix to the Makefile to make it work properly if the build is done
3860 outside the source tree.
3862 23. Added a new extension: a condition to go with recursion. If a conditional
3863 subpattern starts with (?(R) the "true" branch is used if recursion has
3864 happened, whereas the "false" branch is used only at the top level.
3866 24. When there was a very long string of literal characters (over 255 bytes
3867 without UTF support, over 250 bytes with UTF support), the computation of how
3868 much memory was required could be incorrect, leading to segfaults or other
3869 strange effects.
3871 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3872 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3873 there was a subsequent back reference to those brackets. This meant that, for
3874 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3875 possible to check for precisely this case. All we can do is abandon the
3876 optimization if .* occurs inside capturing brackets when there are any back
3877 references whatsoever. (See below for a better fix that came later.)
3879 26. The handling of the optimization for finding the first character of a
3880 non-anchored pattern, and for finding a character that is required later in the
3881 match were failing in some cases. This didn't break the matching; it just
3882 failed to optimize when it could. The way this is done has been re-implemented.
3884 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3886 28. Added a new feature that provides some of the functionality that Perl
3887 provides with (?{...}). The facility is termed a "callout". The way it is done
3888 in PCRE is for the caller to provide an optional function, by setting
3889 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3890 global variable. By default it is unset, which disables all calling out. To get
3891 the function called, the regex must include (?C) at appropriate points. This
3892 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3893 This provides a means of identifying different callout points. When PCRE
3894 reaches such a point in the regex, if pcre_callout has been set, the external
3895 function is called. It is provided with data in a structure called
3896 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3897 matching continues; if it returns a non-zero value, the match at the current
3898 point fails. However, backtracking will occur if possible. [This was changed
3899 later and other features added - see item 49 below.]
3901 29. pcretest is upgraded to test the callout functionality. It provides a
3902 callout function that displays information. By default, it shows the start of
3903 the match and the current position in the text. There are some new data escapes
3904 to vary what happens:
3906 \C+ in addition, show current contents of captured substrings
3907 \C- do not supply a callout function
3908 \C!n return 1 when callout number n is reached
3909 \C!n!m return 1 when callout number n is reached for the mth time
3911 30. If pcregrep was called with the -l option and just a single file name, it
3912 output "<stdin>" if a match was found, instead of the file name.
3914 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3915 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3916 pcre_exec(). This saves a malloc/free per call. The default value of
3917 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3918 when configuring.
3920 32. The default maximum size of a compiled pattern is 64K. There have been a
3921 few cases of people hitting this limit. The code now uses macros to handle the
3922 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3923 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3924 configuring. Tests 2 and 5 work only with 2-byte links because they output
3925 debugging information about compiled patterns.
3927 33. Internal code re-arrangements:
3929 (a) Moved the debugging function for printing out a compiled regex into
3930 its own source file (printint.c) and used #include to pull it into
3931 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3932 separate copies.
3934 (b) Defined the list of op-code names for debugging as a macro in
3935 internal.h so that it is next to the definition of the opcodes.
3937 (c) Defined a table of op-code lengths for simpler skipping along compiled
3938 code. This is again a macro in internal.h so that it is next to the
3939 definition of the opcodes.
3941 34. Added support for recursive calls to individual subpatterns, along the
3942 lines of Robin Houston's patch (but implemented somewhat differently).
3944 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3945 allow it to read and process whole directories in Win32. This code was
3946 contributed by Lionel Fourquaux; it has not been tested by me.
3948 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3949 used to name a group. Names consist of alphanumerics and underscores, and must
3950 be unique. Back references use the syntax (?P=name) and recursive calls use
3951 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3952 numbers. The function pcre_fullinfo() can be used after compilation to extract
3953 a name/number map. There are three relevant calls:
3955 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3956 PCRE_INFO_NAMECOUNT yields the number of entries
3957 PCRE_INFO_NAMETABLE yields a pointer to the map.
3959 The map is a vector of fixed-size entries. The size of each entry depends on
3960 the length of the longest name used. The first two bytes of each entry are the
3961 group number, most significant byte first. There follows the corresponding
3962 name, zero terminated. The names are in alphabetical order.
3964 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3965 case instead of 255. Making it the same both with and without UTF-8 support
3966 means that the same test output works with both.
3968 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3969 calling malloc() with a zero argument.
3971 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3972 optimization. I've improved things by keeping a bitmap of backreferences with
3973 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3974 fact referenced, the optimization can be applied. It is unlikely that a
3975 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3976 the match to follow \n) will appear inside brackets with a number greater than
3977 31, but if it does, any back reference > 31 suppresses the optimization.
3979 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3980 of disabling numbered capturing parentheses. Any opening parenthesis that is
3981 not followed by ? behaves as if it were followed by ?: but named parentheses
3982 can still be used for capturing (and they will acquire numbers in the usual
3983 way).
3985 41. Redesigned the return codes from the match() function into yes/no/error so
3986 that errors can be passed back from deep inside the nested calls. A malloc
3987 failure while inside a recursive subpattern call now causes the
3988 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3990 42. It is now possible to set a limit on the number of times the match()
3991 function is called in a call to pcre_exec(). This facility makes it possible to
3992 limit the amount of recursion and backtracking, though not in a directly
3993 obvious way, because the match() function is used in a number of different
3994 circumstances. The count starts from zero for each position in the subject
3995 string (for non-anchored patterns). The default limit is, for compatibility, a
3996 large number, namely 10 000 000. You can change this in two ways:
3998 (a) When configuring PCRE before making, you can use --with-match-limit=n
3999 to set a default value for the compiled library.
4001 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4002 a different value is set. See 45 below.
4004 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4006 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4007 of things that can be changed at compile time. The first argument specifies
4008 what is wanted and the second points to where the information is to be placed.
4009 The current list of available information is:
4013 The output is an integer that is set to one if UTF-8 support is available;
4014 otherwise it is set to zero.
4018 The output is an integer that it set to the value of the code that is used for
4019 newline. It is either LF (10) or CR (13).
4023 The output is an integer that contains the number of bytes used for internal
4024 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4028 The output is an integer that contains the threshold above which the POSIX
4029 interface uses malloc() for output vectors. See item 31 above.
4033 The output is an unsigned integer that contains the default limit of the number
4034 of match() calls in a pcre_exec() execution. See 42 above.
4036 44. pcretest has been upgraded by the addition of the -C option. This causes it
4037 to extract all the available output from the new pcre_config() function, and to
4038 output it. The program then exits immediately.
4040 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4041 order to support additional features. One way would have been to define
4042 pcre_exec2() (for example) with extra arguments, but this would not have been
4043 extensible, and would also have required all calls to the original function to
4044 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4045 is used for passing in "extra" data from pcre_study().
4047 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4048 contains the following fields:
4050 flags a bitmap indicating which of the following fields are set
4051 study_data opaque data from pcre_study()
4052 match_limit a way of specifying a limit on match() calls for a specific
4053 call to pcre_exec()
4054 callout_data data for callouts (see 49 below)
4056 The flag bits are also defined in pcre.h, and are
4062 The pcre_study() function now returns one of these new pcre_extra blocks, with
4063 the actual study data pointed to by the study_data field, and the
4064 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4065 before. That is, this change is entirely upwards-compatible and requires no
4066 change to existing code.
4068 If you want to pass in additional data to pcre_exec(), you can either place it
4069 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4070 block.
4072 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4073 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4074 times with different match limits, until it finds the minimum value needed for
4075 pcre_exec() to complete. The value is then output. This can be instructive; for
4076 most simple matches the number is quite small, but for pathological cases it
4077 gets very large very quickly.
4079 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4080 returns the size of the data block pointed to by the study_data field in a
4081 pcre_extra block, that is, the value that was passed as the argument to
4082 pcre_malloc() when PCRE was getting memory in which to place the information
4083 created by pcre_study(). The fourth argument should point to a size_t variable.
4084 pcretest has been extended so that this information is shown after a successful
4085 pcre_study() call when information about the compiled regex is being displayed.
4087 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4088 because what follows is always an absolute path. (Later: it turns out that this
4089 is more than cosmetic for MinGW, because it doesn't like empty path
4090 components.)
4092 49. Some changes have been made to the callout feature (see 28 above):
4094 (i) A callout function now has three choices for what it returns:
4096 0 => success, carry on matching
4097 > 0 => failure at this point, but backtrack if possible
4098 < 0 => serious error, return this value from pcre_exec()
4100 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4101 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4102 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4103 use by callout functions. It will never be used by PCRE itself.
4105 (ii) The pcre_extra structure (see 45 above) has a void * field called
4106 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4107 pcre_callout_block structure has a field of the same name. The contents of
4108 the field passed in the pcre_extra structure are passed to the callout
4109 function in the corresponding field in the callout block. This makes it
4110 easier to use the same callout-containing regex from multiple threads. For
4111 testing, the pcretest program has a new data escape
4113 \C*n pass the number n (may be negative) as callout_data
4115 If the callout function in pcretest receives a non-zero value as
4116 callout_data, it returns that value.
4118 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4119 there were some redundant $(CFLAGS) in commands that are now specified as
4120 $(LINK), which already includes $(CFLAGS).
4122 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4123 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4124 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4125 one-byte characters throughout. Note that case-insensitive matching applies
4126 only to characters whose values are less than 256. PCRE doesn't support the
4127 notion of cases for higher-valued characters.
4129 (i) A character class whose characters are all within 0-255 is handled as
4130 a bit map, and the map is inverted for negative classes. Previously, a
4131 character > 255 always failed to match such a class; however it should
4132 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4134 (ii) A negated character class with a single character < 255 is coded as
4135 "not this character" (OP_NOT). This wasn't working properly when the test
4136 character was multibyte, either singly or repeated.
4138 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4139 mode, for example: \x{100}{2,3}.
4141 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4142 singly or repeated) now correctly test multibyte characters. However,
4143 PCRE doesn't recognize any characters with values greater than 255 as
4144 digits, spaces, or word characters. Such characters always match \D, \S,
4145 and \W, and never match \d, \s, or \w.
4147 (v) Classes may now contain characters and character ranges with values
4148 greater than 255. For example: [ab\x{100}-\x{400}].
4150 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4151 PCRE in UTF-8 mode.
4153 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4154 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4155 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4156 value.)
4158 53. The single man page has become too large. I have therefore split it up into
4159 a number of separate man pages. These also give rise to individual HTML pages;
4160 these are now put in a separate directory, and there is an index.html page that
4161 lists them all. Some hyperlinking between the pages has been installed.
4163 54. Added convenience functions for handling named capturing parentheses.
4165 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4166 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4167 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4168 are faulted.
4170 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4171 calling configure. These values are used when compiling the dftables.c program
4172 which is run to generate the source of the default character tables. They
4173 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4174 you will need to set these values.
4176 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4179 Version 3.9 02-Jan-02
4180 ---------------------
4182 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4184 2. If --disable-static was given, the building process failed when trying to
4185 build pcretest and pcregrep. (For some reason it was using libtool to compile
4186 them, which is not right, as they aren't part of the library.)
4189 Version 3.8 18-Dec-01
4190 ---------------------
4192 1. The experimental UTF-8 code was completely screwed up. It was packing the
4193 bytes in the wrong order. How dumb can you get?
4196 Version 3.7 29-Oct-01
4197 ---------------------
4199 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4200 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4201 this didn't happen under Solaris 8, where I normally test things.
4203 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4204 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4205 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4206 this fix an hour or so after the initial 3.7 release.)
4209 Version 3.6 23-Oct-01
4210 ---------------------
4212 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4213 offsets passed as NULL with zero offset count.
4215 2. The config.guess and config.sub files had not been updated when I moved to
4216 the latest autoconf.
4219 Version 3.5 15-Aug-01
4220 ---------------------
4222 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4223 had been forgotten.
4225 2. By using declared but undefined structures, we can avoid using "void"
4226 definitions in pcre.h while keeping the internal definitions of the structures
4227 private.
4229 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4230 user point of view, this means that both static and shared libraries are built
4231 by default, but this can be individually controlled. More of the work of
4232 handling this static/shared cases is now inside libtool instead of PCRE's make
4233 file.
4235 4. The pcretest utility is now installed along with pcregrep because it is
4236 useful for users (to test regexs) and by doing this, it automatically gets
4237 relinked by libtool. The documentation has been turned into a man page, so
4238 there are now .1, .txt, and .html versions in /doc.
4240 5. Upgrades to pcregrep:
4241 (i) Added long-form option names like gnu grep.
4242 (ii) Added --help to list all options with an explanatory phrase.
4243 (iii) Added -r, --recursive to recurse into sub-directories.
4244 (iv) Added -f, --file to read patterns from a file.
4246 6. pcre_exec() was referring to its "code" argument before testing that
4247 argument for NULL (and giving an error if it was NULL).
4249 7. Upgraded Makefile.in to allow for compiling in a different directory from
4250 the source directory.
4252 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4253 options bits, the pointer it was passed was to an int instead of to an unsigned
4254 long int. This mattered only on 64-bit systems.
4256 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4257 generated) instead of pcre.in, which it its source. Also made the same change
4258 in several of the .c files.
4260 10. A new release of gcc defines printf() as a macro, which broke pcretest
4261 because it had an ifdef in the middle of a string argument for printf(). Fixed
4262 by using separate calls to printf().
4264 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4265 script, to force use of CR or LF instead of \n in the source. On non-Unix
4266 systems, the value can be set in config.h.
4268 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4269 absolute limit. Changed the text of the error message to make this clear, and
4270 likewise updated the man page.
4272 13. The limit of 99 on the number of capturing subpatterns has been removed.
4273 The new limit is 65535, which I hope will not be a "real" limit.
4276 Version 3.4 22-Aug-00
4277 ---------------------
4279 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4281 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4284 Version 3.3 01-Aug-00
4285 ---------------------
4287 1. If an octal character was given, but the value was greater than \377, it
4288 was not getting masked to the least significant bits, as documented. This could
4289 lead to crashes in some systems.
4291 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4292 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4294 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4295 These just pass their arguments on to (pcre_free)(), but they are provided
4296 because some uses of PCRE bind it to non-C systems that can call its functions,
4297 but cannot call free() or pcre_free() directly.
4299 4. Add "make test" as a synonym for "make check". Corrected some comments in
4300 the Makefile.
4302 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4303 Makefile.
4305 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4306 command called pgrep for grepping around the active processes.
4308 7. Added the beginnings of support for UTF-8 character strings.
4310 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4311 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4312 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4313 out for the ar command.)
4316 Version 3.2 12-May-00
4317 ---------------------
4319 This is purely a bug fixing release.
4321 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4322 of ZA. This was just one example of several cases that could provoke this bug,
4323 which was introduced by change 9 of version 2.00. The code for breaking
4324 infinite loops after an iteration that matches an empty string was't working
4325 correctly.
4327 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4328 when matched against abbab (for example). After matching an empty string, it
4329 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4330 caused it to match further down the string than it should.
4332 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4333 was there because it doesn't seem to be needed, and it causes trouble on some
4334 systems, as it is not a Standard C header. It has been removed.
4336 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4337 were reported on the Macintosh. The changes were from
4339 while ((c = *(++ptr)) != 0 && c != '\n');
4340 to
4341 while ((c = *(++ptr)) != 0 && c != '\n') ;
4343 Totally extraordinary, but if that's what it takes...
4345 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4346 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4347 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4348 assumes the way PCRE uses memmove() (always moving upwards).
4350 6. PCRE is being used in one environment where strchr() is not available. There
4351 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4352 faster code anyway.
4355 Version 3.1 09-Feb-00
4356 ---------------------
4358 The only change in this release is the fixing of some bugs in Makefile.in for
4359 the "install" target:
4361 (1) It was failing to install pcreposix.h.
4363 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4366 Version 3.0 01-Feb-00
4367 ---------------------
4369 1. Add support for the /+ modifier to perltest (to output $` like it does in
4370 pcretest).
4372 2. Add support for the /g modifier to perltest.
4374 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4375 matches null strings.
4377 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4378 pattern. Perl treats empty patterns specially - it reuses the most recent
4379 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4380 effect.
4382 5. The POSIX interface was broken in that it was just handing over the POSIX
4383 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4384 required a bigger vector, with some working space on the end. This means that
4385 the POSIX wrapper now has to get and free some memory, and copy the results.
4387 6. Added some simple autoconf support, placing the test data and the
4388 documentation in separate directories, re-organizing some of the
4389 information files, and making it build pcre-config (a GNU standard). Also added
4390 libtool support for building PCRE as a shared library, which is now the
4391 default.
4393 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4394 09 are not valid octal constants. Single digits will be used for minor values
4395 less than 10.
4397 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4398 existing programs that set these in the POSIX interface can use PCRE without
4399 modification.
4401 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4402 return all that pcre_info() returns, plus additional data. The pcre_info()
4403 function is retained for compatibility, but is considered to be obsolete.
4405 10. Added experimental recursion feature (?R) to handle one common case that
4406 Perl 5.6 will be able to do with (?p{...}).
4408 11. Added support for POSIX character classes like [:alpha:], which Perl is
4409 adopting.
4412 Version 2.08 31-Aug-99
4413 ----------------------
4415 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4416 trying to match at the startoffset position, but instead was moving forward to
4417 the next newline as if a previous match had failed.
4419 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4420 and could get into a loop if a null string was matched other than at the start
4421 of the subject.
4423 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4424 be distinguished at compile time, and for completeness also added PCRE_DATE.
4426 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4427 in GnuWin32 environments.
4430 Version 2.07 29-Jul-99
4431 ----------------------
4433 1. The documentation is now supplied in plain text form and HTML as well as in
4434 the form of man page sources.
4436 2. C++ compilers don't like assigning (void *) values to other pointer types.
4437 In particular this affects malloc(). Although there is no problem in Standard
4438 C, I've put in casts to keep C++ compilers happy.
4440 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4441 should be (const char *).
4443 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4444 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4445 However, I haven't made this a standard facility. The documentation doesn't
4446 mention it, and the Makefile doesn't support it.
4448 5. The Makefile now contains an "install" target, with editable destinations at
4449 the top of the file. The pcretest program is not installed.
4451 6. pgrep -V now gives the PCRE version number and date.
4453 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4454 causing the entire string to be ignored, instead of just the last character.
4456 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4457 non-matching string, it can take a very, very long time, even for strings of
4458 quite modest length, because of the nested recursion. PCRE now does better in
4459 some of these cases. It does this by remembering the last required literal
4460 character in the pattern, and pre-searching the subject to ensure it is present
4461 before running the real match. In other words, it applies a heuristic to detect
4462 some types of certain failure quickly, and in the above example, if presented
4463 with a string that has no trailing " it gives "no match" very quickly.
4465 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4466 other alternatives are tried instead.
4469 Version 2.06 09-Jun-99
4470 ----------------------
4472 1. Change pcretest's output for amount of store used to show just the code
4473 space, because the remainder (the data block) varies in size between 32-bit and
4474 64-bit systems.
4476 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4477 start matching at. This allows lookbehinds to work when searching for multiple
4478 occurrences in a string.
4480 3. Added additional options to pcretest for testing multiple occurrences:
4482 /+ outputs the rest of the string that follows a match
4483 /g loops for multiple occurrences, using the new startoffset argument
4484 /G loops for multiple occurrences by passing an incremented pointer
4486 4. PCRE wasn't doing the "first character" optimization for patterns starting
4487 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4488 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4489 the letter 'x'. On long subject strings, this gives a significant speed-up.
4492 Version 2.05 21-Apr-99
4493 ----------------------
4495 1. Changed the type of magic_number from int to long int so that it works
4496 properly on 16-bit systems.
4498 2. Fixed a bug which caused patterns starting with .* not to work correctly
4499 when the subject string contained newline characters. PCRE was assuming
4500 anchoring for such patterns in all cases, which is not correct because .* will
4501 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4502 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4503 must be retried after every newline in the subject.
4506 Version 2.04 18-Feb-99
4507 ----------------------
4509 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4510 computation of the store needed to hold the pattern was incorrect (too large).
4511 If such patterns were nested a few deep, this could multiply and become a real
4512 problem.
4514 2. Added /M option to pcretest to show the memory requirement of a specific
4515 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4517 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4518 compiled in such a way that the backtracking after subsequent failure was
4519 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4520 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4523 Version 2.03 02-Feb-99
4524 ----------------------
4526 1. Fixed typo and small mistake in man page.
4528 2. Added 4th condition (GPL supersedes if conflict) and created separate
4529 LICENCE file containing the conditions.
4531 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4532 Perl, that is the internal \ allows the delimiter to be included in the
4533 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4534 the final delimiter, add \ to the end of the pattern (to test the error).
4536 4. Added the convenience functions for extracting substrings after a successful
4537 match. Updated pcretest to make it able to test these functions.
4540 Version 2.02 14-Jan-99
4541 ----------------------
4543 1. Initialized the working variables associated with each extraction so that
4544 their saving and restoring doesn't refer to uninitialized store.
4546 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4547 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4548 fix the problem.
4550 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4551 calls, and wasn't printing the correct value for compiling calls. Increased the
4552 default value of LOOPREPEAT, and the number of significant figures in the
4553 times.
4555 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4557 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4558 a building problem on Windows NT with a FAT file system.
4561 Version 2.01 21-Oct-98
4562 ----------------------
4564 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4565 to character tables built by pcre_maketables() in the current locale. If NULL
4566 is passed, the default tables are used.
4569 Version 2.00 24-Sep-98
4570 ----------------------
4572 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4573 it any more.
4575 2. Allow quantification of (?>) groups, and make it work correctly.
4577 3. The first character computation wasn't working for (?>) groups.
4579 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4580 end of the subject) and add 5.005's \z, which really does match only at the
4581 very end of the subject.
4583 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4585 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4586 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4587 localized options. All options to pcre_study() were also removed.
4589 7. Add other new features from 5.005:
4591 $(?<= positive lookbehind
4592 $(?<! negative lookbehind
4593 (?imsx-imsx) added the unsetting capability
4594 such a setting is global if at outer level; local otherwise
4595 (?imsx-imsx:) non-capturing groups with option setting
4596 (?(cond)re|re) conditional pattern matching
4598 A backreference to itself in a repeated group matches the previous
4599 captured string.
4601 8. General tidying up of studying (both automatic and via "study")
4602 consequential on the addition of new assertions.
4604 9. As in 5.005, unlimited repeated groups that could match an empty substring
4605 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4606 runtime if any iteration does actually match an empty substring.
4608 10. Include the RunTest script in the distribution.
4610 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4611 discrepancies, some of which were old and were also with respect to 5.004. They
4612 have now been fixed.
4615 Version 1.09 28-Apr-98
4616 ----------------------
4618 1. A negated single character class followed by a quantifier with a minimum
4619 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4620 program crashes, or just wrong answers. This did not apply to negated classes
4621 containing more than one character, or to minima other than one.
4624 Version 1.08 27-Mar-98
4625 ----------------------
4627 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4629 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4630 latter must appear before anything that relies on it in the pattern.
4633 Version 1.07 16-Feb-98
4634 ----------------------
4636 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4637 repeat of a potentially empty string).
4640 Version 1.06 23-Jan-98
4641 ----------------------
4643 1. Added Markus Oberhumer's little patches for C++.
4645 2. Literal strings longer than 255 characters were broken.
4648 Version 1.05 23-Dec-97
4649 ----------------------
4651 1. Negated character classes containing more than one character were failing if
4652 PCRE_CASELESS was set at run time.
4655 Version 1.04 19-Dec-97
4656 ----------------------
4658 1. Corrected the man page, where some "const" qualifiers had been omitted.
4660 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4661 input syntax.
4663 3. Fixed memory leak which occurred when a regex with back references was
4664 matched with an offsets vector that wasn't big enough. The temporary memory
4665 that is used in this case wasn't being freed if the match failed.
4667 4. Tidied pcretest to ensure it frees memory that it gets.
4669 5. Temporary memory was being obtained in the case where the passed offsets
4670 vector was exactly big enough.
4672 6. Corrected definition of offsetof() from change 5 below.
4674 7. I had screwed up change 6 below and broken the rules for the use of
4675 setjmp(). Now fixed.
4678 Version 1.03 18-Dec-97
4679 ----------------------
4681 1. A erroneous regex with a missing opening parenthesis was correctly
4682 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4683 on some systems.
4685 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4686 it was reported that one broken compiler failed on the former because "code" is
4687 also an independent variable.
4689 3. The erroneous regex a[]b caused an array overrun reference.
4691 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4692 fail on data ending with that character. (It was going on too far, and checking
4693 the next character, typically a binary zero.) This was specific to the
4694 optimized code for single-character negative classes.
4696 5. Added a contributed patch from the TIN world which does the following:
4698 + Add an undef for memmove, in case the the system defines a macro for it.
4700 + Add a definition of offsetof(), in case there isn't one. (I don't know
4701 the reason behind this - offsetof() is part of the ANSI standard - but
4702 it does no harm).
4704 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4705 most of the places where whitespace preceded '#'. I have given up and
4706 allowed the remaining 2 cases to be at the margin.
4708 + Rename some variables in pcre to eliminate shadowing. This seems very
4709 pedantic, but does no harm, of course.
4711 6. Moved the call to setjmp() into its own function, to get rid of warnings
4712 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4714 7. Constructs such as \d{8,} were compiling into the equivalent of
4715 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4716 outcome, but in this particular case used more store than had been allocated,
4717 which caused the bug to be discovered because it threw up an internal error.
4719 8. The debugging code in both pcre and pcretest for outputting the compiled
4720 form of a regex was going wrong in the case of back references followed by
4721 curly-bracketed repeats.
4724 Version 1.02 12-Dec-97
4725 ----------------------
4727 1. Typos in pcre.3 and comments in the source fixed.
4729 2. Applied a contributed patch to get rid of places where it used to remove
4730 'const' from variables, and fixed some signed/unsigned and uninitialized
4731 variable warnings.
4733 3. Added the "runtest" target to Makefile.
4735 4. Set default compiler flag to -O2 rather than just -O.
4738 Version 1.01 19-Nov-97
4739 ----------------------
4741 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4742 like /([ab]*)*/, that is, for classes with more than one character in them.
4744 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4745 as /((?>a*))*/ (a PCRE_EXTRA facility).
4748 Version 1.00 18-Nov-97
4749 ----------------------
4751 1. Added compile-time macros to support systems such as SunOS4 which don't have
4752 memmove() or strerror() but have other things that can be used instead.
4754 2. Arranged that "make clean" removes the executables.
4757 Version 0.99 27-Oct-97
4758 ----------------------
4760 1. Fixed bug in code for optimizing classes with only one character. It was
4761 initializing a 32-byte map regardless, which could cause it to run off the end
4762 of the memory it had got.
4764 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4767 Version 0.98 22-Oct-97
4768 ----------------------
4770 1. Fixed bug in code for handling temporary memory usage when there are more
4771 back references than supplied space in the ovector. This could cause segfaults.
4774 Version 0.97 21-Oct-97
4775 ----------------------
4777 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4779 2. Optimized negated single characters not to use a bit map.
4781 3. Brought error texts together as macro definitions; clarified some of them;
4782 fixed one that was wrong - it said "range out of order" when it meant "invalid
4783 escape sequence".
4785 4. Changed some char * arguments to const char *.
4787 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4789 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4790 pcretest.
4793 Version 0.96 16-Oct-97
4794 ----------------------
4796 1. Added a simple "pgrep" utility to the distribution.
4798 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4799 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4800 where "ddd" means "one or more decimal digits".
4802 3. Fixed serious bug. If a pattern had a back reference, but the call to
4803 pcre_exec() didn't supply a large enough ovector to record the related
4804 identifying subpattern, the match always failed. PCRE now remembers the number
4805 of the largest back reference, and gets some temporary memory in which to save
4806 the offsets during matching if necessary, in order to ensure that
4807 backreferences always work.
4809 4. Increased the compatibility with Perl in a number of ways:
4811 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4812 to request this handling. The option can be set at compile or exec time.
4814 (b) $ matches before a terminating newline by default; an option
4815 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4816 mode). The option can be set at compile or exec time.
4818 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4819 the same as Perl's. If the decimal number it represents is less than 10
4820 or there aren't that many previous left capturing parentheses, an octal
4821 escape is read. Inside a character class, it's always an octal escape,
4822 even if it is a single digit.
4824 (d) An escaped but undefined alphabetic character is taken as a literal,
4825 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4826 escapes.
4828 (e) {0} is now permitted. (The previous item is removed from the compiled
4829 pattern).
4831 5. Changed all the names of code files so that the basic parts are no longer
4832 than 10 characters, and abolished the teeny "globals.c" file.
4834 6. Changed the handling of character classes; they are now done with a 32-byte
4835 bit map always.
4837 7. Added the -d and /D options to pcretest to make it possible to look at the
4838 internals of compilation without having to recompile pcre.
4841 Version 0.95 23-Sep-97
4842 ----------------------
4844 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4845 \x20 at the start of a run of normal characters. These were being treated as
4846 real characters, instead of the source characters being re-checked.
4849 Version 0.94 18-Sep-97
4850 ----------------------
4852 1. The functions are now thread-safe, with the caveat that the global variables
4853 containing pointers to malloc() and free() or alternative functions are the
4854 same for all threads.
4856 2. Get pcre_study() to generate a bitmap of initial characters for non-
4857 anchored patterns when this is possible, and use it if passed to pcre_exec().
4860 Version 0.93 15-Sep-97
4861 ----------------------
4863 1. /(b)|(:+)/ was computing an incorrect first character.
4865 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4866 but not actually doing anything yet.
4868 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4869 as Perl does (e.g. [-az] or [az-]).
4871 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4872 all possible positions.
4874 5. Split up into different modules to avoid including unneeded functions in a
4875 compiled binary. However, compile and exec are still in one module. The "study"
4876 function is split off.
4878 6. The character tables are now in a separate module whose source is generated
4879 by an auxiliary program - but can then be edited by hand if required. There are
4880 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4881 toupper() in the code.
4883 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4884 make them global. Abolish the function for setting them, as the caller can now
4885 set them directly.
4888 Version 0.92 11-Sep-97
4889 ----------------------
4891 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4892 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4894 2. Caseless matching was not working in character classes if the characters in
4895 the pattern were in upper case.
4897 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4899 4. Make PCRE_ANCHORED public and accept as a compile option.
4901 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4902 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4903 pass them.
4905 6. Give an error if bad option bits passed at compile or run time.
4907 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4908 pcretest to cause it to pass that flag.
4910 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4911 options, and the first character, if set.
4913 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4916 Version 0.91 10-Sep-97
4917 ----------------------
4919 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4920 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4922 2. PCRE was looping on encountering an indefinitely repeated back reference to
4923 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4924 Perl does - treats the match as successful.
4926 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5