ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 971 - (show annotations)
Fri Jun 1 16:29:43 2012 UTC (7 years, 11 months ago) by ph10
File size: 231189 byte(s)
Error occurred while calculating annotation data.
Fix pcregrep build problem when bzip2 but not gzip was enabled.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.31
5 -----------------------------
7 1. Fixing a wrong JIT test case and some compiler warnings.
9 2. Removed a bashism from the RunTest script.
11 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12 to unsigned type, result still unsigned" that was given by an MS compiler
13 on encountering the code "-sizeof(xxx)".
15 4. Partial matching support is added to the JIT compiler.
17 5. Fixed several bugs concerned with partial matching of items that consist
18 of more than one character:
20 (a) /^(..)\1/ did not partially match "aba" because checking references was
21 done on an "all or nothing" basis. This also applied to repeated
22 references.
24 (b) \R did not give a hard partial match if \r was found at the end of the
25 subject.
27 (c) \X did not give a hard partial match after matching one or more
28 characters at the end of the subject.
30 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31 a partial match for the string "\r".
33 (e) When newline was set to CRLF, the metacharacter "." did not recognize
34 a partial match for a CR character at the end of the subject string.
36 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37 running pcretest, the text "(JIT)" added to the output whenever JIT is
38 actually used to run the match.
40 7. Individual JIT compile options can be set in pcretest by following -s+[+]
41 or /S+[+] with a digit between 1 and 7.
43 8. OP_NOT now supports any UTF character not just single-byte ones.
45 9. (*MARK) control verb is now supported by the JIT compiler.
47 10. The command "./RunTest list" lists the available tests without actually
48 running any of them. (Because I keep forgetting what they all are.)
52 12. Applied a (slightly modified) user-supplied patch that improves performance
53 when the heap is used for recursion (compiled with --disable-stack-for-
54 recursion). Instead of malloc and free for each heap frame each time a
55 logical recursion happens, frames are retained on a chain and re-used where
56 possible. This sometimes gives as much as 30% improvement.
58 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59 call.
61 14. As documented, (*COMMIT) is now confined to within a positive assertion.
63 15. It is now possible to link pcretest with libedit as an alternative to
64 libreadline.
66 16. (*COMMIT) control verb is now supported by the JIT compiler.
68 17. The Unicode data tables have been updated to Unicode 6.1.0.
70 18. Added --file-list option to pcregrep.
72 19. Added binary file support to pcregrep, including the -a, --binary-files,
73 -I, and --text options.
75 20. The madvise function is renamed for posix_madvise for QNX compatibility
76 reasons. Fixed by Giuseppe D'Angelo.
78 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
80 22. Changed the option for creating symbolic links for 16-bit man pages from
81 -s to -sf so that re-installing does not cause issues.
83 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
85 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86 matches in certain environments (the workspace was not being correctly
87 retained). Also added to pcre_dfa_exec() a simple plausibility check on
88 some of the workspace data at the beginning of a restart.
90 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
91 was not doing so when it should - probably a typo introduced by SVN 528
92 (change 8.10/14).
94 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
95 \w+ when the character tables indicated that \x{c4} was a word character.
96 There were several related cases, all because the tests for doing a table
97 lookup were testing for characters less than 127 instead of 255.
99 27. If a pattern contains capturing parentheses that are not used in a match,
100 their slots in the ovector are set to -1. For those that are higher than
101 any matched groups, this happens at the end of processing. In the case when
102 there were back references that the ovector was too small to contain
103 (causing temporary malloc'd memory to be used during matching), and the
104 highest capturing number was not used, memory off the end of the ovector
105 was incorrectly being set to -1. (It was using the size of the temporary
106 memory instead of the true size.)
108 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
109 ovector size, it uses memory at the end of the block that it has got.
111 29. Check for an overlong MARK name and give an error at compile time. The
112 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
114 30. JIT compiler update.
116 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
117 Rill for the patch.
119 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
121 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
123 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
124 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
125 was enabled.
128 Version 8.30 04-February-2012
129 -----------------------------
131 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
132 name is defined in ctype.h.
134 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
135 only in quite long subpatterns.
137 3. Removed the function pcre_info(), which has been obsolete and deprecated
138 since it was replaced by pcre_fullinfo() in February 2000.
140 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
141 match a (*MARK), and the match failed at the start of the subject, a
142 reference to memory before the start of the subject could occur. This bug
143 was introduced by fix 17 of release 8.21.
145 5. A reference to an unset group with zero minimum repetition was giving
146 totally wrong answers (in non-JavaScript-compatibility mode). For example,
147 /(another)?(\1?)test/ matched against "hello world test". This bug was
148 introduced in release 8.13.
150 6. Add support for 16-bit character strings (a large amount of work involving
151 many changes and refactorings).
153 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
154 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
155 from a file.
157 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
158 rounding is not applied in this particular case).
160 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
161 if they appear, or are escaped, in patterns.
163 10. Get rid of a number of -Wunused-but-set-variable warnings.
165 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
166 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
167 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
168 also returns the mark "x". This bug applied to capturing parentheses,
169 non-capturing parentheses, and atomic parentheses. It also applied to some
170 assertions.
172 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
173 information out of configure.ac instead of relying on pcre.h.generic, which
174 is not stored in the repository.
176 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
177 -lreadline.
179 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
181 15. Applied Graycode's patch to put the top-level frame on the stack rather
182 than the heap when not using the stack for recursion. This gives a
183 performance improvement in many cases when recursion is not deep.
185 16. Experimental code added to "pcretest -C" to output the stack frame size.
188 Version 8.21 12-Dec-2011
189 ------------------------
191 1. Updating the JIT compiler.
193 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
194 are added as well.
196 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
197 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
198 calling _pcre_jit_exec. Some extra comments are added.
200 4. (*MARK) settings inside atomic groups that do not contain any capturing
201 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
202 was introduced by change 18 for 8.20.
204 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
205 ECMA-262 standard.
207 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
208 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
209 This bug was probably introduced by change 9 of 8.13.
211 7. While fixing 6 above, I noticed that a number of other items were being
212 incorrectly rejected as "not fixed length". This arose partly because newer
213 opcodes had not been added to the fixed-length checking code. I have (a)
214 corrected the bug and added tests for these items, and (b) arranged for an
215 error to occur if an unknown opcode is encountered while checking for fixed
216 length instead of just assuming "not fixed length". The items that were
217 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
218 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
219 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
221 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
222 being incorrectly compiled and would have given unpredicatble results.
224 9. A possessively repeated subpattern with minimum repeat count greater than
225 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
226 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
227 the first (A) could occur when it should not.
229 10. Add a cast and remove a redundant test from the code.
231 11. JIT should use pcre_malloc/pcre_free for allocation.
233 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
234 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
235 is anything other than /usr, -L is still shown).
237 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
239 14. Perl does not support \N without a following name in a [] class; PCRE now
240 also gives an error.
242 15. If a forward reference was repeated with an upper limit of around 2000,
243 it caused the error "internal error: overran compiling workspace". The
244 maximum number of forward references (including repeats) was limited by the
245 internal workspace, and dependent on the LINK_SIZE. The code has been
246 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
247 the default depends on LINK_SIZE. There is a new upper limit (for safety)
248 of around 200,000 forward references. While doing this, I also speeded up
249 the filling in of repeated forward references.
251 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
252 incorrectly expecting the subject to contain another "a" after the start.
254 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
255 in the match, the SKIP should be ignored. This was not happening; instead
256 the SKIP was being treated as NOMATCH. For patterns such as
257 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
258 tested.
260 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
261 now much more compatible with Perl, in particular in cases where the result
262 is a non-match for a non-anchored pattern. For example, if
263 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
264 "m", where previously it did not return a name. A side effect of this
265 change is that for partial matches, the last encountered mark name is
266 returned, as for non matches. A number of tests that were previously not
267 Perl-compatible have been moved into the Perl-compatible test files. The
268 refactoring has had the pleasing side effect of removing one argument from
269 the match() function, thus reducing its stack requirements.
271 19. If the /S+ option was used in pcretest to study a pattern using JIT,
272 subsequent uses of /S (without +) incorrectly behaved like /S+.
274 21. Retrieve executable code size support for the JIT compiler and fixing
275 some warnings.
277 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
278 not work when the shorter character appeared right at the end of the
279 subject string.
281 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
282 systems.
284 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
285 output it when the /M option is used in pcretest.
287 25. The CheckMan script was not being included in the distribution. Also, added
288 an explicit "perl" to run Perl scripts from the PrepareRelease script
289 because this is reportedly needed in Windows.
291 26. If study data was being save in a file and studying had not found a set of
292 "starts with" bytes for the pattern, the data written to the file (though
293 never used) was taken from uninitialized memory and so caused valgrind to
294 complain.
296 27. Updated RunTest.bat as provided by Sheri Pierce.
298 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
300 29. Computation of memory usage for the table of capturing group names was
301 giving an unnecessarily large value.
304 Version 8.20 21-Oct-2011
305 ------------------------
307 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
308 a POSIX class. After further experiments with Perl, which convinced me that
309 Perl has bugs and confusions, a closing square bracket is no longer allowed
310 in a POSIX name. This bug also affected patterns with classes that started
311 with full stops.
313 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
314 captured substring, but while checking the failing first alternative,
315 substring 1 is temporarily captured. If the output vector supplied to
316 pcre_exec() was not big enough for this capture, the yield of the function
317 was still zero ("insufficient space for captured substrings"). This cannot
318 be totally fixed without adding another stack variable, which seems a lot
319 of expense for a edge case. However, I have improved the situation in cases
320 such as /(a)(b)x|abc/ matched against "abc", where the return code
321 indicates that fewer than the maximum number of slots in the ovector have
322 been set.
324 3. Related to (2) above: when there are more back references in a pattern than
325 slots in the output vector, pcre_exec() uses temporary memory during
326 matching, and copies in the captures as far as possible afterwards. It was
327 using the entire output vector, but this conflicts with the specification
328 that only 2/3 is used for passing back captured substrings. Now it uses
329 only the first 2/3, for compatibility. This is, of course, another edge
330 case.
332 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
333 main code base, and can be used by building with --enable-jit. When this is
334 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
335 runtime --no-jit option is given.
337 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
338 ovector, the return from the function was zero, implying that there were
339 other matches that did not fit. The correct "exactly full" value is now
340 returned.
342 6. If a subpattern that was called recursively or as a subroutine contained
343 (*PRUNE) or any other control that caused it to give a non-standard return,
344 invalid errors such as "Error -26 (nested recursion at the same subject
345 position)" or even infinite loops could occur.
347 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
348 computing the minimum length on reaching *ACCEPT, and so ended up with the
349 wrong value of 1 rather than 0. Further investigation indicates that
350 computing a minimum subject length in the presence of *ACCEPT is difficult
351 (think back references, subroutine calls), and so I have changed the code
352 so that no minimum is registered for a pattern that contains *ACCEPT.
354 8. If (*THEN) was present in the first (true) branch of a conditional group,
355 it was not handled as intended. [But see 16 below.]
357 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
358 Sheri Pierce.
360 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
361 the first byte in a match must be "a".
363 11. Change 17 for 8.13 increased the recursion depth for patterns like
364 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
365 pattern contains any instances of (*THEN). If it does not, the old
366 optimizations are restored. It would be nice to do this on a per-group
367 basis, but at the moment that is not feasible.
369 12. In some environments, the output of pcretest -C is CRLF terminated. This
370 broke RunTest's code that checks for the link size. A single white space
371 character after the value is now allowed for.
373 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
374 For "fr", it uses the Windows-specific input and output files.
376 14. If (*THEN) appeared in a group that was called recursively or as a
377 subroutine, it did not work as intended. [But see next item.]
379 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
380 pattern fragments (but not containing any | characters). If A and B are
381 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
382 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
383 D. In other words, Perl considers parentheses that do not contain any |
384 characters to be part of a surrounding alternative, whereas PCRE was
385 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
386 differently. PCRE now behaves in the same way as Perl, except in the case
387 of subroutine/recursion calls such as (?1) which have in any case always
388 been different (but PCRE had them first :-).
390 16. Related to 15 above: Perl does not treat the | in a conditional group as
391 creating alternatives. Such a group is treated in the same way as an
392 ordinary group without any | characters when processing (*THEN). PCRE has
393 been changed to match Perl's behaviour.
395 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
396 RunGrepTest script failed.
398 18. Change 22 for version 13 caused atomic groups to use more stack. This is
399 inevitable for groups that contain captures, but it can lead to a lot of
400 stack use in large patterns. The old behaviour has been restored for atomic
401 groups that do not contain any capturing parentheses.
403 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
404 suppress the check for a minimum subject length at run time. (If it was
405 given to pcre_exec() or pcre_dfa_exec() it did work.)
407 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
408 fail to work when decoding hex characters in data strings in EBCDIC
409 environments.
411 21. It appears that in at least one Mac OS environment, the isxdigit() function
412 is implemented as a macro that evaluates to its argument more than once,
413 contravening the C 90 Standard (I haven't checked a later standard). There
414 was an instance in pcretest which caused it to go wrong when processing
415 \x{...} escapes in subject strings. The has been rewritten to avoid using
416 things like p++ in the argument of isxdigit().
419 Version 8.13 16-Aug-2011
420 ------------------------
422 1. The Unicode data tables have been updated to Unicode 6.0.0.
424 2. Two minor typos in pcre_internal.h have been fixed.
426 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
427 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
428 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
430 4. There were a number of related bugs in the code for matching backrefences
431 caselessly in UTF-8 mode when codes for the characters concerned were
432 different numbers of bytes. For example, U+023A and U+2C65 are an upper
433 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
434 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
435 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
436 2-byte code at the end of the subject (it thought there wasn't enough data
437 left).
439 5. Comprehensive information about what went wrong is now returned by
440 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
441 as the output vector has at least 2 elements. The offset of the start of
442 the failing character and a reason code are placed in the vector.
444 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
445 now returned is for the first byte of the failing character, instead of the
446 last byte inspected. This is an incompatible change, but I hope it is small
447 enough not to be a problem. It makes the returned offset consistent with
448 pcre_exec() and pcre_dfa_exec().
450 7. pcretest now gives a text phrase as well as the error number when
451 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
452 failure, the offset and reason code are output.
454 8. When \R was used with a maximizing quantifier it failed to skip backwards
455 over a \r\n pair if the subsequent match failed. Instead, it just skipped
456 back over a single character (\n). This seems wrong (because it treated the
457 two characters as a single entity when going forwards), conflicts with the
458 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
459 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
460 has been changed.
462 9. Some internal refactoring has changed the processing so that the handling
463 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
464 time (the PCRE_DOTALL option was changed this way some time ago: version
465 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
466 which was always a bit of a fudge. It also means that there is one less
467 argument for the match() function, which reduces its stack requirements
468 slightly. This change also fixes an incompatibility with Perl: the pattern
469 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
471 10. More internal refactoring has drastically reduced the number of recursive
472 calls to match() for possessively repeated groups such as (abc)++ when
473 using pcre_exec().
475 11. While implementing 10, a number of bugs in the handling of groups were
476 discovered and fixed:
478 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
479 (a|)*(?1) gave a compile-time internal error.
480 ((a|)+)+ did not notice that the outer group could match an empty string.
481 (^a|^)+ was not marked as anchored.
482 (.*a|.*)+ was not marked as matching at start or after a newline.
484 12. Yet more internal refactoring has removed another argument from the match()
485 function. Special calls to this function are now indicated by setting a
486 value in a variable in the "match data" data block.
488 13. Be more explicit in pcre_study() instead of relying on "default" for
489 opcodes that mean there is no starting character; this means that when new
490 ones are added and accidentally left out of pcre_study(), testing should
491 pick them up.
493 14. The -s option of pcretest has been documented for ages as being an old
494 synonym of -m (show memory usage). I have changed it to mean "force study
495 for every regex", that is, assume /S for every regex. This is similar to -i
496 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
497 using it. It makes it easier to run collections of tests with and without
498 study enabled, and thereby test pcre_study() more easily. All the standard
499 tests are now run with and without -s (but some patterns can be marked as
500 "never study" - see 20 below).
502 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
503 restoration of the capturing data to the outer values was not happening
504 correctly.
506 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
507 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
508 pattern had matched an empty string, and so incorrectly returned a no
509 match.
511 17. There was optimizing code for the last branch of non-capturing parentheses,
512 and also for the obeyed branch of a conditional subexpression, which used
513 tail recursion to cut down on stack usage. Unfortunately, now that there is
514 the possibility of (*THEN) occurring in these branches, tail recursion is
515 no longer possible because the return has to be checked for (*THEN). These
516 two optimizations have therefore been removed. [But see 8.20/11 above.]
518 18. If a pattern containing \R was studied, it was assumed that \R always
519 matched two bytes, thus causing the minimum subject length to be
520 incorrectly computed because \R can also match just one byte.
522 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
523 was incorrectly computed.
525 20. If /S is present twice on a test pattern in pcretest input, it now
526 *disables* studying, thereby overriding the use of -s on the command line
527 (see 14 above). This is necessary for one or two tests to keep the output
528 identical in both cases.
530 21. When (*ACCEPT) was used in an assertion that matched an empty string and
531 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
533 22. When an atomic group that contained a capturing parenthesis was
534 successfully matched, but the branch in which it appeared failed, the
535 capturing was not being forgotten if a higher numbered group was later
536 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
537 group 1 to "a", when in fact it should be unset. This applied to multi-
538 branched capturing and non-capturing groups, repeated or not, and also to
539 positive assertions (capturing in negative assertions does not happen
540 in PCRE) and also to nested atomic groups.
542 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
543 subject after a captured substring, to make it easier to tell which of a
544 number of identical substrings has been captured.
546 24. The way atomic groups are processed by pcre_exec() has been changed so that
547 if they are repeated, backtracking one repetition now resets captured
548 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
549 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
550 "aaa". Previously, it would have been "a". As part of this code
551 refactoring, the way recursive calls are handled has also been changed.
553 25. If an assertion condition captured any substrings, they were not passed
554 back unless some other capturing happened later. For example, if
555 (?(?=(a))a) was matched against "a", no capturing was returned.
557 26. When studying a pattern that contained subroutine calls or assertions,
558 the code for finding the minimum length of a possible match was handling
559 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
560 group 1 called group 2 while simultaneously a separate group 2 called group
561 1). A stack overflow occurred in this case. I have fixed this by limiting
562 the recursion depth to 10.
564 27. Updated RunTest.bat in the distribution to the version supplied by Tom
565 Fortmann. This supports explicit test numbers on the command line, and has
566 argument validation and error reporting.
568 28. An instance of \X with an unlimited repeat could fail if at any point the
569 first character it looked at was a mark character.
571 29. Some minor code refactoring concerning Unicode properties and scripts
572 should reduce the stack requirement of match() slightly.
574 30. Added the '=' option to pcretest to check the setting of unused capturing
575 slots at the end of the pattern, which are documented as being -1, but are
576 not included in the return count.
578 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
579 compiled something random. Now it gives a compile-time error (as does
580 Perl).
582 32. A *MARK encountered during the processing of a positive assertion is now
583 recorded and passed back (compatible with Perl).
585 33. If --only-matching or --colour was set on a pcregrep call whose pattern
586 had alternative anchored branches, the search for a second match in a line
587 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
588 matched the line "0102" twice. The same bug affected patterns that started
589 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
590 twice.
592 34. Previously, PCRE did not allow quantification of assertions. However, Perl
593 does, and because of capturing effects, quantifying parenthesized
594 assertions may at times be useful. Quantifiers are now allowed for
595 parenthesized assertions.
597 35. A minor code tidy in pcre_compile() when checking options for \R usage.
599 36. \g was being checked for fancy things in a character class, when it should
600 just be a literal "g".
602 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
603 appearance of a nested POSIX class supersedes an apparent external class.
604 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
605 unescaped square brackets may also appear as part of class names. For
606 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
607 more like Perl. (But see 8.20/1 above.)
609 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
610 was because it thought it was \N{name}, which is not supported).
612 39. Add minix to OS list not supporting the -S option in pcretest.
614 40. PCRE tries to detect cases of infinite recursion at compile time, but it
615 cannot analyze patterns in sufficient detail to catch mutual recursions
616 such as ((?1))((?2)). There is now a runtime test that gives an error if a
617 subgroup is called recursively as a subpattern for a second time at the
618 same position in the subject string. In previous releases this might have
619 been caught by the recursion limit, or it might have run out of stack.
621 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
622 happen only once. PCRE was, however incorrectly giving a compile time error
623 "recursive call could loop indefinitely" because it cannot analyze the
624 pattern in sufficient detail. The compile time test no longer happens when
625 PCRE is compiling a conditional subpattern, but actual runaway loops are
626 now caught at runtime (see 40 above).
628 42. It seems that Perl allows any characters other than a closing parenthesis
629 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
630 has been changed to be the same.
632 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
633 as not to get warnings when autogen.sh is called. Also changed
634 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
636 44. To help people who use pcregrep to scan files containing exceedingly long
637 lines, the following changes have been made:
639 (a) The default value of the buffer size parameter has been increased from
640 8K to 20K. (The actual buffer used is three times this size.)
642 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
643 PCRE is built.
645 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
646 to be set at run time.
648 (d) Numerical values in pcregrep options can be followed by K or M, for
649 example --buffer-size=50K.
651 (e) If a line being scanned overflows pcregrep's buffer, an error is now
652 given and the return code is set to 2.
654 45. Add a pointer to the latest mark to the callout data block.
656 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
657 partial match of an empty string instead of no match. This was specific to
658 the use of ".".
660 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
661 complete match instead of a partial match. This bug was dependent on both
662 the PCRE_UTF8 and PCRE_DOTALL options being set.
664 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
665 starting byte set, because \b was not being ignored.
668 Version 8.12 15-Jan-2011
669 ------------------------
671 1. Fixed some typos in the markup of the man pages, and wrote a script that
672 checks for such things as part of the documentation building process.
674 2. On a big-endian 64-bit system, pcregrep did not correctly process the
675 --match-limit and --recursion-limit options (added for 8.11). In
676 particular, this made one of the standard tests fail. (The integer value
677 went into the wrong half of a long int.)
679 3. If the --colour option was given to pcregrep with -v (invert match), it
680 did strange things, either producing crazy output, or crashing. It should,
681 of course, ignore a request for colour when reporting lines that do not
682 match.
684 4. Another pcregrep bug caused similar problems if --colour was specified with
685 -M (multiline) and the pattern match finished with a line ending.
687 5. In pcregrep, when a pattern that ended with a literal newline sequence was
688 matched in multiline mode, the following line was shown as part of the
689 match. This seems wrong, so I have changed it.
691 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
692 the check for further matches in the same line (so they could be coloured)
693 to overrun the end of the current line. If another match was found, it was
694 incorrectly shown (and then shown again when found in the next line).
696 7. If pcregrep was compiled under Windows, there was a reference to the
697 function pcregrep_exit() before it was defined. I am assuming this was
698 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
699 reported by a user. I've moved the definition above the reference.
702 Version 8.11 10-Dec-2010
703 ------------------------
705 1. (*THEN) was not working properly if there were untried alternatives prior
706 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
707 backtracked to try for "b" instead of moving to the next alternative branch
708 at the same level (in this case, to look for "c"). The Perl documentation
709 is clear that when (*THEN) is backtracked onto, it goes to the "next
710 alternative in the innermost enclosing group".
712 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
713 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
714 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
715 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
716 (*THEN).
718 3. If \s appeared in a character class, it removed the VT character from
719 the class, even if it had been included by some previous item, for example
720 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
721 of \s, but is part of the POSIX "space" class.)
723 4. A partial match never returns an empty string (because you can always
724 match an empty string at the end of the subject); however the checking for
725 an empty string was starting at the "start of match" point. This has been
726 changed to the "earliest inspected character" point, because the returned
727 data for a partial match starts at this character. This means that, for
728 example, /(?<=abc)def/ gives a partial match for the subject "abc"
729 (previously it gave "no match").
731 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
732 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
733 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
734 has an implication that the given string is incomplete (because a partial
735 match is preferred over a full match). For this reason, these items now
736 give a partial match in this situation. [Aside: previously, the one case
737 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
738 match rather than a full match, which was wrong by the old rules, but is
739 now correct.]
741 6. There was a bug in the handling of #-introduced comments, recognized when
742 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
743 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
744 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
745 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
746 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
747 places in pcre_compile().
749 7. Related to (6) above, when pcre_compile() was skipping #-introduced
750 comments when looking ahead for named forward references to subpatterns,
751 the only newline sequence it recognized was NL. It now handles newlines
752 according to the set newline convention.
754 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
755 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
756 cater for a lack of strerror(). These oversights have been fixed.
758 9. Added --match-limit and --recursion-limit to pcregrep.
760 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
762 11. When the -o option was used, pcregrep was setting a return code of 1, even
763 when matches were found, and --line-buffered was not being honoured.
765 12. Added an optional parentheses number to the -o and --only-matching options
766 of pcregrep.
768 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
769 can match an empty string. The code to do it in pcretest and pcredemo
770 needed fixing:
772 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
773 only one byte after an empty string match just before CRLF (this case
774 just got forgotten; "any" and "anycrlf" were OK).
776 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
777 mode when an empty string match preceded an ASCII character followed by
778 a non-ASCII character. (The code for advancing by one character rather
779 than one byte was nonsense.)
781 (c) The pcredemo.c sample program did not have any code at all to handle
782 the cases when CRLF is a valid newline sequence.
784 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
785 as a starting offset was within the subject string. There is now a new
786 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
787 negative or greater than the length of the string. In order to test this,
788 pcretest is extended to allow the setting of negative starting offsets.
790 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
791 starting offset points to the beginning of a UTF-8 character was
792 unnecessarily clumsy. I tidied it up.
794 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
795 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
797 17. Nobody had reported that the --include_dir option, which was added in
798 release 7.7 should have been called --include-dir (hyphen, not underscore)
799 for compatibility with GNU grep. I have changed it to --include-dir, but
800 left --include_dir as an undocumented synonym, and the same for
801 --exclude-dir, though that is not available in GNU grep, at least as of
802 release 2.5.4.
804 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
805 characters from a string of bytes) have been redefined so as not to use
806 loops, in order to improve performance in some environments. At the same
807 time, I abstracted some of the common code into auxiliary macros to save
808 repetition (this should not affect the compiled code).
810 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
811 compile-time error is now given if \c is not followed by an ASCII
812 character, that is, a byte less than 128. (In EBCDIC mode, the code is
813 different, and any byte value is allowed.)
815 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
816 START_OPTIMIZE option, which is now allowed at compile time - but just
817 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
818 to pcregrep and other applications that have no direct access to PCRE
819 options. The new /Y option in pcretest sets this option when calling
820 pcre_compile().
822 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
823 back references. Groups containing recursive back references were forced to
824 be atomic by that change, but in the case of named groups, the amount of
825 memory required was incorrectly computed, leading to "Failed: internal
826 error: code overflow". This has been fixed.
828 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
829 pcretest.c, to avoid build problems in some Borland environments.
832 Version 8.10 25-Jun-2010
833 ------------------------
835 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
836 THEN.
838 2. (*ACCEPT) was not working when inside an atomic group.
840 3. Inside a character class, \B is treated as a literal by default, but
841 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
842 causes the error). The code is unchanged, but I tidied the documentation.
844 4. Inside a character class, PCRE always treated \R and \X as literals,
845 whereas Perl faults them if its -w option is set. I have changed PCRE so
846 that it faults them when PCRE_EXTRA is set.
848 5. Added support for \N, which always matches any character other than
849 newline. (It is the same as "." when PCRE_DOTALL is not set.)
851 6. When compiling pcregrep with newer versions of gcc which may have
852 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
853 declared with attribute warn_unused_result" were given. Just casting the
854 result to (void) does not stop the warnings; a more elaborate fudge is
855 needed. I've used a macro to implement this.
857 7. Minor change to pcretest.c to avoid a compiler warning.
859 8. Added four artifical Unicode properties to help with an option to make
860 \s etc use properties (see next item). The new properties are: Xan
861 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
863 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
864 use Unicode properties. (*UCP) at the start of a pattern can be used to set
865 this option. Modified pcretest to add /W to test this facility. Added
866 REG_UCP to make it available via the POSIX interface.
868 10. Added --line-buffered to pcregrep.
870 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
871 studied, and the match started with a letter with a code point greater than
872 127 whose first byte was different to the first byte of the other case of
873 the letter, the other case of this starting letter was not recognized
874 (#976).
876 12. If a pattern that was studied started with a repeated Unicode property
877 test, for example, \p{Nd}+, there was the theoretical possibility of
878 setting up an incorrect bitmap of starting bytes, but fortunately it could
879 not have actually happened in practice until change 8 above was made (it
880 added property types that matched character-matching opcodes).
882 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
883 possible starting bytes for non-anchored patterns.
885 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
886 \R, and also a number of cases that involve Unicode properties, both
887 explicit and implicit when PCRE_UCP is set.
889 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
890 input, it could crash or give wrong results if characters with values
891 greater than 0xc0 were present in the subject string. (Detail: it assumed
892 UTF-8 input when processing these items.)
894 16. Added a lot of (int) casts to avoid compiler warnings in systems where
895 size_t is 64-bit (#991).
897 17. Added a check for running out of memory when PCRE is compiled with
898 --disable-stack-for-recursion (#990).
900 18. If the last data line in a file for pcretest does not have a newline on
901 the end, a newline was missing in the output.
903 19. The default pcre_chartables.c file recognizes only ASCII characters (values
904 less than 128) in its various bitmaps. However, there is a facility for
905 generating tables according to the current locale when PCRE is compiled. It
906 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
907 characters, are recognized by isspace() and therefore were getting set in
908 these tables, and indeed these tables seem to approximate to ISO 8859. This
909 caused a problem in UTF-8 mode when pcre_study() was used to create a list
910 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
911 which of course cannot start UTF-8 characters. I have changed the code so
912 that only real ASCII characters (less than 128) and the correct starting
913 bytes for UTF-8 encodings are set for characters greater than 127 when in
914 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
915 altogether.)
917 20. Added the /T option to pcretest so as to be able to run tests with non-
918 standard character tables, thus making it possible to include the tests
919 used for 19 above in the standard set of tests.
921 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
922 reference to a subpattern the other side of a comment that contains an
923 opening parenthesis caused either an internal compiling error, or a
924 reference to the wrong subpattern.
927 Version 8.02 19-Mar-2010
928 ------------------------
930 1. The Unicode data tables have been updated to Unicode 5.2.0.
932 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
933 configured.
935 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
936 original author of that file, following a query about its status.
938 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
939 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
941 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
942 quantifier applied to a forward-referencing subroutine call, could compile
943 incorrect code or give the error "internal error: previously-checked
944 referenced subpattern not found".
946 6. Both MS Visual Studio and Symbian OS have problems with initializing
947 variables to point to external functions. For these systems, therefore,
948 pcre_malloc etc. are now initialized to local functions that call the
949 relevant global functions.
951 7. There were two entries missing in the vectors called coptable and poptable
952 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
953 I've fixed the data, and added a kludgy way of testing at compile time that
954 the lengths are correct (equal to the number of opcodes).
956 8. Following on from 7, I added a similar kludge to check the length of the
957 eint vector in pcreposix.c.
959 9. Error texts for pcre_compile() are held as one long string to avoid too
960 much relocation at load time. To find a text, the string is searched,
961 counting zeros. There was no check for running off the end of the string,
962 which could happen if a new error number was added without updating the
963 string.
965 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
967 11. \K was not working if it appeared in an atomic group or in a group that
968 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
969 \K is "not well defined" if used in an assertion. PCRE now accepts it if
970 the assertion is positive, but not if it is negative.
972 12. Change 11 fortuitously reduced the size of the stack frame used in the
973 "match()" function of pcre_exec.c by one pointer. Forthcoming
974 implementation of support for (*MARK) will need an extra pointer on the
975 stack; I have reserved it now, so that the stack frame size does not
976 decrease.
978 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
979 item in branch that calls a recursion is a subroutine call - as in the
980 second branch in the above example - was incorrectly given the compile-
981 time error "recursive call could loop indefinitely" because pcre_compile()
982 was not correctly checking the subroutine for matching a non-empty string.
984 14. The checks for overrunning compiling workspace could trigger after an
985 overrun had occurred. This is a "should never occur" error, but it can be
986 triggered by pathological patterns such as hundreds of nested parentheses.
987 The checks now trigger 100 bytes before the end of the workspace.
989 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
992 Version 8.01 19-Jan-2010
993 ------------------------
995 1. If a pattern contained a conditional subpattern with only one branch (in
996 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
997 computed the wrong minimum data length (which is of course zero for such
998 subpatterns). This could cause incorrect "no match" results.
1000 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1001 the pattern is reset in the first branch, pcre_compile() failed with
1002 "internal error: code overflow at offset...". This happened only when
1003 the reset was to the original external option setting. (An optimization
1004 abstracts leading options settings into an external setting, which was the
1005 cause of this.)
1007 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1008 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1009 assertion pattern did not match (meaning that the assertion was true), it
1010 was incorrectly treated as false if the SKIP had been reached during the
1011 matching. This also applied to assertions used as conditions.
1013 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1014 assertion subpattern, including such a pattern used as a condition,
1015 unpredictable results occurred, instead of the error return
1018 5. The C++ GlobalReplace function was not working like Perl for the special
1019 situation when an empty string is matched. It now does the fancy magic
1020 stuff that is necessary.
1022 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1023 removed. (These were left over from very, very early versions of PCRE.)
1025 7. Some cosmetic changes to the code to make life easier when compiling it
1026 as part of something else:
1028 (a) Change DEBUG to PCRE_DEBUG.
1030 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1031 called "current" as "current_branch", to prevent a collision with the
1032 Linux macro when compiled as a kernel module.
1034 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1035 prevent a collision with the Linux macro when compiled as a kernel
1036 module.
1038 8. In pcre_compile() there are some checks for integer overflows that used to
1039 cast potentially large values to (double). This has been changed to that
1040 when building, a check for int64_t is made, and if it is found, it is used
1041 instead, thus avoiding the use of floating point arithmetic. (There is no
1042 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1043 double.
1045 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1046 2005 (difference between two addresses compared to an unsigned value).
1048 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1049 custom one, because of the following reported problem in Windows:
1051 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1052 under Win32.
1053 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1054 therefore missing the function definition.
1055 - The compiler thus generates a "C" signature for the test function.
1056 - The linker fails to find the "C" function.
1057 - PCRE fails to configure if asked to do so against libbz2.
1059 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1060 messages were output:
1062 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1063 rerunning libtoolize, to keep the correct libtool macros in-tree.
1064 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1066 I have done both of these things.
1068 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1069 most of the time, it *can* run out if it is given a pattern that contains a
1070 runaway infinite recursion. I updated the discussion in the pcrestack man
1071 page.
1073 13. Now that we have gone to the x.xx style of version numbers, the minor
1074 version may start with zero. Using 08 or 09 is a bad idea because users
1075 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1076 interpreted as invalid octal numbers. I've updated the previous comment in
1077 configure.ac, and also added a check that gives an error if 08 or 09 are
1078 used.
1080 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1081 causing partial matching to fail when the end of the subject matched \W
1082 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1084 15. There were some discrepancies between the declarations in pcre_internal.h
1085 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1086 their definitions. The declarations used "const uschar *" and the
1087 definitions used USPTR. Even though USPTR is normally defined as "const
1088 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1089 reported that: "This difference in casting confuses some C++ compilers, for
1090 example, SunCC recognizes above declarations as different functions and
1091 generates broken code for hbpcre." I have changed the declarations to use
1092 USPTR.
1094 16. GNU libtool is named differently on some systems. The autogen.sh script now
1095 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1096 (FreeBSD).
1098 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1099 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1100 comment: "Figure out how to create a longlong from a string: strtoll and
1101 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1102 instance, but it only takes 2 args instead of 3!"
1104 18. A subtle bug concerned with back references has been fixed by a change of
1105 specification, with a corresponding code fix. A pattern such as
1106 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1107 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1108 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1109 same bug. Such groups have to be quantified to be useful, or contained
1110 inside another quantified group. (If there's no repetition, the reference
1111 can never match.) The problem arises because, having left the group and
1112 moved on to the rest of the pattern, a later failure that backtracks into
1113 the group uses the captured value from the final iteration of the group
1114 rather than the correct earlier one. I have fixed this in PCRE by forcing
1115 any group that contains a reference to itself to be an atomic group; that
1116 is, there cannot be any backtracking into it once it has completed. This is
1117 similar to recursive and subroutine calls.
1120 Version 8.00 19-Oct-09
1121 ----------------------
1123 1. The table for translating pcre_compile() error codes into POSIX error codes
1124 was out-of-date, and there was no check on the pcre_compile() error code
1125 being within the table. This could lead to an OK return being given in
1126 error.
1128 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1129 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1130 in a Windows environment.
1132 3. The pcregrep --count option prints the count for each file even when it is
1133 zero, as does GNU grep. However, pcregrep was also printing all files when
1134 --files-with-matches was added. Now, when both options are given, it prints
1135 counts only for those files that have at least one match. (GNU grep just
1136 prints the file name in this circumstance, but including the count seems
1137 more useful - otherwise, why use --count?) Also ensured that the
1138 combination -clh just lists non-zero counts, with no names.
1140 4. The long form of the pcregrep -F option was incorrectly implemented as
1141 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1142 but it seems right to fix it, and I didn't think it was worth preserving
1143 the old behaviour.
1145 5. The command line items --regex=pattern and --regexp=pattern were not
1146 recognized by pcregrep, which required --regex pattern or --regexp pattern
1147 (with a space rather than an '='). The man page documented the '=' forms,
1148 which are compatible with GNU grep; these now work.
1150 6. No libpcreposix.pc file was created for pkg-config; there was just
1151 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1153 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1154 when UCP support is not needed, by modifying the Python script that
1155 generates it from Unicode data files. This should not matter if the module
1156 is correctly used as a library, but I received one complaint about 50K of
1157 unwanted data. My guess is that the person linked everything into his
1158 program rather than using a library. Anyway, it does no harm.
1160 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1161 was a minimum greater than 1 for a wide character in a possessive
1162 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1163 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1164 character. Chaos in the form of incorrect output or a compiling loop could
1165 result.
1167 9. The restrictions on what a pattern can contain when partial matching is
1168 requested for pcre_exec() have been removed. All patterns can now be
1169 partially matched by this function. In addition, if there are at least two
1170 slots in the offset vector, the offset of the earliest inspected character
1171 for the match and the offset of the end of the subject are set in them when
1172 PCRE_ERROR_PARTIAL is returned.
1174 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1175 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1176 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1177 and may be more useful for multi-segment matching.
1179 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1180 used to be given if ever the end of the subject was reached; now it is
1181 given only if matching could not proceed because another character was
1182 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1183 string "Z", which now yields "no match" instead of "partial match". In the
1184 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1185 final character ended with (*FAIL).
1187 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1188 if the pattern had a "must contain" character that was already found in the
1189 earlier partial match, unless partial matching was again requested. For
1190 example, with the pattern /dog.(body)?/, the "must contain" character is
1191 "g". If the first part-match was for the string "dog", restarting with
1192 "sbody" failed. This bug has been fixed.
1194 13. The string returned by pcre_dfa_exec() after a partial match has been
1195 changed so that it starts at the first inspected character rather than the
1196 first character of the match. This makes a difference only if the pattern
1197 starts with a lookbehind assertion or \b or \B (\K is not supported by
1198 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1199 matching functions compatible, and I think it's the right thing to do.
1201 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1202 so that the demonstration program is easily available in environments where
1203 PCRE has not been installed from source.
1205 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1206 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1207 library.
1209 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1210 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1211 is not the first non-POSIX option to be added. Clearly some people find
1212 these options useful.
1214 17. If a caller to the POSIX matching function regexec() passes a non-zero
1215 value for nmatch with a NULL value for pmatch, the value of
1216 nmatch is forced to zero.
1218 18. RunGrepTest did not have a test for the availability of the -u option of
1219 the diff command, as RunTest does. It now checks in the same way as
1220 RunTest, and also checks for the -b option.
1222 19. If an odd number of negated classes containing just a single character
1223 interposed, within parentheses, between a forward reference to a named
1224 subpattern and the definition of the subpattern, compilation crashed with
1225 an internal error, complaining that it could not find the referenced
1226 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1227 [The bug was that it was starting one character too far in when skipping
1228 over the character class, thus treating the ] as data rather than
1229 terminating the class. This meant it could skip too much.]
1231 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1232 /g option in pcretest when the pattern contains \K, which makes it possible
1233 to have an empty string match not at the start, even when the pattern is
1234 anchored. Updated pcretest and pcredemo to use this option.
1236 21. If the maximum number of capturing subpatterns in a recursion was greater
1237 than the maximum at the outer level, the higher number was returned, but
1238 with unset values at the outer level. The correct (outer level) value is
1239 now given.
1241 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1242 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1243 make it do so. The string so far is captured, making this feature
1244 compatible with Perl.
1246 23. The tests have been re-organized, adding tests 11 and 12, to make it
1247 possible to check the Perl 5.10 features against Perl 5.10.
1249 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1250 pattern matches a fixed length string. PCRE did not allow this; now it
1251 does. Neither allows recursion.
1253 25. I finally figured out how to implement a request to provide the minimum
1254 length of subject string that was needed in order to match a given pattern.
1255 (It was back references and recursion that I had previously got hung up
1256 on.) This code has now been added to pcre_study(); it finds a lower bound
1257 to the length of subject needed. It is not necessarily the greatest lower
1258 bound, but using it to avoid searching strings that are too short does give
1259 some useful speed-ups. The value is available to calling programs via
1260 pcre_fullinfo().
1262 26. While implementing 25, I discovered to my embarrassment that pcretest had
1263 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1264 study optimizations had never been tested with that matching function.
1265 Oops. What is worse, even when it was passed study data, there was a bug in
1266 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1267 were also very few tests of studied patterns with pcre_dfa_exec().
1269 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1270 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1271 on the other side of the coin, they are no longer allowed to have different
1272 names, because these cannot be distinguished in PCRE, and this has caused
1273 confusion. (This is a difference from Perl.)
1275 28. When duplicate subpattern names are present (necessarily with different
1276 numbers, as required by 27 above), and a test is made by name in a
1277 conditional pattern, either for a subpattern having been matched, or for
1278 recursion in such a pattern, all the associated numbered subpatterns are
1279 tested, and the overall condition is true if the condition is true for any
1280 one of them. This is the way Perl works, and is also more like the way
1281 testing by number works.
1284 Version 7.9 11-Apr-09
1285 ---------------------
1287 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1288 (pcretest), all targets were linked against these libraries. This included
1289 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1290 libraries. This caused unwanted dependencies to be created. This problem
1291 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1292 pcretest is linked with readline.
1294 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1295 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1296 moved outside it again, because FALSE and TRUE are already defined in AIX,
1297 but BOOL is not.
1299 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1300 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1302 4. The pcregrep documentation said spaces were inserted as well as colons (or
1303 hyphens) following file names and line numbers when outputting matching
1304 lines. This is not true; no spaces are inserted. I have also clarified the
1305 wording for the --colour (or --color) option.
1307 5. In pcregrep, when --colour was used with -o, the list of matching strings
1308 was not coloured; this is different to GNU grep, so I have changed it to be
1309 the same.
1311 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1312 each matching line was coloured. Now it goes on to look for further matches
1313 of any of the test patterns, which is the same behaviour as GNU grep.
1315 7. A pattern that could match an empty string could cause pcregrep to loop; it
1316 doesn't make sense to accept an empty string match in pcregrep, so I have
1317 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1318 seems to be how GNU grep behaves.
1320 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1321 start or after a newline", because the conditional assertion was not being
1322 correctly handled. The rule now is that both the assertion and what follows
1323 in the first alternative must satisfy the test.
1325 9. If auto-callout was enabled in a pattern with a conditional group whose
1326 condition was an assertion, PCRE could crash during matching, both with
1327 pcre_exec() and pcre_dfa_exec().
1329 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1330 used for matching.
1332 11. Unicode property support in character classes was not working for
1333 characters (bytes) greater than 127 when not in UTF-8 mode.
1335 12. Added the -M command line option to pcretest.
1337 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1339 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1341 16. Added comments and documentation about mis-use of no_arg in the C++
1342 wrapper.
1344 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1345 from Martin Jerabek that uses macro names for all relevant character and
1346 string constants.
1348 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1349 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1350 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1351 these, but not everybody uses configure.
1353 19. A conditional group that had only one branch was not being correctly
1354 recognized as an item that could match an empty string. This meant that an
1355 enclosing group might also not be so recognized, causing infinite looping
1356 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1357 with the subject "ab", where knowledge that the repeated group can match
1358 nothing is needed in order to break the loop.
1360 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1361 exec(), but without supplying a callout function, matching went wrong.
1363 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1364 leak if the size of the offset vector was greater than 30. When the vector
1365 is smaller, the saved offsets during recursion go onto a local stack
1366 vector, but for larger vectors malloc() is used. It was failing to free
1367 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1368 error, in fact).
1370 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1371 heapframe that is used only when UTF-8 support is enabled. This caused no
1372 problem, but was untidy.
1374 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1375 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1376 included within another project.
1378 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1379 slightly modified by me:
1381 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1382 not building pcregrep.
1384 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1385 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1387 25. Forward references, both numeric and by name, in patterns that made use of
1388 duplicate group numbers, could behave incorrectly or give incorrect errors,
1389 because when scanning forward to find the reference group, PCRE was not
1390 taking into account the duplicate group numbers. A pattern such as
1391 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1393 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1394 the feature of a custom pointer more persuasive (as requested by a user).
1396 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1397 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1398 pre-defined.
1400 28. Added support for (*UTF8) at the start of a pattern.
1402 29. Arrange for flags added by the "release type" setting in CMake to be shown
1403 in the configuration summary.
1406 Version 7.8 05-Sep-08
1407 ---------------------
1409 1. Replaced UCP searching code with optimized version as implemented for Ad
1410 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1411 stage table and inline lookup instead of a function, giving speed ups of 2
1412 to 5 times on some simple patterns that I tested. Permission was given to
1413 distribute the MultiStage2.py script that generates the tables (it's not in
1414 the tarball, but is in the Subversion repository).
1416 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1417 scripts.
1419 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1420 a group with a zero qualifier. The result of the study could be incorrect,
1421 or the function might crash, depending on the pattern.
1423 4. Caseless matching was not working for non-ASCII characters in back
1424 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1425 It now works when Unicode Property Support is available.
1427 5. In pcretest, an escape such as \x{de} in the data was always generating
1428 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1429 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1430 truncation.
1432 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1434 7. Added two (int) casts to pcregrep when printing the difference of two
1435 pointers, in case they are 64-bit values.
1437 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1438 test 2 if it fails.
1440 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1441 and a #define of that name to empty if it is not externally set. This is to
1442 allow users of MSVC to set it if necessary.
1444 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1445 the convenience functions in the pcre_get.c source file.
1447 11. An option change at the start of a pattern that had top-level alternatives
1448 could cause overwriting and/or a crash. This command provoked a crash in
1449 some environments:
1451 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1453 This potential security problem was recorded as CVE-2008-2371.
1455 12. For a pattern where the match had to start at the beginning or immediately
1456 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1457 pcre_dfa_exec() could read past the end of the passed subject if there was
1458 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1459 pcretest so that it places the subject at the end of its malloc-ed buffer.
1461 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1462 exec() might read past the end of the data buffer in UTF-8 mode.
1464 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1465 the data contained the byte 0x85 as part of a UTF-8 character within its
1466 first line. This applied both to normal and DFA matching.
1468 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1469 /^[^d]*?$/8 failed to match "abc".
1471 16. Added a missing copyright notice to pcrecpp_internal.h.
1473 17. Make it more clear in the documentation that values returned from
1474 pcre_exec() in ovector are byte offsets, not character counts.
1476 18. Tidied a few places to stop certain compilers from issuing warnings.
1478 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1479 supplied by Stefan Weber. I made a further small update for 7.8 because
1480 there is a change of source arrangements: the pcre_searchfuncs.c module is
1481 replaced by pcre_ucd.c.
1484 Version 7.7 07-May-08
1485 ---------------------
1487 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1488 a string to a long long, pretend we don't even have a long long." This is
1489 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1491 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1492 pre-7.6 versions, which defined a global no_arg variable instead of putting
1493 it in the RE class. (See also #8 below.)
1495 3. Remove a line of dead code, identified by coverity and reported by Nuno
1496 Lopes.
1498 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1500 (1) The include/exclude patterns were being applied to the whole pathnames
1501 of files, instead of just to the final components.
1503 (2) If there was more than one level of directory, the subdirectories were
1504 skipped unless they satisfied the include/exclude conditions. This is
1505 inconsistent with GNU grep (and could even be seen as contrary to the
1506 pcregrep specification - which I improved to make it absolutely clear).
1507 The action now is always to scan all levels of directory, and just
1508 apply the include/exclude patterns to regular files.
1510 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1511 --exclude_dir in the tests to avoid scanning .svn directories.
1513 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1514 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1515 doesn't support NULs in patterns.
1517 7. Added some missing "const"s to declarations of static tables in
1518 pcre_compile.c and pcre_dfa_exec.c.
1520 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1521 caused by fix #2 above. (Subsequently also a second patch to fix the
1522 first patch. And a third patch - this was a messy problem.)
1524 9. Applied Craig's patch to remove the use of push_back().
1526 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1527 matching function regexec().
1529 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1530 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1531 references. PCRE supports relative numbers with this syntax (I don't think
1532 Oniguruma does).
1534 12. Previously, a group with a zero repeat such as (...){0} was completely
1535 omitted from the compiled regex. However, this means that if the group
1536 was called as a subroutine from elsewhere in the pattern, things went wrong
1537 (an internal error was given). Such groups are now left in the compiled
1538 pattern, with a new opcode that causes them to be skipped at execution
1539 time.
1541 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1542 to the way PCRE behaves:
1544 (a) A lone ] character is dis-allowed (Perl treats it as data).
1546 (b) A back reference to an unmatched subpattern matches an empty string
1547 (Perl fails the current match path).
1549 (c) A data ] in a character class must be notated as \] because if the
1550 first data character in a class is ], it defines an empty class. (In
1551 Perl it is not possible to have an empty class.) The empty class []
1552 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1553 The negative empty class [^] matches any one character, independently
1554 of the DOTALL setting.
1556 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1557 non-existent subpattern following a character class starting with ']' and
1558 containing () gave an internal compiling error instead of "reference to
1559 non-existent subpattern". Fortunately, when the pattern did exist, the
1560 compiled code was correct. (When scanning forwards to check for the
1561 existencd of the subpattern, it was treating the data ']' as terminating
1562 the class, so got the count wrong. When actually compiling, the reference
1563 was subsequently set up correctly.)
1565 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1566 it was being rejected as not supported by pcre_dfa_exec(), even though
1567 other assertions are supported. I have made pcre_dfa_exec() support
1568 (*FAIL).
1570 16. The implementation of 13c above involved the invention of a new opcode,
1571 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1572 cannot be changed at match time, I realized I could make a small
1573 improvement to matching performance by compiling OP_ALLANY instead of
1574 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1575 on the OP_ANY path.
1577 17. Compiling pcretest on Windows with readline support failed without the
1578 following two fixes: (1) Make the unistd.h include conditional on
1579 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1581 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1582 ncurses library to be included for pcretest when ReadLine support is
1583 requested, but also to allow for it to be overridden. This patch came from
1584 Daniel Bergström.
1586 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1587 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1588 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1589 spotting this.
1592 Version 7.6 28-Jan-08
1593 ---------------------
1595 1. A character class containing a very large number of characters with
1596 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1597 overflow.
1599 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1600 HAVE_LONG_LONG is not defined.
1602 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1603 bring them up to date and include new features. This patch includes:
1605 - Fixed PH's badly added libz and libbz2 support.
1606 - Fixed a problem with static linking.
1607 - Added pcredemo. [But later removed - see 7 below.]
1608 - Fixed dftables problem and added an option.
1609 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1611 - Added readline support for pcretest.
1612 - Added an listing of the option settings after cmake has run.
1614 4. A user submitted a patch to Makefile that makes it easy to create
1615 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1616 Makefile.am that cause it to include this special target, without
1617 affecting anything else. Note that the same mingw target plus all
1618 the other distribution libraries and programs are now supported
1619 when configuring with CMake (see 6 below) instead of with
1620 Configure/Make.
1622 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1623 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1624 exported in the Windows port". It has not yet been confirmed that the patch
1625 solves the problem, but it does no harm.
1627 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1628 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1629 with CMake, and also correct the comment about stack recursion.
1631 7. Remove the automatic building of pcredemo from the ./configure system and
1632 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1633 of a program that users should build themselves after PCRE is installed, so
1634 building it automatically is not really right. What is more, it gave
1635 trouble in some build environments.
1637 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1640 Version 7.5 10-Jan-08
1641 ---------------------
1643 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1644 values in parens when parsing an RE using the C++ wrapper."
1646 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1647 Characters greater than 255 were excluded from the class instead of being
1648 included.
1650 3. The same bug as (2) above applied to negated POSIX classes such as
1651 [:^space:].
1653 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1654 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1655 I have changed it.
1657 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1658 first named subpattern) and a construct such as (?&a) would reference the
1659 first named subpattern whose name started with "a" (in other words, the
1660 length check was missing). Both these problems are fixed. "Subpattern name
1661 expected" is now given for (?&) (a zero-length name), and this patch also
1662 makes it give the same error for \k'' (previously it complained that that
1663 was a reference to a non-existent subpattern).
1665 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1666 this is right because (?- can be followed by option settings as well as by
1667 digits. I have, however, made the messages clearer.
1669 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1670 than the number used in the conditional) now cause a compile-time error.
1671 This is actually not compatible with Perl, which accepts such patterns, but
1672 treats the conditional as always being FALSE (as PCRE used to), but it
1673 seems to me that giving a diagnostic is better.
1675 8. Change "alphameric" to the more common word "alphanumeric" in comments
1676 and messages.
1678 9. Fix two occurrences of "backslash" in comments that should have been
1679 "backspace".
1681 10. Remove two redundant lines of code that can never be obeyed (their function
1682 was moved elsewhere).
1684 11. The program that makes PCRE's Unicode character property table had a bug
1685 which caused it to generate incorrect table entries for sequences of
1686 characters that have the same character type, but are in different scripts.
1687 It amalgamated them into a single range, with the script of the first of
1688 them. In other words, some characters were in the wrong script. There were
1689 thirteen such cases, affecting characters in the following ranges:
1691 U+002b0 - U+002c1
1692 U+0060c - U+0060d
1693 U+0061e - U+00612
1694 U+0064b - U+0065e
1695 U+0074d - U+0076d
1696 U+01800 - U+01805
1697 U+01d00 - U+01d77
1698 U+01d9b - U+01dbf
1699 U+0200b - U+0200f
1700 U+030fc - U+030fe
1701 U+03260 - U+0327f
1702 U+0fb46 - U+0fbb1
1703 U+10450 - U+1049d
1705 12. The -o option (show only the matching part of a line) for pcregrep was not
1706 compatible with GNU grep in that, if there was more than one match in a
1707 line, it showed only the first of them. It now behaves in the same way as
1708 GNU grep.
1710 13. If the -o and -v options were combined for pcregrep, it printed a blank
1711 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1712 does the same. The return code can be used to tell if there were any
1713 non-matching lines.
1715 14. Added --file-offsets and --line-offsets to pcregrep.
1717 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1718 infinitely looping recursion. The bug was that positive lookaheads were not
1719 being skipped when checking for a possible empty match (negative lookaheads
1720 and both kinds of lookbehind were skipped).
1722 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1723 inclusion of <windows.h> to before rather than after the definition of
1724 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1726 17. Specifying a possessive quantifier with a specific limit for a Unicode
1727 character property caused pcre_compile() to compile bad code, which led at
1728 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1729 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1730 caused the error; without that there was no problem.
1732 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1734 19. Added --enable-pcretest-libreadline.
1736 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1737 RE::GlobalReplace(). As a result, the number of replacements returned was
1738 double what it should be. I removed one of the increments, but Craig sent a
1739 later patch that removed the other one (the right fix) and added unit tests
1740 that check the return values (which was not done before).
1742 21. Several CMake things:
1744 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1745 the names libpcre and libpcreposix, not just pcre and pcreposix.
1747 (2) The above change means that pcretest and pcregrep are now correctly
1748 linked with the newly-built libraries, not previously installed ones.
1752 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1753 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1754 UTF-8 newline character). The key issue is that the pattern starts .*;
1755 this means that the match must be either at the beginning, or after a
1756 newline. The bug was in the code for advancing after a failed match and
1757 checking that the new position followed a newline. It was not taking
1758 account of UTF-8 characters correctly.
1760 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1761 character classes. PCRE was not treating the sequence [:...:] as a
1762 character class unless the ... were all letters. Perl, however, seems to
1763 allow any characters between [: and :], though of course it rejects as
1764 unknown any "names" that contain non-letters, because all the known class
1765 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1766 for example, whereas PCRE did not - it did not recognize a POSIX character
1767 class. This seemed a bit dangerous, so the code has been changed to be
1768 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1769 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1770 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1771 Perl does, and where it didn't before.
1773 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1774 Windows environments %n is disabled by default.
1777 Version 7.4 21-Sep-07
1778 ---------------------
1780 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1781 means that a class such as [\s] counted as "explicit reference to CR or
1782 LF". That isn't really right - the whole point of the change was to try to
1783 help when there was an actual mention of one of the two characters. So now
1784 the change happens only if \r or \n (or a literal CR or LF) character is
1785 encountered.
1787 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1788 of both had grown to the point where there were only 3 bits left.
1789 Fortunately, there was spare space in the data structure, and so I have
1790 moved the internal flags into a new 16-bit field to free up more option
1791 bits.
1793 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1794 but did not set the internal JCHANGED flag - either of these is enough to
1795 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1796 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1797 start sets both bits.
1799 4. Added options (at build time, compile time, exec time) to change \R from
1800 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1802 5. doc/pcresyntax.html was missing from the distribution.
1804 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1805 compatibility, even though it is no longer used.
1807 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1808 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1809 windows.h file is present (where different names are used). [This was
1810 reversed later after testing - see 16 below.]
1812 8. Changed all #include <config.h> to #include "config.h". There were also
1813 some further <pcre.h> cases that I changed to "pcre.h".
1815 9. When pcregrep was used with the --colour option, it missed the line ending
1816 sequence off the lines that it output.
1818 10. It was pointed out to me that arrays of string pointers cause lots of
1819 relocations when a shared library is dynamically loaded. A technique of
1820 using a single long string with a table of offsets can drastically reduce
1821 these. I have refactored PCRE in four places to do this. The result is
1822 dramatic:
1824 Originally: 290
1825 After changing UCP table: 187
1826 After changing error message table: 43
1827 After changing table of "verbs" 36
1828 After changing table of Posix names 22
1830 Thanks to the folks working on Gregex for glib for this insight.
1832 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1833 unicode-properties was also set.
1835 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1837 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1838 checked only for CRLF.
1840 14. Added casts to pcretest.c to avoid compiler warnings.
1842 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1844 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1845 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1846 entirely. This removes changes made in 7 above.
1848 17. The CMake files have been updated, and there is now more information about
1849 building with CMake in the NON-UNIX-USE document.
1852 Version 7.3 28-Aug-07
1853 ---------------------
1855 1. In the rejigging of the build system that eventually resulted in 7.1, the
1856 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1857 brackets there is not right, since it causes compilers to look for an
1858 installed pcre.h, not the version that is in the source that is being
1859 compiled (which of course may be different). I have changed it back to:
1861 #include "pcre.h"
1863 I have a vague recollection that the change was concerned with compiling in
1864 different directories, but in the new build system, that is taken care of
1865 by the VPATH setting the Makefile.
1867 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1868 when the subject happened to end in the byte 0x85 (e.g. if the last
1869 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1870 characters but of course it shouldn't be taken as a newline when it is part
1871 of another character. The bug was that, for an unlimited repeat of . in
1872 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1873 characters when looking for a newline.
1875 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1877 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1878 in debug output.
1880 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1881 long printing in the pcrecpp unittest when running under MinGW.
1883 6. ESC_K was left out of the EBCDIC table.
1885 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1886 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1887 limit also applies to "virtual nesting" when a pattern is recursive, and in
1888 this case 1000 isn't so big. I have been able to remove this limit at the
1889 expense of backing off one optimization in certain circumstances. Normally,
1890 when pcre_exec() would call its internal match() function recursively and
1891 immediately return the result unconditionally, it uses a "tail recursion"
1892 feature to save stack. However, when a subpattern that can match an empty
1893 string has an unlimited repetition quantifier, it no longer makes this
1894 optimization. That gives it a stack frame in which to save the data for
1895 checking that an empty string has been matched. Previously this was taken
1896 from the 1000-entry workspace that had been reserved. So now there is no
1897 explicit limit, but more stack is used.
1899 8. Applied Daniel's patches to solve problems with the import/export magic
1900 syntax that is required for Windows, and which was going wrong for the
1901 pcreposix and pcrecpp parts of the library. These were overlooked when this
1902 problem was solved for the main library.
1904 9. There were some crude static tests to avoid integer overflow when computing
1905 the size of patterns that contain repeated groups with explicit upper
1906 limits. As the maximum quantifier is 65535, the maximum group length was
1907 set at 30,000 so that the product of these two numbers did not overflow a
1908 32-bit integer. However, it turns out that people want to use groups that
1909 are longer than 30,000 bytes (though not repeat them that many times).
1910 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1911 made it possible to implement the integer overflow checks in a much more
1912 dynamic way, which I have now done. The artificial limitation on group
1913 length has been removed - we now have only the limit on the total length of
1914 the compiled pattern, which depends on the LINK_SIZE setting.
1916 10. Fixed a bug in the documentation for get/copy named substring when
1917 duplicate names are permitted. If none of the named substrings are set, the
1918 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1919 empty string.
1921 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1922 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1923 because the ] is interpreted as the first data character and the
1924 terminating ] is not found. PCRE has been made compatible with Perl in this
1925 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1926 cause memory overwriting.
1928 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1929 string has been matched (to stop an infinite loop). It was not recognizing
1930 a conditional subpattern that could match an empty string if that
1931 subpattern was within another subpattern. For example, it looped when
1932 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1933 condition was not nested. This bug has been fixed.
1935 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1936 past the start of the subject in the presence of bytes with the top bit
1937 set, for example "\x8aBCD".
1939 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1940 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1942 14. Optimized (?!) to (*FAIL).
1944 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1945 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1946 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1947 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1948 does: it's just the validity check that is more restrictive.
1950 16. Inserted checks for integer overflows during escape sequence (backslash)
1951 processing, and also fixed erroneous offset values for syntax errors during
1952 backslash processing.
1954 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1955 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1957 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1958 caused an overrun.
1960 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1961 something other than just ASCII characters) inside a group that had an
1962 unlimited repeat caused a loop at compile time (while checking to see
1963 whether the group could match an empty string).
1965 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1966 [\P{Any}] did so. (Error in the code for printing property names.)
1968 21. An orphan \E inside a character class could cause a crash.
1970 22. A repeated capturing bracket such as (A)? could cause a wild memory
1971 reference during compilation.
1973 23. There are several functions in pcre_compile() that scan along a compiled
1974 expression for various reasons (e.g. to see if it's fixed length for look
1975 behind). There were bugs in these functions when a repeated \p or \P was
1976 present in the pattern. These operators have additional parameters compared
1977 with \d, etc, and these were not being taken into account when moving along
1978 the compiled data. Specifically:
1980 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1981 length.
1983 (b) An item such as \pL+ within a repeated group could cause crashes or
1984 loops.
1986 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1987 "reference to non-existent subpattern" error.
1989 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1991 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1992 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1994 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1996 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1997 character were causing crashes (broken optimization).
1999 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2000 \p or \P) caused a compile-time loop.
2002 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2003 break. For example, the unstudied pattern [\r\n]A does not match the string
2004 "\r\nA" because change 7.0/46 below moves the current point on by two
2005 characters after failing to match at the start. However, the pattern \nA
2006 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2007 the same is true. There doesn't seem any very clean way out of this, but
2008 what I have chosen to do makes the common cases work: PCRE now takes note
2009 of whether there can be an explicit match for \r or \n anywhere in the
2010 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2011 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2012 pattern has explicit CR or LF references.
2014 29. Added (*CR) etc for changing newline setting at start of pattern.
2017 Version 7.2 19-Jun-07
2018 ---------------------
2020 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2021 which is apparently normally available under Windows.
2023 2. Re-jig the pcregrep tests with different newline settings in an attempt
2024 to make them independent of the local environment's newline setting.
2026 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2028 4. Some of the "internals" tests were previously cut out when the link size
2029 was not 2, because the output contained actual offsets. The recent new
2030 "Z" feature of pcretest means that these can be cut out, making the tests
2031 usable with all link sizes.
2033 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2034 stack recursion. This gives a massive performance boost under BSD, but just
2035 a small improvement under Linux. However, it saves one field in the frame
2036 in all cases.
2038 6. Added more features from the forthcoming Perl 5.10:
2040 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2041 recursion call. It refers to the nth most recently opened parentheses.
2043 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2044 to be opened parentheses.
2046 (c) Conditions that refer to capturing parentheses can be specified
2047 relatively, for example, (?(-2)... or (?(+3)...
2049 (d) \K resets the start of the current match so that everything before
2050 is not part of it.
2052 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2054 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2055 reference syntax.
2057 (g) (?| introduces a group in which the numbering of parentheses in each
2058 alternative starts with the same number.
2060 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2062 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2065 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2066 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2067 for detecting groups that can match an empty string.
2069 9. A pattern with a very large number of alternatives (more than several
2070 hundred) was running out of internal workspace during the pre-compile
2071 phase, where pcre_compile() figures out how much memory will be needed. A
2072 bit of new cunning has reduced the workspace needed for groups with
2073 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2074 workspace instead of running out of the 4096 that are available.
2076 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2078 11. Applied patch from Google to remove an optimization that didn't quite work.
2079 The report of the bug said:
2081 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2082 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2083 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2085 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2086 it matched the wrong number of bytes.
2089 Version 7.1 24-Apr-07
2090 ---------------------
2092 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2093 that is more "standard", making use of automake and other Autotools. There
2094 is some re-arrangement of the files and adjustment of comments consequent
2095 on this.
2097 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2098 for recursive directory scanning broke on some systems because the files
2099 are not scanned in any specific order and on different systems the order
2100 was different. A call to "sort" has been inserted into RunGrepTest for the
2101 approprate test as a short-term fix. In the longer term there may be an
2102 alternative.
2104 3. I had an email from Eric Raymond about problems translating some of PCRE's
2105 man pages to HTML (despite the fact that I distribute HTML pages, some
2106 people do their own conversions for various reasons). The problems
2107 concerned the use of low-level troff macros .br and .in. I have therefore
2108 removed all such uses from the man pages (some were redundant, some could
2109 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2110 HTML has been updated to handle .nf/.fi and to complain if it encounters
2111 .br or .in.
2113 4. Updated comments in configure.ac that get placed in config.h.in and also
2114 arranged for config.h to be included in the distribution, with the name
2115 config.h.generic, for the benefit of those who have to compile without
2116 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2118 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2119 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2120 makevp.bat for the current PCRE, using the additional files
2121 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2123 6. A Windows user reported a minor discrepancy with test 2, which turned out
2124 to be caused by a trailing space on an input line that had got lost in his
2125 copy. The trailing space was an accident, so I've just removed it.
2127 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2128 that is needed.
2130 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2131 as "const" (a) because they are and (b) because it helps the PHP
2132 maintainers who have recently made a script to detect big data structures
2133 in the php code that should be moved to the .rodata section. I remembered
2134 to update Builducptable as well, so it won't revert if ucptable.h is ever
2135 re-created.
2137 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2138 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2139 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2140 support is not required. This saves 1.5-2K of code, which is important in
2141 some applications.
2143 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2144 so as not to refer to the tables, even though these functions will never be
2145 called when UTF-8 support is disabled. Otherwise there are problems with a
2146 shared library.
2148 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2150 (a) It was defining its arguments as char * instead of void *.
2152 (b) It was assuming that all moves were upwards in memory; this was true
2153 a long time ago when I wrote it, but is no longer the case.
2155 The emulated memove() is provided for those environments that have neither
2156 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2157 is clearly not the case, as these two bugs were recently reported.
2159 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2160 and Detrail to create the HTML documentation, the .txt form of the man
2161 pages, and it removes trailing spaces from listed files. It also creates
2162 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2163 case, it wraps all the #defines with #ifndefs. This script should be run
2164 before "make dist".
2166 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2167 with Unicode property support.
2169 (a) For a maximizing quantifier, if the two different cases of the
2170 character were of different lengths in their UTF-8 codings (there are
2171 some cases like this - I found 11), and the matching function had to
2172 back up over a mixture of the two cases, it incorrectly assumed they
2173 were both the same length.
2175 (b) When PCRE was configured to use the heap rather than the stack for
2176 recursion during matching, it was not correctly preserving the data for
2177 the other case of a UTF-8 character when checking ahead for a match
2178 while processing a minimizing repeat. If the check also involved
2179 matching a wide character, but failed, corruption could cause an
2180 erroneous result when trying to check for a repeat of the original
2181 character.
2183 13. Some tidying changes to the testing mechanism:
2185 (a) The RunTest script now detects the internal link size and whether there
2186 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2187 values substituted by "configure". (The RunGrepTest script already did
2188 this for UTF-8.) The configure.ac script no longer substitutes the
2189 relevant variables.
2191 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2192 with length and offset values. This means that the output is different
2193 for different internal link sizes. Test 2 is skipped for link sizes
2194 other than 2 because of this, bypassing the problem. Unfortunately,
2195 there was also a test in test 3 (the locale tests) that used /B and
2196 failed for link sizes other than 2. Rather than cut the whole test out,
2197 I have added a new /Z option to pcretest that replaces the length and
2198 offset values with spaces. This is now used to make test 3 independent
2199 of link size. (Test 2 will be tidied up later.)
2201 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2202 segmentation fault instead of returning the appropriate error message.
2204 15. In multiline mode when the newline sequence was set to "any", the pattern
2205 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2206 This doesn't seem right; it now treats the CRLF combination as the line
2207 ending, and so does not match in that case. It's only a pattern such as ^$
2208 that would hit this one: something like ^ABC$ would have failed after \r
2209 and then tried again after \r\n.
2211 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2212 in an attempt to make files that differ only in their line terminators
2213 compare equal. This works on Linux.
2215 17. Under certain error circumstances pcregrep might try to free random memory
2216 as it exited. This is now fixed, thanks to valgrind.
2218 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2219 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2220 was because its rules for how to advance for /g after matching an empty
2221 string at the end of a line did not allow for this case. They now check for
2222 it specially.
2224 20. pcretest is supposed to handle patterns and data of any length, by
2225 extending its buffers when necessary. It was getting this wrong when the
2226 buffer for a data line had to be extended.
2228 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2229 CRLF as a newline sequence.
2231 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2232 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2233 I have nevertheless tidied it up.
2235 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2237 24. Added a man page for pcre-config.
2240 Version 7.0 19-Dec-06
2241 ---------------------
2243 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2244 moving to gcc 4.1.1.
2246 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2247 sys/time.h, which is documented as needed for this function. It doesn't
2248 seem to matter on Linux, but it showed up on some releases of OS X.
2250 3. It seems that there are systems where bytes whose values are greater than
2251 127 match isprint() in the "C" locale. The "C" locale should be the
2252 default when a C program starts up. In most systems, only ASCII printing
2253 characters match isprint(). This difference caused the output from pcretest
2254 to vary, making some of the tests fail. I have changed pcretest so that:
2256 (a) When it is outputting text in the compiled version of a pattern, bytes
2257 other than 32-126 are always shown as hex escapes.
2259 (b) When it is outputting text that is a matched part of a subject string,
2260 it does the same, unless a different locale has been set for the match
2261 (using the /L modifier). In this case, it uses isprint() to decide.
2263 4. Fixed a major bug that caused incorrect computation of the amount of memory
2264 required for a compiled pattern when options that changed within the
2265 pattern affected the logic of the preliminary scan that determines the
2266 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2267 that the computed length was too small. The symptoms of this bug were
2268 either the PCRE error "internal error: code overflow" from pcre_compile(),
2269 or a glibc crash with a message such as "pcretest: free(): invalid next
2270 size (fast)". Examples of patterns that provoked this bug (shown in
2271 pcretest format) are:
2273 /(?-x: )/x
2274 /(?x)(?-x: \s*#\s*)/
2275 /((?i)[\x{c0}])/8
2276 /(?i:[\x{c0}])/8
2278 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2279 is now done differently.
2281 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2282 wrapper classes; (b) implement a new function in the C++ scanner that is
2283 more efficient than the old way of doing things because it avoids levels of
2284 recursion in the regex matching; (c) add a paragraph to the documentation
2285 for the FullMatch() function.
2287 6. The escape sequence \n was being treated as whatever was defined as
2288 "newline". Not only was this contrary to the documentation, which states
2289 that \n is character 10 (hex 0A), but it also went horribly wrong when
2290 "newline" was defined as CRLF. This has been fixed.
2292 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2293 was being set to -1 for the "end of line" case (supposedly a value that no
2294 character can have). Though this value is never used (the check for end of
2295 line is "zero bytes in current character"), it caused compiler complaints.
2296 I've changed it to 0xffffffff.
2298 8. In pcre_version.c, the version string was being built by a sequence of
2299 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2300 string (as it is for production releases) called a macro with an empty
2301 argument. The C standard says the result of this is undefined. The gcc
2302 compiler treats it as an empty string (which was what was wanted) but it is
2303 reported that Visual C gives an error. The source has been hacked around to
2304 avoid this problem.
2306 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2307 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2308 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2309 of them did).
2311 10. Originally, pcretest opened its input and output without "b"; then I was
2312 told that "b" was needed in some environments, so it was added for release
2313 5.0 to both the input and output. (It makes no difference on Unix-like
2314 systems.) Later I was told that it is wrong for the input on Windows. I've
2315 now abstracted the modes into two macros, to make it easier to fiddle with
2316 them, and removed "b" from the input mode under Windows.
2318 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2320 12. Added -help and --help to pcretest as an official way of being reminded
2321 of the options.
2323 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2324 and pcrecpp.cc because they annoy compilers at high warning levels.
2326 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2328 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2329 scripts are not C programs :-) and which was not noticed because it works
2330 on Linux.
2332 16. pcretest is supposed to handle any length of pattern and data line (as one
2333 line or as a continued sequence of lines) by extending its input buffer if
2334 necessary. This feature was broken for very long pattern lines, leading to
2335 a string of junk being passed to pcre_compile() if the pattern was longer
2336 than about 50K.
2338 17. I have done a major re-factoring of the way pcre_compile() computes the
2339 amount of memory needed for a compiled pattern. Previously, there was code
2340 that made a preliminary scan of the pattern in order to do this. That was
2341 OK when PCRE was new, but as the facilities have expanded, it has become
2342 harder and harder to keep it in step with the real compile phase, and there
2343 have been a number of bugs (see for example, 4 above). I have now found a
2344 cunning way of running the real compile function in a "fake" mode that
2345 enables it to compute how much memory it would need, while actually only
2346 ever using a few hundred bytes of working memory and without too many
2347 tests of the mode. This should make future maintenance and development
2348 easier. A side effect of this work is that the limit of 200 on the nesting
2349 depth of parentheses has been removed (though this was never a serious
2350 limitation, I suspect). However, there is a downside: pcre_compile() now
2351 runs more slowly than before (30% or more, depending on the pattern). I
2352 hope this isn't a big issue. There is no effect on runtime performance.
2354 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2355 newline (only possible for the last line of a file) and it was a
2356 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2358 19. Added additional timing features to pcretest. (1) The -tm option now times
2359 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2360 separate command line item, by a number that specifies the number of
2361 repeats to use when timing. The default is 50000; this gives better
2362 precision, but takes uncomfortably long for very large patterns.
2364 20. Extended pcre_study() to be more clever in cases where a branch of a
2365 subpattern has no definite first character. For example, (a*|b*)[cd] would
2366 previously give no result from pcre_study(). Now it recognizes that the
2367 first character must be a, b, c, or d.
2369 21. There was an incorrect error "recursive call could loop indefinitely" if
2370 a subpattern (or the entire pattern) that was being tested for matching an
2371 empty string contained only one non-empty item after a nested subpattern.
2372 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2373 incorrectly, because the \d was being skipped in the check.
2375 22. The pcretest program now has a new pattern option /B and a command line
2376 option -b, which is equivalent to adding /B to every pattern. This causes
2377 it to show the compiled bytecode, without the additional information that
2378 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2379 is the same as /B/I).
2381 23. A new optimization is now able automatically to treat some sequences such
2382 as a*b as a*+b. More specifically, if something simple (such as a character
2383 or a simple class like \d) has an unlimited quantifier, and is followed by
2384 something that cannot possibly match the quantified thing, the quantifier
2385 is automatically "possessified".
2387 24. A recursive reference to a subpattern whose number was greater than 39
2388 went wrong under certain circumstances in UTF-8 mode. This bug could also
2389 have affected the operation of pcre_study().
2391 25. Realized that a little bit of performance could be had by replacing
2392 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2394 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2396 27. Possessive quantifiers such as a++ were previously implemented by turning
2397 them into atomic groups such as ($>a+). Now they have their own opcodes,
2398 which improves performance. This includes the automatically created ones
2399 from 23 above.
2401 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2402 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2403 the first matched character to be a colon. This applied both to named and
2404 numbered groups.
2406 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2408 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2409 building PCRE on Windows easier, so I have included it in the distribution.
2411 31. There is now a check in pcretest against a ridiculously large number being
2412 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2413 loop, the loop is abandoned.
2415 32. Forward references to subpatterns in conditions such as (?(2)...) where
2416 subpattern 2 is defined later cause pcre_compile() to search forwards in
2417 the pattern for the relevant set of parentheses. This search went wrong
2418 when there were unescaped parentheses in a character class, parentheses
2419 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2421 33. "Subroutine" calls and backreferences were previously restricted to
2422 referencing subpatterns earlier in the regex. This restriction has now
2423 been removed.
2425 34. Added a number of extra features that are going to be in Perl 5.10. On the
2426 whole, these are just syntactic alternatives for features that PCRE had
2427 previously implemented using the Python syntax or my own invention. The
2428 other formats are all retained for compatibility.
2430 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2431 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2432 also .NET compatible.
2434 (b) A recursion or subroutine call to a named group can now be defined as
2435 (?&name) as well as (?P>name).
2437 (c) A backreference to a named group can now be defined as \k<name> or
2438 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2439 5.10, are also .NET compatible.
2441 (d) A conditional reference to a named group can now use the syntax
2442 (?(<name>) or (?('name') as well as (?(name).
2444 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2445 groups (named and numbered) that are never evaluated inline, but can be
2446 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2447 is always false. There may be only one alternative in such a group.
2449 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2450 as the simple (?(R). The condition is true only if the most recent
2451 recursion is that of the given number or name. It does not search out
2452 through the entire recursion stack.
2454 (g) The escape \gN or \g{N} has been added, where N is a positive or
2455 negative number, specifying an absolute or relative reference.
2457 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2458 some "unreachable code" warnings.
2460 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2461 things, this adds five new scripts.
2463 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2464 There were also incompatibilities regarding the handling of \Q..\E inside
2465 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2466 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2468 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2469 matches an empty string, and forcibly breaks the loop. There were bugs in
2470 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2471 against aaaa the result was just "a" rather than "aaaa", for example. Two
2472 separate and independent bugs (that affected different cases) have been
2473 fixed.
2475 39. Refactored the code to abolish the use of different opcodes for small
2476 capturing bracket numbers. This is a tidy that I avoided doing when I
2477 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2478 The new approach is not only tidier, it makes it possible to reduce the
2479 memory needed to fix the previous bug (38).
2481 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2482 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2483 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2484 mode.
2486 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2487 report.
2489 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2490 copy construction and assignment in the C++ wrapper.
2492 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2493 couple of bugs that could have given wrong results in the "--newline=crlf"
2494 case.
2496 44. Added a number of casts and did some reorganization of signed/unsigned int
2497 variables following suggestions from Dair Grant. Also renamed the variable
2498 "this" as "item" because it is a C++ keyword.
2500 45. Arranged for dftables to add
2502 #include "pcre_internal.h"
2504 to pcre_chartables.c because without it, gcc 4.x may remove the array
2505 definition from the final binary if PCRE is built into a static library and
2506 dead code stripping is activated.
2508 46. For an unanchored pattern, if a match attempt fails at the start of a
2509 newline sequence, and the newline setting is CRLF or ANY, and the next two
2510 characters are CRLF, advance by two characters instead of one.
2513 Version 6.7 04-Jul-06
2514 ---------------------
2516 1. In order to handle tests when input lines are enormously long, pcretest has
2517 been re-factored so that it automatically extends its buffers when
2518 necessary. The code is crude, but this _is_ just a test program. The
2519 default size has been increased from 32K to 50K.
2521 2. The code in pcre_study() was using the value of the re argument before
2522 testing it for NULL. (Of course, in any sensible call of the function, it
2523 won't be NULL.)
2525 3. The memmove() emulation function in pcre_internal.h, which is used on
2526 systems that lack both memmove() and bcopy() - that is, hardly ever -
2527 was missing a "static" storage class specifier.
2529 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2530 containing an extended class (one that cannot be represented by a bitmap
2531 because it contains high-valued characters or Unicode property items, e.g.
2532 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2533 pattern, but PCRE should not loop if you do not (it no longer does).
2534 [Detail: two cases were found: (a) a repeated subpattern containing an
2535 extended class; (b) a recursive reference to a subpattern that followed a
2536 previous extended class. It wasn't skipping over the extended class
2537 correctly when UTF-8 mode was not set.]
2539 5. A negated single-character class was not being recognized as fixed-length
2540 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2541 compile error "lookbehind assertion is not fixed length".
2543 6. The RunPerlTest auxiliary script was showing an unexpected difference
2544 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2545 write a Perl script that can interpret lines of an input file either as
2546 byte characters or as UTF-8, which is what "perltest" was being required to
2547 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2548 can't do is switch easily at run time between having the "use utf8;" pragma
2549 or not. In the end, I fudged it by using the RunPerlTest script to insert
2550 "use utf8;" explicitly for the UTF-8 tests.
2552 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2553 the end of the subject string, contrary to the documentation and to what
2554 Perl does. This was true of both matching functions. Now it matches only at
2555 the start of the subject and immediately after *internal* newlines.
2557 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2558 a pointer to an int instead of a pointer to an unsigned long int. This
2559 caused problems on 64-bit systems.
2561 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2562 instance of the 'standard' template library not being so standard".
2564 10. There was no check on the number of named subpatterns nor the maximum
2565 length of a subpattern name. The product of these values is used to compute
2566 the size of the memory block for a compiled pattern. By supplying a very
2567 long subpattern name and a large number of named subpatterns, the size
2568 computation could be caused to overflow. This is now prevented by limiting
2569 the length of names to 32 characters, and the number of named subpatterns
2570 to 10,000.
2572 11. Subpatterns that are repeated with specific counts have to be replicated in
2573 the compiled pattern. The size of memory for this was computed from the
2574 length of the subpattern and the repeat count. The latter is limited to
2575 65535, but there was no limit on the former, meaning that integer overflow
2576 could in principle occur. The compiled length of a repeated subpattern is
2577 now limited to 30,000 bytes in order to prevent this.
2579 12. Added the optional facility to have named substrings with the same name.
2581 13. Added the ability to use a named substring as a condition, using the
2582 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2583 are numbers (not recommended). Forward references are permitted.
2585 14. Added forward references in named backreferences (if you see what I mean).
2587 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2588 pattern could run off the end of the subject. For example, the pattern
2589 "(?s)(.{1,5})"8 did this with the subject "ab".
2591 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2592 PCRE_CASELESS was set when matching characters that were quantified with ?
2593 or *.
2595 17. A character class other than a single negated character that had a minimum
2596 but no maximum quantifier - for example [ab]{6,} - was not handled
2597 correctly by pce_dfa_exec(). It would match only one character.
2599 18. A valid (though odd) pattern that looked like a POSIX character
2600 class but used an invalid character after [ (for example [[,abc,]]) caused
2601 pcre_compile() to give the error "Failed: internal error: code overflow" or
2602 in some cases to crash with a glibc free() error. This could even happen if
2603 the pattern terminated after [[ but there just happened to be a sequence of
2604 letters, a binary zero, and a closing ] in the memory that followed.
2606 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2607 over the years. Originally (before any Unicode support), just the bottom 8
2608 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2609 output from "man perlunicode" includes this:
2611 The regular expression compiler produces polymorphic opcodes. That
2612 is, the pattern adapts to the data and automatically switches to
2613 the Unicode character scheme when presented with Unicode data--or
2614 instead uses a traditional byte scheme when presented with byte
2615 data.
2617 Sadly, a wide octal escape does not cause a switch, and in a string with
2618 no other multibyte characters, these octal escapes are treated as before.
2619 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2620 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2621 Unicode string.
2623 I have not perpetrated such confusion in PCRE. Up till now, it took just
2624 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2625 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2626 translate to the appropriate multibyte character.
2628 29. Applied some refactoring to reduce the number of warnings from Microsoft
2629 and Borland compilers. This has included removing the fudge introduced
2630 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2631 a warning about an unused variable.
2633 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2634 characters since release 4.0, because Perl (from release 5.004) does not.
2635 [Or at least, is documented not to: some releases seem to be in conflict
2636 with the documentation.] However, when a pattern was studied with
2637 pcre_study() and all its branches started with \s, PCRE still included VT
2638 as a possible starting character. Of course, this did no harm; it just
2639 caused an unnecessary match attempt.
2641 22. Removed a now-redundant internal flag bit that recorded the fact that case
2642 dependency changed within the pattern. This was once needed for "required
2643 byte" processing, but is no longer used. This recovers a now-scarce options
2644 bit. Also moved the least significant internal flag bit to the most-
2645 significant bit of the word, which was not previously used (hangover from
2646 the days when it was an int rather than a uint) to free up another bit for
2647 the future.
2649 23. Added support for CRLF line endings as well as CR and LF. As well as the
2650 default being selectable at build time, it can now be changed at runtime
2651 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2652 specify that it is scanning data with non-default line endings.
2654 24. Changed the definition of CXXLINK to make it agree with the definition of
2655 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2657 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2658 recursions. This makes a big different to stack usage for some patterns.
2660 26. If a subpattern containing a named recursion or subroutine reference such
2661 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2662 the space required for the compiled pattern went wrong and gave too small a
2663 value. Depending on the environment, this could lead to "Failed: internal
2664 error: code overflow at offset 49" or "glibc detected double free or
2665 corruption" errors.
2667 27. Applied patches from Google (a) to support the new newline modes and (b) to
2668 advance over multibyte UTF-8 characters in GlobalReplace.
2670 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2671 difference for some implementation of PCRE in some Windows version.
2673 29. Added some extra testing facilities to pcretest:
2675 \q<number> in a data line sets the "match limit" value
2676 \Q<number> in a data line sets the "match recursion limt" value
2677 -S <number> sets the stack size, where <number> is in megabytes
2679 The -S option isn't available for Windows.
2682 Version 6.6 06-Feb-06
2683 ---------------------
2685 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2686 in pcreposix.h. I have copied the definition from pcre.h.
2688 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2689 because pcre.h is no longer a built file.
2691 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2692 not normally included in the compiled code.
2695 Version 6.5 01-Feb-06
2696 ---------------------
2698 1. When using the partial match feature with pcre_dfa_exec(), it was not
2699 anchoring the second and subsequent partial matches at the new starting
2700 point. This could lead to incorrect results. For example, with the pattern
2701 /1234/, partially matching against "123" and then "a4" gave a match.
2703 2. Changes to pcregrep:
2705 (a) All non-match returns from pcre_exec() were being treated as failures
2706 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2707 error message is output. Some extra information is given for the
2709 probably the only errors that are likely to be caused by users (by
2710 specifying a regex that has nested indefinite repeats, for instance).
2711 If there are more than 20 of these errors, pcregrep is abandoned.
2713 (b) A binary zero was treated as data while matching, but terminated the
2714 output line if it was written out. This has been fixed: binary zeroes
2715 are now no different to any other data bytes.
2717 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2718 used to set a locale for matching. The --locale=xxxx long option has
2719 been added (no short equivalent) to specify a locale explicitly on the
2720 pcregrep command, overriding the environment variables.
2722 (d) When -B was used with -n, some line numbers in the output were one less
2723 than they should have been.
2725 (e) Added the -o (--only-matching) option.
2727 (f) If -A or -C was used with -c (count only), some lines of context were
2728 accidentally printed for the final match.
2730 (g) Added the -H (--with-filename) option.
2732 (h) The combination of options -rh failed to suppress file names for files
2733 that were found from directory arguments.
2735 (i) Added the -D (--devices) and -d (--directories) options.
2737 (j) Added the -F (--fixed-strings) option.
2739 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2741 (l) Added the --colo(u)r option.
2743 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2744 is not present by default.
2746 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2747 items such as (?R) or (?1), when the recursion could match a number of
2748 alternatives. If it matched one of the alternatives, but subsequently,
2749 outside the recursion, there was a failure, the code tried to back up into
2750 the recursion. However, because of the way PCRE is implemented, this is not
2751 possible, and the result was an incorrect result from the match.
2753 In order to prevent this happening, the specification of recursion has
2754 been changed so that all such subpatterns are automatically treated as
2755 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2757 4. I had overlooked the fact that, in some locales, there are characters for
2758 which isalpha() is true but neither isupper() nor islower() are true. In
2759 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2760 and ordfeminine) are like this. This affected the treatment of \w and \W
2761 when they appeared in character classes, but not when they appeared outside
2762 a character class. The bit map for "word" characters is now created
2763 separately from the results of isalnum() instead of just taking it from the
2764 upper, lower, and digit maps. (Plus the underscore character, of course.)
2766 5. The above bug also affected the handling of POSIX character classes such as
2767 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2768 permanent tables. Instead, the bit maps for such a class were previously
2769 created as the appropriate unions of the upper, lower, and digit bitmaps.
2770 Now they are created by subtraction from the [[:word:]] class, which has
2771 its own bitmap.
2773 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2774 It is created by subtracting the vertical space characters (\x09, \x0a,
2775 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2776 subtraction was done in the overall bitmap for a character class, meaning
2777 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2778 be recognized. This bug has been fixed.
2780 7. Patches from the folks at Google:
2782 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2783 real life, but is still worth protecting against".
2785 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2786 regular expressions".
2788 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2789 have it.
2791 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2792 "configure" and the latter not, in order to fix a problem somebody had
2793 with compiling the Arg class on HP-UX.
2795 (e) Improve the error-handling of the C++ wrapper a little bit.
2797 (f) New tests for checking recursion limiting.
2799 8. The pcre_memmove() function, which is used only if the environment does not
2800 have a standard memmove() function (and is therefore rarely compiled),
2801 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2802 returning a result (though PCRE never actually uses the result).
2804 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2805 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2806 returned instead of calling malloc() with an overflowing number that would
2807 most likely cause subsequent chaos.
2809 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2811 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2812 with this option is matched, the nmatch and pmatch options of regexec() are
2813 ignored.
2815 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2816 provided in case anyone wants to the the POSIX interface with UTF-8
2817 strings.
2819 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2820 C++ linking (needed for some HP-UX environments).
2822 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2823 (unused parameter) and in the pcre_printint() function (omitted "default"
2824 switch label when the default is to do nothing).
2826 15. Added some code to make it possible, when PCRE is compiled as a C++
2827 library, to replace subject pointers for pcre_exec() with a smart pointer
2828 class, thus making it possible to process discontinuous strings.
2830 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2831 much the same function. They were added by different people who were trying
2832 to make PCRE easy to compile on non-Unix systems. It has been suggested
2833 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2834 for compiling on Windows systems. I have therefore replaced it with
2835 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2836 defaults to "extern" for C or "extern C" for C++, which works fine on
2837 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2838 SCOPE with something explicit in config.h. In addition:
2840 (a) pcreposix.h still had just "extern" instead of either of these macros;
2841 I have replaced it with PCRE_DATA_SCOPE.
2843 (b) Functions such as _pcre_xclass(), which are internal to the library,
2844 but external in the C sense, all had PCRE_EXPORT in their definitions.
2845 This is apparently wrong for the Windows case, so I have removed it.
2846 (It makes no difference on Unix-like systems.)
2848 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2849 of recursive calls to match(). This is different to MATCH_LIMIT because
2850 that limits the total number of calls to match(), not all of which increase
2851 the depth of recursion. Limiting the recursion depth limits the amount of
2852 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2853 when PCRE is compiled, and changed at run time. A patch from Google adds
2854 this functionality to the C++ interface.
2856 18. Changes to the handling of Unicode character properties:
2858 (a) Updated the table to Unicode 4.1.0.
2860 (b) Recognize characters that are not in the table as "Cn" (undefined).
2862 (c) I revised the way the table is implemented to a much improved format
2863 which includes recognition of ranges. It now supports the ranges that
2864 are defined in UnicodeData.txt, and it also amalgamates other
2865 characters into ranges. This has reduced the number of entries in the
2866 table from around 16,000 to around 3,000, thus reducing its size
2867 considerably. I realized I did not need to use a tree structure after
2868 all - a binary chop search is just as efficient. Having reduced the
2869 number of entries, I extended their size from 6 bytes to 8 bytes to
2870 allow for more data.
2872 (d) Added support for Unicode script names via properties such as \p{Han}.
2874 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2875 matching that character.
2877 20. When matching a repeated Unicode property with a minimum greater than zero,
2878 (for example \pL{2,}), PCRE could look past the end of the subject if it
2879 reached it while seeking the minimum number of characters. This could
2880 happen only if some of the characters were more than one byte long, because
2881 there is a check for at least the minimum number of bytes.
2883 21. Refactored the implementation of \p and \P so as to be more general, to
2884 allow for more different types of property in future. This has changed the
2885 compiled form incompatibly. Anybody with saved compiled patterns that use
2886 \p or \P will have to recompile them.
2888 22. Added "Any" and "L&" to the supported property types.
2890 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2891 but give a compile time error if the value is greater than 0xff.
2893 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2894 accidentally not being installed or uninstalled.
2896 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2897 made were to insert the current release number. This seemed silly, because
2898 it made things harder for people building PCRE on systems that don't run
2899 "configure". I have turned pcre.h into a distributed file, no longer built
2900 by "configure", with the version identification directly included. There is
2901 no longer a pcre.h.in file.
2903 However, this change necessitated a change to the pcre-config script as
2904 well. It is built from pcre-config.in, and one of the substitutions was the
2905 release number. I have updated configure.ac so that ./configure now finds
2906 the release number by grepping pcre.h.
2908 26. Added the ability to run the tests under valgrind.
2911 Version 6.4 05-Sep-05
2912 ---------------------
2914 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2915 "--" to be printed when multiple files were scanned, even when none of the
2916 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2917 consider it to be a bug, and have restored the previous behaviour.
2919 2. A couple of code tidies to get rid of compiler warnings.
2921 3. The pcretest program used to cheat by referring to symbols in the library
2922 whose names begin with _pcre_. These are internal symbols that are not
2923 really supposed to be visible externally, and in some environments it is
2924 possible to suppress them. The cheating is now confined to including
2925 certain files from the library's source, which is a bit cleaner.
2927 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2928 file's purpose clearer.
2930 5. Reorganized pcre_ucp_findchar().
2933 Version 6.3 15-Aug-05
2934 ---------------------
2936 1. The file libpcre.pc.in did not have general read permission in the tarball.
2938 2. There were some problems when building without C++ support:
2940 (a) If C++ support was not built, "make install" and "make test" still
2941 tried to test it.
2943 (b) There were problems when the value of CXX was explicitly set. Some
2944 changes have been made to try to fix these, and ...
2946 (c) --disable-cpp can now be used to explicitly disable C++ support.
2948 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2949 backslash in a target when C++ was disabled. This confuses some
2950 versions of "make", apparently. Using an intermediate variable solves
2951 this. (Same for CPP_LOBJ.)
2953 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2954 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2955 necessary on certain architectures.
2957 4. Added a setting of -export-symbols-regex to the link command to remove
2958 those symbols that are exported in the C sense, but actually are local
2959 within the library, and not documented. Their names all begin with
2960 "_pcre_". This is not a perfect job, because (a) we have to except some
2961 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2962 available (and never for static libraries). I have made a note to try to
2963 find a way round (a) in the future.
2966 Version 6.2 01-Aug-05
2967 ---------------------
2969 1. There was no test for integer overflow of quantifier values. A construction
2970 such as {1111111111111111} would give undefined results. What is worse, if
2971 a minimum quantifier for a parenthesized subpattern overflowed and became
2972 negative, the calculation of the memory size went wrong. This could have
2973 led to memory overwriting.
2975 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2977 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2978 operating environments where this matters.
2980 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2981 PCRE options from within the C++ wrapper.
2983 5. Named capturing subpatterns were not being correctly counted when a pattern
2984 was compiled. This caused two problems: (a) If there were more than 100
2985 such subpatterns, the calculation of the memory needed for the whole
2986 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2987 back references of the form \12, where the number was greater than 9, were
2988 not recognized as back references, even though there were sufficient
2989 previous subpatterns.
2991 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2992 versions of gcc, e.g. 2.95.4.
2995 Version 6.1 21-Jun-05
2996 ---------------------
2998 1. There was one reference to the variable "posix" in pcretest.c that was not
2999 surrounded by "#if !defined NOPOSIX".
3001 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3002 the cross-check on the old pcre_info() function, for the benefit of the
3003 cut-down version of PCRE that is currently imported into Exim.
3005 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3006 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3007 patterns that start (?i) but I don't think that matters. The use of (?i) is
3008 just an example; this all applies to the other options as well.
3010 4. Since libtool seems to echo the compile commands it is issuing, the output
3011 from "make" can be reduced a bit by putting "@" in front of each libtool
3012 compile command.
3014 5. Patch from the folks at Google for configure.in to be a bit more thorough
3015 in checking for a suitable C++ installation before trying to compile the
3016 C++ stuff. This should fix a reported problem when a compiler was present,
3017 but no suitable headers.
3019 6. The man pages all had just "PCRE" as their title. I have changed them to
3020 be the relevant file name. I have also arranged that these names are
3021 retained in the file doc/pcre.txt, which is a concatenation in text format
3022 of all the man pages except the little individual ones for each function.
3024 7. The NON-UNIX-USE file had not been updated for the different set of source
3025 files that come with release 6. I also added a few comments about the C++
3026 wrapper.
3029 Version 6.0 07-Jun-05
3030 ---------------------
3032 1. Some minor internal re-organization to help with my DFA experiments.
3034 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3035 didn't matter for the library itself when fully configured, but did matter
3036 when compiling without UCP support, or within Exim, where the ucp files are
3037 not imported.
3039 3. Refactoring of the library code to split up the various functions into
3040 different source modules. The addition of the new DFA matching code (see
3041 below) to a single monolithic source would have made it really too
3042 unwieldy, quite apart from causing all the code to be include in a
3043 statically linked application, when only some functions are used. This is
3044 relevant even without the DFA addition now that patterns can be compiled in
3045 one application and matched in another.
3047 The downside of splitting up is that there have to be some external
3048 functions and data tables that are used internally in different modules of
3049 the library but which are not part of the API. These have all had their
3050 names changed to start with "_pcre_" so that they are unlikely to clash
3051 with other external names.
3053 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3054 a different (DFA) algorithm. Although it is slower than the original
3055 function, it does have some advantages for certain types of matching
3056 problem.
3058 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3059 including restarting after a partial match.
3061 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3062 defined when compiling for Windows was sent to me. I have put it into the
3063 code, though I have no means of testing or verifying it.
3065 7. Added the pcre_refcount() auxiliary function.
3067 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3068 match before or at the first newline in the subject string. In pcretest,
3069 the /f option on a pattern can be used to set this.
3071 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3072 would behave wrongly. This has been present in PCRE since release 4.0.
3074 10. A number of changes to the pcregrep command:
3076 (a) Refactored how -x works; insert ^(...)$ instead of setting
3077 PCRE_ANCHORED and checking the length, in preparation for adding
3078 something similar for -w.
3080 (b) Added the -w (match as a word) option.
3082 (c) Refactored the way lines are read and buffered so as to have more
3083 than one at a time available.
3085 (d) Implemented a pcregrep test script.
3087 (e) Added the -M (multiline match) option. This allows patterns to match
3088 over several lines of the subject. The buffering ensures that at least
3089 8K, or the rest of the document (whichever is the shorter) is available
3090 for matching (and similarly the previous 8K for lookbehind assertions).
3092 (f) Changed the --help output so that it now says
3094 -w, --word-regex(p)
3096 instead of two lines, one with "regex" and the other with "regexp"
3097 because that confused at least one person since the short forms are the
3098 same. (This required a bit of code, as the output is generated
3099 automatically from a table. It wasn't just a text change.)
3101 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3102 option but starts with a hyphen. Could be a pattern or a path name
3103 starting with a hyphen, for instance.
3105 (h) "-" can be given as a file name to represent stdin.
3107 (i) When file names are being printed, "(standard input)" is used for
3108 the standard input, for compatibility with GNU grep. Previously
3109 "<stdin>" was used.
3111 (j) The option --label=xxx can be used to supply a name to be used for
3112 stdin when file names are being printed. There is no short form.
3114 (k) Re-factored the options decoding logic because we are going to add
3115 two more options that take data. Such options can now be given in four
3116 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3118 (l) Added the -A, -B, and -C options for requesting that lines of context
3119 around matches be printed.
3121 (m) Added the -L option to print the names of files that do not contain
3122 any matching lines, that is, the complement of -l.
3124 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3125 continue to scan other files.
3127 (o) The -s option was incorrectly implemented. For compatibility with other
3128 greps, it now suppresses the error message for a non-existent or non-
3129 accessible file (but not the return code). There is a new option called
3130 -q that suppresses the output of matching lines, which was what -s was
3131 previously doing.
3133 (p) Added --include and --exclude options to specify files for inclusion
3134 and exclusion when recursing.
3136 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3137 Hopefully, it now does.
3139 12. Missing cast in pcre_study().
3141 13. Added an "uninstall" target to the makefile.
3143 14. Replaced "extern" in the function prototypes in Makefile.in with
3144 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3145 world, but is set differently for Windows.
3147 15. Added a second compiling function called pcre_compile2(). The only
3148 difference is that it has an extra argument, which is a pointer to an
3149 integer error code. When there is a compile-time failure, this is set
3150 non-zero, in addition to the error test pointer being set to point to an
3151 error message. The new argument may be NULL if no error number is required
3152 (but then you may as well call pcre_compile(), which is now just a
3153 wrapper). This facility is provided because some applications need a
3154 numeric error indication, but it has also enabled me to tidy up the way
3155 compile-time errors are handled in the POSIX wrapper.
3157 16. Added VPATH=.libs to the makefile; this should help when building with one
3158 prefix path and installing with another. (Or so I'm told by someone who
3159 knows more about this stuff than I do.)
3161 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3162 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3163 match everything, including newlines. This is not POSIX-compatible, but
3164 somebody wanted the feature. From pcretest it can be activated by using
3165 both the P and the s flags.
3167 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3169 19. libpcre.pc was being incorrectly installed as executable.
3171 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3172 it now also looks for '\r' so that it will work unmodified on Windows.
3174 21. Added Google's contributed C++ wrapper to the distribution.
3176 22. Added some untidy missing memory free() calls in pcretest, to keep
3177 Electric Fence happy when testing.
3181 Version 5.0 13-Sep-04
3182 ---------------------
3184 1. Internal change: literal characters are no longer packed up into items
3185 containing multiple characters in a single byte-string. Each character
3186 is now matched using a separate opcode. However, there may be more than one
3187 byte in the character in UTF-8 mode.
3189 2. The pcre_callout_block structure has two new fields: pattern_position and
3190 next_item_length. These contain the offset in the pattern to the next match
3191 item, and its length, respectively.
3193 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3194 insertion of callouts before each pattern item. Added the /C option to
3195 pcretest to make use of this.
3197 4. On the advice of a Windows user, the lines
3199 #if defined(_WIN32) || defined(WIN32)
3200 _setmode( _fileno( stdout ), 0x8000 );
3201 #endif /* defined(_WIN32) || defined(WIN32) */
3203 have been added to the source of pcretest. This apparently does useful
3204 magic in relation to line terminators.
3206 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3207 for the benefit of those environments where the "b" makes a difference.
3209 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3210 to know about it. I have put a hack into configure.in that adds in code
3211 to set GCC=yes if CC=icc. This seems to end up at a point in the
3212 generated configure script that is early enough to affect the setting of
3213 compiler options, which is what is needed, but I have no means of testing
3214 whether it really works. (The user who reported this had patched the
3215 generated configure script, which of course I cannot do.)
3217 LATER: After change 22 below (new libtool files), the configure script
3218 seems to know about icc (and also ecc). Therefore, I have commented out
3219 this hack in configure.in.
3221 7. Added support for pkg-config (2 patches were sent in).
3223 8. Negated POSIX character classes that used a combination of internal tables
3224 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3225 [[:^ascii]]. Typically, they would match almost any characters. The other
3226 POSIX classes were not broken in this way.
3228 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3229 to find the match, as PCRE was deluded into thinking that the match had to
3230 start at the start point or following a newline. The same bug applied to
3231 patterns with negative forward assertions or any backward assertions
3232 preceding ".*" at the start, unless the pattern required a fixed first
3233 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3235 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3236 starting at the last subject character, bytes beyond the end of the subject
3237 string were read.
3239 11. Renamed the variable "class" as "classbits" to make life easier for C++
3240 users. (Previously there was a macro definition, but it apparently wasn't
3241 enough.)
3243 12. Added the new field "tables" to the extra data so that tables can be passed
3244 in at exec time, or the internal tables can be re-selected. This allows
3245 a compiled regex to be saved and re-used at a later time by a different
3246 program that might have everything at different addresses.
3248 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3249 -R library as well as a -L library.
3251 14. The debugging options of pcretest (-d on the command line or D on a
3252 pattern) showed incorrect output for anything following an extended class
3253 that contained multibyte characters and which was followed by a quantifier.
3255 15. Added optional support for general category Unicode character properties
3256 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3257 support. It adds about 90K to the size of the library. The meanings of the
3258 inbuilt class escapes such as \d and \s have NOT been changed.
3260 16. Updated pcredemo.c to include calls to free() to release the memory for the
3261 compiled pattern.
3263 17. The generated file chartables.c was being created in the source directory
3264 instead of in the building directory. This caused the build to fail if the
3265 source directory was different from the building directory, and was
3266 read-only.
3268 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3269 file. No doubt somebody will tell me if they don't make sense... Also added
3270 Dan Mooney's comments about building on OpenVMS.
3272 19. Added support for partial matching via the PCRE_PARTIAL option for
3273 pcre_exec() and the \P data escape in pcretest.
3275 20. Extended pcretest with 3 new pattern features:
3277 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3278 write the compiled pattern to the file whose name is "rest-of-line".
3279 This is a straight binary dump of the data, with the saved pointer to
3280 the character tables forced to be NULL. The study data, if any, is
3281 written too. After writing, pcretest reads a new pattern.
3283 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3284 compiled pattern from the given file. There must not be any
3285 occurrences of "<" in the file name (pretty unlikely); if there are,
3286 pcretest will instead treat the initial "<" as a pattern delimiter.
3287 After reading in the pattern, pcretest goes on to read data lines as
3288 usual.
3290 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3291 and 16-bit fields in a compiled pattern, to simulate a pattern that
3292 was compiled on a host of opposite endianness.
3294 21. The pcre-exec() function can now cope with patterns that were compiled on
3295 hosts of opposite endianness, with this restriction:
3297 As for any compiled expression that is saved and used later, the tables
3298 pointer field cannot be preserved; the extra_data field in the arguments
3299 to pcre_exec() should be used to pass in a tables address if a value
3300 other than the default internal tables were used at compile time.
3302 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3303 now diagnosed as an error. Previously, most of the time, a negative number
3304 would have been treated as zero, but if in addition "ovector" was passed as
3305 NULL, a crash could occur.
3307 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3308 new versions from the libtool 1.5 distribution (the last one is a copy of
3309 a file called libtool.m4). This seems to have fixed the need to patch
3310 "configure" to support Darwin 1.3 (which I used to do). However, I still
3311 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3312 workstation).
3314 24. Changed the PCRE licence to be the more standard "BSD" licence.
3317 Version 4.5 01-Dec-03
3318 ---------------------
3320 1. There has been some re-arrangement of the code for the match() function so
3321 that it can be compiled in a version that does not call itself recursively.
3322 Instead, it keeps those local variables that need separate instances for
3323 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3324 needs to "recurse". Keeping track of where control must go is done by means
3325 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3326 hide most of the details from the main code, and operates only if
3327 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3328 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3329 operating.
3331 To make it easier for callers to provide specially tailored get/free
3332 functions for this usage, two new functions, pcre_stack_malloc, and
3333 pcre_stack_free, are used. They are always called in strict stacking order,
3334 and the size of block requested is always the same.
3336 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3337 PCRE has been compiled to use the stack or the heap for recursion. The
3338 -C option of pcretest uses this to show which version is compiled.
3340 A new data escape \S, is added to pcretest; it causes the amounts of store
3341 obtained and freed by both kinds of malloc/free at match time to be added
3342 to the output.
3344 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3345 what's available on my current Linux desktop machine.
3347 3. When matching a UTF-8 string, the test for a valid string at the start has
3348 been extended. If start_offset is not zero, PCRE now checks that it points
3349 to a byte that is the start of a UTF-8 character. If not, it returns
3350 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3351 this is necessary because there may be backward assertions in the pattern.
3352 When matching the same subject several times, it may save resources to use
3353 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3355 4. The code for checking the validity of UTF-8 strings has been tightened so
3356 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3357 containing "overlong sequences".
3359 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3360 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3361 should have been "&", but it just so happened that all the cases this let
3362 through by mistake were picked up later in the function.
3364 6. I had used a variable called "isblank" - this is a C99 function, causing
3365 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3367 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3368 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3369 so the version is shown; (c) stop "make test" echoing "./RunTest".
3371 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3373 9. The prototype for memmove() for systems that don't have it was using
3374 size_t, but the inclusion of the header that defines size_t was later. I've
3375 moved the #includes for the C headers earlier to avoid this.
3377 10. Added some adjustments to the code to make it easier to compiler on certain
3378 special systems:
3380 (a) Some "const" qualifiers were missing.
3381 (b) Added the macro EXPORT before all exported functions; by default this
3382 is defined to be empty.
3383 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3384 that it reads its output file name as an argument instead of writing
3385 to the standard output and assuming this can be redirected.
3387 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3388 class containing characters with values greater than 255, PCRE compilation
3389 went into a loop.
3391 12. A recursive reference to a subpattern that was within another subpattern
3392 that had a minimum quantifier of zero caused PCRE to crash. For example,
3393 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3394 recursion. If the recursively-called subpattern itself had a zero repeat,
3395 that was OK.
3397 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3398 buffer into which it was copied (for escape processing) was still set at
3399 1024, so long lines caused crashes.
3401 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3402 "internal error: code overflow...". This applied to any character class
3403 that was followed by a possessive quantifier.
3405 15. Modified the Makefile to add libpcre.la as a prerequisite for
3406 libpcreposix.la because I was told this is needed for a parallel build to
3407 work.
3409 16. If a pattern that contained .* following optional items at the start was
3410 studied, the wrong optimizing data was generated, leading to matching
3411 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3412 matching string must start with a or b or c. The correct conclusion for
3413 this pattern is that a match can start with any character.
3416 Version 4.4 13-Aug-03
3417 ---------------------
3419 1. In UTF-8 mode, a character class containing characters with values between
3420 127 and 255 was not handled correctly if the compiled pattern was studied.
3421 In fixing this, I have also improved the studying algorithm for such
3422 classes (slightly).
3424 2. Three internal functions had redundant arguments passed to them. Removal
3425 might give a very teeny performance improvement.
3427 3. Documentation bug: the value of the capture_top field in a callout is *one
3428 more than* the number of the hightest numbered captured substring.
3430 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3431 in incorrectly linking with a previously installed version. They now link
3432 explicitly with libpcre.la.
3434 5. configure.in no longer needs to recognize Cygwin specially.
3436 6. A problem in pcre.in for Windows platforms is fixed.
3438 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3439 pcretest, it used to include the size of the study block as part of its
3440 output. Unfortunately, the structure contains a field that has a different
3441 size on different hardware architectures. This meant that the tests that
3442 showed this size failed. As the block is currently always of a fixed size,
3443 this information isn't actually particularly useful in pcretest output, so
3444 I have just removed it.
3446 8. Three pre-processor statements accidentally did not start in column 1.
3447 Sadly, there are *still* compilers around that complain, even though
3448 standard C has not required this for well over a decade. Sigh.
3450 9. In pcretest, the code for checking callouts passed small integers in the
3451 callout_data field, which is a void * field. However, some picky compilers
3452 complained about the casts involved for this on 64-bit systems. Now
3453 pcretest passes the address of the small integer instead, which should get
3454 rid of the warnings.
3456 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3457 both compile and run time, and gives an error if an invalid UTF-8 sequence
3458 is found. There is a option for disabling this check in cases where the
3459 string is known to be correct and/or the maximum performance is wanted.
3461 11. In response to a bug report, I changed one line in Makefile.in from
3463 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3464 to
3465 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3467 to look similar to other lines, but I have no way of telling whether this
3468 is the right thing to do, as I do not use Windows. No doubt I'll get told
3469 if it's wrong...
3472 Version 4.3 21-May-03
3473 ---------------------
3475 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3476 Makefile.
3478 2. Some refactoring to improve the quality of the code:
3480 (i) The utf8_table... variables are now declared "const".
3482 (ii) The code for \cx, which used the "case flipping" table to upper case
3483 lower case letters, now just substracts 32. This is ASCII-specific,
3484 but the whole concept of \cx is ASCII-specific, so it seems
3485 reasonable.
3487 (iii) PCRE was using its character types table to recognize decimal and
3488 hexadecimal digits in the pattern. This is silly, because it handles
3489 only 0-9, a-f, and A-F, but the character types table is locale-
3490 specific, which means strange things might happen. A private
3491 table is now used for this - though it costs 256 bytes, a table is
3492 much faster than multiple explicit tests. Of course, the standard
3493 character types table is still used for matching digits in subject
3494 strings against \d.
3496 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3497 ending in _t are). So I've renamed it as ESC_tee.
3499 3. The first argument for regexec() in the POSIX wrapper should have been
3500 defined as "const".
3502 4. Changed pcretest to use malloc() for its buffers so that they can be
3503 Electric Fenced for debugging.
3505 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3506 to read one or more bytes before the start of the subject string. Often this
3507 had no effect on PCRE's behaviour, but in some circumstances it could
3508 provoke a segmentation fault.
3510 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3511 to try to read one or more bytes before the start of the subject string.
3513 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3514 UTF-8 support could misbehave in various ways if the subject string
3515 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3516 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3517 back over UTF-8 characters.)
3520 Version 4.2 14-Apr-03
3521 ---------------------
3523 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3525 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3526 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3527 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3528 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3529 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3531 Note: automatic setting of the BUILD variables is not yet working
3532 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3533 compile-time but not at link-time
3534 [LINK]: use for linking executables only
3535 make different versions for Windows and non-Windows
3536 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3537 libraries
3538 [LINK_FOR_BUILD]: new variable
3539 [OBJEXT]: use throughout
3540 [EXEEXT]: use throughout
3541 <winshared>: new target
3542 <wininstall>: new target
3543 <dftables.o>: use native compiler
3544 <dftables>: use native linker
3545 <install>: handle Windows platform correctly
3546 <clean>: ditto
3547 <check>: ditto
3548 copy DLL to top builddir before testing
3550 As part of these changes, -no-undefined was removed again. This was reported
3551 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3552 in any case.
3554 3. Some tidies to get rid of compiler warnings:
3556 . In the match_data structure, match_limit was an unsigned long int, whereas
3557 match_call_count was an int. I've made them both unsigned long ints.
3559 . In pcretest the fact that a const uschar * doesn't automatically cast to
3560 a void * provoked a warning.
3562 . Turning on some more compiler warnings threw up some "shadow" variables
3563 and a few more missing casts.
3565 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3566 option, a class that contained a single character with a value between 128
3567 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3569 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3570 option, a class that contained several characters, but with at least one
3571 whose value was between 128 and 255 caused PCRE to crash.
3574 Version 4.1 12-Mar-03
3575 ---------------------
3577 1. Compiling with gcc -pedantic found a couple of places where casts were
3578 needed, and a string in dftables.c that was longer than standard compilers are
3579 required to support.
3581 2. Compiling with Sun's compiler found a few more places where the code could
3582 be tidied up in order to avoid warnings.
3584 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3585 first of these names is deprecated in the latest Autoconf in favour of the name
3586 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3587 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3588 analogy I have changed it to CFLAGS_FOR_BUILD.
3590 4. Added -no-undefined to the linking command in the Makefile, because this is
3591 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3592 linking step for the pcreposix library.
3594 5. PCRE was failing to diagnose the case of two named groups with the same
3595 name.
3597 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3598 literal character that is needed in the subject for a match, and scans along to
3599 ensure that it is present before embarking on the full matching process. This
3600 saves time in cases of nested unlimited repeats that are never going to match.
3601 Problem: the scan can take a lot of time if the subject is very long (e.g.
3602 megabytes), thus penalizing straightforward matches. It is now done only if the
3603 amount of subject to be scanned is less than 1000 bytes.
3605 7. A lesser problem with the same optimization is that it was recording the
3606 first character of an anchored pattern as "needed", thus provoking a search
3607 right along the subject, even when the first match of the pattern was going to
3608 fail. The "needed" character is now not set for anchored patterns, unless it
3609 follows something in the pattern that is of non-fixed length. Thus, it still
3610 fulfils its original purpose of finding quick non-matches in cases of nested
3611 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3614 Version 4.0 17-Feb-03
3615 ---------------------
3617 1. If a comment in an extended regex that started immediately after a meta-item
3618 extended to the end of string, PCRE compiled incorrect data. This could lead to
3619 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3621 2. Moved to autoconf 2.53 and libtool 1.4.2.
3623 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3624 the special perltest8 script is no longer needed - all the tests can be run
3625 from a single perltest script.
3627 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3628 by \s. It has now been removed in PCRE. This means it isn't recognized as
3629 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3630 class [:space:] *does* include VT, thereby creating a mess.
3632 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3633 space and tab.
3635 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3636 its new features into the main test script, reducing the number of scripts.
3638 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3639 were backward compatible, and made the (?i) apply to the whole pattern, as if
3640 /i were given. Now it behaves more logically, and applies the option setting
3641 only to what follows. PCRE has been changed to follow suit. However, if it
3642 finds options settings right at the start of the pattern, it extracts them into
3643 the global options, as before. Thus, they show up in the info data.
3645 8. Added support for the \Q...\E escape sequence. Characters in between are
3646 treated as literals. This is slightly different from Perl in that $ and @ are
3647 also handled as literals inside the quotes. In Perl, they will cause variable
3648 interpolation. Note the following examples:
3650 Pattern PCRE matches Perl matches
3652 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3653 \Qabc\$xyz\E abc\$xyz abc\$xyz
3654 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3656 For compatibility with Perl, \Q...\E sequences are recognized inside character
3657 classes as well as outside them.
3659 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3660 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3661 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3662 signed/unsigned warnings.
3664 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3665 option for pcretest, so I've replaced it by a simple function that does just
3666 that job.
3668 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3669 "pcregrep -".
3671 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3672 Java package. This provides some syntactic sugar for simple cases of what my
3673 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3674 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3675 item, you can use this simplified notation. Note that only makes sense with
3676 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3677 greediness, whatever the setting of the PCRE_UNGREEDY option.
3679 13. A change of greediness default within a pattern was not taking effect at
3680 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3681 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3682 was abstracted outside.
3684 14. PCRE now supports the \G assertion. It is true when the current matching
3685 position is at the start point of the match. This differs from \A when the
3686 starting offset is non-zero. Used with the /g option of pcretest (or similar
3687 code), it works in the same way as it does for Perl's /g option. If all
3688 alternatives of a regex begin with \G, the expression is anchored to the start
3689 match position, and the "anchored" flag is set in the compiled expression.
3691 15. Some bugs concerning the handling of certain option changes within patterns
3692 have been fixed. These applied to options other than (?ims). For example,
3693 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3694 been the other way round. Some of this was related to change 7 above.
3696 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3697 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3698 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3699 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3701 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3702 mode. Unlike ".", it always matches newline, whatever the setting of
3703 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3704 assertions. Perl allows it, but it doesn't (in general) work because it can't
3705 calculate the length of the lookbehind. At least, that's the case for Perl
3706 5.8.0 - I've been told they are going to document that it doesn't work in
3707 future.
3709 18. Added an error diagnosis for escapes that PCRE does not support: these are
3710 \L, \l, \N, \P, \p, \U, \u, and \X.
3712 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3713 reading past the end of the pattern in cases such as /[abcd/.
3715 20. PCRE was getting more memory than necessary for patterns with classes that
3716 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3718 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3719 compiling PCRE for use with Virtual Pascal.
3721 22. Small fix to the Makefile to make it work properly if the build is done
3722 outside the source tree.
3724 23. Added a new extension: a condition to go with recursion. If a conditional
3725 subpattern starts with (?(R) the "true" branch is used if recursion has
3726 happened, whereas the "false" branch is used only at the top level.
3728 24. When there was a very long string of literal characters (over 255 bytes
3729 without UTF support, over 250 bytes with UTF support), the computation of how
3730 much memory was required could be incorrect, leading to segfaults or other
3731 strange effects.
3733 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3734 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3735 there was a subsequent back reference to those brackets. This meant that, for
3736 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3737 possible to check for precisely this case. All we can do is abandon the
3738 optimization if .* occurs inside capturing brackets when there are any back
3739 references whatsoever. (See below for a better fix that came later.)
3741 26. The handling of the optimization for finding the first character of a
3742 non-anchored pattern, and for finding a character that is required later in the
3743 match were failing in some cases. This didn't break the matching; it just
3744 failed to optimize when it could. The way this is done has been re-implemented.
3746 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3748 28. Added a new feature that provides some of the functionality that Perl
3749 provides with (?{...}). The facility is termed a "callout". The way it is done
3750 in PCRE is for the caller to provide an optional function, by setting
3751 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3752 global variable. By default it is unset, which disables all calling out. To get
3753 the function called, the regex must include (?C) at appropriate points. This
3754 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3755 This provides a means of identifying different callout points. When PCRE
3756 reaches such a point in the regex, if pcre_callout has been set, the external
3757 function is called. It is provided with data in a structure called
3758 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3759 matching continues; if it returns a non-zero value, the match at the current
3760 point fails. However, backtracking will occur if possible. [This was changed
3761 later and other features added - see item 49 below.]
3763 29. pcretest is upgraded to test the callout functionality. It provides a
3764 callout function that displays information. By default, it shows the start of
3765 the match and the current position in the text. There are some new data escapes
3766 to vary what happens:
3768 \C+ in addition, show current contents of captured substrings
3769 \C- do not supply a callout function
3770 \C!n return 1 when callout number n is reached
3771 \C!n!m return 1 when callout number n is reached for the mth time
3773 30. If pcregrep was called with the -l option and just a single file name, it
3774 output "<stdin>" if a match was found, instead of the file name.
3776 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3777 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3778 pcre_exec(). This saves a malloc/free per call. The default value of
3779 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3780 when configuring.
3782 32. The default maximum size of a compiled pattern is 64K. There have been a
3783 few cases of people hitting this limit. The code now uses macros to handle the
3784 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3785 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3786 configuring. Tests 2 and 5 work only with 2-byte links because they output
3787 debugging information about compiled patterns.
3789 33. Internal code re-arrangements:
3791 (a) Moved the debugging function for printing out a compiled regex into
3792 its own source file (printint.c) and used #include to pull it into
3793 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3794 separate copies.
3796 (b) Defined the list of op-code names for debugging as a macro in
3797 internal.h so that it is next to the definition of the opcodes.
3799 (c) Defined a table of op-code lengths for simpler skipping along compiled
3800 code. This is again a macro in internal.h so that it is next to the
3801 definition of the opcodes.
3803 34. Added support for recursive calls to individual subpatterns, along the
3804 lines of Robin Houston's patch (but implemented somewhat differently).
3806 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3807 allow it to read and process whole directories in Win32. This code was
3808 contributed by Lionel Fourquaux; it has not been tested by me.
3810 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3811 used to name a group. Names consist of alphanumerics and underscores, and must
3812 be unique. Back references use the syntax (?P=name) and recursive calls use
3813 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3814 numbers. The function pcre_fullinfo() can be used after compilation to extract
3815 a name/number map. There are three relevant calls:
3817 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3818 PCRE_INFO_NAMECOUNT yields the number of entries
3819 PCRE_INFO_NAMETABLE yields a pointer to the map.
3821 The map is a vector of fixed-size entries. The size of each entry depends on
3822 the length of the longest name used. The first two bytes of each entry are the
3823 group number, most significant byte first. There follows the corresponding
3824 name, zero terminated. The names are in alphabetical order.
3826 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3827 case instead of 255. Making it the same both with and without UTF-8 support
3828 means that the same test output works with both.
3830 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3831 calling malloc() with a zero argument.
3833 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3834 optimization. I've improved things by keeping a bitmap of backreferences with
3835 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3836 fact referenced, the optimization can be applied. It is unlikely that a
3837 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3838 the match to follow \n) will appear inside brackets with a number greater than
3839 31, but if it does, any back reference > 31 suppresses the optimization.
3841 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3842 of disabling numbered capturing parentheses. Any opening parenthesis that is
3843 not followed by ? behaves as if it were followed by ?: but named parentheses
3844 can still be used for capturing (and they will acquire numbers in the usual
3845 way).
3847 41. Redesigned the return codes from the match() function into yes/no/error so
3848 that errors can be passed back from deep inside the nested calls. A malloc
3849 failure while inside a recursive subpattern call now causes the
3850 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3852 42. It is now possible to set a limit on the number of times the match()
3853 function is called in a call to pcre_exec(). This facility makes it possible to
3854 limit the amount of recursion and backtracking, though not in a directly
3855 obvious way, because the match() function is used in a number of different
3856 circumstances. The count starts from zero for each position in the subject
3857 string (for non-anchored patterns). The default limit is, for compatibility, a
3858 large number, namely 10 000 000. You can change this in two ways:
3860 (a) When configuring PCRE before making, you can use --with-match-limit=n
3861 to set a default value for the compiled library.
3863 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3864 a different value is set. See 45 below.
3866 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3868 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3869 of things that can be changed at compile time. The first argument specifies
3870 what is wanted and the second points to where the information is to be placed.
3871 The current list of available information is:
3875 The output is an integer that is set to one if UTF-8 support is available;
3876 otherwise it is set to zero.
3880 The output is an integer that it set to the value of the code that is used for
3881 newline. It is either LF (10) or CR (13).
3885 The output is an integer that contains the number of bytes used for internal
3886 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3890 The output is an integer that contains the threshold above which the POSIX
3891 interface uses malloc() for output vectors. See item 31 above.
3895 The output is an unsigned integer that contains the default limit of the number
3896 of match() calls in a pcre_exec() execution. See 42 above.
3898 44. pcretest has been upgraded by the addition of the -C option. This causes it
3899 to extract all the available output from the new pcre_config() function, and to
3900 output it. The program then exits immediately.
3902 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3903 order to support additional features. One way would have been to define
3904 pcre_exec2() (for example) with extra arguments, but this would not have been
3905 extensible, and would also have required all calls to the original function to
3906 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3907 is used for passing in "extra" data from pcre_study().
3909 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3910 contains the following fields:
3912 flags a bitmap indicating which of the following fields are set
3913 study_data opaque data from pcre_study()
3914 match_limit a way of specifying a limit on match() calls for a specific
3915 call to pcre_exec()
3916 callout_data data for callouts (see 49 below)
3918 The flag bits are also defined in pcre.h, and are
3924 The pcre_study() function now returns one of these new pcre_extra blocks, with
3925 the actual study data pointed to by the study_data field, and the
3926 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3927 before. That is, this change is entirely upwards-compatible and requires no
3928 change to existing code.
3930 If you want to pass in additional data to pcre_exec(), you can either place it
3931 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3932 block.
3934 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3935 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3936 times with different match limits, until it finds the minimum value needed for
3937 pcre_exec() to complete. The value is then output. This can be instructive; for
3938 most simple matches the number is quite small, but for pathological cases it
3939 gets very large very quickly.
3941 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3942 returns the size of the data block pointed to by the study_data field in a
3943 pcre_extra block, that is, the value that was passed as the argument to
3944 pcre_malloc() when PCRE was getting memory in which to place the information
3945 created by pcre_study(). The fourth argument should point to a size_t variable.
3946 pcretest has been extended so that this information is shown after a successful
3947 pcre_study() call when information about the compiled regex is being displayed.
3949 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3950 because what follows is always an absolute path. (Later: it turns out that this
3951 is more than cosmetic for MinGW, because it doesn't like empty path
3952 components.)
3954 49. Some changes have been made to the callout feature (see 28 above):
3956 (i) A callout function now has three choices for what it returns:
3958 0 => success, carry on matching
3959 > 0 => failure at this point, but backtrack if possible
3960 < 0 => serious error, return this value from pcre_exec()
3962 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3963 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3964 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3965 use by callout functions. It will never be used by PCRE itself.
3967 (ii) The pcre_extra structure (see 45 above) has a void * field called
3968 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3969 pcre_callout_block structure has a field of the same name. The contents of
3970 the field passed in the pcre_extra structure are passed to the callout
3971 function in the corresponding field in the callout block. This makes it
3972 easier to use the same callout-containing regex from multiple threads. For
3973 testing, the pcretest program has a new data escape
3975 \C*n pass the number n (may be negative) as callout_data
3977 If the callout function in pcretest receives a non-zero value as
3978 callout_data, it returns that value.
3980 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3981 there were some redundant $(CFLAGS) in commands that are now specified as
3982 $(LINK), which already includes $(CFLAGS).
3984 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3985 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3986 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3987 one-byte characters throughout. Note that case-insensitive matching applies
3988 only to characters whose values are less than 256. PCRE doesn't support the
3989 notion of cases for higher-valued characters.
3991 (i) A character class whose characters are all within 0-255 is handled as
3992 a bit map, and the map is inverted for negative classes. Previously, a
3993 character > 255 always failed to match such a class; however it should
3994 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3996 (ii) A negated character class with a single character < 255 is coded as
3997 "not this character" (OP_NOT). This wasn't working properly when the test
3998 character was multibyte, either singly or repeated.
4000 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4001 mode, for example: \x{100}{2,3}.
4003 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4004 singly or repeated) now correctly test multibyte characters. However,
4005 PCRE doesn't recognize any characters with values greater than 255 as
4006 digits, spaces, or word characters. Such characters always match \D, \S,
4007 and \W, and never match \d, \s, or \w.
4009 (v) Classes may now contain characters and character ranges with values
4010 greater than 255. For example: [ab\x{100}-\x{400}].
4012 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4013 PCRE in UTF-8 mode.
4015 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4016 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4017 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4018 value.)
4020 53. The single man page has become too large. I have therefore split it up into
4021 a number of separate man pages. These also give rise to individual HTML pages;
4022 these are now put in a separate directory, and there is an index.html page that
4023 lists them all. Some hyperlinking between the pages has been installed.
4025 54. Added convenience functions for handling named capturing parentheses.
4027 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4028 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4029 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4030 are faulted.
4032 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4033 calling configure. These values are used when compiling the dftables.c program
4034 which is run to generate the source of the default character tables. They
4035 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4036 you will need to set these values.
4038 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4041 Version 3.9 02-Jan-02
4042 ---------------------
4044 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4046 2. If --disable-static was given, the building process failed when trying to
4047 build pcretest and pcregrep. (For some reason it was using libtool to compile
4048 them, which is not right, as they aren't part of the library.)
4051 Version 3.8 18-Dec-01
4052 ---------------------
4054 1. The experimental UTF-8 code was completely screwed up. It was packing the
4055 bytes in the wrong order. How dumb can you get?
4058 Version 3.7 29-Oct-01
4059 ---------------------
4061 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4062 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4063 this didn't happen under Solaris 8, where I normally test things.
4065 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4066 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4067 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4068 this fix an hour or so after the initial 3.7 release.)
4071 Version 3.6 23-Oct-01
4072 ---------------------
4074 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4075 offsets passed as NULL with zero offset count.
4077 2. The config.guess and config.sub files had not been updated when I moved to
4078 the latest autoconf.
4081 Version 3.5 15-Aug-01
4082 ---------------------
4084 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4085 had been forgotten.
4087 2. By using declared but undefined structures, we can avoid using "void"
4088 definitions in pcre.h while keeping the internal definitions of the structures
4089 private.
4091 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4092 user point of view, this means that both static and shared libraries are built
4093 by default, but this can be individually controlled. More of the work of
4094 handling this static/shared cases is now inside libtool instead of PCRE's make
4095 file.
4097 4. The pcretest utility is now installed along with pcregrep because it is
4098 useful for users (to test regexs) and by doing this, it automatically gets
4099 relinked by libtool. The documentation has been turned into a man page, so
4100 there are now .1, .txt, and .html versions in /doc.
4102 5. Upgrades to pcregrep:
4103 (i) Added long-form option names like gnu grep.
4104 (ii) Added --help to list all options with an explanatory phrase.
4105 (iii) Added -r, --recursive to recurse into sub-directories.
4106 (iv) Added -f, --file to read patterns from a file.
4108 6. pcre_exec() was referring to its "code" argument before testing that
4109 argument for NULL (and giving an error if it was NULL).
4111 7. Upgraded Makefile.in to allow for compiling in a different directory from
4112 the source directory.
4114 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4115 options bits, the pointer it was passed was to an int instead of to an unsigned
4116 long int. This mattered only on 64-bit systems.
4118 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4119 generated) instead of pcre.in, which it its source. Also made the same change
4120 in several of the .c files.
4122 10. A new release of gcc defines printf() as a macro, which broke pcretest
4123 because it had an ifdef in the middle of a string argument for printf(). Fixed
4124 by using separate calls to printf().
4126 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4127 script, to force use of CR or LF instead of \n in the source. On non-Unix
4128 systems, the value can be set in config.h.
4130 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4131 absolute limit. Changed the text of the error message to make this clear, and
4132 likewise updated the man page.
4134 13. The limit of 99 on the number of capturing subpatterns has been removed.
4135 The new limit is 65535, which I hope will not be a "real" limit.
4138 Version 3.4 22-Aug-00
4139 ---------------------
4141 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4143 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4146 Version 3.3 01-Aug-00
4147 ---------------------
4149 1. If an octal character was given, but the value was greater than \377, it
4150 was not getting masked to the least significant bits, as documented. This could
4151 lead to crashes in some systems.
4153 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4154 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4156 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4157 These just pass their arguments on to (pcre_free)(), but they are provided
4158 because some uses of PCRE bind it to non-C systems that can call its functions,
4159 but cannot call free() or pcre_free() directly.
4161 4. Add "make test" as a synonym for "make check". Corrected some comments in
4162 the Makefile.
4164 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4165 Makefile.
4167 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4168 command called pgrep for grepping around the active processes.
4170 7. Added the beginnings of support for UTF-8 character strings.
4172 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4173 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4174 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4175 out for the ar command.)
4178 Version 3.2 12-May-00
4179 ---------------------
4181 This is purely a bug fixing release.
4183 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4184 of ZA. This was just one example of several cases that could provoke this bug,
4185 which was introduced by change 9 of version 2.00. The code for breaking
4186 infinite loops after an iteration that matches an empty string was't working
4187 correctly.
4189 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4190 when matched against abbab (for example). After matching an empty string, it
4191 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4192 caused it to match further down the string than it should.
4194 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4195 was there because it doesn't seem to be needed, and it causes trouble on some
4196 systems, as it is not a Standard C header. It has been removed.
4198 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4199 were reported on the Macintosh. The changes were from
4201 while ((c = *(++ptr)) != 0 && c != '\n');
4202 to
4203 while ((c = *(++ptr)) != 0 && c != '\n') ;
4205 Totally extraordinary, but if that's what it takes...
4207 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4208 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4209 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4210 assumes the way PCRE uses memmove() (always moving upwards).
4212 6. PCRE is being used in one environment where strchr() is not available. There
4213 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4214 faster code anyway.
4217 Version 3.1 09-Feb-00
4218 ---------------------
4220 The only change in this release is the fixing of some bugs in Makefile.in for
4221 the "install" target:
4223 (1) It was failing to install pcreposix.h.
4225 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4228 Version 3.0 01-Feb-00
4229 ---------------------
4231 1. Add support for the /+ modifier to perltest (to output $` like it does in
4232 pcretest).
4234 2. Add support for the /g modifier to perltest.
4236 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4237 matches null strings.
4239 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4240 pattern. Perl treats empty patterns specially - it reuses the most recent
4241 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4242 effect.
4244 5. The POSIX interface was broken in that it was just handing over the POSIX
4245 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4246 required a bigger vector, with some working space on the end. This means that
4247 the POSIX wrapper now has to get and free some memory, and copy the results.
4249 6. Added some simple autoconf support, placing the test data and the
4250 documentation in separate directories, re-organizing some of the
4251 information files, and making it build pcre-config (a GNU standard). Also added
4252 libtool support for building PCRE as a shared library, which is now the
4253 default.
4255 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4256 09 are not valid octal constants. Single digits will be used for minor values
4257 less than 10.
4259 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4260 existing programs that set these in the POSIX interface can use PCRE without
4261 modification.
4263 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4264 return all that pcre_info() returns, plus additional data. The pcre_info()
4265 function is retained for compatibility, but is considered to be obsolete.
4267 10. Added experimental recursion feature (?R) to handle one common case that
4268 Perl 5.6 will be able to do with (?p{...}).
4270 11. Added support for POSIX character classes like [:alpha:], which Perl is
4271 adopting.
4274 Version 2.08 31-Aug-99
4275 ----------------------
4277 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4278 trying to match at the startoffset position, but instead was moving forward to
4279 the next newline as if a previous match had failed.
4281 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4282 and could get into a loop if a null string was matched other than at the start
4283 of the subject.
4285 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4286 be distinguished at compile time, and for completeness also added PCRE_DATE.
4288 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4289 in GnuWin32 environments.
4292 Version 2.07 29-Jul-99
4293 ----------------------
4295 1. The documentation is now supplied in plain text form and HTML as well as in
4296 the form of man page sources.
4298 2. C++ compilers don't like assigning (void *) values to other pointer types.
4299 In particular this affects malloc(). Although there is no problem in Standard
4300 C, I've put in casts to keep C++ compilers happy.
4302 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4303 should be (const char *).
4305 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4306 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4307 However, I haven't made this a standard facility. The documentation doesn't
4308 mention it, and the Makefile doesn't support it.
4310 5. The Makefile now contains an "install" target, with editable destinations at
4311 the top of the file. The pcretest program is not installed.
4313 6. pgrep -V now gives the PCRE version number and date.
4315 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4316 causing the entire string to be ignored, instead of just the last character.
4318 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4319 non-matching string, it can take a very, very long time, even for strings of
4320 quite modest length, because of the nested recursion. PCRE now does better in
4321 some of these cases. It does this by remembering the last required literal
4322 character in the pattern, and pre-searching the subject to ensure it is present
4323 before running the real match. In other words, it applies a heuristic to detect
4324 some types of certain failure quickly, and in the above example, if presented
4325 with a string that has no trailing " it gives "no match" very quickly.
4327 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4328 other alternatives are tried instead.
4331 Version 2.06 09-Jun-99
4332 ----------------------
4334 1. Change pcretest's output for amount of store used to show just the code
4335 space, because the remainder (the data block) varies in size between 32-bit and
4336 64-bit systems.
4338 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4339 start matching at. This allows lookbehinds to work when searching for multiple
4340 occurrences in a string.
4342 3. Added additional options to pcretest for testing multiple occurrences:
4344 /+ outputs the rest of the string that follows a match
4345 /g loops for multiple occurrences, using the new startoffset argument
4346 /G loops for multiple occurrences by passing an incremented pointer
4348 4. PCRE wasn't doing the "first character" optimization for patterns starting
4349 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4350 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4351 the letter 'x'. On long subject strings, this gives a significant speed-up.
4354 Version 2.05 21-Apr-99
4355 ----------------------
4357 1. Changed the type of magic_number from int to long int so that it works
4358 properly on 16-bit systems.
4360 2. Fixed a bug which caused patterns starting with .* not to work correctly
4361 when the subject string contained newline characters. PCRE was assuming
4362 anchoring for such patterns in all cases, which is not correct because .* will
4363 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4364 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4365 must be retried after every newline in the subject.
4368 Version 2.04 18-Feb-99
4369 ----------------------
4371 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4372 computation of the store needed to hold the pattern was incorrect (too large).
4373 If such patterns were nested a few deep, this could multiply and become a real
4374 problem.
4376 2. Added /M option to pcretest to show the memory requirement of a specific
4377 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4379 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4380 compiled in such a way that the backtracking after subsequent failure was
4381 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4382 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4385 Version 2.03 02-Feb-99
4386 ----------------------
4388 1. Fixed typo and small mistake in man page.
4390 2. Added 4th condition (GPL supersedes if conflict) and created separate
4391 LICENCE file containing the conditions.
4393 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4394 Perl, that is the internal \ allows the delimiter to be included in the
4395 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4396 the final delimiter, add \ to the end of the pattern (to test the error).
4398 4. Added the convenience functions for extracting substrings after a successful
4399 match. Updated pcretest to make it able to test these functions.
4402 Version 2.02 14-Jan-99
4403 ----------------------
4405 1. Initialized the working variables associated with each extraction so that
4406 their saving and restoring doesn't refer to uninitialized store.
4408 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4409 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4410 fix the problem.
4412 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4413 calls, and wasn't printing the correct value for compiling calls. Increased the
4414 default value of LOOPREPEAT, and the number of significant figures in the
4415 times.
4417 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4419 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4420 a building problem on Windows NT with a FAT file system.
4423 Version 2.01 21-Oct-98
4424 ----------------------
4426 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4427 to character tables built by pcre_maketables() in the current locale. If NULL
4428 is passed, the default tables are used.
4431 Version 2.00 24-Sep-98
4432 ----------------------
4434 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4435 it any more.
4437 2. Allow quantification of (?>) groups, and make it work correctly.
4439 3. The first character computation wasn't working for (?>) groups.
4441 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4442 end of the subject) and add 5.005's \z, which really does match only at the
4443 very end of the subject.
4445 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4447 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4448 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4449 localized options. All options to pcre_study() were also removed.
4451 7. Add other new features from 5.005:
4453 $(?<= positive lookbehind
4454 $(?<! negative lookbehind
4455 (?imsx-imsx) added the unsetting capability
4456 such a setting is global if at outer level; local otherwise
4457 (?imsx-imsx:) non-capturing groups with option setting
4458 (?(cond)re|re) conditional pattern matching
4460 A backreference to itself in a repeated group matches the previous
4461 captured string.
4463 8. General tidying up of studying (both automatic and via "study")
4464 consequential on the addition of new assertions.
4466 9. As in 5.005, unlimited repeated groups that could match an empty substring
4467 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4468 runtime if any iteration does actually match an empty substring.
4470 10. Include the RunTest script in the distribution.
4472 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4473 discrepancies, some of which were old and were also with respect to 5.004. They
4474 have now been fixed.
4477 Version 1.09 28-Apr-98
4478 ----------------------
4480 1. A negated single character class followed by a quantifier with a minimum
4481 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4482 program crashes, or just wrong answers. This did not apply to negated classes
4483 containing more than one character, or to minima other than one.
4486 Version 1.08 27-Mar-98
4487 ----------------------
4489 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4491 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4492 latter must appear before anything that relies on it in the pattern.
4495 Version 1.07 16-Feb-98
4496 ----------------------
4498 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4499 repeat of a potentially empty string).
4502 Version 1.06 23-Jan-98
4503 ----------------------
4505 1. Added Markus Oberhumer's little patches for C++.
4507 2. Literal strings longer than 255 characters were broken.
4510 Version 1.05 23-Dec-97
4511 ----------------------
4513 1. Negated character classes containing more than one character were failing if
4514 PCRE_CASELESS was set at run time.
4517 Version 1.04 19-Dec-97
4518 ----------------------
4520 1. Corrected the man page, where some "const" qualifiers had been omitted.
4522 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4523 input syntax.
4525 3. Fixed memory leak which occurred when a regex with back references was
4526 matched with an offsets vector that wasn't big enough. The temporary memory
4527 that is used in this case wasn't being freed if the match failed.
4529 4. Tidied pcretest to ensure it frees memory that it gets.
4531 5. Temporary memory was being obtained in the case where the passed offsets
4532 vector was exactly big enough.
4534 6. Corrected definition of offsetof() from change 5 below.
4536 7. I had screwed up change 6 below and broken the rules for the use of
4537 setjmp(). Now fixed.
4540 Version 1.03 18-Dec-97
4541 ----------------------
4543 1. A erroneous regex with a missing opening parenthesis was correctly
4544 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4545 on some systems.
4547 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4548 it was reported that one broken compiler failed on the former because "code" is
4549 also an independent variable.
4551 3. The erroneous regex a[]b caused an array overrun reference.
4553 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4554 fail on data ending with that character. (It was going on too far, and checking
4555 the next character, typically a binary zero.) This was specific to the
4556 optimized code for single-character negative classes.
4558 5. Added a contributed patch from the TIN world which does the following:
4560 + Add an undef for memmove, in case the the system defines a macro for it.
4562 + Add a definition of offsetof(), in case there isn't one. (I don't know
4563 the reason behind this - offsetof() is part of the ANSI standard - but
4564 it does no harm).
4566 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4567 most of the places where whitespace preceded '#'. I have given up and
4568 allowed the remaining 2 cases to be at the margin.
4570 + Rename some variables in pcre to eliminate shadowing. This seems very
4571 pedantic, but does no harm, of course.
4573 6. Moved the call to setjmp() into its own function, to get rid of warnings
4574 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4576 7. Constructs such as \d{8,} were compiling into the equivalent of
4577 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4578 outcome, but in this particular case used more store than had been allocated,
4579 which caused the bug to be discovered because it threw up an internal error.
4581 8. The debugging code in both pcre and pcretest for outputting the compiled
4582 form of a regex was going wrong in the case of back references followed by
4583 curly-bracketed repeats.
4586 Version 1.02 12-Dec-97
4587 ----------------------
4589 1. Typos in pcre.3 and comments in the source fixed.
4591 2. Applied a contributed patch to get rid of places where it used to remove
4592 'const' from variables, and fixed some signed/unsigned and uninitialized
4593 variable warnings.
4595 3. Added the "runtest" target to Makefile.
4597 4. Set default compiler flag to -O2 rather than just -O.
4600 Version 1.01 19-Nov-97
4601 ----------------------
4603 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4604 like /([ab]*)*/, that is, for classes with more than one character in them.
4606 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4607 as /((?>a*))*/ (a PCRE_EXTRA facility).
4610 Version 1.00 18-Nov-97
4611 ----------------------
4613 1. Added compile-time macros to support systems such as SunOS4 which don't have
4614 memmove() or strerror() but have other things that can be used instead.
4616 2. Arranged that "make clean" removes the executables.
4619 Version 0.99 27-Oct-97
4620 ----------------------
4622 1. Fixed bug in code for optimizing classes with only one character. It was
4623 initializing a 32-byte map regardless, which could cause it to run off the end
4624 of the memory it had got.
4626 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4629 Version 0.98 22-Oct-97
4630 ----------------------
4632 1. Fixed bug in code for handling temporary memory usage when there are more
4633 back references than supplied space in the ovector. This could cause segfaults.
4636 Version 0.97 21-Oct-97
4637 ----------------------
4639 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4641 2. Optimized negated single characters not to use a bit map.
4643 3. Brought error texts together as macro definitions; clarified some of them;
4644 fixed one that was wrong - it said "range out of order" when it meant "invalid
4645 escape sequence".
4647 4. Changed some char * arguments to const char *.
4649 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4651 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4652 pcretest.
4655 Version 0.96 16-Oct-97
4656 ----------------------
4658 1. Added a simple "pgrep" utility to the distribution.
4660 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4661 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4662 where "ddd" means "one or more decimal digits".
4664 3. Fixed serious bug. If a pattern had a back reference, but the call to
4665 pcre_exec() didn't supply a large enough ovector to record the related
4666 identifying subpattern, the match always failed. PCRE now remembers the number
4667 of the largest back reference, and gets some temporary memory in which to save
4668 the offsets during matching if necessary, in order to ensure that
4669 backreferences always work.
4671 4. Increased the compatibility with Perl in a number of ways:
4673 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4674 to request this handling. The option can be set at compile or exec time.
4676 (b) $ matches before a terminating newline by default; an option
4677 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4678 mode). The option can be set at compile or exec time.
4680 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4681 the same as Perl's. If the decimal number it represents is less than 10
4682 or there aren't that many previous left capturing parentheses, an octal
4683 escape is read. Inside a character class, it's always an octal escape,
4684 even if it is a single digit.
4686 (d) An escaped but undefined alphabetic character is taken as a literal,
4687 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4688 escapes.
4690 (e) {0} is now permitted. (The previous item is removed from the compiled
4691 pattern).
4693 5. Changed all the names of code files so that the basic parts are no longer
4694 than 10 characters, and abolished the teeny "globals.c" file.
4696 6. Changed the handling of character classes; they are now done with a 32-byte
4697 bit map always.
4699 7. Added the -d and /D options to pcretest to make it possible to look at the
4700 internals of compilation without having to recompile pcre.
4703 Version 0.95 23-Sep-97
4704 ----------------------
4706 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4707 \x20 at the start of a run of normal characters. These were being treated as
4708 real characters, instead of the source characters being re-checked.
4711 Version 0.94 18-Sep-97
4712 ----------------------
4714 1. The functions are now thread-safe, with the caveat that the global variables
4715 containing pointers to malloc() and free() or alternative functions are the
4716 same for all threads.
4718 2. Get pcre_study() to generate a bitmap of initial characters for non-
4719 anchored patterns when this is possible, and use it if passed to pcre_exec().
4722 Version 0.93 15-Sep-97
4723 ----------------------
4725 1. /(b)|(:+)/ was computing an incorrect first character.
4727 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4728 but not actually doing anything yet.
4730 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4731 as Perl does (e.g. [-az] or [az-]).
4733 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4734 all possible positions.
4736 5. Split up into different modules to avoid including unneeded functions in a
4737 compiled binary. However, compile and exec are still in one module. The "study"
4738 function is split off.
4740 6. The character tables are now in a separate module whose source is generated
4741 by an auxiliary program - but can then be edited by hand if required. There are
4742 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4743 toupper() in the code.
4745 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4746 make them global. Abolish the function for setting them, as the caller can now
4747 set them directly.
4750 Version 0.92 11-Sep-97
4751 ----------------------
4753 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4754 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4756 2. Caseless matching was not working in character classes if the characters in
4757 the pattern were in upper case.
4759 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4761 4. Make PCRE_ANCHORED public and accept as a compile option.
4763 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4764 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4765 pass them.
4767 6. Give an error if bad option bits passed at compile or run time.
4769 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4770 pcretest to cause it to pass that flag.
4772 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4773 options, and the first character, if set.
4775 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4778 Version 0.91 10-Sep-97
4779 ----------------------
4781 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4782 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4784 2. PCRE was looping on encountering an indefinitely repeated back reference to
4785 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4786 Perl does - treats the match as successful.
4788 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5