ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 970 - (show annotations)
Sun May 27 05:28:19 2012 UTC (8 years ago) by zherczeg
File size: 231013 byte(s)
Error occurred while calculating annotation data.
Variable renamings in the PCRE-JIT compiler
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.
124 Version 8.30 04-February-2012
125 -----------------------------
127 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
128 name is defined in ctype.h.
130 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
131 only in quite long subpatterns.
133 3. Removed the function pcre_info(), which has been obsolete and deprecated
134 since it was replaced by pcre_fullinfo() in February 2000.
136 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
137 match a (*MARK), and the match failed at the start of the subject, a
138 reference to memory before the start of the subject could occur. This bug
139 was introduced by fix 17 of release 8.21.
141 5. A reference to an unset group with zero minimum repetition was giving
142 totally wrong answers (in non-JavaScript-compatibility mode). For example,
143 /(another)?(\1?)test/ matched against "hello world test". This bug was
144 introduced in release 8.13.
146 6. Add support for 16-bit character strings (a large amount of work involving
147 many changes and refactorings).
149 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
150 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
151 from a file.
153 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
154 rounding is not applied in this particular case).
156 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
157 if they appear, or are escaped, in patterns.
159 10. Get rid of a number of -Wunused-but-set-variable warnings.
161 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
162 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
163 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
164 also returns the mark "x". This bug applied to capturing parentheses,
165 non-capturing parentheses, and atomic parentheses. It also applied to some
166 assertions.
168 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
169 information out of configure.ac instead of relying on pcre.h.generic, which
170 is not stored in the repository.
172 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
173 -lreadline.
175 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
177 15. Applied Graycode's patch to put the top-level frame on the stack rather
178 than the heap when not using the stack for recursion. This gives a
179 performance improvement in many cases when recursion is not deep.
181 16. Experimental code added to "pcretest -C" to output the stack frame size.
184 Version 8.21 12-Dec-2011
185 ------------------------
187 1. Updating the JIT compiler.
189 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
190 are added as well.
192 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
193 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
194 calling _pcre_jit_exec. Some extra comments are added.
196 4. (*MARK) settings inside atomic groups that do not contain any capturing
197 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
198 was introduced by change 18 for 8.20.
200 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
201 ECMA-262 standard.
203 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
204 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
205 This bug was probably introduced by change 9 of 8.13.
207 7. While fixing 6 above, I noticed that a number of other items were being
208 incorrectly rejected as "not fixed length". This arose partly because newer
209 opcodes had not been added to the fixed-length checking code. I have (a)
210 corrected the bug and added tests for these items, and (b) arranged for an
211 error to occur if an unknown opcode is encountered while checking for fixed
212 length instead of just assuming "not fixed length". The items that were
213 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
214 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
215 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
217 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
218 being incorrectly compiled and would have given unpredicatble results.
220 9. A possessively repeated subpattern with minimum repeat count greater than
221 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
222 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
223 the first (A) could occur when it should not.
225 10. Add a cast and remove a redundant test from the code.
227 11. JIT should use pcre_malloc/pcre_free for allocation.
229 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
230 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
231 is anything other than /usr, -L is still shown).
233 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
235 14. Perl does not support \N without a following name in a [] class; PCRE now
236 also gives an error.
238 15. If a forward reference was repeated with an upper limit of around 2000,
239 it caused the error "internal error: overran compiling workspace". The
240 maximum number of forward references (including repeats) was limited by the
241 internal workspace, and dependent on the LINK_SIZE. The code has been
242 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
243 the default depends on LINK_SIZE. There is a new upper limit (for safety)
244 of around 200,000 forward references. While doing this, I also speeded up
245 the filling in of repeated forward references.
247 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
248 incorrectly expecting the subject to contain another "a" after the start.
250 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
251 in the match, the SKIP should be ignored. This was not happening; instead
252 the SKIP was being treated as NOMATCH. For patterns such as
253 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
254 tested.
256 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
257 now much more compatible with Perl, in particular in cases where the result
258 is a non-match for a non-anchored pattern. For example, if
259 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
260 "m", where previously it did not return a name. A side effect of this
261 change is that for partial matches, the last encountered mark name is
262 returned, as for non matches. A number of tests that were previously not
263 Perl-compatible have been moved into the Perl-compatible test files. The
264 refactoring has had the pleasing side effect of removing one argument from
265 the match() function, thus reducing its stack requirements.
267 19. If the /S+ option was used in pcretest to study a pattern using JIT,
268 subsequent uses of /S (without +) incorrectly behaved like /S+.
270 21. Retrieve executable code size support for the JIT compiler and fixing
271 some warnings.
273 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
274 not work when the shorter character appeared right at the end of the
275 subject string.
277 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
278 systems.
280 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
281 output it when the /M option is used in pcretest.
283 25. The CheckMan script was not being included in the distribution. Also, added
284 an explicit "perl" to run Perl scripts from the PrepareRelease script
285 because this is reportedly needed in Windows.
287 26. If study data was being save in a file and studying had not found a set of
288 "starts with" bytes for the pattern, the data written to the file (though
289 never used) was taken from uninitialized memory and so caused valgrind to
290 complain.
292 27. Updated RunTest.bat as provided by Sheri Pierce.
294 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
296 29. Computation of memory usage for the table of capturing group names was
297 giving an unnecessarily large value.
300 Version 8.20 21-Oct-2011
301 ------------------------
303 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
304 a POSIX class. After further experiments with Perl, which convinced me that
305 Perl has bugs and confusions, a closing square bracket is no longer allowed
306 in a POSIX name. This bug also affected patterns with classes that started
307 with full stops.
309 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
310 captured substring, but while checking the failing first alternative,
311 substring 1 is temporarily captured. If the output vector supplied to
312 pcre_exec() was not big enough for this capture, the yield of the function
313 was still zero ("insufficient space for captured substrings"). This cannot
314 be totally fixed without adding another stack variable, which seems a lot
315 of expense for a edge case. However, I have improved the situation in cases
316 such as /(a)(b)x|abc/ matched against "abc", where the return code
317 indicates that fewer than the maximum number of slots in the ovector have
318 been set.
320 3. Related to (2) above: when there are more back references in a pattern than
321 slots in the output vector, pcre_exec() uses temporary memory during
322 matching, and copies in the captures as far as possible afterwards. It was
323 using the entire output vector, but this conflicts with the specification
324 that only 2/3 is used for passing back captured substrings. Now it uses
325 only the first 2/3, for compatibility. This is, of course, another edge
326 case.
328 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
329 main code base, and can be used by building with --enable-jit. When this is
330 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
331 runtime --no-jit option is given.
333 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
334 ovector, the return from the function was zero, implying that there were
335 other matches that did not fit. The correct "exactly full" value is now
336 returned.
338 6. If a subpattern that was called recursively or as a subroutine contained
339 (*PRUNE) or any other control that caused it to give a non-standard return,
340 invalid errors such as "Error -26 (nested recursion at the same subject
341 position)" or even infinite loops could occur.
343 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
344 computing the minimum length on reaching *ACCEPT, and so ended up with the
345 wrong value of 1 rather than 0. Further investigation indicates that
346 computing a minimum subject length in the presence of *ACCEPT is difficult
347 (think back references, subroutine calls), and so I have changed the code
348 so that no minimum is registered for a pattern that contains *ACCEPT.
350 8. If (*THEN) was present in the first (true) branch of a conditional group,
351 it was not handled as intended. [But see 16 below.]
353 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
354 Sheri Pierce.
356 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
357 the first byte in a match must be "a".
359 11. Change 17 for 8.13 increased the recursion depth for patterns like
360 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
361 pattern contains any instances of (*THEN). If it does not, the old
362 optimizations are restored. It would be nice to do this on a per-group
363 basis, but at the moment that is not feasible.
365 12. In some environments, the output of pcretest -C is CRLF terminated. This
366 broke RunTest's code that checks for the link size. A single white space
367 character after the value is now allowed for.
369 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
370 For "fr", it uses the Windows-specific input and output files.
372 14. If (*THEN) appeared in a group that was called recursively or as a
373 subroutine, it did not work as intended. [But see next item.]
375 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
376 pattern fragments (but not containing any | characters). If A and B are
377 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
378 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
379 D. In other words, Perl considers parentheses that do not contain any |
380 characters to be part of a surrounding alternative, whereas PCRE was
381 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
382 differently. PCRE now behaves in the same way as Perl, except in the case
383 of subroutine/recursion calls such as (?1) which have in any case always
384 been different (but PCRE had them first :-).
386 16. Related to 15 above: Perl does not treat the | in a conditional group as
387 creating alternatives. Such a group is treated in the same way as an
388 ordinary group without any | characters when processing (*THEN). PCRE has
389 been changed to match Perl's behaviour.
391 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
392 RunGrepTest script failed.
394 18. Change 22 for version 13 caused atomic groups to use more stack. This is
395 inevitable for groups that contain captures, but it can lead to a lot of
396 stack use in large patterns. The old behaviour has been restored for atomic
397 groups that do not contain any capturing parentheses.
399 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
400 suppress the check for a minimum subject length at run time. (If it was
401 given to pcre_exec() or pcre_dfa_exec() it did work.)
403 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
404 fail to work when decoding hex characters in data strings in EBCDIC
405 environments.
407 21. It appears that in at least one Mac OS environment, the isxdigit() function
408 is implemented as a macro that evaluates to its argument more than once,
409 contravening the C 90 Standard (I haven't checked a later standard). There
410 was an instance in pcretest which caused it to go wrong when processing
411 \x{...} escapes in subject strings. The has been rewritten to avoid using
412 things like p++ in the argument of isxdigit().
415 Version 8.13 16-Aug-2011
416 ------------------------
418 1. The Unicode data tables have been updated to Unicode 6.0.0.
420 2. Two minor typos in pcre_internal.h have been fixed.
422 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
423 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
424 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
426 4. There were a number of related bugs in the code for matching backrefences
427 caselessly in UTF-8 mode when codes for the characters concerned were
428 different numbers of bytes. For example, U+023A and U+2C65 are an upper
429 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
430 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
431 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
432 2-byte code at the end of the subject (it thought there wasn't enough data
433 left).
435 5. Comprehensive information about what went wrong is now returned by
436 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
437 as the output vector has at least 2 elements. The offset of the start of
438 the failing character and a reason code are placed in the vector.
440 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
441 now returned is for the first byte of the failing character, instead of the
442 last byte inspected. This is an incompatible change, but I hope it is small
443 enough not to be a problem. It makes the returned offset consistent with
444 pcre_exec() and pcre_dfa_exec().
446 7. pcretest now gives a text phrase as well as the error number when
447 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
448 failure, the offset and reason code are output.
450 8. When \R was used with a maximizing quantifier it failed to skip backwards
451 over a \r\n pair if the subsequent match failed. Instead, it just skipped
452 back over a single character (\n). This seems wrong (because it treated the
453 two characters as a single entity when going forwards), conflicts with the
454 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
455 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
456 has been changed.
458 9. Some internal refactoring has changed the processing so that the handling
459 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
460 time (the PCRE_DOTALL option was changed this way some time ago: version
461 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
462 which was always a bit of a fudge. It also means that there is one less
463 argument for the match() function, which reduces its stack requirements
464 slightly. This change also fixes an incompatibility with Perl: the pattern
465 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
467 10. More internal refactoring has drastically reduced the number of recursive
468 calls to match() for possessively repeated groups such as (abc)++ when
469 using pcre_exec().
471 11. While implementing 10, a number of bugs in the handling of groups were
472 discovered and fixed:
474 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
475 (a|)*(?1) gave a compile-time internal error.
476 ((a|)+)+ did not notice that the outer group could match an empty string.
477 (^a|^)+ was not marked as anchored.
478 (.*a|.*)+ was not marked as matching at start or after a newline.
480 12. Yet more internal refactoring has removed another argument from the match()
481 function. Special calls to this function are now indicated by setting a
482 value in a variable in the "match data" data block.
484 13. Be more explicit in pcre_study() instead of relying on "default" for
485 opcodes that mean there is no starting character; this means that when new
486 ones are added and accidentally left out of pcre_study(), testing should
487 pick them up.
489 14. The -s option of pcretest has been documented for ages as being an old
490 synonym of -m (show memory usage). I have changed it to mean "force study
491 for every regex", that is, assume /S for every regex. This is similar to -i
492 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
493 using it. It makes it easier to run collections of tests with and without
494 study enabled, and thereby test pcre_study() more easily. All the standard
495 tests are now run with and without -s (but some patterns can be marked as
496 "never study" - see 20 below).
498 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
499 restoration of the capturing data to the outer values was not happening
500 correctly.
502 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
503 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
504 pattern had matched an empty string, and so incorrectly returned a no
505 match.
507 17. There was optimizing code for the last branch of non-capturing parentheses,
508 and also for the obeyed branch of a conditional subexpression, which used
509 tail recursion to cut down on stack usage. Unfortunately, now that there is
510 the possibility of (*THEN) occurring in these branches, tail recursion is
511 no longer possible because the return has to be checked for (*THEN). These
512 two optimizations have therefore been removed. [But see 8.20/11 above.]
514 18. If a pattern containing \R was studied, it was assumed that \R always
515 matched two bytes, thus causing the minimum subject length to be
516 incorrectly computed because \R can also match just one byte.
518 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
519 was incorrectly computed.
521 20. If /S is present twice on a test pattern in pcretest input, it now
522 *disables* studying, thereby overriding the use of -s on the command line
523 (see 14 above). This is necessary for one or two tests to keep the output
524 identical in both cases.
526 21. When (*ACCEPT) was used in an assertion that matched an empty string and
527 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
529 22. When an atomic group that contained a capturing parenthesis was
530 successfully matched, but the branch in which it appeared failed, the
531 capturing was not being forgotten if a higher numbered group was later
532 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
533 group 1 to "a", when in fact it should be unset. This applied to multi-
534 branched capturing and non-capturing groups, repeated or not, and also to
535 positive assertions (capturing in negative assertions does not happen
536 in PCRE) and also to nested atomic groups.
538 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
539 subject after a captured substring, to make it easier to tell which of a
540 number of identical substrings has been captured.
542 24. The way atomic groups are processed by pcre_exec() has been changed so that
543 if they are repeated, backtracking one repetition now resets captured
544 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
545 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
546 "aaa". Previously, it would have been "a". As part of this code
547 refactoring, the way recursive calls are handled has also been changed.
549 25. If an assertion condition captured any substrings, they were not passed
550 back unless some other capturing happened later. For example, if
551 (?(?=(a))a) was matched against "a", no capturing was returned.
553 26. When studying a pattern that contained subroutine calls or assertions,
554 the code for finding the minimum length of a possible match was handling
555 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
556 group 1 called group 2 while simultaneously a separate group 2 called group
557 1). A stack overflow occurred in this case. I have fixed this by limiting
558 the recursion depth to 10.
560 27. Updated RunTest.bat in the distribution to the version supplied by Tom
561 Fortmann. This supports explicit test numbers on the command line, and has
562 argument validation and error reporting.
564 28. An instance of \X with an unlimited repeat could fail if at any point the
565 first character it looked at was a mark character.
567 29. Some minor code refactoring concerning Unicode properties and scripts
568 should reduce the stack requirement of match() slightly.
570 30. Added the '=' option to pcretest to check the setting of unused capturing
571 slots at the end of the pattern, which are documented as being -1, but are
572 not included in the return count.
574 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
575 compiled something random. Now it gives a compile-time error (as does
576 Perl).
578 32. A *MARK encountered during the processing of a positive assertion is now
579 recorded and passed back (compatible with Perl).
581 33. If --only-matching or --colour was set on a pcregrep call whose pattern
582 had alternative anchored branches, the search for a second match in a line
583 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
584 matched the line "0102" twice. The same bug affected patterns that started
585 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
586 twice.
588 34. Previously, PCRE did not allow quantification of assertions. However, Perl
589 does, and because of capturing effects, quantifying parenthesized
590 assertions may at times be useful. Quantifiers are now allowed for
591 parenthesized assertions.
593 35. A minor code tidy in pcre_compile() when checking options for \R usage.
595 36. \g was being checked for fancy things in a character class, when it should
596 just be a literal "g".
598 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
599 appearance of a nested POSIX class supersedes an apparent external class.
600 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
601 unescaped square brackets may also appear as part of class names. For
602 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
603 more like Perl. (But see 8.20/1 above.)
605 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
606 was because it thought it was \N{name}, which is not supported).
608 39. Add minix to OS list not supporting the -S option in pcretest.
610 40. PCRE tries to detect cases of infinite recursion at compile time, but it
611 cannot analyze patterns in sufficient detail to catch mutual recursions
612 such as ((?1))((?2)). There is now a runtime test that gives an error if a
613 subgroup is called recursively as a subpattern for a second time at the
614 same position in the subject string. In previous releases this might have
615 been caught by the recursion limit, or it might have run out of stack.
617 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
618 happen only once. PCRE was, however incorrectly giving a compile time error
619 "recursive call could loop indefinitely" because it cannot analyze the
620 pattern in sufficient detail. The compile time test no longer happens when
621 PCRE is compiling a conditional subpattern, but actual runaway loops are
622 now caught at runtime (see 40 above).
624 42. It seems that Perl allows any characters other than a closing parenthesis
625 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
626 has been changed to be the same.
628 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
629 as not to get warnings when autogen.sh is called. Also changed
630 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
632 44. To help people who use pcregrep to scan files containing exceedingly long
633 lines, the following changes have been made:
635 (a) The default value of the buffer size parameter has been increased from
636 8K to 20K. (The actual buffer used is three times this size.)
638 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
639 PCRE is built.
641 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
642 to be set at run time.
644 (d) Numerical values in pcregrep options can be followed by K or M, for
645 example --buffer-size=50K.
647 (e) If a line being scanned overflows pcregrep's buffer, an error is now
648 given and the return code is set to 2.
650 45. Add a pointer to the latest mark to the callout data block.
652 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
653 partial match of an empty string instead of no match. This was specific to
654 the use of ".".
656 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
657 complete match instead of a partial match. This bug was dependent on both
658 the PCRE_UTF8 and PCRE_DOTALL options being set.
660 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
661 starting byte set, because \b was not being ignored.
664 Version 8.12 15-Jan-2011
665 ------------------------
667 1. Fixed some typos in the markup of the man pages, and wrote a script that
668 checks for such things as part of the documentation building process.
670 2. On a big-endian 64-bit system, pcregrep did not correctly process the
671 --match-limit and --recursion-limit options (added for 8.11). In
672 particular, this made one of the standard tests fail. (The integer value
673 went into the wrong half of a long int.)
675 3. If the --colour option was given to pcregrep with -v (invert match), it
676 did strange things, either producing crazy output, or crashing. It should,
677 of course, ignore a request for colour when reporting lines that do not
678 match.
680 4. Another pcregrep bug caused similar problems if --colour was specified with
681 -M (multiline) and the pattern match finished with a line ending.
683 5. In pcregrep, when a pattern that ended with a literal newline sequence was
684 matched in multiline mode, the following line was shown as part of the
685 match. This seems wrong, so I have changed it.
687 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
688 the check for further matches in the same line (so they could be coloured)
689 to overrun the end of the current line. If another match was found, it was
690 incorrectly shown (and then shown again when found in the next line).
692 7. If pcregrep was compiled under Windows, there was a reference to the
693 function pcregrep_exit() before it was defined. I am assuming this was
694 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
695 reported by a user. I've moved the definition above the reference.
698 Version 8.11 10-Dec-2010
699 ------------------------
701 1. (*THEN) was not working properly if there were untried alternatives prior
702 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
703 backtracked to try for "b" instead of moving to the next alternative branch
704 at the same level (in this case, to look for "c"). The Perl documentation
705 is clear that when (*THEN) is backtracked onto, it goes to the "next
706 alternative in the innermost enclosing group".
708 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
709 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
710 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
711 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
712 (*THEN).
714 3. If \s appeared in a character class, it removed the VT character from
715 the class, even if it had been included by some previous item, for example
716 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
717 of \s, but is part of the POSIX "space" class.)
719 4. A partial match never returns an empty string (because you can always
720 match an empty string at the end of the subject); however the checking for
721 an empty string was starting at the "start of match" point. This has been
722 changed to the "earliest inspected character" point, because the returned
723 data for a partial match starts at this character. This means that, for
724 example, /(?<=abc)def/ gives a partial match for the subject "abc"
725 (previously it gave "no match").
727 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
728 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
729 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
730 has an implication that the given string is incomplete (because a partial
731 match is preferred over a full match). For this reason, these items now
732 give a partial match in this situation. [Aside: previously, the one case
733 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
734 match rather than a full match, which was wrong by the old rules, but is
735 now correct.]
737 6. There was a bug in the handling of #-introduced comments, recognized when
738 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
739 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
740 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
741 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
742 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
743 places in pcre_compile().
745 7. Related to (6) above, when pcre_compile() was skipping #-introduced
746 comments when looking ahead for named forward references to subpatterns,
747 the only newline sequence it recognized was NL. It now handles newlines
748 according to the set newline convention.
750 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
751 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
752 cater for a lack of strerror(). These oversights have been fixed.
754 9. Added --match-limit and --recursion-limit to pcregrep.
756 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
758 11. When the -o option was used, pcregrep was setting a return code of 1, even
759 when matches were found, and --line-buffered was not being honoured.
761 12. Added an optional parentheses number to the -o and --only-matching options
762 of pcregrep.
764 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
765 can match an empty string. The code to do it in pcretest and pcredemo
766 needed fixing:
768 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
769 only one byte after an empty string match just before CRLF (this case
770 just got forgotten; "any" and "anycrlf" were OK).
772 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
773 mode when an empty string match preceded an ASCII character followed by
774 a non-ASCII character. (The code for advancing by one character rather
775 than one byte was nonsense.)
777 (c) The pcredemo.c sample program did not have any code at all to handle
778 the cases when CRLF is a valid newline sequence.
780 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
781 as a starting offset was within the subject string. There is now a new
782 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
783 negative or greater than the length of the string. In order to test this,
784 pcretest is extended to allow the setting of negative starting offsets.
786 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
787 starting offset points to the beginning of a UTF-8 character was
788 unnecessarily clumsy. I tidied it up.
790 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
791 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
793 17. Nobody had reported that the --include_dir option, which was added in
794 release 7.7 should have been called --include-dir (hyphen, not underscore)
795 for compatibility with GNU grep. I have changed it to --include-dir, but
796 left --include_dir as an undocumented synonym, and the same for
797 --exclude-dir, though that is not available in GNU grep, at least as of
798 release 2.5.4.
800 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
801 characters from a string of bytes) have been redefined so as not to use
802 loops, in order to improve performance in some environments. At the same
803 time, I abstracted some of the common code into auxiliary macros to save
804 repetition (this should not affect the compiled code).
806 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
807 compile-time error is now given if \c is not followed by an ASCII
808 character, that is, a byte less than 128. (In EBCDIC mode, the code is
809 different, and any byte value is allowed.)
811 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
812 START_OPTIMIZE option, which is now allowed at compile time - but just
813 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
814 to pcregrep and other applications that have no direct access to PCRE
815 options. The new /Y option in pcretest sets this option when calling
816 pcre_compile().
818 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
819 back references. Groups containing recursive back references were forced to
820 be atomic by that change, but in the case of named groups, the amount of
821 memory required was incorrectly computed, leading to "Failed: internal
822 error: code overflow". This has been fixed.
824 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
825 pcretest.c, to avoid build problems in some Borland environments.
828 Version 8.10 25-Jun-2010
829 ------------------------
831 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
832 THEN.
834 2. (*ACCEPT) was not working when inside an atomic group.
836 3. Inside a character class, \B is treated as a literal by default, but
837 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
838 causes the error). The code is unchanged, but I tidied the documentation.
840 4. Inside a character class, PCRE always treated \R and \X as literals,
841 whereas Perl faults them if its -w option is set. I have changed PCRE so
842 that it faults them when PCRE_EXTRA is set.
844 5. Added support for \N, which always matches any character other than
845 newline. (It is the same as "." when PCRE_DOTALL is not set.)
847 6. When compiling pcregrep with newer versions of gcc which may have
848 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
849 declared with attribute warn_unused_result" were given. Just casting the
850 result to (void) does not stop the warnings; a more elaborate fudge is
851 needed. I've used a macro to implement this.
853 7. Minor change to pcretest.c to avoid a compiler warning.
855 8. Added four artifical Unicode properties to help with an option to make
856 \s etc use properties (see next item). The new properties are: Xan
857 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
859 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
860 use Unicode properties. (*UCP) at the start of a pattern can be used to set
861 this option. Modified pcretest to add /W to test this facility. Added
862 REG_UCP to make it available via the POSIX interface.
864 10. Added --line-buffered to pcregrep.
866 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
867 studied, and the match started with a letter with a code point greater than
868 127 whose first byte was different to the first byte of the other case of
869 the letter, the other case of this starting letter was not recognized
870 (#976).
872 12. If a pattern that was studied started with a repeated Unicode property
873 test, for example, \p{Nd}+, there was the theoretical possibility of
874 setting up an incorrect bitmap of starting bytes, but fortunately it could
875 not have actually happened in practice until change 8 above was made (it
876 added property types that matched character-matching opcodes).
878 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
879 possible starting bytes for non-anchored patterns.
881 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
882 \R, and also a number of cases that involve Unicode properties, both
883 explicit and implicit when PCRE_UCP is set.
885 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
886 input, it could crash or give wrong results if characters with values
887 greater than 0xc0 were present in the subject string. (Detail: it assumed
888 UTF-8 input when processing these items.)
890 16. Added a lot of (int) casts to avoid compiler warnings in systems where
891 size_t is 64-bit (#991).
893 17. Added a check for running out of memory when PCRE is compiled with
894 --disable-stack-for-recursion (#990).
896 18. If the last data line in a file for pcretest does not have a newline on
897 the end, a newline was missing in the output.
899 19. The default pcre_chartables.c file recognizes only ASCII characters (values
900 less than 128) in its various bitmaps. However, there is a facility for
901 generating tables according to the current locale when PCRE is compiled. It
902 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
903 characters, are recognized by isspace() and therefore were getting set in
904 these tables, and indeed these tables seem to approximate to ISO 8859. This
905 caused a problem in UTF-8 mode when pcre_study() was used to create a list
906 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
907 which of course cannot start UTF-8 characters. I have changed the code so
908 that only real ASCII characters (less than 128) and the correct starting
909 bytes for UTF-8 encodings are set for characters greater than 127 when in
910 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
911 altogether.)
913 20. Added the /T option to pcretest so as to be able to run tests with non-
914 standard character tables, thus making it possible to include the tests
915 used for 19 above in the standard set of tests.
917 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
918 reference to a subpattern the other side of a comment that contains an
919 opening parenthesis caused either an internal compiling error, or a
920 reference to the wrong subpattern.
923 Version 8.02 19-Mar-2010
924 ------------------------
926 1. The Unicode data tables have been updated to Unicode 5.2.0.
928 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
929 configured.
931 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
932 original author of that file, following a query about its status.
934 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
935 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
937 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
938 quantifier applied to a forward-referencing subroutine call, could compile
939 incorrect code or give the error "internal error: previously-checked
940 referenced subpattern not found".
942 6. Both MS Visual Studio and Symbian OS have problems with initializing
943 variables to point to external functions. For these systems, therefore,
944 pcre_malloc etc. are now initialized to local functions that call the
945 relevant global functions.
947 7. There were two entries missing in the vectors called coptable and poptable
948 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
949 I've fixed the data, and added a kludgy way of testing at compile time that
950 the lengths are correct (equal to the number of opcodes).
952 8. Following on from 7, I added a similar kludge to check the length of the
953 eint vector in pcreposix.c.
955 9. Error texts for pcre_compile() are held as one long string to avoid too
956 much relocation at load time. To find a text, the string is searched,
957 counting zeros. There was no check for running off the end of the string,
958 which could happen if a new error number was added without updating the
959 string.
961 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
963 11. \K was not working if it appeared in an atomic group or in a group that
964 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
965 \K is "not well defined" if used in an assertion. PCRE now accepts it if
966 the assertion is positive, but not if it is negative.
968 12. Change 11 fortuitously reduced the size of the stack frame used in the
969 "match()" function of pcre_exec.c by one pointer. Forthcoming
970 implementation of support for (*MARK) will need an extra pointer on the
971 stack; I have reserved it now, so that the stack frame size does not
972 decrease.
974 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
975 item in branch that calls a recursion is a subroutine call - as in the
976 second branch in the above example - was incorrectly given the compile-
977 time error "recursive call could loop indefinitely" because pcre_compile()
978 was not correctly checking the subroutine for matching a non-empty string.
980 14. The checks for overrunning compiling workspace could trigger after an
981 overrun had occurred. This is a "should never occur" error, but it can be
982 triggered by pathological patterns such as hundreds of nested parentheses.
983 The checks now trigger 100 bytes before the end of the workspace.
985 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
988 Version 8.01 19-Jan-2010
989 ------------------------
991 1. If a pattern contained a conditional subpattern with only one branch (in
992 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
993 computed the wrong minimum data length (which is of course zero for such
994 subpatterns). This could cause incorrect "no match" results.
996 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
997 the pattern is reset in the first branch, pcre_compile() failed with
998 "internal error: code overflow at offset...". This happened only when
999 the reset was to the original external option setting. (An optimization
1000 abstracts leading options settings into an external setting, which was the
1001 cause of this.)
1003 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1004 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1005 assertion pattern did not match (meaning that the assertion was true), it
1006 was incorrectly treated as false if the SKIP had been reached during the
1007 matching. This also applied to assertions used as conditions.
1009 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1010 assertion subpattern, including such a pattern used as a condition,
1011 unpredictable results occurred, instead of the error return
1014 5. The C++ GlobalReplace function was not working like Perl for the special
1015 situation when an empty string is matched. It now does the fancy magic
1016 stuff that is necessary.
1018 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1019 removed. (These were left over from very, very early versions of PCRE.)
1021 7. Some cosmetic changes to the code to make life easier when compiling it
1022 as part of something else:
1024 (a) Change DEBUG to PCRE_DEBUG.
1026 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1027 called "current" as "current_branch", to prevent a collision with the
1028 Linux macro when compiled as a kernel module.
1030 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1031 prevent a collision with the Linux macro when compiled as a kernel
1032 module.
1034 8. In pcre_compile() there are some checks for integer overflows that used to
1035 cast potentially large values to (double). This has been changed to that
1036 when building, a check for int64_t is made, and if it is found, it is used
1037 instead, thus avoiding the use of floating point arithmetic. (There is no
1038 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1039 double.
1041 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1042 2005 (difference between two addresses compared to an unsigned value).
1044 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1045 custom one, because of the following reported problem in Windows:
1047 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1048 under Win32.
1049 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1050 therefore missing the function definition.
1051 - The compiler thus generates a "C" signature for the test function.
1052 - The linker fails to find the "C" function.
1053 - PCRE fails to configure if asked to do so against libbz2.
1055 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1056 messages were output:
1058 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1059 rerunning libtoolize, to keep the correct libtool macros in-tree.
1060 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1062 I have done both of these things.
1064 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1065 most of the time, it *can* run out if it is given a pattern that contains a
1066 runaway infinite recursion. I updated the discussion in the pcrestack man
1067 page.
1069 13. Now that we have gone to the x.xx style of version numbers, the minor
1070 version may start with zero. Using 08 or 09 is a bad idea because users
1071 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1072 interpreted as invalid octal numbers. I've updated the previous comment in
1073 configure.ac, and also added a check that gives an error if 08 or 09 are
1074 used.
1076 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1077 causing partial matching to fail when the end of the subject matched \W
1078 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1080 15. There were some discrepancies between the declarations in pcre_internal.h
1081 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1082 their definitions. The declarations used "const uschar *" and the
1083 definitions used USPTR. Even though USPTR is normally defined as "const
1084 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1085 reported that: "This difference in casting confuses some C++ compilers, for
1086 example, SunCC recognizes above declarations as different functions and
1087 generates broken code for hbpcre." I have changed the declarations to use
1088 USPTR.
1090 16. GNU libtool is named differently on some systems. The autogen.sh script now
1091 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1092 (FreeBSD).
1094 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1095 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1096 comment: "Figure out how to create a longlong from a string: strtoll and
1097 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1098 instance, but it only takes 2 args instead of 3!"
1100 18. A subtle bug concerned with back references has been fixed by a change of
1101 specification, with a corresponding code fix. A pattern such as
1102 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1103 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1104 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1105 same bug. Such groups have to be quantified to be useful, or contained
1106 inside another quantified group. (If there's no repetition, the reference
1107 can never match.) The problem arises because, having left the group and
1108 moved on to the rest of the pattern, a later failure that backtracks into
1109 the group uses the captured value from the final iteration of the group
1110 rather than the correct earlier one. I have fixed this in PCRE by forcing
1111 any group that contains a reference to itself to be an atomic group; that
1112 is, there cannot be any backtracking into it once it has completed. This is
1113 similar to recursive and subroutine calls.
1116 Version 8.00 19-Oct-09
1117 ----------------------
1119 1. The table for translating pcre_compile() error codes into POSIX error codes
1120 was out-of-date, and there was no check on the pcre_compile() error code
1121 being within the table. This could lead to an OK return being given in
1122 error.
1124 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1125 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1126 in a Windows environment.
1128 3. The pcregrep --count option prints the count for each file even when it is
1129 zero, as does GNU grep. However, pcregrep was also printing all files when
1130 --files-with-matches was added. Now, when both options are given, it prints
1131 counts only for those files that have at least one match. (GNU grep just
1132 prints the file name in this circumstance, but including the count seems
1133 more useful - otherwise, why use --count?) Also ensured that the
1134 combination -clh just lists non-zero counts, with no names.
1136 4. The long form of the pcregrep -F option was incorrectly implemented as
1137 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1138 but it seems right to fix it, and I didn't think it was worth preserving
1139 the old behaviour.
1141 5. The command line items --regex=pattern and --regexp=pattern were not
1142 recognized by pcregrep, which required --regex pattern or --regexp pattern
1143 (with a space rather than an '='). The man page documented the '=' forms,
1144 which are compatible with GNU grep; these now work.
1146 6. No libpcreposix.pc file was created for pkg-config; there was just
1147 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1149 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1150 when UCP support is not needed, by modifying the Python script that
1151 generates it from Unicode data files. This should not matter if the module
1152 is correctly used as a library, but I received one complaint about 50K of
1153 unwanted data. My guess is that the person linked everything into his
1154 program rather than using a library. Anyway, it does no harm.
1156 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1157 was a minimum greater than 1 for a wide character in a possessive
1158 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1159 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1160 character. Chaos in the form of incorrect output or a compiling loop could
1161 result.
1163 9. The restrictions on what a pattern can contain when partial matching is
1164 requested for pcre_exec() have been removed. All patterns can now be
1165 partially matched by this function. In addition, if there are at least two
1166 slots in the offset vector, the offset of the earliest inspected character
1167 for the match and the offset of the end of the subject are set in them when
1168 PCRE_ERROR_PARTIAL is returned.
1170 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1171 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1172 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1173 and may be more useful for multi-segment matching.
1175 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1176 used to be given if ever the end of the subject was reached; now it is
1177 given only if matching could not proceed because another character was
1178 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1179 string "Z", which now yields "no match" instead of "partial match". In the
1180 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1181 final character ended with (*FAIL).
1183 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1184 if the pattern had a "must contain" character that was already found in the
1185 earlier partial match, unless partial matching was again requested. For
1186 example, with the pattern /dog.(body)?/, the "must contain" character is
1187 "g". If the first part-match was for the string "dog", restarting with
1188 "sbody" failed. This bug has been fixed.
1190 13. The string returned by pcre_dfa_exec() after a partial match has been
1191 changed so that it starts at the first inspected character rather than the
1192 first character of the match. This makes a difference only if the pattern
1193 starts with a lookbehind assertion or \b or \B (\K is not supported by
1194 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1195 matching functions compatible, and I think it's the right thing to do.
1197 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1198 so that the demonstration program is easily available in environments where
1199 PCRE has not been installed from source.
1201 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1202 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1203 library.
1205 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1206 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1207 is not the first non-POSIX option to be added. Clearly some people find
1208 these options useful.
1210 17. If a caller to the POSIX matching function regexec() passes a non-zero
1211 value for nmatch with a NULL value for pmatch, the value of
1212 nmatch is forced to zero.
1214 18. RunGrepTest did not have a test for the availability of the -u option of
1215 the diff command, as RunTest does. It now checks in the same way as
1216 RunTest, and also checks for the -b option.
1218 19. If an odd number of negated classes containing just a single character
1219 interposed, within parentheses, between a forward reference to a named
1220 subpattern and the definition of the subpattern, compilation crashed with
1221 an internal error, complaining that it could not find the referenced
1222 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1223 [The bug was that it was starting one character too far in when skipping
1224 over the character class, thus treating the ] as data rather than
1225 terminating the class. This meant it could skip too much.]
1227 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1228 /g option in pcretest when the pattern contains \K, which makes it possible
1229 to have an empty string match not at the start, even when the pattern is
1230 anchored. Updated pcretest and pcredemo to use this option.
1232 21. If the maximum number of capturing subpatterns in a recursion was greater
1233 than the maximum at the outer level, the higher number was returned, but
1234 with unset values at the outer level. The correct (outer level) value is
1235 now given.
1237 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1238 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1239 make it do so. The string so far is captured, making this feature
1240 compatible with Perl.
1242 23. The tests have been re-organized, adding tests 11 and 12, to make it
1243 possible to check the Perl 5.10 features against Perl 5.10.
1245 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1246 pattern matches a fixed length string. PCRE did not allow this; now it
1247 does. Neither allows recursion.
1249 25. I finally figured out how to implement a request to provide the minimum
1250 length of subject string that was needed in order to match a given pattern.
1251 (It was back references and recursion that I had previously got hung up
1252 on.) This code has now been added to pcre_study(); it finds a lower bound
1253 to the length of subject needed. It is not necessarily the greatest lower
1254 bound, but using it to avoid searching strings that are too short does give
1255 some useful speed-ups. The value is available to calling programs via
1256 pcre_fullinfo().
1258 26. While implementing 25, I discovered to my embarrassment that pcretest had
1259 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1260 study optimizations had never been tested with that matching function.
1261 Oops. What is worse, even when it was passed study data, there was a bug in
1262 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1263 were also very few tests of studied patterns with pcre_dfa_exec().
1265 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1266 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1267 on the other side of the coin, they are no longer allowed to have different
1268 names, because these cannot be distinguished in PCRE, and this has caused
1269 confusion. (This is a difference from Perl.)
1271 28. When duplicate subpattern names are present (necessarily with different
1272 numbers, as required by 27 above), and a test is made by name in a
1273 conditional pattern, either for a subpattern having been matched, or for
1274 recursion in such a pattern, all the associated numbered subpatterns are
1275 tested, and the overall condition is true if the condition is true for any
1276 one of them. This is the way Perl works, and is also more like the way
1277 testing by number works.
1280 Version 7.9 11-Apr-09
1281 ---------------------
1283 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1284 (pcretest), all targets were linked against these libraries. This included
1285 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1286 libraries. This caused unwanted dependencies to be created. This problem
1287 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1288 pcretest is linked with readline.
1290 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1291 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1292 moved outside it again, because FALSE and TRUE are already defined in AIX,
1293 but BOOL is not.
1295 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1296 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1298 4. The pcregrep documentation said spaces were inserted as well as colons (or
1299 hyphens) following file names and line numbers when outputting matching
1300 lines. This is not true; no spaces are inserted. I have also clarified the
1301 wording for the --colour (or --color) option.
1303 5. In pcregrep, when --colour was used with -o, the list of matching strings
1304 was not coloured; this is different to GNU grep, so I have changed it to be
1305 the same.
1307 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1308 each matching line was coloured. Now it goes on to look for further matches
1309 of any of the test patterns, which is the same behaviour as GNU grep.
1311 7. A pattern that could match an empty string could cause pcregrep to loop; it
1312 doesn't make sense to accept an empty string match in pcregrep, so I have
1313 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1314 seems to be how GNU grep behaves.
1316 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1317 start or after a newline", because the conditional assertion was not being
1318 correctly handled. The rule now is that both the assertion and what follows
1319 in the first alternative must satisfy the test.
1321 9. If auto-callout was enabled in a pattern with a conditional group whose
1322 condition was an assertion, PCRE could crash during matching, both with
1323 pcre_exec() and pcre_dfa_exec().
1325 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1326 used for matching.
1328 11. Unicode property support in character classes was not working for
1329 characters (bytes) greater than 127 when not in UTF-8 mode.
1331 12. Added the -M command line option to pcretest.
1333 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1335 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1337 16. Added comments and documentation about mis-use of no_arg in the C++
1338 wrapper.
1340 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1341 from Martin Jerabek that uses macro names for all relevant character and
1342 string constants.
1344 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1345 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1346 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1347 these, but not everybody uses configure.
1349 19. A conditional group that had only one branch was not being correctly
1350 recognized as an item that could match an empty string. This meant that an
1351 enclosing group might also not be so recognized, causing infinite looping
1352 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1353 with the subject "ab", where knowledge that the repeated group can match
1354 nothing is needed in order to break the loop.
1356 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1357 exec(), but without supplying a callout function, matching went wrong.
1359 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1360 leak if the size of the offset vector was greater than 30. When the vector
1361 is smaller, the saved offsets during recursion go onto a local stack
1362 vector, but for larger vectors malloc() is used. It was failing to free
1363 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1364 error, in fact).
1366 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1367 heapframe that is used only when UTF-8 support is enabled. This caused no
1368 problem, but was untidy.
1370 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1371 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1372 included within another project.
1374 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1375 slightly modified by me:
1377 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1378 not building pcregrep.
1380 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1381 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1383 25. Forward references, both numeric and by name, in patterns that made use of
1384 duplicate group numbers, could behave incorrectly or give incorrect errors,
1385 because when scanning forward to find the reference group, PCRE was not
1386 taking into account the duplicate group numbers. A pattern such as
1387 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1389 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1390 the feature of a custom pointer more persuasive (as requested by a user).
1392 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1393 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1394 pre-defined.
1396 28. Added support for (*UTF8) at the start of a pattern.
1398 29. Arrange for flags added by the "release type" setting in CMake to be shown
1399 in the configuration summary.
1402 Version 7.8 05-Sep-08
1403 ---------------------
1405 1. Replaced UCP searching code with optimized version as implemented for Ad
1406 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1407 stage table and inline lookup instead of a function, giving speed ups of 2
1408 to 5 times on some simple patterns that I tested. Permission was given to
1409 distribute the MultiStage2.py script that generates the tables (it's not in
1410 the tarball, but is in the Subversion repository).
1412 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1413 scripts.
1415 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1416 a group with a zero qualifier. The result of the study could be incorrect,
1417 or the function might crash, depending on the pattern.
1419 4. Caseless matching was not working for non-ASCII characters in back
1420 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1421 It now works when Unicode Property Support is available.
1423 5. In pcretest, an escape such as \x{de} in the data was always generating
1424 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1425 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1426 truncation.
1428 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1430 7. Added two (int) casts to pcregrep when printing the difference of two
1431 pointers, in case they are 64-bit values.
1433 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1434 test 2 if it fails.
1436 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1437 and a #define of that name to empty if it is not externally set. This is to
1438 allow users of MSVC to set it if necessary.
1440 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1441 the convenience functions in the pcre_get.c source file.
1443 11. An option change at the start of a pattern that had top-level alternatives
1444 could cause overwriting and/or a crash. This command provoked a crash in
1445 some environments:
1447 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1449 This potential security problem was recorded as CVE-2008-2371.
1451 12. For a pattern where the match had to start at the beginning or immediately
1452 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1453 pcre_dfa_exec() could read past the end of the passed subject if there was
1454 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1455 pcretest so that it places the subject at the end of its malloc-ed buffer.
1457 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1458 exec() might read past the end of the data buffer in UTF-8 mode.
1460 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1461 the data contained the byte 0x85 as part of a UTF-8 character within its
1462 first line. This applied both to normal and DFA matching.
1464 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1465 /^[^d]*?$/8 failed to match "abc".
1467 16. Added a missing copyright notice to pcrecpp_internal.h.
1469 17. Make it more clear in the documentation that values returned from
1470 pcre_exec() in ovector are byte offsets, not character counts.
1472 18. Tidied a few places to stop certain compilers from issuing warnings.
1474 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1475 supplied by Stefan Weber. I made a further small update for 7.8 because
1476 there is a change of source arrangements: the pcre_searchfuncs.c module is
1477 replaced by pcre_ucd.c.
1480 Version 7.7 07-May-08
1481 ---------------------
1483 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1484 a string to a long long, pretend we don't even have a long long." This is
1485 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1487 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1488 pre-7.6 versions, which defined a global no_arg variable instead of putting
1489 it in the RE class. (See also #8 below.)
1491 3. Remove a line of dead code, identified by coverity and reported by Nuno
1492 Lopes.
1494 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1496 (1) The include/exclude patterns were being applied to the whole pathnames
1497 of files, instead of just to the final components.
1499 (2) If there was more than one level of directory, the subdirectories were
1500 skipped unless they satisfied the include/exclude conditions. This is
1501 inconsistent with GNU grep (and could even be seen as contrary to the
1502 pcregrep specification - which I improved to make it absolutely clear).
1503 The action now is always to scan all levels of directory, and just
1504 apply the include/exclude patterns to regular files.
1506 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1507 --exclude_dir in the tests to avoid scanning .svn directories.
1509 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1510 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1511 doesn't support NULs in patterns.
1513 7. Added some missing "const"s to declarations of static tables in
1514 pcre_compile.c and pcre_dfa_exec.c.
1516 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1517 caused by fix #2 above. (Subsequently also a second patch to fix the
1518 first patch. And a third patch - this was a messy problem.)
1520 9. Applied Craig's patch to remove the use of push_back().
1522 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1523 matching function regexec().
1525 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1526 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1527 references. PCRE supports relative numbers with this syntax (I don't think
1528 Oniguruma does).
1530 12. Previously, a group with a zero repeat such as (...){0} was completely
1531 omitted from the compiled regex. However, this means that if the group
1532 was called as a subroutine from elsewhere in the pattern, things went wrong
1533 (an internal error was given). Such groups are now left in the compiled
1534 pattern, with a new opcode that causes them to be skipped at execution
1535 time.
1537 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1538 to the way PCRE behaves:
1540 (a) A lone ] character is dis-allowed (Perl treats it as data).
1542 (b) A back reference to an unmatched subpattern matches an empty string
1543 (Perl fails the current match path).
1545 (c) A data ] in a character class must be notated as \] because if the
1546 first data character in a class is ], it defines an empty class. (In
1547 Perl it is not possible to have an empty class.) The empty class []
1548 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1549 The negative empty class [^] matches any one character, independently
1550 of the DOTALL setting.
1552 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1553 non-existent subpattern following a character class starting with ']' and
1554 containing () gave an internal compiling error instead of "reference to
1555 non-existent subpattern". Fortunately, when the pattern did exist, the
1556 compiled code was correct. (When scanning forwards to check for the
1557 existencd of the subpattern, it was treating the data ']' as terminating
1558 the class, so got the count wrong. When actually compiling, the reference
1559 was subsequently set up correctly.)
1561 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1562 it was being rejected as not supported by pcre_dfa_exec(), even though
1563 other assertions are supported. I have made pcre_dfa_exec() support
1564 (*FAIL).
1566 16. The implementation of 13c above involved the invention of a new opcode,
1567 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1568 cannot be changed at match time, I realized I could make a small
1569 improvement to matching performance by compiling OP_ALLANY instead of
1570 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1571 on the OP_ANY path.
1573 17. Compiling pcretest on Windows with readline support failed without the
1574 following two fixes: (1) Make the unistd.h include conditional on
1575 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1577 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1578 ncurses library to be included for pcretest when ReadLine support is
1579 requested, but also to allow for it to be overridden. This patch came from
1580 Daniel Bergström.
1582 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1583 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1584 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1585 spotting this.
1588 Version 7.6 28-Jan-08
1589 ---------------------
1591 1. A character class containing a very large number of characters with
1592 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1593 overflow.
1595 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1596 HAVE_LONG_LONG is not defined.
1598 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1599 bring them up to date and include new features. This patch includes:
1601 - Fixed PH's badly added libz and libbz2 support.
1602 - Fixed a problem with static linking.
1603 - Added pcredemo. [But later removed - see 7 below.]
1604 - Fixed dftables problem and added an option.
1605 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1607 - Added readline support for pcretest.
1608 - Added an listing of the option settings after cmake has run.
1610 4. A user submitted a patch to Makefile that makes it easy to create
1611 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1612 Makefile.am that cause it to include this special target, without
1613 affecting anything else. Note that the same mingw target plus all
1614 the other distribution libraries and programs are now supported
1615 when configuring with CMake (see 6 below) instead of with
1616 Configure/Make.
1618 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1619 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1620 exported in the Windows port". It has not yet been confirmed that the patch
1621 solves the problem, but it does no harm.
1623 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1624 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1625 with CMake, and also correct the comment about stack recursion.
1627 7. Remove the automatic building of pcredemo from the ./configure system and
1628 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1629 of a program that users should build themselves after PCRE is installed, so
1630 building it automatically is not really right. What is more, it gave
1631 trouble in some build environments.
1633 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1636 Version 7.5 10-Jan-08
1637 ---------------------
1639 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1640 values in parens when parsing an RE using the C++ wrapper."
1642 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1643 Characters greater than 255 were excluded from the class instead of being
1644 included.
1646 3. The same bug as (2) above applied to negated POSIX classes such as
1647 [:^space:].
1649 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1650 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1651 I have changed it.
1653 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1654 first named subpattern) and a construct such as (?&a) would reference the
1655 first named subpattern whose name started with "a" (in other words, the
1656 length check was missing). Both these problems are fixed. "Subpattern name
1657 expected" is now given for (?&) (a zero-length name), and this patch also
1658 makes it give the same error for \k'' (previously it complained that that
1659 was a reference to a non-existent subpattern).
1661 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1662 this is right because (?- can be followed by option settings as well as by
1663 digits. I have, however, made the messages clearer.
1665 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1666 than the number used in the conditional) now cause a compile-time error.
1667 This is actually not compatible with Perl, which accepts such patterns, but
1668 treats the conditional as always being FALSE (as PCRE used to), but it
1669 seems to me that giving a diagnostic is better.
1671 8. Change "alphameric" to the more common word "alphanumeric" in comments
1672 and messages.
1674 9. Fix two occurrences of "backslash" in comments that should have been
1675 "backspace".
1677 10. Remove two redundant lines of code that can never be obeyed (their function
1678 was moved elsewhere).
1680 11. The program that makes PCRE's Unicode character property table had a bug
1681 which caused it to generate incorrect table entries for sequences of
1682 characters that have the same character type, but are in different scripts.
1683 It amalgamated them into a single range, with the script of the first of
1684 them. In other words, some characters were in the wrong script. There were
1685 thirteen such cases, affecting characters in the following ranges:
1687 U+002b0 - U+002c1
1688 U+0060c - U+0060d
1689 U+0061e - U+00612
1690 U+0064b - U+0065e
1691 U+0074d - U+0076d
1692 U+01800 - U+01805
1693 U+01d00 - U+01d77
1694 U+01d9b - U+01dbf
1695 U+0200b - U+0200f
1696 U+030fc - U+030fe
1697 U+03260 - U+0327f
1698 U+0fb46 - U+0fbb1
1699 U+10450 - U+1049d
1701 12. The -o option (show only the matching part of a line) for pcregrep was not
1702 compatible with GNU grep in that, if there was more than one match in a
1703 line, it showed only the first of them. It now behaves in the same way as
1704 GNU grep.
1706 13. If the -o and -v options were combined for pcregrep, it printed a blank
1707 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1708 does the same. The return code can be used to tell if there were any
1709 non-matching lines.
1711 14. Added --file-offsets and --line-offsets to pcregrep.
1713 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1714 infinitely looping recursion. The bug was that positive lookaheads were not
1715 being skipped when checking for a possible empty match (negative lookaheads
1716 and both kinds of lookbehind were skipped).
1718 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1719 inclusion of <windows.h> to before rather than after the definition of
1720 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1722 17. Specifying a possessive quantifier with a specific limit for a Unicode
1723 character property caused pcre_compile() to compile bad code, which led at
1724 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1725 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1726 caused the error; without that there was no problem.
1728 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1730 19. Added --enable-pcretest-libreadline.
1732 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1733 RE::GlobalReplace(). As a result, the number of replacements returned was
1734 double what it should be. I removed one of the increments, but Craig sent a
1735 later patch that removed the other one (the right fix) and added unit tests
1736 that check the return values (which was not done before).
1738 21. Several CMake things:
1740 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1741 the names libpcre and libpcreposix, not just pcre and pcreposix.
1743 (2) The above change means that pcretest and pcregrep are now correctly
1744 linked with the newly-built libraries, not previously installed ones.
1748 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1749 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1750 UTF-8 newline character). The key issue is that the pattern starts .*;
1751 this means that the match must be either at the beginning, or after a
1752 newline. The bug was in the code for advancing after a failed match and
1753 checking that the new position followed a newline. It was not taking
1754 account of UTF-8 characters correctly.
1756 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1757 character classes. PCRE was not treating the sequence [:...:] as a
1758 character class unless the ... were all letters. Perl, however, seems to
1759 allow any characters between [: and :], though of course it rejects as
1760 unknown any "names" that contain non-letters, because all the known class
1761 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1762 for example, whereas PCRE did not - it did not recognize a POSIX character
1763 class. This seemed a bit dangerous, so the code has been changed to be
1764 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1765 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1766 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1767 Perl does, and where it didn't before.
1769 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1770 Windows environments %n is disabled by default.
1773 Version 7.4 21-Sep-07
1774 ---------------------
1776 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1777 means that a class such as [\s] counted as "explicit reference to CR or
1778 LF". That isn't really right - the whole point of the change was to try to
1779 help when there was an actual mention of one of the two characters. So now
1780 the change happens only if \r or \n (or a literal CR or LF) character is
1781 encountered.
1783 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1784 of both had grown to the point where there were only 3 bits left.
1785 Fortunately, there was spare space in the data structure, and so I have
1786 moved the internal flags into a new 16-bit field to free up more option
1787 bits.
1789 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1790 but did not set the internal JCHANGED flag - either of these is enough to
1791 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1792 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1793 start sets both bits.
1795 4. Added options (at build time, compile time, exec time) to change \R from
1796 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1798 5. doc/pcresyntax.html was missing from the distribution.
1800 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1801 compatibility, even though it is no longer used.
1803 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1804 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1805 windows.h file is present (where different names are used). [This was
1806 reversed later after testing - see 16 below.]
1808 8. Changed all #include <config.h> to #include "config.h". There were also
1809 some further <pcre.h> cases that I changed to "pcre.h".
1811 9. When pcregrep was used with the --colour option, it missed the line ending
1812 sequence off the lines that it output.
1814 10. It was pointed out to me that arrays of string pointers cause lots of
1815 relocations when a shared library is dynamically loaded. A technique of
1816 using a single long string with a table of offsets can drastically reduce
1817 these. I have refactored PCRE in four places to do this. The result is
1818 dramatic:
1820 Originally: 290
1821 After changing UCP table: 187
1822 After changing error message table: 43
1823 After changing table of "verbs" 36
1824 After changing table of Posix names 22
1826 Thanks to the folks working on Gregex for glib for this insight.
1828 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1829 unicode-properties was also set.
1831 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1833 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1834 checked only for CRLF.
1836 14. Added casts to pcretest.c to avoid compiler warnings.
1838 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1840 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1841 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1842 entirely. This removes changes made in 7 above.
1844 17. The CMake files have been updated, and there is now more information about
1845 building with CMake in the NON-UNIX-USE document.
1848 Version 7.3 28-Aug-07
1849 ---------------------
1851 1. In the rejigging of the build system that eventually resulted in 7.1, the
1852 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1853 brackets there is not right, since it causes compilers to look for an
1854 installed pcre.h, not the version that is in the source that is being
1855 compiled (which of course may be different). I have changed it back to:
1857 #include "pcre.h"
1859 I have a vague recollection that the change was concerned with compiling in
1860 different directories, but in the new build system, that is taken care of
1861 by the VPATH setting the Makefile.
1863 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1864 when the subject happened to end in the byte 0x85 (e.g. if the last
1865 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1866 characters but of course it shouldn't be taken as a newline when it is part
1867 of another character. The bug was that, for an unlimited repeat of . in
1868 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1869 characters when looking for a newline.
1871 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1873 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1874 in debug output.
1876 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1877 long printing in the pcrecpp unittest when running under MinGW.
1879 6. ESC_K was left out of the EBCDIC table.
1881 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1882 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1883 limit also applies to "virtual nesting" when a pattern is recursive, and in
1884 this case 1000 isn't so big. I have been able to remove this limit at the
1885 expense of backing off one optimization in certain circumstances. Normally,
1886 when pcre_exec() would call its internal match() function recursively and
1887 immediately return the result unconditionally, it uses a "tail recursion"
1888 feature to save stack. However, when a subpattern that can match an empty
1889 string has an unlimited repetition quantifier, it no longer makes this
1890 optimization. That gives it a stack frame in which to save the data for
1891 checking that an empty string has been matched. Previously this was taken
1892 from the 1000-entry workspace that had been reserved. So now there is no
1893 explicit limit, but more stack is used.
1895 8. Applied Daniel's patches to solve problems with the import/export magic
1896 syntax that is required for Windows, and which was going wrong for the
1897 pcreposix and pcrecpp parts of the library. These were overlooked when this
1898 problem was solved for the main library.
1900 9. There were some crude static tests to avoid integer overflow when computing
1901 the size of patterns that contain repeated groups with explicit upper
1902 limits. As the maximum quantifier is 65535, the maximum group length was
1903 set at 30,000 so that the product of these two numbers did not overflow a
1904 32-bit integer. However, it turns out that people want to use groups that
1905 are longer than 30,000 bytes (though not repeat them that many times).
1906 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1907 made it possible to implement the integer overflow checks in a much more
1908 dynamic way, which I have now done. The artificial limitation on group
1909 length has been removed - we now have only the limit on the total length of
1910 the compiled pattern, which depends on the LINK_SIZE setting.
1912 10. Fixed a bug in the documentation for get/copy named substring when
1913 duplicate names are permitted. If none of the named substrings are set, the
1914 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1915 empty string.
1917 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1918 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1919 because the ] is interpreted as the first data character and the
1920 terminating ] is not found. PCRE has been made compatible with Perl in this
1921 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1922 cause memory overwriting.
1924 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1925 string has been matched (to stop an infinite loop). It was not recognizing
1926 a conditional subpattern that could match an empty string if that
1927 subpattern was within another subpattern. For example, it looped when
1928 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1929 condition was not nested. This bug has been fixed.
1931 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1932 past the start of the subject in the presence of bytes with the top bit
1933 set, for example "\x8aBCD".
1935 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1936 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1938 14. Optimized (?!) to (*FAIL).
1940 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1941 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1942 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1943 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1944 does: it's just the validity check that is more restrictive.
1946 16. Inserted checks for integer overflows during escape sequence (backslash)
1947 processing, and also fixed erroneous offset values for syntax errors during
1948 backslash processing.
1950 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1951 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1953 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1954 caused an overrun.
1956 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1957 something other than just ASCII characters) inside a group that had an
1958 unlimited repeat caused a loop at compile time (while checking to see
1959 whether the group could match an empty string).
1961 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1962 [\P{Any}] did so. (Error in the code for printing property names.)
1964 21. An orphan \E inside a character class could cause a crash.
1966 22. A repeated capturing bracket such as (A)? could cause a wild memory
1967 reference during compilation.
1969 23. There are several functions in pcre_compile() that scan along a compiled
1970 expression for various reasons (e.g. to see if it's fixed length for look
1971 behind). There were bugs in these functions when a repeated \p or \P was
1972 present in the pattern. These operators have additional parameters compared
1973 with \d, etc, and these were not being taken into account when moving along
1974 the compiled data. Specifically:
1976 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1977 length.
1979 (b) An item such as \pL+ within a repeated group could cause crashes or
1980 loops.
1982 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1983 "reference to non-existent subpattern" error.
1985 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1987 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1988 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1990 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1992 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1993 character were causing crashes (broken optimization).
1995 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1996 \p or \P) caused a compile-time loop.
1998 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1999 break. For example, the unstudied pattern [\r\n]A does not match the string
2000 "\r\nA" because change 7.0/46 below moves the current point on by two
2001 characters after failing to match at the start. However, the pattern \nA
2002 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2003 the same is true. There doesn't seem any very clean way out of this, but
2004 what I have chosen to do makes the common cases work: PCRE now takes note
2005 of whether there can be an explicit match for \r or \n anywhere in the
2006 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2007 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2008 pattern has explicit CR or LF references.
2010 29. Added (*CR) etc for changing newline setting at start of pattern.
2013 Version 7.2 19-Jun-07
2014 ---------------------
2016 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2017 which is apparently normally available under Windows.
2019 2. Re-jig the pcregrep tests with different newline settings in an attempt
2020 to make them independent of the local environment's newline setting.
2022 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2024 4. Some of the "internals" tests were previously cut out when the link size
2025 was not 2, because the output contained actual offsets. The recent new
2026 "Z" feature of pcretest means that these can be cut out, making the tests
2027 usable with all link sizes.
2029 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2030 stack recursion. This gives a massive performance boost under BSD, but just
2031 a small improvement under Linux. However, it saves one field in the frame
2032 in all cases.
2034 6. Added more features from the forthcoming Perl 5.10:
2036 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2037 recursion call. It refers to the nth most recently opened parentheses.
2039 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2040 to be opened parentheses.
2042 (c) Conditions that refer to capturing parentheses can be specified
2043 relatively, for example, (?(-2)... or (?(+3)...
2045 (d) \K resets the start of the current match so that everything before
2046 is not part of it.
2048 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2050 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2051 reference syntax.
2053 (g) (?| introduces a group in which the numbering of parentheses in each
2054 alternative starts with the same number.
2056 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2058 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2061 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2062 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2063 for detecting groups that can match an empty string.
2065 9. A pattern with a very large number of alternatives (more than several
2066 hundred) was running out of internal workspace during the pre-compile
2067 phase, where pcre_compile() figures out how much memory will be needed. A
2068 bit of new cunning has reduced the workspace needed for groups with
2069 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2070 workspace instead of running out of the 4096 that are available.
2072 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2074 11. Applied patch from Google to remove an optimization that didn't quite work.
2075 The report of the bug said:
2077 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2078 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2079 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2081 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2082 it matched the wrong number of bytes.
2085 Version 7.1 24-Apr-07
2086 ---------------------
2088 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2089 that is more "standard", making use of automake and other Autotools. There
2090 is some re-arrangement of the files and adjustment of comments consequent
2091 on this.
2093 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2094 for recursive directory scanning broke on some systems because the files
2095 are not scanned in any specific order and on different systems the order
2096 was different. A call to "sort" has been inserted into RunGrepTest for the
2097 approprate test as a short-term fix. In the longer term there may be an
2098 alternative.
2100 3. I had an email from Eric Raymond about problems translating some of PCRE's
2101 man pages to HTML (despite the fact that I distribute HTML pages, some
2102 people do their own conversions for various reasons). The problems
2103 concerned the use of low-level troff macros .br and .in. I have therefore
2104 removed all such uses from the man pages (some were redundant, some could
2105 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2106 HTML has been updated to handle .nf/.fi and to complain if it encounters
2107 .br or .in.
2109 4. Updated comments in configure.ac that get placed in config.h.in and also
2110 arranged for config.h to be included in the distribution, with the name
2111 config.h.generic, for the benefit of those who have to compile without
2112 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2114 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2115 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2116 makevp.bat for the current PCRE, using the additional files
2117 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2119 6. A Windows user reported a minor discrepancy with test 2, which turned out
2120 to be caused by a trailing space on an input line that had got lost in his
2121 copy. The trailing space was an accident, so I've just removed it.
2123 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2124 that is needed.
2126 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2127 as "const" (a) because they are and (b) because it helps the PHP
2128 maintainers who have recently made a script to detect big data structures
2129 in the php code that should be moved to the .rodata section. I remembered
2130 to update Builducptable as well, so it won't revert if ucptable.h is ever
2131 re-created.
2133 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2134 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2135 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2136 support is not required. This saves 1.5-2K of code, which is important in
2137 some applications.
2139 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2140 so as not to refer to the tables, even though these functions will never be
2141 called when UTF-8 support is disabled. Otherwise there are problems with a
2142 shared library.
2144 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2146 (a) It was defining its arguments as char * instead of void *.
2148 (b) It was assuming that all moves were upwards in memory; this was true
2149 a long time ago when I wrote it, but is no longer the case.
2151 The emulated memove() is provided for those environments that have neither
2152 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2153 is clearly not the case, as these two bugs were recently reported.
2155 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2156 and Detrail to create the HTML documentation, the .txt form of the man
2157 pages, and it removes trailing spaces from listed files. It also creates
2158 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2159 case, it wraps all the #defines with #ifndefs. This script should be run
2160 before "make dist".
2162 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2163 with Unicode property support.
2165 (a) For a maximizing quantifier, if the two different cases of the
2166 character were of different lengths in their UTF-8 codings (there are
2167 some cases like this - I found 11), and the matching function had to
2168 back up over a mixture of the two cases, it incorrectly assumed they
2169 were both the same length.
2171 (b) When PCRE was configured to use the heap rather than the stack for
2172 recursion during matching, it was not correctly preserving the data for
2173 the other case of a UTF-8 character when checking ahead for a match
2174 while processing a minimizing repeat. If the check also involved
2175 matching a wide character, but failed, corruption could cause an
2176 erroneous result when trying to check for a repeat of the original
2177 character.
2179 13. Some tidying changes to the testing mechanism:
2181 (a) The RunTest script now detects the internal link size and whether there
2182 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2183 values substituted by "configure". (The RunGrepTest script already did
2184 this for UTF-8.) The configure.ac script no longer substitutes the
2185 relevant variables.
2187 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2188 with length and offset values. This means that the output is different
2189 for different internal link sizes. Test 2 is skipped for link sizes
2190 other than 2 because of this, bypassing the problem. Unfortunately,
2191 there was also a test in test 3 (the locale tests) that used /B and
2192 failed for link sizes other than 2. Rather than cut the whole test out,
2193 I have added a new /Z option to pcretest that replaces the length and
2194 offset values with spaces. This is now used to make test 3 independent
2195 of link size. (Test 2 will be tidied up later.)
2197 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2198 segmentation fault instead of returning the appropriate error message.
2200 15. In multiline mode when the newline sequence was set to "any", the pattern
2201 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2202 This doesn't seem right; it now treats the CRLF combination as the line
2203 ending, and so does not match in that case. It's only a pattern such as ^$
2204 that would hit this one: something like ^ABC$ would have failed after \r
2205 and then tried again after \r\n.
2207 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2208 in an attempt to make files that differ only in their line terminators
2209 compare equal. This works on Linux.
2211 17. Under certain error circumstances pcregrep might try to free random memory
2212 as it exited. This is now fixed, thanks to valgrind.
2214 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2215 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2216 was because its rules for how to advance for /g after matching an empty
2217 string at the end of a line did not allow for this case. They now check for
2218 it specially.
2220 20. pcretest is supposed to handle patterns and data of any length, by
2221 extending its buffers when necessary. It was getting this wrong when the
2222 buffer for a data line had to be extended.
2224 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2225 CRLF as a newline sequence.
2227 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2228 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2229 I have nevertheless tidied it up.
2231 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2233 24. Added a man page for pcre-config.
2236 Version 7.0 19-Dec-06
2237 ---------------------
2239 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2240 moving to gcc 4.1.1.
2242 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2243 sys/time.h, which is documented as needed for this function. It doesn't
2244 seem to matter on Linux, but it showed up on some releases of OS X.
2246 3. It seems that there are systems where bytes whose values are greater than
2247 127 match isprint() in the "C" locale. The "C" locale should be the
2248 default when a C program starts up. In most systems, only ASCII printing
2249 characters match isprint(). This difference caused the output from pcretest
2250 to vary, making some of the tests fail. I have changed pcretest so that:
2252 (a) When it is outputting text in the compiled version of a pattern, bytes
2253 other than 32-126 are always shown as hex escapes.
2255 (b) When it is outputting text that is a matched part of a subject string,
2256 it does the same, unless a different locale has been set for the match
2257 (using the /L modifier). In this case, it uses isprint() to decide.
2259 4. Fixed a major bug that caused incorrect computation of the amount of memory
2260 required for a compiled pattern when options that changed within the
2261 pattern affected the logic of the preliminary scan that determines the
2262 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2263 that the computed length was too small. The symptoms of this bug were
2264 either the PCRE error "internal error: code overflow" from pcre_compile(),
2265 or a glibc crash with a message such as "pcretest: free(): invalid next
2266 size (fast)". Examples of patterns that provoked this bug (shown in
2267 pcretest format) are:
2269 /(?-x: )/x
2270 /(?x)(?-x: \s*#\s*)/
2271 /((?i)[\x{c0}])/8
2272 /(?i:[\x{c0}])/8
2274 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2275 is now done differently.
2277 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2278 wrapper classes; (b) implement a new function in the C++ scanner that is
2279 more efficient than the old way of doing things because it avoids levels of
2280 recursion in the regex matching; (c) add a paragraph to the documentation
2281 for the FullMatch() function.
2283 6. The escape sequence \n was being treated as whatever was defined as
2284 "newline". Not only was this contrary to the documentation, which states
2285 that \n is character 10 (hex 0A), but it also went horribly wrong when
2286 "newline" was defined as CRLF. This has been fixed.
2288 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2289 was being set to -1 for the "end of line" case (supposedly a value that no
2290 character can have). Though this value is never used (the check for end of
2291 line is "zero bytes in current character"), it caused compiler complaints.
2292 I've changed it to 0xffffffff.
2294 8. In pcre_version.c, the version string was being built by a sequence of
2295 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2296 string (as it is for production releases) called a macro with an empty
2297 argument. The C standard says the result of this is undefined. The gcc
2298 compiler treats it as an empty string (which was what was wanted) but it is
2299 reported that Visual C gives an error. The source has been hacked around to
2300 avoid this problem.
2302 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2303 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2304 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2305 of them did).
2307 10. Originally, pcretest opened its input and output without "b"; then I was
2308 told that "b" was needed in some environments, so it was added for release
2309 5.0 to both the input and output. (It makes no difference on Unix-like
2310 systems.) Later I was told that it is wrong for the input on Windows. I've
2311 now abstracted the modes into two macros, to make it easier to fiddle with
2312 them, and removed "b" from the input mode under Windows.
2314 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2316 12. Added -help and --help to pcretest as an official way of being reminded
2317 of the options.
2319 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2320 and pcrecpp.cc because they annoy compilers at high warning levels.
2322 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2324 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2325 scripts are not C programs :-) and which was not noticed because it works
2326 on Linux.
2328 16. pcretest is supposed to handle any length of pattern and data line (as one
2329 line or as a continued sequence of lines) by extending its input buffer if
2330 necessary. This feature was broken for very long pattern lines, leading to
2331 a string of junk being passed to pcre_compile() if the pattern was longer
2332 than about 50K.
2334 17. I have done a major re-factoring of the way pcre_compile() computes the
2335 amount of memory needed for a compiled pattern. Previously, there was code
2336 that made a preliminary scan of the pattern in order to do this. That was
2337 OK when PCRE was new, but as the facilities have expanded, it has become
2338 harder and harder to keep it in step with the real compile phase, and there
2339 have been a number of bugs (see for example, 4 above). I have now found a
2340 cunning way of running the real compile function in a "fake" mode that
2341 enables it to compute how much memory it would need, while actually only
2342 ever using a few hundred bytes of working memory and without too many
2343 tests of the mode. This should make future maintenance and development
2344 easier. A side effect of this work is that the limit of 200 on the nesting
2345 depth of parentheses has been removed (though this was never a serious
2346 limitation, I suspect). However, there is a downside: pcre_compile() now
2347 runs more slowly than before (30% or more, depending on the pattern). I
2348 hope this isn't a big issue. There is no effect on runtime performance.
2350 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2351 newline (only possible for the last line of a file) and it was a
2352 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2354 19. Added additional timing features to pcretest. (1) The -tm option now times
2355 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2356 separate command line item, by a number that specifies the number of
2357 repeats to use when timing. The default is 50000; this gives better
2358 precision, but takes uncomfortably long for very large patterns.
2360 20. Extended pcre_study() to be more clever in cases where a branch of a
2361 subpattern has no definite first character. For example, (a*|b*)[cd] would
2362 previously give no result from pcre_study(). Now it recognizes that the
2363 first character must be a, b, c, or d.
2365 21. There was an incorrect error "recursive call could loop indefinitely" if
2366 a subpattern (or the entire pattern) that was being tested for matching an
2367 empty string contained only one non-empty item after a nested subpattern.
2368 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2369 incorrectly, because the \d was being skipped in the check.
2371 22. The pcretest program now has a new pattern option /B and a command line
2372 option -b, which is equivalent to adding /B to every pattern. This causes
2373 it to show the compiled bytecode, without the additional information that
2374 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2375 is the same as /B/I).
2377 23. A new optimization is now able automatically to treat some sequences such
2378 as a*b as a*+b. More specifically, if something simple (such as a character
2379 or a simple class like \d) has an unlimited quantifier, and is followed by
2380 something that cannot possibly match the quantified thing, the quantifier
2381 is automatically "possessified".
2383 24. A recursive reference to a subpattern whose number was greater than 39
2384 went wrong under certain circumstances in UTF-8 mode. This bug could also
2385 have affected the operation of pcre_study().
2387 25. Realized that a little bit of performance could be had by replacing
2388 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2390 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2392 27. Possessive quantifiers such as a++ were previously implemented by turning
2393 them into atomic groups such as ($>a+). Now they have their own opcodes,
2394 which improves performance. This includes the automatically created ones
2395 from 23 above.
2397 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2398 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2399 the first matched character to be a colon. This applied both to named and
2400 numbered groups.
2402 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2404 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2405 building PCRE on Windows easier, so I have included it in the distribution.
2407 31. There is now a check in pcretest against a ridiculously large number being
2408 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2409 loop, the loop is abandoned.
2411 32. Forward references to subpatterns in conditions such as (?(2)...) where
2412 subpattern 2 is defined later cause pcre_compile() to search forwards in
2413 the pattern for the relevant set of parentheses. This search went wrong
2414 when there were unescaped parentheses in a character class, parentheses
2415 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2417 33. "Subroutine" calls and backreferences were previously restricted to
2418 referencing subpatterns earlier in the regex. This restriction has now
2419 been removed.
2421 34. Added a number of extra features that are going to be in Perl 5.10. On the
2422 whole, these are just syntactic alternatives for features that PCRE had
2423 previously implemented using the Python syntax or my own invention. The
2424 other formats are all retained for compatibility.
2426 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2427 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2428 also .NET compatible.
2430 (b) A recursion or subroutine call to a named group can now be defined as
2431 (?&name) as well as (?P>name).
2433 (c) A backreference to a named group can now be defined as \k<name> or
2434 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2435 5.10, are also .NET compatible.
2437 (d) A conditional reference to a named group can now use the syntax
2438 (?(<name>) or (?('name') as well as (?(name).
2440 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2441 groups (named and numbered) that are never evaluated inline, but can be
2442 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2443 is always false. There may be only one alternative in such a group.
2445 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2446 as the simple (?(R). The condition is true only if the most recent
2447 recursion is that of the given number or name. It does not search out
2448 through the entire recursion stack.
2450 (g) The escape \gN or \g{N} has been added, where N is a positive or
2451 negative number, specifying an absolute or relative reference.
2453 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2454 some "unreachable code" warnings.
2456 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2457 things, this adds five new scripts.
2459 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2460 There were also incompatibilities regarding the handling of \Q..\E inside
2461 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2462 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2464 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2465 matches an empty string, and forcibly breaks the loop. There were bugs in
2466 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2467 against aaaa the result was just "a" rather than "aaaa", for example. Two
2468 separate and independent bugs (that affected different cases) have been
2469 fixed.
2471 39. Refactored the code to abolish the use of different opcodes for small
2472 capturing bracket numbers. This is a tidy that I avoided doing when I
2473 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2474 The new approach is not only tidier, it makes it possible to reduce the
2475 memory needed to fix the previous bug (38).
2477 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2478 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2479 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2480 mode.
2482 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2483 report.
2485 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2486 copy construction and assignment in the C++ wrapper.
2488 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2489 couple of bugs that could have given wrong results in the "--newline=crlf"
2490 case.
2492 44. Added a number of casts and did some reorganization of signed/unsigned int
2493 variables following suggestions from Dair Grant. Also renamed the variable
2494 "this" as "item" because it is a C++ keyword.
2496 45. Arranged for dftables to add
2498 #include "pcre_internal.h"
2500 to pcre_chartables.c because without it, gcc 4.x may remove the array
2501 definition from the final binary if PCRE is built into a static library and
2502 dead code stripping is activated.
2504 46. For an unanchored pattern, if a match attempt fails at the start of a
2505 newline sequence, and the newline setting is CRLF or ANY, and the next two
2506 characters are CRLF, advance by two characters instead of one.
2509 Version 6.7 04-Jul-06
2510 ---------------------
2512 1. In order to handle tests when input lines are enormously long, pcretest has
2513 been re-factored so that it automatically extends its buffers when
2514 necessary. The code is crude, but this _is_ just a test program. The
2515 default size has been increased from 32K to 50K.
2517 2. The code in pcre_study() was using the value of the re argument before
2518 testing it for NULL. (Of course, in any sensible call of the function, it
2519 won't be NULL.)
2521 3. The memmove() emulation function in pcre_internal.h, which is used on
2522 systems that lack both memmove() and bcopy() - that is, hardly ever -
2523 was missing a "static" storage class specifier.
2525 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2526 containing an extended class (one that cannot be represented by a bitmap
2527 because it contains high-valued characters or Unicode property items, e.g.
2528 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2529 pattern, but PCRE should not loop if you do not (it no longer does).
2530 [Detail: two cases were found: (a) a repeated subpattern containing an
2531 extended class; (b) a recursive reference to a subpattern that followed a
2532 previous extended class. It wasn't skipping over the extended class
2533 correctly when UTF-8 mode was not set.]
2535 5. A negated single-character class was not being recognized as fixed-length
2536 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2537 compile error "lookbehind assertion is not fixed length".
2539 6. The RunPerlTest auxiliary script was showing an unexpected difference
2540 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2541 write a Perl script that can interpret lines of an input file either as
2542 byte characters or as UTF-8, which is what "perltest" was being required to
2543 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2544 can't do is switch easily at run time between having the "use utf8;" pragma
2545 or not. In the end, I fudged it by using the RunPerlTest script to insert
2546 "use utf8;" explicitly for the UTF-8 tests.
2548 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2549 the end of the subject string, contrary to the documentation and to what
2550 Perl does. This was true of both matching functions. Now it matches only at
2551 the start of the subject and immediately after *internal* newlines.
2553 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2554 a pointer to an int instead of a pointer to an unsigned long int. This
2555 caused problems on 64-bit systems.
2557 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2558 instance of the 'standard' template library not being so standard".
2560 10. There was no check on the number of named subpatterns nor the maximum
2561 length of a subpattern name. The product of these values is used to compute
2562 the size of the memory block for a compiled pattern. By supplying a very
2563 long subpattern name and a large number of named subpatterns, the size
2564 computation could be caused to overflow. This is now prevented by limiting
2565 the length of names to 32 characters, and the number of named subpatterns
2566 to 10,000.
2568 11. Subpatterns that are repeated with specific counts have to be replicated in
2569 the compiled pattern. The size of memory for this was computed from the
2570 length of the subpattern and the repeat count. The latter is limited to
2571 65535, but there was no limit on the former, meaning that integer overflow
2572 could in principle occur. The compiled length of a repeated subpattern is
2573 now limited to 30,000 bytes in order to prevent this.
2575 12. Added the optional facility to have named substrings with the same name.
2577 13. Added the ability to use a named substring as a condition, using the
2578 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2579 are numbers (not recommended). Forward references are permitted.
2581 14. Added forward references in named backreferences (if you see what I mean).
2583 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2584 pattern could run off the end of the subject. For example, the pattern
2585 "(?s)(.{1,5})"8 did this with the subject "ab".
2587 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2588 PCRE_CASELESS was set when matching characters that were quantified with ?
2589 or *.
2591 17. A character class other than a single negated character that had a minimum
2592 but no maximum quantifier - for example [ab]{6,} - was not handled
2593 correctly by pce_dfa_exec(). It would match only one character.
2595 18. A valid (though odd) pattern that looked like a POSIX character
2596 class but used an invalid character after [ (for example [[,abc,]]) caused
2597 pcre_compile() to give the error "Failed: internal error: code overflow" or
2598 in some cases to crash with a glibc free() error. This could even happen if
2599 the pattern terminated after [[ but there just happened to be a sequence of
2600 letters, a binary zero, and a closing ] in the memory that followed.
2602 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2603 over the years. Originally (before any Unicode support), just the bottom 8
2604 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2605 output from "man perlunicode" includes this:
2607 The regular expression compiler produces polymorphic opcodes. That
2608 is, the pattern adapts to the data and automatically switches to
2609 the Unicode character scheme when presented with Unicode data--or
2610 instead uses a traditional byte scheme when presented with byte
2611 data.
2613 Sadly, a wide octal escape does not cause a switch, and in a string with
2614 no other multibyte characters, these octal escapes are treated as before.
2615 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2616 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2617 Unicode string.
2619 I have not perpetrated such confusion in PCRE. Up till now, it took just
2620 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2621 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2622 translate to the appropriate multibyte character.
2624 29. Applied some refactoring to reduce the number of warnings from Microsoft
2625 and Borland compilers. This has included removing the fudge introduced
2626 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2627 a warning about an unused variable.
2629 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2630 characters since release 4.0, because Perl (from release 5.004) does not.
2631 [Or at least, is documented not to: some releases seem to be in conflict
2632 with the documentation.] However, when a pattern was studied with
2633 pcre_study() and all its branches started with \s, PCRE still included VT
2634 as a possible starting character. Of course, this did no harm; it just
2635 caused an unnecessary match attempt.
2637 22. Removed a now-redundant internal flag bit that recorded the fact that case
2638 dependency changed within the pattern. This was once needed for "required
2639 byte" processing, but is no longer used. This recovers a now-scarce options
2640 bit. Also moved the least significant internal flag bit to the most-
2641 significant bit of the word, which was not previously used (hangover from
2642 the days when it was an int rather than a uint) to free up another bit for
2643 the future.
2645 23. Added support for CRLF line endings as well as CR and LF. As well as the
2646 default being selectable at build time, it can now be changed at runtime
2647 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2648 specify that it is scanning data with non-default line endings.
2650 24. Changed the definition of CXXLINK to make it agree with the definition of
2651 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2653 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2654 recursions. This makes a big different to stack usage for some patterns.
2656 26. If a subpattern containing a named recursion or subroutine reference such
2657 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2658 the space required for the compiled pattern went wrong and gave too small a
2659 value. Depending on the environment, this could lead to "Failed: internal
2660 error: code overflow at offset 49" or "glibc detected double free or
2661 corruption" errors.
2663 27. Applied patches from Google (a) to support the new newline modes and (b) to
2664 advance over multibyte UTF-8 characters in GlobalReplace.
2666 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2667 difference for some implementation of PCRE in some Windows version.
2669 29. Added some extra testing facilities to pcretest:
2671 \q<number> in a data line sets the "match limit" value
2672 \Q<number> in a data line sets the "match recursion limt" value
2673 -S <number> sets the stack size, where <number> is in megabytes
2675 The -S option isn't available for Windows.
2678 Version 6.6 06-Feb-06
2679 ---------------------
2681 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2682 in pcreposix.h. I have copied the definition from pcre.h.
2684 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2685 because pcre.h is no longer a built file.
2687 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2688 not normally included in the compiled code.
2691 Version 6.5 01-Feb-06
2692 ---------------------
2694 1. When using the partial match feature with pcre_dfa_exec(), it was not
2695 anchoring the second and subsequent partial matches at the new starting
2696 point. This could lead to incorrect results. For example, with the pattern
2697 /1234/, partially matching against "123" and then "a4" gave a match.
2699 2. Changes to pcregrep:
2701 (a) All non-match returns from pcre_exec() were being treated as failures
2702 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2703 error message is output. Some extra information is given for the
2705 probably the only errors that are likely to be caused by users (by
2706 specifying a regex that has nested indefinite repeats, for instance).
2707 If there are more than 20 of these errors, pcregrep is abandoned.
2709 (b) A binary zero was treated as data while matching, but terminated the
2710 output line if it was written out. This has been fixed: binary zeroes
2711 are now no different to any other data bytes.
2713 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2714 used to set a locale for matching. The --locale=xxxx long option has
2715 been added (no short equivalent) to specify a locale explicitly on the
2716 pcregrep command, overriding the environment variables.
2718 (d) When -B was used with -n, some line numbers in the output were one less
2719 than they should have been.
2721 (e) Added the -o (--only-matching) option.
2723 (f) If -A or -C was used with -c (count only), some lines of context were
2724 accidentally printed for the final match.
2726 (g) Added the -H (--with-filename) option.
2728 (h) The combination of options -rh failed to suppress file names for files
2729 that were found from directory arguments.
2731 (i) Added the -D (--devices) and -d (--directories) options.
2733 (j) Added the -F (--fixed-strings) option.
2735 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2737 (l) Added the --colo(u)r option.
2739 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2740 is not present by default.
2742 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2743 items such as (?R) or (?1), when the recursion could match a number of
2744 alternatives. If it matched one of the alternatives, but subsequently,
2745 outside the recursion, there was a failure, the code tried to back up into
2746 the recursion. However, because of the way PCRE is implemented, this is not
2747 possible, and the result was an incorrect result from the match.
2749 In order to prevent this happening, the specification of recursion has
2750 been changed so that all such subpatterns are automatically treated as
2751 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2753 4. I had overlooked the fact that, in some locales, there are characters for
2754 which isalpha() is true but neither isupper() nor islower() are true. In
2755 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2756 and ordfeminine) are like this. This affected the treatment of \w and \W
2757 when they appeared in character classes, but not when they appeared outside
2758 a character class. The bit map for "word" characters is now created
2759 separately from the results of isalnum() instead of just taking it from the
2760 upper, lower, and digit maps. (Plus the underscore character, of course.)
2762 5. The above bug also affected the handling of POSIX character classes such as
2763 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2764 permanent tables. Instead, the bit maps for such a class were previously
2765 created as the appropriate unions of the upper, lower, and digit bitmaps.
2766 Now they are created by subtraction from the [[:word:]] class, which has
2767 its own bitmap.
2769 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2770 It is created by subtracting the vertical space characters (\x09, \x0a,
2771 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2772 subtraction was done in the overall bitmap for a character class, meaning
2773 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2774 be recognized. This bug has been fixed.
2776 7. Patches from the folks at Google:
2778 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2779 real life, but is still worth protecting against".
2781 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2782 regular expressions".
2784 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2785 have it.
2787 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2788 "configure" and the latter not, in order to fix a problem somebody had
2789 with compiling the Arg class on HP-UX.
2791 (e) Improve the error-handling of the C++ wrapper a little bit.
2793 (f) New tests for checking recursion limiting.
2795 8. The pcre_memmove() function, which is used only if the environment does not
2796 have a standard memmove() function (and is therefore rarely compiled),
2797 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2798 returning a result (though PCRE never actually uses the result).
2800 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2801 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2802 returned instead of calling malloc() with an overflowing number that would
2803 most likely cause subsequent chaos.
2805 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2807 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2808 with this option is matched, the nmatch and pmatch options of regexec() are
2809 ignored.
2811 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2812 provided in case anyone wants to the the POSIX interface with UTF-8
2813 strings.
2815 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2816 C++ linking (needed for some HP-UX environments).
2818 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2819 (unused parameter) and in the pcre_printint() function (omitted "default"
2820 switch label when the default is to do nothing).
2822 15. Added some code to make it possible, when PCRE is compiled as a C++
2823 library, to replace subject pointers for pcre_exec() with a smart pointer
2824 class, thus making it possible to process discontinuous strings.
2826 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2827 much the same function. They were added by different people who were trying
2828 to make PCRE easy to compile on non-Unix systems. It has been suggested
2829 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2830 for compiling on Windows systems. I have therefore replaced it with
2831 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2832 defaults to "extern" for C or "extern C" for C++, which works fine on
2833 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2834 SCOPE with something explicit in config.h. In addition:
2836 (a) pcreposix.h still had just "extern" instead of either of these macros;
2837 I have replaced it with PCRE_DATA_SCOPE.
2839 (b) Functions such as _pcre_xclass(), which are internal to the library,
2840 but external in the C sense, all had PCRE_EXPORT in their definitions.
2841 This is apparently wrong for the Windows case, so I have removed it.
2842 (It makes no difference on Unix-like systems.)
2844 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2845 of recursive calls to match(). This is different to MATCH_LIMIT because
2846 that limits the total number of calls to match(), not all of which increase
2847 the depth of recursion. Limiting the recursion depth limits the amount of
2848 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2849 when PCRE is compiled, and changed at run time. A patch from Google adds
2850 this functionality to the C++ interface.
2852 18. Changes to the handling of Unicode character properties:
2854 (a) Updated the table to Unicode 4.1.0.
2856 (b) Recognize characters that are not in the table as "Cn" (undefined).
2858 (c) I revised the way the table is implemented to a much improved format
2859 which includes recognition of ranges. It now supports the ranges that
2860 are defined in UnicodeData.txt, and it also amalgamates other
2861 characters into ranges. This has reduced the number of entries in the
2862 table from around 16,000 to around 3,000, thus reducing its size
2863 considerably. I realized I did not need to use a tree structure after
2864 all - a binary chop search is just as efficient. Having reduced the
2865 number of entries, I extended their size from 6 bytes to 8 bytes to
2866 allow for more data.
2868 (d) Added support for Unicode script names via properties such as \p{Han}.
2870 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2871 matching that character.
2873 20. When matching a repeated Unicode property with a minimum greater than zero,
2874 (for example \pL{2,}), PCRE could look past the end of the subject if it
2875 reached it while seeking the minimum number of characters. This could
2876 happen only if some of the characters were more than one byte long, because
2877 there is a check for at least the minimum number of bytes.
2879 21. Refactored the implementation of \p and \P so as to be more general, to
2880 allow for more different types of property in future. This has changed the
2881 compiled form incompatibly. Anybody with saved compiled patterns that use
2882 \p or \P will have to recompile them.
2884 22. Added "Any" and "L&" to the supported property types.
2886 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2887 but give a compile time error if the value is greater than 0xff.
2889 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2890 accidentally not being installed or uninstalled.
2892 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2893 made were to insert the current release number. This seemed silly, because
2894 it made things harder for people building PCRE on systems that don't run
2895 "configure". I have turned pcre.h into a distributed file, no longer built
2896 by "configure", with the version identification directly included. There is
2897 no longer a pcre.h.in file.
2899 However, this change necessitated a change to the pcre-config script as
2900 well. It is built from pcre-config.in, and one of the substitutions was the
2901 release number. I have updated configure.ac so that ./configure now finds
2902 the release number by grepping pcre.h.
2904 26. Added the ability to run the tests under valgrind.
2907 Version 6.4 05-Sep-05
2908 ---------------------
2910 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2911 "--" to be printed when multiple files were scanned, even when none of the
2912 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2913 consider it to be a bug, and have restored the previous behaviour.
2915 2. A couple of code tidies to get rid of compiler warnings.
2917 3. The pcretest program used to cheat by referring to symbols in the library
2918 whose names begin with _pcre_. These are internal symbols that are not
2919 really supposed to be visible externally, and in some environments it is
2920 possible to suppress them. The cheating is now confined to including
2921 certain files from the library's source, which is a bit cleaner.
2923 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2924 file's purpose clearer.
2926 5. Reorganized pcre_ucp_findchar().
2929 Version 6.3 15-Aug-05
2930 ---------------------
2932 1. The file libpcre.pc.in did not have general read permission in the tarball.
2934 2. There were some problems when building without C++ support:
2936 (a) If C++ support was not built, "make install" and "make test" still
2937 tried to test it.
2939 (b) There were problems when the value of CXX was explicitly set. Some
2940 changes have been made to try to fix these, and ...
2942 (c) --disable-cpp can now be used to explicitly disable C++ support.
2944 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2945 backslash in a target when C++ was disabled. This confuses some
2946 versions of "make", apparently. Using an intermediate variable solves
2947 this. (Same for CPP_LOBJ.)
2949 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2950 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2951 necessary on certain architectures.
2953 4. Added a setting of -export-symbols-regex to the link command to remove
2954 those symbols that are exported in the C sense, but actually are local
2955 within the library, and not documented. Their names all begin with
2956 "_pcre_". This is not a perfect job, because (a) we have to except some
2957 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2958 available (and never for static libraries). I have made a note to try to
2959 find a way round (a) in the future.
2962 Version 6.2 01-Aug-05
2963 ---------------------
2965 1. There was no test for integer overflow of quantifier values. A construction
2966 such as {1111111111111111} would give undefined results. What is worse, if
2967 a minimum quantifier for a parenthesized subpattern overflowed and became
2968 negative, the calculation of the memory size went wrong. This could have
2969 led to memory overwriting.
2971 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2973 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2974 operating environments where this matters.
2976 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2977 PCRE options from within the C++ wrapper.
2979 5. Named capturing subpatterns were not being correctly counted when a pattern
2980 was compiled. This caused two problems: (a) If there were more than 100
2981 such subpatterns, the calculation of the memory needed for the whole
2982 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2983 back references of the form \12, where the number was greater than 9, were
2984 not recognized as back references, even though there were sufficient
2985 previous subpatterns.
2987 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2988 versions of gcc, e.g. 2.95.4.
2991 Version 6.1 21-Jun-05
2992 ---------------------
2994 1. There was one reference to the variable "posix" in pcretest.c that was not
2995 surrounded by "#if !defined NOPOSIX".
2997 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2998 the cross-check on the old pcre_info() function, for the benefit of the
2999 cut-down version of PCRE that is currently imported into Exim.
3001 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3002 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3003 patterns that start (?i) but I don't think that matters. The use of (?i) is
3004 just an example; this all applies to the other options as well.
3006 4. Since libtool seems to echo the compile commands it is issuing, the output
3007 from "make" can be reduced a bit by putting "@" in front of each libtool
3008 compile command.
3010 5. Patch from the folks at Google for configure.in to be a bit more thorough
3011 in checking for a suitable C++ installation before trying to compile the
3012 C++ stuff. This should fix a reported problem when a compiler was present,
3013 but no suitable headers.
3015 6. The man pages all had just "PCRE" as their title. I have changed them to
3016 be the relevant file name. I have also arranged that these names are
3017 retained in the file doc/pcre.txt, which is a concatenation in text format
3018 of all the man pages except the little individual ones for each function.
3020 7. The NON-UNIX-USE file had not been updated for the different set of source
3021 files that come with release 6. I also added a few comments about the C++
3022 wrapper.
3025 Version 6.0 07-Jun-05
3026 ---------------------
3028 1. Some minor internal re-organization to help with my DFA experiments.
3030 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3031 didn't matter for the library itself when fully configured, but did matter
3032 when compiling without UCP support, or within Exim, where the ucp files are
3033 not imported.
3035 3. Refactoring of the library code to split up the various functions into
3036 different source modules. The addition of the new DFA matching code (see
3037 below) to a single monolithic source would have made it really too
3038 unwieldy, quite apart from causing all the code to be include in a
3039 statically linked application, when only some functions are used. This is
3040 relevant even without the DFA addition now that patterns can be compiled in
3041 one application and matched in another.
3043 The downside of splitting up is that there have to be some external
3044 functions and data tables that are used internally in different modules of
3045 the library but which are not part of the API. These have all had their
3046 names changed to start with "_pcre_" so that they are unlikely to clash
3047 with other external names.
3049 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3050 a different (DFA) algorithm. Although it is slower than the original
3051 function, it does have some advantages for certain types of matching
3052 problem.
3054 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3055 including restarting after a partial match.
3057 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3058 defined when compiling for Windows was sent to me. I have put it into the
3059 code, though I have no means of testing or verifying it.
3061 7. Added the pcre_refcount() auxiliary function.
3063 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3064 match before or at the first newline in the subject string. In pcretest,
3065 the /f option on a pattern can be used to set this.
3067 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3068 would behave wrongly. This has been present in PCRE since release 4.0.
3070 10. A number of changes to the pcregrep command:
3072 (a) Refactored how -x works; insert ^(...)$ instead of setting
3073 PCRE_ANCHORED and checking the length, in preparation for adding
3074 something similar for -w.
3076 (b) Added the -w (match as a word) option.
3078 (c) Refactored the way lines are read and buffered so as to have more
3079 than one at a time available.
3081 (d) Implemented a pcregrep test script.
3083 (e) Added the -M (multiline match) option. This allows patterns to match
3084 over several lines of the subject. The buffering ensures that at least
3085 8K, or the rest of the document (whichever is the shorter) is available
3086 for matching (and similarly the previous 8K for lookbehind assertions).
3088 (f) Changed the --help output so that it now says
3090 -w, --word-regex(p)
3092 instead of two lines, one with "regex" and the other with "regexp"
3093 because that confused at least one person since the short forms are the
3094 same. (This required a bit of code, as the output is generated
3095 automatically from a table. It wasn't just a text change.)
3097 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3098 option but starts with a hyphen. Could be a pattern or a path name
3099 starting with a hyphen, for instance.
3101 (h) "-" can be given as a file name to represent stdin.
3103 (i) When file names are being printed, "(standard input)" is used for
3104 the standard input, for compatibility with GNU grep. Previously
3105 "<stdin>" was used.
3107 (j) The option --label=xxx can be used to supply a name to be used for
3108 stdin when file names are being printed. There is no short form.
3110 (k) Re-factored the options decoding logic because we are going to add
3111 two more options that take data. Such options can now be given in four
3112 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3114 (l) Added the -A, -B, and -C options for requesting that lines of context
3115 around matches be printed.
3117 (m) Added the -L option to print the names of files that do not contain
3118 any matching lines, that is, the complement of -l.
3120 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3121 continue to scan other files.
3123 (o) The -s option was incorrectly implemented. For compatibility with other
3124 greps, it now suppresses the error message for a non-existent or non-
3125 accessible file (but not the return code). There is a new option called
3126 -q that suppresses the output of matching lines, which was what -s was
3127 previously doing.
3129 (p) Added --include and --exclude options to specify files for inclusion
3130 and exclusion when recursing.
3132 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3133 Hopefully, it now does.
3135 12. Missing cast in pcre_study().
3137 13. Added an "uninstall" target to the makefile.
3139 14. Replaced "extern" in the function prototypes in Makefile.in with
3140 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3141 world, but is set differently for Windows.
3143 15. Added a second compiling function called pcre_compile2(). The only
3144 difference is that it has an extra argument, which is a pointer to an
3145 integer error code. When there is a compile-time failure, this is set
3146 non-zero, in addition to the error test pointer being set to point to an
3147 error message. The new argument may be NULL if no error number is required
3148 (but then you may as well call pcre_compile(), which is now just a
3149 wrapper). This facility is provided because some applications need a
3150 numeric error indication, but it has also enabled me to tidy up the way
3151 compile-time errors are handled in the POSIX wrapper.
3153 16. Added VPATH=.libs to the makefile; this should help when building with one
3154 prefix path and installing with another. (Or so I'm told by someone who
3155 knows more about this stuff than I do.)
3157 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3158 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3159 match everything, including newlines. This is not POSIX-compatible, but
3160 somebody wanted the feature. From pcretest it can be activated by using
3161 both the P and the s flags.
3163 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3165 19. libpcre.pc was being incorrectly installed as executable.
3167 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3168 it now also looks for '\r' so that it will work unmodified on Windows.
3170 21. Added Google's contributed C++ wrapper to the distribution.
3172 22. Added some untidy missing memory free() calls in pcretest, to keep
3173 Electric Fence happy when testing.
3177 Version 5.0 13-Sep-04
3178 ---------------------
3180 1. Internal change: literal characters are no longer packed up into items
3181 containing multiple characters in a single byte-string. Each character
3182 is now matched using a separate opcode. However, there may be more than one
3183 byte in the character in UTF-8 mode.
3185 2. The pcre_callout_block structure has two new fields: pattern_position and
3186 next_item_length. These contain the offset in the pattern to the next match
3187 item, and its length, respectively.
3189 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3190 insertion of callouts before each pattern item. Added the /C option to
3191 pcretest to make use of this.
3193 4. On the advice of a Windows user, the lines
3195 #if defined(_WIN32) || defined(WIN32)
3196 _setmode( _fileno( stdout ), 0x8000 );
3197 #endif /* defined(_WIN32) || defined(WIN32) */
3199 have been added to the source of pcretest. This apparently does useful
3200 magic in relation to line terminators.
3202 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3203 for the benefit of those environments where the "b" makes a difference.
3205 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3206 to know about it. I have put a hack into configure.in that adds in code
3207 to set GCC=yes if CC=icc. This seems to end up at a point in the
3208 generated configure script that is early enough to affect the setting of
3209 compiler options, which is what is needed, but I have no means of testing
3210 whether it really works. (The user who reported this had patched the
3211 generated configure script, which of course I cannot do.)
3213 LATER: After change 22 below (new libtool files), the configure script
3214 seems to know about icc (and also ecc). Therefore, I have commented out
3215 this hack in configure.in.
3217 7. Added support for pkg-config (2 patches were sent in).
3219 8. Negated POSIX character classes that used a combination of internal tables
3220 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3221 [[:^ascii]]. Typically, they would match almost any characters. The other
3222 POSIX classes were not broken in this way.
3224 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3225 to find the match, as PCRE was deluded into thinking that the match had to
3226 start at the start point or following a newline. The same bug applied to
3227 patterns with negative forward assertions or any backward assertions
3228 preceding ".*" at the start, unless the pattern required a fixed first
3229 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3231 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3232 starting at the last subject character, bytes beyond the end of the subject
3233 string were read.
3235 11. Renamed the variable "class" as "classbits" to make life easier for C++
3236 users. (Previously there was a macro definition, but it apparently wasn't
3237 enough.)
3239 12. Added the new field "tables" to the extra data so that tables can be passed
3240 in at exec time, or the internal tables can be re-selected. This allows
3241 a compiled regex to be saved and re-used at a later time by a different
3242 program that might have everything at different addresses.
3244 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3245 -R library as well as a -L library.
3247 14. The debugging options of pcretest (-d on the command line or D on a
3248 pattern) showed incorrect output for anything following an extended class
3249 that contained multibyte characters and which was followed by a quantifier.
3251 15. Added optional support for general category Unicode character properties
3252 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3253 support. It adds about 90K to the size of the library. The meanings of the
3254 inbuilt class escapes such as \d and \s have NOT been changed.
3256 16. Updated pcredemo.c to include calls to free() to release the memory for the
3257 compiled pattern.
3259 17. The generated file chartables.c was being created in the source directory
3260 instead of in the building directory. This caused the build to fail if the
3261 source directory was different from the building directory, and was
3262 read-only.
3264 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3265 file. No doubt somebody will tell me if they don't make sense... Also added
3266 Dan Mooney's comments about building on OpenVMS.
3268 19. Added support for partial matching via the PCRE_PARTIAL option for
3269 pcre_exec() and the \P data escape in pcretest.
3271 20. Extended pcretest with 3 new pattern features:
3273 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3274 write the compiled pattern to the file whose name is "rest-of-line".
3275 This is a straight binary dump of the data, with the saved pointer to
3276 the character tables forced to be NULL. The study data, if any, is
3277 written too. After writing, pcretest reads a new pattern.
3279 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3280 compiled pattern from the given file. There must not be any
3281 occurrences of "<" in the file name (pretty unlikely); if there are,
3282 pcretest will instead treat the initial "<" as a pattern delimiter.
3283 After reading in the pattern, pcretest goes on to read data lines as
3284 usual.
3286 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3287 and 16-bit fields in a compiled pattern, to simulate a pattern that
3288 was compiled on a host of opposite endianness.
3290 21. The pcre-exec() function can now cope with patterns that were compiled on
3291 hosts of opposite endianness, with this restriction:
3293 As for any compiled expression that is saved and used later, the tables
3294 pointer field cannot be preserved; the extra_data field in the arguments
3295 to pcre_exec() should be used to pass in a tables address if a value
3296 other than the default internal tables were used at compile time.
3298 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3299 now diagnosed as an error. Previously, most of the time, a negative number
3300 would have been treated as zero, but if in addition "ovector" was passed as
3301 NULL, a crash could occur.
3303 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3304 new versions from the libtool 1.5 distribution (the last one is a copy of
3305 a file called libtool.m4). This seems to have fixed the need to patch
3306 "configure" to support Darwin 1.3 (which I used to do). However, I still
3307 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3308 workstation).
3310 24. Changed the PCRE licence to be the more standard "BSD" licence.
3313 Version 4.5 01-Dec-03
3314 ---------------------
3316 1. There has been some re-arrangement of the code for the match() function so
3317 that it can be compiled in a version that does not call itself recursively.
3318 Instead, it keeps those local variables that need separate instances for
3319 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3320 needs to "recurse". Keeping track of where control must go is done by means
3321 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3322 hide most of the details from the main code, and operates only if
3323 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3324 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3325 operating.
3327 To make it easier for callers to provide specially tailored get/free
3328 functions for this usage, two new functions, pcre_stack_malloc, and
3329 pcre_stack_free, are used. They are always called in strict stacking order,
3330 and the size of block requested is always the same.
3332 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3333 PCRE has been compiled to use the stack or the heap for recursion. The
3334 -C option of pcretest uses this to show which version is compiled.
3336 A new data escape \S, is added to pcretest; it causes the amounts of store
3337 obtained and freed by both kinds of malloc/free at match time to be added
3338 to the output.
3340 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3341 what's available on my current Linux desktop machine.
3343 3. When matching a UTF-8 string, the test for a valid string at the start has
3344 been extended. If start_offset is not zero, PCRE now checks that it points
3345 to a byte that is the start of a UTF-8 character. If not, it returns
3346 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3347 this is necessary because there may be backward assertions in the pattern.
3348 When matching the same subject several times, it may save resources to use
3349 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3351 4. The code for checking the validity of UTF-8 strings has been tightened so
3352 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3353 containing "overlong sequences".
3355 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3356 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3357 should have been "&", but it just so happened that all the cases this let
3358 through by mistake were picked up later in the function.
3360 6. I had used a variable called "isblank" - this is a C99 function, causing
3361 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3363 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3364 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3365 so the version is shown; (c) stop "make test" echoing "./RunTest".
3367 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3369 9. The prototype for memmove() for systems that don't have it was using
3370 size_t, but the inclusion of the header that defines size_t was later. I've
3371 moved the #includes for the C headers earlier to avoid this.
3373 10. Added some adjustments to the code to make it easier to compiler on certain
3374 special systems:
3376 (a) Some "const" qualifiers were missing.
3377 (b) Added the macro EXPORT before all exported functions; by default this
3378 is defined to be empty.
3379 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3380 that it reads its output file name as an argument instead of writing
3381 to the standard output and assuming this can be redirected.
3383 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3384 class containing characters with values greater than 255, PCRE compilation
3385 went into a loop.
3387 12. A recursive reference to a subpattern that was within another subpattern
3388 that had a minimum quantifier of zero caused PCRE to crash. For example,
3389 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3390 recursion. If the recursively-called subpattern itself had a zero repeat,
3391 that was OK.
3393 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3394 buffer into which it was copied (for escape processing) was still set at
3395 1024, so long lines caused crashes.
3397 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3398 "internal error: code overflow...". This applied to any character class
3399 that was followed by a possessive quantifier.
3401 15. Modified the Makefile to add libpcre.la as a prerequisite for
3402 libpcreposix.la because I was told this is needed for a parallel build to
3403 work.
3405 16. If a pattern that contained .* following optional items at the start was
3406 studied, the wrong optimizing data was generated, leading to matching
3407 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3408 matching string must start with a or b or c. The correct conclusion for
3409 this pattern is that a match can start with any character.
3412 Version 4.4 13-Aug-03
3413 ---------------------
3415 1. In UTF-8 mode, a character class containing characters with values between
3416 127 and 255 was not handled correctly if the compiled pattern was studied.
3417 In fixing this, I have also improved the studying algorithm for such
3418 classes (slightly).
3420 2. Three internal functions had redundant arguments passed to them. Removal
3421 might give a very teeny performance improvement.
3423 3. Documentation bug: the value of the capture_top field in a callout is *one
3424 more than* the number of the hightest numbered captured substring.
3426 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3427 in incorrectly linking with a previously installed version. They now link
3428 explicitly with libpcre.la.
3430 5. configure.in no longer needs to recognize Cygwin specially.
3432 6. A problem in pcre.in for Windows platforms is fixed.
3434 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3435 pcretest, it used to include the size of the study block as part of its
3436 output. Unfortunately, the structure contains a field that has a different
3437 size on different hardware architectures. This meant that the tests that
3438 showed this size failed. As the block is currently always of a fixed size,
3439 this information isn't actually particularly useful in pcretest output, so
3440 I have just removed it.
3442 8. Three pre-processor statements accidentally did not start in column 1.
3443 Sadly, there are *still* compilers around that complain, even though
3444 standard C has not required this for well over a decade. Sigh.
3446 9. In pcretest, the code for checking callouts passed small integers in the
3447 callout_data field, which is a void * field. However, some picky compilers
3448 complained about the casts involved for this on 64-bit systems. Now
3449 pcretest passes the address of the small integer instead, which should get
3450 rid of the warnings.
3452 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3453 both compile and run time, and gives an error if an invalid UTF-8 sequence
3454 is found. There is a option for disabling this check in cases where the
3455 string is known to be correct and/or the maximum performance is wanted.
3457 11. In response to a bug report, I changed one line in Makefile.in from
3459 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3460 to
3461 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3463 to look similar to other lines, but I have no way of telling whether this
3464 is the right thing to do, as I do not use Windows. No doubt I'll get told
3465 if it's wrong...
3468 Version 4.3 21-May-03
3469 ---------------------
3471 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3472 Makefile.
3474 2. Some refactoring to improve the quality of the code:
3476 (i) The utf8_table... variables are now declared "const".
3478 (ii) The code for \cx, which used the "case flipping" table to upper case
3479 lower case letters, now just substracts 32. This is ASCII-specific,
3480 but the whole concept of \cx is ASCII-specific, so it seems
3481 reasonable.
3483 (iii) PCRE was using its character types table to recognize decimal and
3484 hexadecimal digits in the pattern. This is silly, because it handles
3485 only 0-9, a-f, and A-F, but the character types table is locale-
3486 specific, which means strange things might happen. A private
3487 table is now used for this - though it costs 256 bytes, a table is
3488 much faster than multiple explicit tests. Of course, the standard
3489 character types table is still used for matching digits in subject
3490 strings against \d.
3492 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3493 ending in _t are). So I've renamed it as ESC_tee.
3495 3. The first argument for regexec() in the POSIX wrapper should have been
3496 defined as "const".
3498 4. Changed pcretest to use malloc() for its buffers so that they can be
3499 Electric Fenced for debugging.
3501 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3502 to read one or more bytes before the start of the subject string. Often this
3503 had no effect on PCRE's behaviour, but in some circumstances it could
3504 provoke a segmentation fault.
3506 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3507 to try to read one or more bytes before the start of the subject string.
3509 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3510 UTF-8 support could misbehave in various ways if the subject string
3511 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3512 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3513 back over UTF-8 characters.)
3516 Version 4.2 14-Apr-03
3517 ---------------------
3519 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3521 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3522 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3523 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3524 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3525 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3527 Note: automatic setting of the BUILD variables is not yet working
3528 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3529 compile-time but not at link-time
3530 [LINK]: use for linking executables only
3531 make different versions for Windows and non-Windows
3532 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3533 libraries
3534 [LINK_FOR_BUILD]: new variable
3535 [OBJEXT]: use throughout
3536 [EXEEXT]: use throughout
3537 <winshared>: new target
3538 <wininstall>: new target
3539 <dftables.o>: use native compiler
3540 <dftables>: use native linker
3541 <install>: handle Windows platform correctly
3542 <clean>: ditto
3543 <check>: ditto
3544 copy DLL to top builddir before testing
3546 As part of these changes, -no-undefined was removed again. This was reported
3547 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3548 in any case.
3550 3. Some tidies to get rid of compiler warnings:
3552 . In the match_data structure, match_limit was an unsigned long int, whereas
3553 match_call_count was an int. I've made them both unsigned long ints.
3555 . In pcretest the fact that a const uschar * doesn't automatically cast to
3556 a void * provoked a warning.
3558 . Turning on some more compiler warnings threw up some "shadow" variables
3559 and a few more missing casts.
3561 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3562 option, a class that contained a single character with a value between 128
3563 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3565 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3566 option, a class that contained several characters, but with at least one
3567 whose value was between 128 and 255 caused PCRE to crash.
3570 Version 4.1 12-Mar-03
3571 ---------------------
3573 1. Compiling with gcc -pedantic found a couple of places where casts were
3574 needed, and a string in dftables.c that was longer than standard compilers are
3575 required to support.
3577 2. Compiling with Sun's compiler found a few more places where the code could
3578 be tidied up in order to avoid warnings.
3580 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3581 first of these names is deprecated in the latest Autoconf in favour of the name
3582 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3583 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3584 analogy I have changed it to CFLAGS_FOR_BUILD.
3586 4. Added -no-undefined to the linking command in the Makefile, because this is
3587 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3588 linking step for the pcreposix library.
3590 5. PCRE was failing to diagnose the case of two named groups with the same
3591 name.
3593 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3594 literal character that is needed in the subject for a match, and scans along to
3595 ensure that it is present before embarking on the full matching process. This
3596 saves time in cases of nested unlimited repeats that are never going to match.
3597 Problem: the scan can take a lot of time if the subject is very long (e.g.
3598 megabytes), thus penalizing straightforward matches. It is now done only if the
3599 amount of subject to be scanned is less than 1000 bytes.
3601 7. A lesser problem with the same optimization is that it was recording the
3602 first character of an anchored pattern as "needed", thus provoking a search
3603 right along the subject, even when the first match of the pattern was going to
3604 fail. The "needed" character is now not set for anchored patterns, unless it
3605 follows something in the pattern that is of non-fixed length. Thus, it still
3606 fulfils its original purpose of finding quick non-matches in cases of nested
3607 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3610 Version 4.0 17-Feb-03
3611 ---------------------
3613 1. If a comment in an extended regex that started immediately after a meta-item
3614 extended to the end of string, PCRE compiled incorrect data. This could lead to
3615 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3617 2. Moved to autoconf 2.53 and libtool 1.4.2.
3619 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3620 the special perltest8 script is no longer needed - all the tests can be run
3621 from a single perltest script.
3623 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3624 by \s. It has now been removed in PCRE. This means it isn't recognized as
3625 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3626 class [:space:] *does* include VT, thereby creating a mess.
3628 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3629 space and tab.
3631 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3632 its new features into the main test script, reducing the number of scripts.
3634 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3635 were backward compatible, and made the (?i) apply to the whole pattern, as if
3636 /i were given. Now it behaves more logically, and applies the option setting
3637 only to what follows. PCRE has been changed to follow suit. However, if it
3638 finds options settings right at the start of the pattern, it extracts them into
3639 the global options, as before. Thus, they show up in the info data.
3641 8. Added support for the \Q...\E escape sequence. Characters in between are
3642 treated as literals. This is slightly different from Perl in that $ and @ are
3643 also handled as literals inside the quotes. In Perl, they will cause variable
3644 interpolation. Note the following examples:
3646 Pattern PCRE matches Perl matches
3648 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3649 \Qabc\$xyz\E abc\$xyz abc\$xyz
3650 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3652 For compatibility with Perl, \Q...\E sequences are recognized inside character
3653 classes as well as outside them.
3655 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3656 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3657 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3658 signed/unsigned warnings.
3660 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3661 option for pcretest, so I've replaced it by a simple function that does just
3662 that job.
3664 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3665 "pcregrep -".
3667 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3668 Java package. This provides some syntactic sugar for simple cases of what my
3669 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3670 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3671 item, you can use this simplified notation. Note that only makes sense with
3672 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3673 greediness, whatever the setting of the PCRE_UNGREEDY option.
3675 13. A change of greediness default within a pattern was not taking effect at
3676 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3677 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3678 was abstracted outside.
3680 14. PCRE now supports the \G assertion. It is true when the current matching
3681 position is at the start point of the match. This differs from \A when the
3682 starting offset is non-zero. Used with the /g option of pcretest (or similar
3683 code), it works in the same way as it does for Perl's /g option. If all
3684 alternatives of a regex begin with \G, the expression is anchored to the start
3685 match position, and the "anchored" flag is set in the compiled expression.
3687 15. Some bugs concerning the handling of certain option changes within patterns
3688 have been fixed. These applied to options other than (?ims). For example,
3689 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3690 been the other way round. Some of this was related to change 7 above.
3692 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3693 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3694 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3695 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3697 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3698 mode. Unlike ".", it always matches newline, whatever the setting of
3699 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3700 assertions. Perl allows it, but it doesn't (in general) work because it can't
3701 calculate the length of the lookbehind. At least, that's the case for Perl
3702 5.8.0 - I've been told they are going to document that it doesn't work in
3703 future.
3705 18. Added an error diagnosis for escapes that PCRE does not support: these are
3706 \L, \l, \N, \P, \p, \U, \u, and \X.
3708 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3709 reading past the end of the pattern in cases such as /[abcd/.
3711 20. PCRE was getting more memory than necessary for patterns with classes that
3712 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3714 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3715 compiling PCRE for use with Virtual Pascal.
3717 22. Small fix to the Makefile to make it work properly if the build is done
3718 outside the source tree.
3720 23. Added a new extension: a condition to go with recursion. If a conditional
3721 subpattern starts with (?(R) the "true" branch is used if recursion has
3722 happened, whereas the "false" branch is used only at the top level.
3724 24. When there was a very long string of literal characters (over 255 bytes
3725 without UTF support, over 250 bytes with UTF support), the computation of how
3726 much memory was required could be incorrect, leading to segfaults or other
3727 strange effects.
3729 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3730 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3731 there was a subsequent back reference to those brackets. This meant that, for
3732 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3733 possible to check for precisely this case. All we can do is abandon the
3734 optimization if .* occurs inside capturing brackets when there are any back
3735 references whatsoever. (See below for a better fix that came later.)
3737 26. The handling of the optimization for finding the first character of a
3738 non-anchored pattern, and for finding a character that is required later in the
3739 match were failing in some cases. This didn't break the matching; it just
3740 failed to optimize when it could. The way this is done has been re-implemented.
3742 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3744 28. Added a new feature that provides some of the functionality that Perl
3745 provides with (?{...}). The facility is termed a "callout". The way it is done
3746 in PCRE is for the caller to provide an optional function, by setting
3747 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3748 global variable. By default it is unset, which disables all calling out. To get
3749 the function called, the regex must include (?C) at appropriate points. This
3750 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3751 This provides a means of identifying different callout points. When PCRE
3752 reaches such a point in the regex, if pcre_callout has been set, the external
3753 function is called. It is provided with data in a structure called
3754 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3755 matching continues; if it returns a non-zero value, the match at the current
3756 point fails. However, backtracking will occur if possible. [This was changed
3757 later and other features added - see item 49 below.]
3759 29. pcretest is upgraded to test the callout functionality. It provides a
3760 callout function that displays information. By default, it shows the start of
3761 the match and the current position in the text. There are some new data escapes
3762 to vary what happens:
3764 \C+ in addition, show current contents of captured substrings
3765 \C- do not supply a callout function
3766 \C!n return 1 when callout number n is reached
3767 \C!n!m return 1 when callout number n is reached for the mth time
3769 30. If pcregrep was called with the -l option and just a single file name, it
3770 output "<stdin>" if a match was found, instead of the file name.
3772 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3773 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3774 pcre_exec(). This saves a malloc/free per call. The default value of
3775 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3776 when configuring.
3778 32. The default maximum size of a compiled pattern is 64K. There have been a
3779 few cases of people hitting this limit. The code now uses macros to handle the
3780 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3781 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3782 configuring. Tests 2 and 5 work only with 2-byte links because they output
3783 debugging information about compiled patterns.
3785 33. Internal code re-arrangements:
3787 (a) Moved the debugging function for printing out a compiled regex into
3788 its own source file (printint.c) and used #include to pull it into
3789 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3790 separate copies.
3792 (b) Defined the list of op-code names for debugging as a macro in
3793 internal.h so that it is next to the definition of the opcodes.
3795 (c) Defined a table of op-code lengths for simpler skipping along compiled
3796 code. This is again a macro in internal.h so that it is next to the
3797 definition of the opcodes.
3799 34. Added support for recursive calls to individual subpatterns, along the
3800 lines of Robin Houston's patch (but implemented somewhat differently).
3802 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3803 allow it to read and process whole directories in Win32. This code was
3804 contributed by Lionel Fourquaux; it has not been tested by me.
3806 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3807 used to name a group. Names consist of alphanumerics and underscores, and must
3808 be unique. Back references use the syntax (?P=name) and recursive calls use
3809 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3810 numbers. The function pcre_fullinfo() can be used after compilation to extract
3811 a name/number map. There are three relevant calls:
3813 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3814 PCRE_INFO_NAMECOUNT yields the number of entries
3815 PCRE_INFO_NAMETABLE yields a pointer to the map.
3817 The map is a vector of fixed-size entries. The size of each entry depends on
3818 the length of the longest name used. The first two bytes of each entry are the
3819 group number, most significant byte first. There follows the corresponding
3820 name, zero terminated. The names are in alphabetical order.
3822 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3823 case instead of 255. Making it the same both with and without UTF-8 support
3824 means that the same test output works with both.
3826 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3827 calling malloc() with a zero argument.
3829 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3830 optimization. I've improved things by keeping a bitmap of backreferences with
3831 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3832 fact referenced, the optimization can be applied. It is unlikely that a
3833 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3834 the match to follow \n) will appear inside brackets with a number greater than
3835 31, but if it does, any back reference > 31 suppresses the optimization.
3837 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3838 of disabling numbered capturing parentheses. Any opening parenthesis that is
3839 not followed by ? behaves as if it were followed by ?: but named parentheses
3840 can still be used for capturing (and they will acquire numbers in the usual
3841 way).
3843 41. Redesigned the return codes from the match() function into yes/no/error so
3844 that errors can be passed back from deep inside the nested calls. A malloc
3845 failure while inside a recursive subpattern call now causes the
3846 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3848 42. It is now possible to set a limit on the number of times the match()
3849 function is called in a call to pcre_exec(). This facility makes it possible to
3850 limit the amount of recursion and backtracking, though not in a directly
3851 obvious way, because the match() function is used in a number of different
3852 circumstances. The count starts from zero for each position in the subject
3853 string (for non-anchored patterns). The default limit is, for compatibility, a
3854 large number, namely 10 000 000. You can change this in two ways:
3856 (a) When configuring PCRE before making, you can use --with-match-limit=n
3857 to set a default value for the compiled library.
3859 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3860 a different value is set. See 45 below.
3862 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3864 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3865 of things that can be changed at compile time. The first argument specifies
3866 what is wanted and the second points to where the information is to be placed.
3867 The current list of available information is:
3871 The output is an integer that is set to one if UTF-8 support is available;
3872 otherwise it is set to zero.
3876 The output is an integer that it set to the value of the code that is used for
3877 newline. It is either LF (10) or CR (13).
3881 The output is an integer that contains the number of bytes used for internal
3882 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3886 The output is an integer that contains the threshold above which the POSIX
3887 interface uses malloc() for output vectors. See item 31 above.
3891 The output is an unsigned integer that contains the default limit of the number
3892 of match() calls in a pcre_exec() execution. See 42 above.
3894 44. pcretest has been upgraded by the addition of the -C option. This causes it
3895 to extract all the available output from the new pcre_config() function, and to
3896 output it. The program then exits immediately.
3898 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3899 order to support additional features. One way would have been to define
3900 pcre_exec2() (for example) with extra arguments, but this would not have been
3901 extensible, and would also have required all calls to the original function to
3902 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3903 is used for passing in "extra" data from pcre_study().
3905 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3906 contains the following fields:
3908 flags a bitmap indicating which of the following fields are set
3909 study_data opaque data from pcre_study()
3910 match_limit a way of specifying a limit on match() calls for a specific
3911 call to pcre_exec()
3912 callout_data data for callouts (see 49 below)
3914 The flag bits are also defined in pcre.h, and are
3920 The pcre_study() function now returns one of these new pcre_extra blocks, with
3921 the actual study data pointed to by the study_data field, and the
3922 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3923 before. That is, this change is entirely upwards-compatible and requires no
3924 change to existing code.
3926 If you want to pass in additional data to pcre_exec(), you can either place it
3927 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3928 block.
3930 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3931 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3932 times with different match limits, until it finds the minimum value needed for
3933 pcre_exec() to complete. The value is then output. This can be instructive; for
3934 most simple matches the number is quite small, but for pathological cases it
3935 gets very large very quickly.
3937 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3938 returns the size of the data block pointed to by the study_data field in a
3939 pcre_extra block, that is, the value that was passed as the argument to
3940 pcre_malloc() when PCRE was getting memory in which to place the information
3941 created by pcre_study(). The fourth argument should point to a size_t variable.
3942 pcretest has been extended so that this information is shown after a successful
3943 pcre_study() call when information about the compiled regex is being displayed.
3945 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3946 because what follows is always an absolute path. (Later: it turns out that this
3947 is more than cosmetic for MinGW, because it doesn't like empty path
3948 components.)
3950 49. Some changes have been made to the callout feature (see 28 above):
3952 (i) A callout function now has three choices for what it returns:
3954 0 => success, carry on matching
3955 > 0 => failure at this point, but backtrack if possible
3956 < 0 => serious error, return this value from pcre_exec()
3958 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3959 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3960 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3961 use by callout functions. It will never be used by PCRE itself.
3963 (ii) The pcre_extra structure (see 45 above) has a void * field called
3964 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3965 pcre_callout_block structure has a field of the same name. The contents of
3966 the field passed in the pcre_extra structure are passed to the callout
3967 function in the corresponding field in the callout block. This makes it
3968 easier to use the same callout-containing regex from multiple threads. For
3969 testing, the pcretest program has a new data escape
3971 \C*n pass the number n (may be negative) as callout_data
3973 If the callout function in pcretest receives a non-zero value as
3974 callout_data, it returns that value.
3976 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3977 there were some redundant $(CFLAGS) in commands that are now specified as
3978 $(LINK), which already includes $(CFLAGS).
3980 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3981 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3982 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3983 one-byte characters throughout. Note that case-insensitive matching applies
3984 only to characters whose values are less than 256. PCRE doesn't support the
3985 notion of cases for higher-valued characters.
3987 (i) A character class whose characters are all within 0-255 is handled as
3988 a bit map, and the map is inverted for negative classes. Previously, a
3989 character > 255 always failed to match such a class; however it should
3990 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3992 (ii) A negated character class with a single character < 255 is coded as
3993 "not this character" (OP_NOT). This wasn't working properly when the test
3994 character was multibyte, either singly or repeated.
3996 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3997 mode, for example: \x{100}{2,3}.
3999 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4000 singly or repeated) now correctly test multibyte characters. However,
4001 PCRE doesn't recognize any characters with values greater than 255 as
4002 digits, spaces, or word characters. Such characters always match \D, \S,
4003 and \W, and never match \d, \s, or \w.
4005 (v) Classes may now contain characters and character ranges with values
4006 greater than 255. For example: [ab\x{100}-\x{400}].
4008 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4009 PCRE in UTF-8 mode.
4011 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4012 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4013 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4014 value.)
4016 53. The single man page has become too large. I have therefore split it up into
4017 a number of separate man pages. These also give rise to individual HTML pages;
4018 these are now put in a separate directory, and there is an index.html page that
4019 lists them all. Some hyperlinking between the pages has been installed.
4021 54. Added convenience functions for handling named capturing parentheses.
4023 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4024 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4025 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4026 are faulted.
4028 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4029 calling configure. These values are used when compiling the dftables.c program
4030 which is run to generate the source of the default character tables. They
4031 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4032 you will need to set these values.
4034 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4037 Version 3.9 02-Jan-02
4038 ---------------------
4040 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4042 2. If --disable-static was given, the building process failed when trying to
4043 build pcretest and pcregrep. (For some reason it was using libtool to compile
4044 them, which is not right, as they aren't part of the library.)
4047 Version 3.8 18-Dec-01
4048 ---------------------
4050 1. The experimental UTF-8 code was completely screwed up. It was packing the
4051 bytes in the wrong order. How dumb can you get?
4054 Version 3.7 29-Oct-01
4055 ---------------------
4057 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4058 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4059 this didn't happen under Solaris 8, where I normally test things.
4061 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4062 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4063 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4064 this fix an hour or so after the initial 3.7 release.)
4067 Version 3.6 23-Oct-01
4068 ---------------------
4070 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4071 offsets passed as NULL with zero offset count.
4073 2. The config.guess and config.sub files had not been updated when I moved to
4074 the latest autoconf.
4077 Version 3.5 15-Aug-01
4078 ---------------------
4080 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4081 had been forgotten.
4083 2. By using declared but undefined structures, we can avoid using "void"
4084 definitions in pcre.h while keeping the internal definitions of the structures
4085 private.
4087 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4088 user point of view, this means that both static and shared libraries are built
4089 by default, but this can be individually controlled. More of the work of
4090 handling this static/shared cases is now inside libtool instead of PCRE's make
4091 file.
4093 4. The pcretest utility is now installed along with pcregrep because it is
4094 useful for users (to test regexs) and by doing this, it automatically gets
4095 relinked by libtool. The documentation has been turned into a man page, so
4096 there are now .1, .txt, and .html versions in /doc.
4098 5. Upgrades to pcregrep:
4099 (i) Added long-form option names like gnu grep.
4100 (ii) Added --help to list all options with an explanatory phrase.
4101 (iii) Added -r, --recursive to recurse into sub-directories.
4102 (iv) Added -f, --file to read patterns from a file.
4104 6. pcre_exec() was referring to its "code" argument before testing that
4105 argument for NULL (and giving an error if it was NULL).
4107 7. Upgraded Makefile.in to allow for compiling in a different directory from
4108 the source directory.
4110 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4111 options bits, the pointer it was passed was to an int instead of to an unsigned
4112 long int. This mattered only on 64-bit systems.
4114 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4115 generated) instead of pcre.in, which it its source. Also made the same change
4116 in several of the .c files.
4118 10. A new release of gcc defines printf() as a macro, which broke pcretest
4119 because it had an ifdef in the middle of a string argument for printf(). Fixed
4120 by using separate calls to printf().
4122 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4123 script, to force use of CR or LF instead of \n in the source. On non-Unix
4124 systems, the value can be set in config.h.
4126 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4127 absolute limit. Changed the text of the error message to make this clear, and
4128 likewise updated the man page.
4130 13. The limit of 99 on the number of capturing subpatterns has been removed.
4131 The new limit is 65535, which I hope will not be a "real" limit.
4134 Version 3.4 22-Aug-00
4135 ---------------------
4137 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4139 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4142 Version 3.3 01-Aug-00
4143 ---------------------
4145 1. If an octal character was given, but the value was greater than \377, it
4146 was not getting masked to the least significant bits, as documented. This could
4147 lead to crashes in some systems.
4149 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4150 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4152 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4153 These just pass their arguments on to (pcre_free)(), but they are provided
4154 because some uses of PCRE bind it to non-C systems that can call its functions,
4155 but cannot call free() or pcre_free() directly.
4157 4. Add "make test" as a synonym for "make check". Corrected some comments in
4158 the Makefile.
4160 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4161 Makefile.
4163 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4164 command called pgrep for grepping around the active processes.
4166 7. Added the beginnings of support for UTF-8 character strings.
4168 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4169 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4170 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4171 out for the ar command.)
4174 Version 3.2 12-May-00
4175 ---------------------
4177 This is purely a bug fixing release.
4179 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4180 of ZA. This was just one example of several cases that could provoke this bug,
4181 which was introduced by change 9 of version 2.00. The code for breaking
4182 infinite loops after an iteration that matches an empty string was't working
4183 correctly.
4185 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4186 when matched against abbab (for example). After matching an empty string, it
4187 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4188 caused it to match further down the string than it should.
4190 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4191 was there because it doesn't seem to be needed, and it causes trouble on some
4192 systems, as it is not a Standard C header. It has been removed.
4194 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4195 were reported on the Macintosh. The changes were from
4197 while ((c = *(++ptr)) != 0 && c != '\n');
4198 to
4199 while ((c = *(++ptr)) != 0 && c != '\n') ;
4201 Totally extraordinary, but if that's what it takes...
4203 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4204 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4205 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4206 assumes the way PCRE uses memmove() (always moving upwards).
4208 6. PCRE is being used in one environment where strchr() is not available. There
4209 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4210 faster code anyway.
4213 Version 3.1 09-Feb-00
4214 ---------------------
4216 The only change in this release is the fixing of some bugs in Makefile.in for
4217 the "install" target:
4219 (1) It was failing to install pcreposix.h.
4221 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4224 Version 3.0 01-Feb-00
4225 ---------------------
4227 1. Add support for the /+ modifier to perltest (to output $` like it does in
4228 pcretest).
4230 2. Add support for the /g modifier to perltest.
4232 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4233 matches null strings.
4235 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4236 pattern. Perl treats empty patterns specially - it reuses the most recent
4237 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4238 effect.
4240 5. The POSIX interface was broken in that it was just handing over the POSIX
4241 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4242 required a bigger vector, with some working space on the end. This means that
4243 the POSIX wrapper now has to get and free some memory, and copy the results.
4245 6. Added some simple autoconf support, placing the test data and the
4246 documentation in separate directories, re-organizing some of the
4247 information files, and making it build pcre-config (a GNU standard). Also added
4248 libtool support for building PCRE as a shared library, which is now the
4249 default.
4251 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4252 09 are not valid octal constants. Single digits will be used for minor values
4253 less than 10.
4255 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4256 existing programs that set these in the POSIX interface can use PCRE without
4257 modification.
4259 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4260 return all that pcre_info() returns, plus additional data. The pcre_info()
4261 function is retained for compatibility, but is considered to be obsolete.
4263 10. Added experimental recursion feature (?R) to handle one common case that
4264 Perl 5.6 will be able to do with (?p{...}).
4266 11. Added support for POSIX character classes like [:alpha:], which Perl is
4267 adopting.
4270 Version 2.08 31-Aug-99
4271 ----------------------
4273 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4274 trying to match at the startoffset position, but instead was moving forward to
4275 the next newline as if a previous match had failed.
4277 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4278 and could get into a loop if a null string was matched other than at the start
4279 of the subject.
4281 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4282 be distinguished at compile time, and for completeness also added PCRE_DATE.
4284 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4285 in GnuWin32 environments.
4288 Version 2.07 29-Jul-99
4289 ----------------------
4291 1. The documentation is now supplied in plain text form and HTML as well as in
4292 the form of man page sources.
4294 2. C++ compilers don't like assigning (void *) values to other pointer types.
4295 In particular this affects malloc(). Although there is no problem in Standard
4296 C, I've put in casts to keep C++ compilers happy.
4298 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4299 should be (const char *).
4301 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4302 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4303 However, I haven't made this a standard facility. The documentation doesn't
4304 mention it, and the Makefile doesn't support it.
4306 5. The Makefile now contains an "install" target, with editable destinations at
4307 the top of the file. The pcretest program is not installed.
4309 6. pgrep -V now gives the PCRE version number and date.
4311 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4312 causing the entire string to be ignored, instead of just the last character.
4314 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4315 non-matching string, it can take a very, very long time, even for strings of
4316 quite modest length, because of the nested recursion. PCRE now does better in
4317 some of these cases. It does this by remembering the last required literal
4318 character in the pattern, and pre-searching the subject to ensure it is present
4319 before running the real match. In other words, it applies a heuristic to detect
4320 some types of certain failure quickly, and in the above example, if presented
4321 with a string that has no trailing " it gives "no match" very quickly.
4323 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4324 other alternatives are tried instead.
4327 Version 2.06 09-Jun-99
4328 ----------------------
4330 1. Change pcretest's output for amount of store used to show just the code
4331 space, because the remainder (the data block) varies in size between 32-bit and
4332 64-bit systems.
4334 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4335 start matching at. This allows lookbehinds to work when searching for multiple
4336 occurrences in a string.
4338 3. Added additional options to pcretest for testing multiple occurrences:
4340 /+ outputs the rest of the string that follows a match
4341 /g loops for multiple occurrences, using the new startoffset argument
4342 /G loops for multiple occurrences by passing an incremented pointer
4344 4. PCRE wasn't doing the "first character" optimization for patterns starting
4345 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4346 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4347 the letter 'x'. On long subject strings, this gives a significant speed-up.
4350 Version 2.05 21-Apr-99
4351 ----------------------
4353 1. Changed the type of magic_number from int to long int so that it works
4354 properly on 16-bit systems.
4356 2. Fixed a bug which caused patterns starting with .* not to work correctly
4357 when the subject string contained newline characters. PCRE was assuming
4358 anchoring for such patterns in all cases, which is not correct because .* will
4359 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4360 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4361 must be retried after every newline in the subject.
4364 Version 2.04 18-Feb-99
4365 ----------------------
4367 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4368 computation of the store needed to hold the pattern was incorrect (too large).
4369 If such patterns were nested a few deep, this could multiply and become a real
4370 problem.
4372 2. Added /M option to pcretest to show the memory requirement of a specific
4373 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4375 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4376 compiled in such a way that the backtracking after subsequent failure was
4377 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4378 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4381 Version 2.03 02-Feb-99
4382 ----------------------
4384 1. Fixed typo and small mistake in man page.
4386 2. Added 4th condition (GPL supersedes if conflict) and created separate
4387 LICENCE file containing the conditions.
4389 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4390 Perl, that is the internal \ allows the delimiter to be included in the
4391 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4392 the final delimiter, add \ to the end of the pattern (to test the error).
4394 4. Added the convenience functions for extracting substrings after a successful
4395 match. Updated pcretest to make it able to test these functions.
4398 Version 2.02 14-Jan-99
4399 ----------------------
4401 1. Initialized the working variables associated with each extraction so that
4402 their saving and restoring doesn't refer to uninitialized store.
4404 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4405 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4406 fix the problem.
4408 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4409 calls, and wasn't printing the correct value for compiling calls. Increased the
4410 default value of LOOPREPEAT, and the number of significant figures in the
4411 times.
4413 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4415 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4416 a building problem on Windows NT with a FAT file system.
4419 Version 2.01 21-Oct-98
4420 ----------------------
4422 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4423 to character tables built by pcre_maketables() in the current locale. If NULL
4424 is passed, the default tables are used.
4427 Version 2.00 24-Sep-98
4428 ----------------------
4430 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4431 it any more.
4433 2. Allow quantification of (?>) groups, and make it work correctly.
4435 3. The first character computation wasn't working for (?>) groups.
4437 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4438 end of the subject) and add 5.005's \z, which really does match only at the
4439 very end of the subject.
4441 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4443 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4444 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4445 localized options. All options to pcre_study() were also removed.
4447 7. Add other new features from 5.005:
4449 $(?<= positive lookbehind
4450 $(?<! negative lookbehind
4451 (?imsx-imsx) added the unsetting capability
4452 such a setting is global if at outer level; local otherwise
4453 (?imsx-imsx:) non-capturing groups with option setting
4454 (?(cond)re|re) conditional pattern matching
4456 A backreference to itself in a repeated group matches the previous
4457 captured string.
4459 8. General tidying up of studying (both automatic and via "study")
4460 consequential on the addition of new assertions.
4462 9. As in 5.005, unlimited repeated groups that could match an empty substring
4463 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4464 runtime if any iteration does actually match an empty substring.
4466 10. Include the RunTest script in the distribution.
4468 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4469 discrepancies, some of which were old and were also with respect to 5.004. They
4470 have now been fixed.
4473 Version 1.09 28-Apr-98
4474 ----------------------
4476 1. A negated single character class followed by a quantifier with a minimum
4477 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4478 program crashes, or just wrong answers. This did not apply to negated classes
4479 containing more than one character, or to minima other than one.
4482 Version 1.08 27-Mar-98
4483 ----------------------
4485 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4487 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4488 latter must appear before anything that relies on it in the pattern.
4491 Version 1.07 16-Feb-98
4492 ----------------------
4494 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4495 repeat of a potentially empty string).
4498 Version 1.06 23-Jan-98
4499 ----------------------
4501 1. Added Markus Oberhumer's little patches for C++.
4503 2. Literal strings longer than 255 characters were broken.
4506 Version 1.05 23-Dec-97
4507 ----------------------
4509 1. Negated character classes containing more than one character were failing if
4510 PCRE_CASELESS was set at run time.
4513 Version 1.04 19-Dec-97
4514 ----------------------
4516 1. Corrected the man page, where some "const" qualifiers had been omitted.
4518 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4519 input syntax.
4521 3. Fixed memory leak which occurred when a regex with back references was
4522 matched with an offsets vector that wasn't big enough. The temporary memory
4523 that is used in this case wasn't being freed if the match failed.
4525 4. Tidied pcretest to ensure it frees memory that it gets.
4527 5. Temporary memory was being obtained in the case where the passed offsets
4528 vector was exactly big enough.
4530 6. Corrected definition of offsetof() from change 5 below.
4532 7. I had screwed up change 6 below and broken the rules for the use of
4533 setjmp(). Now fixed.
4536 Version 1.03 18-Dec-97
4537 ----------------------
4539 1. A erroneous regex with a missing opening parenthesis was correctly
4540 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4541 on some systems.
4543 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4544 it was reported that one broken compiler failed on the former because "code" is
4545 also an independent variable.
4547 3. The erroneous regex a[]b caused an array overrun reference.
4549 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4550 fail on data ending with that character. (It was going on too far, and checking
4551 the next character, typically a binary zero.) This was specific to the
4552 optimized code for single-character negative classes.
4554 5. Added a contributed patch from the TIN world which does the following:
4556 + Add an undef for memmove, in case the the system defines a macro for it.
4558 + Add a definition of offsetof(), in case there isn't one. (I don't know
4559 the reason behind this - offsetof() is part of the ANSI standard - but
4560 it does no harm).
4562 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4563 most of the places where whitespace preceded '#'. I have given up and
4564 allowed the remaining 2 cases to be at the margin.
4566 + Rename some variables in pcre to eliminate shadowing. This seems very
4567 pedantic, but does no harm, of course.
4569 6. Moved the call to setjmp() into its own function, to get rid of warnings
4570 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4572 7. Constructs such as \d{8,} were compiling into the equivalent of
4573 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4574 outcome, but in this particular case used more store than had been allocated,
4575 which caused the bug to be discovered because it threw up an internal error.
4577 8. The debugging code in both pcre and pcretest for outputting the compiled
4578 form of a regex was going wrong in the case of back references followed by
4579 curly-bracketed repeats.
4582 Version 1.02 12-Dec-97
4583 ----------------------
4585 1. Typos in pcre.3 and comments in the source fixed.
4587 2. Applied a contributed patch to get rid of places where it used to remove
4588 'const' from variables, and fixed some signed/unsigned and uninitialized
4589 variable warnings.
4591 3. Added the "runtest" target to Makefile.
4593 4. Set default compiler flag to -O2 rather than just -O.
4596 Version 1.01 19-Nov-97
4597 ----------------------
4599 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4600 like /([ab]*)*/, that is, for classes with more than one character in them.
4602 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4603 as /((?>a*))*/ (a PCRE_EXTRA facility).
4606 Version 1.00 18-Nov-97
4607 ----------------------
4609 1. Added compile-time macros to support systems such as SunOS4 which don't have
4610 memmove() or strerror() but have other things that can be used instead.
4612 2. Arranged that "make clean" removes the executables.
4615 Version 0.99 27-Oct-97
4616 ----------------------
4618 1. Fixed bug in code for optimizing classes with only one character. It was
4619 initializing a 32-byte map regardless, which could cause it to run off the end
4620 of the memory it had got.
4622 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4625 Version 0.98 22-Oct-97
4626 ----------------------
4628 1. Fixed bug in code for handling temporary memory usage when there are more
4629 back references than supplied space in the ovector. This could cause segfaults.
4632 Version 0.97 21-Oct-97
4633 ----------------------
4635 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4637 2. Optimized negated single characters not to use a bit map.
4639 3. Brought error texts together as macro definitions; clarified some of them;
4640 fixed one that was wrong - it said "range out of order" when it meant "invalid
4641 escape sequence".
4643 4. Changed some char * arguments to const char *.
4645 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4647 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4648 pcretest.
4651 Version 0.96 16-Oct-97
4652 ----------------------
4654 1. Added a simple "pgrep" utility to the distribution.
4656 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4657 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4658 where "ddd" means "one or more decimal digits".
4660 3. Fixed serious bug. If a pattern had a back reference, but the call to
4661 pcre_exec() didn't supply a large enough ovector to record the related
4662 identifying subpattern, the match always failed. PCRE now remembers the number
4663 of the largest back reference, and gets some temporary memory in which to save
4664 the offsets during matching if necessary, in order to ensure that
4665 backreferences always work.
4667 4. Increased the compatibility with Perl in a number of ways:
4669 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4670 to request this handling. The option can be set at compile or exec time.
4672 (b) $ matches before a terminating newline by default; an option
4673 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4674 mode). The option can be set at compile or exec time.
4676 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4677 the same as Perl's. If the decimal number it represents is less than 10
4678 or there aren't that many previous left capturing parentheses, an octal
4679 escape is read. Inside a character class, it's always an octal escape,
4680 even if it is a single digit.
4682 (d) An escaped but undefined alphabetic character is taken as a literal,
4683 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4684 escapes.
4686 (e) {0} is now permitted. (The previous item is removed from the compiled
4687 pattern).
4689 5. Changed all the names of code files so that the basic parts are no longer
4690 than 10 characters, and abolished the teeny "globals.c" file.
4692 6. Changed the handling of character classes; they are now done with a 32-byte
4693 bit map always.
4695 7. Added the -d and /D options to pcretest to make it possible to look at the
4696 internals of compilation without having to recompile pcre.
4699 Version 0.95 23-Sep-97
4700 ----------------------
4702 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4703 \x20 at the start of a run of normal characters. These were being treated as
4704 real characters, instead of the source characters being re-checked.
4707 Version 0.94 18-Sep-97
4708 ----------------------
4710 1. The functions are now thread-safe, with the caveat that the global variables
4711 containing pointers to malloc() and free() or alternative functions are the
4712 same for all threads.
4714 2. Get pcre_study() to generate a bitmap of initial characters for non-
4715 anchored patterns when this is possible, and use it if passed to pcre_exec().
4718 Version 0.93 15-Sep-97
4719 ----------------------
4721 1. /(b)|(:+)/ was computing an incorrect first character.
4723 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4724 but not actually doing anything yet.
4726 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4727 as Perl does (e.g. [-az] or [az-]).
4729 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4730 all possible positions.
4732 5. Split up into different modules to avoid including unneeded functions in a
4733 compiled binary. However, compile and exec are still in one module. The "study"
4734 function is split off.
4736 6. The character tables are now in a separate module whose source is generated
4737 by an auxiliary program - but can then be edited by hand if required. There are
4738 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4739 toupper() in the code.
4741 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4742 make them global. Abolish the function for setting them, as the caller can now
4743 set them directly.
4746 Version 0.92 11-Sep-97
4747 ----------------------
4749 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4750 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4752 2. Caseless matching was not working in character classes if the characters in
4753 the pattern were in upper case.
4755 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4757 4. Make PCRE_ANCHORED public and accept as a compile option.
4759 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4760 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4761 pass them.
4763 6. Give an error if bad option bits passed at compile or run time.
4765 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4766 pcretest to cause it to pass that flag.
4768 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4769 options, and the first character, if set.
4771 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4774 Version 0.91 10-Sep-97
4775 ----------------------
4777 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4778 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4780 2. PCRE was looping on encountering an indefinitely repeated back reference to
4781 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4782 Perl does - treats the match as successful.
4784 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5