/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 944 - (show annotations)
Tue Feb 28 16:49:21 2012 UTC (3 years, 4 months ago) by ph10
File size: 228579 byte(s)
Error occurred while calculating annotation data.
Added --file-list to pcregrep.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.31
5 -----------------------------
6
7 1. Fixing a wrong JIT test case and some compiler warnings.
8
9 2. Removed a bashism from the RunTest script.
10
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)".
14
15 4. Partial matching support is added to the JIT compiler.
16
17 5. Fixed several bugs concerned with partial matching of items that consist
18 of more than one character:
19
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.
23
24 (b) \R did not give a hard partial match if \r was found at the end of the
25 subject.
26
27 (c) \X did not give a hard partial match after matching one or more
28 characters at the end of the subject.
29
30 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31 a partial match for the string "\r".
32
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.
35
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.
39
40 7. Individual JIT compile options can be set in pcretest by following -s+[+]
41 or /S+[+] with a digit between 1 and 7.
42
43 8. OP_NOT now supports any UTF character not just single-byte ones.
44
45 9. (*MARK) control verb is now supported by the JIT compiler.
46
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.)
49
50 11. Add PCRE_INFO_MAXLOOKBEHIND.
51
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.
57
58 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59 call.
60
61 14. As documented, (*COMMIT) is now confined to within a positive assertion.
62
63 15. It is now possible to link pcretest with libedit as an alternative to
64 libreadline.
65
66 16. (*COMMIT) control verb is now supported by the JIT compiler.
67
68 17. The Unicode data tables have been updated to Unicode 6.1.0.
69
70 18. Added --file-list option to pcregrep.
71
72
73 Version 8.30 04-February-2012
74 -----------------------------
75
76 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
77 name is defined in ctype.h.
78
79 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
80 only in quite long subpatterns.
81
82 3. Removed the function pcre_info(), which has been obsolete and deprecated
83 since it was replaced by pcre_fullinfo() in February 2000.
84
85 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
86 match a (*MARK), and the match failed at the start of the subject, a
87 reference to memory before the start of the subject could occur. This bug
88 was introduced by fix 17 of release 8.21.
89
90 5. A reference to an unset group with zero minimum repetition was giving
91 totally wrong answers (in non-JavaScript-compatibility mode). For example,
92 /(another)?(\1?)test/ matched against "hello world test". This bug was
93 introduced in release 8.13.
94
95 6. Add support for 16-bit character strings (a large amount of work involving
96 many changes and refactorings).
97
98 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
99 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
100 from a file.
101
102 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
103 rounding is not applied in this particular case).
104
105 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
106 if they appear, or are escaped, in patterns.
107
108 10. Get rid of a number of -Wunused-but-set-variable warnings.
109
110 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
111 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
112 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
113 also returns the mark "x". This bug applied to capturing parentheses,
114 non-capturing parentheses, and atomic parentheses. It also applied to some
115 assertions.
116
117 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
118 information out of configure.ac instead of relying on pcre.h.generic, which
119 is not stored in the repository.
120
121 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
122 -lreadline.
123
124 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
125
126 15. Applied Graycode's patch to put the top-level frame on the stack rather
127 than the heap when not using the stack for recursion. This gives a
128 performance improvement in many cases when recursion is not deep.
129
130 16. Experimental code added to "pcretest -C" to output the stack frame size.
131
132
133 Version 8.21 12-Dec-2011
134 ------------------------
135
136 1. Updating the JIT compiler.
137
138 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
139 are added as well.
140
141 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
142 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
143 calling _pcre_jit_exec. Some extra comments are added.
144
145 4. (*MARK) settings inside atomic groups that do not contain any capturing
146 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
147 was introduced by change 18 for 8.20.
148
149 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
150 ECMA-262 standard.
151
152 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
153 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
154 This bug was probably introduced by change 9 of 8.13.
155
156 7. While fixing 6 above, I noticed that a number of other items were being
157 incorrectly rejected as "not fixed length". This arose partly because newer
158 opcodes had not been added to the fixed-length checking code. I have (a)
159 corrected the bug and added tests for these items, and (b) arranged for an
160 error to occur if an unknown opcode is encountered while checking for fixed
161 length instead of just assuming "not fixed length". The items that were
162 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
163 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
164 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
165
166 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
167 being incorrectly compiled and would have given unpredicatble results.
168
169 9. A possessively repeated subpattern with minimum repeat count greater than
170 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
171 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
172 the first (A) could occur when it should not.
173
174 10. Add a cast and remove a redundant test from the code.
175
176 11. JIT should use pcre_malloc/pcre_free for allocation.
177
178 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
179 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
180 is anything other than /usr, -L is still shown).
181
182 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
183
184 14. Perl does not support \N without a following name in a [] class; PCRE now
185 also gives an error.
186
187 15. If a forward reference was repeated with an upper limit of around 2000,
188 it caused the error "internal error: overran compiling workspace". The
189 maximum number of forward references (including repeats) was limited by the
190 internal workspace, and dependent on the LINK_SIZE. The code has been
191 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
192 the default depends on LINK_SIZE. There is a new upper limit (for safety)
193 of around 200,000 forward references. While doing this, I also speeded up
194 the filling in of repeated forward references.
195
196 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
197 incorrectly expecting the subject to contain another "a" after the start.
198
199 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
200 in the match, the SKIP should be ignored. This was not happening; instead
201 the SKIP was being treated as NOMATCH. For patterns such as
202 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
203 tested.
204
205 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
206 now much more compatible with Perl, in particular in cases where the result
207 is a non-match for a non-anchored pattern. For example, if
208 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
209 "m", where previously it did not return a name. A side effect of this
210 change is that for partial matches, the last encountered mark name is
211 returned, as for non matches. A number of tests that were previously not
212 Perl-compatible have been moved into the Perl-compatible test files. The
213 refactoring has had the pleasing side effect of removing one argument from
214 the match() function, thus reducing its stack requirements.
215
216 19. If the /S+ option was used in pcretest to study a pattern using JIT,
217 subsequent uses of /S (without +) incorrectly behaved like /S+.
218
219 21. Retrieve executable code size support for the JIT compiler and fixing
220 some warnings.
221
222 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
223 not work when the shorter character appeared right at the end of the
224 subject string.
225
226 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
227 systems.
228
229 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
230 output it when the /M option is used in pcretest.
231
232 25. The CheckMan script was not being included in the distribution. Also, added
233 an explicit "perl" to run Perl scripts from the PrepareRelease script
234 because this is reportedly needed in Windows.
235
236 26. If study data was being save in a file and studying had not found a set of
237 "starts with" bytes for the pattern, the data written to the file (though
238 never used) was taken from uninitialized memory and so caused valgrind to
239 complain.
240
241 27. Updated RunTest.bat as provided by Sheri Pierce.
242
243 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
244
245 29. Computation of memory usage for the table of capturing group names was
246 giving an unnecessarily large value.
247
248
249 Version 8.20 21-Oct-2011
250 ------------------------
251
252 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
253 a POSIX class. After further experiments with Perl, which convinced me that
254 Perl has bugs and confusions, a closing square bracket is no longer allowed
255 in a POSIX name. This bug also affected patterns with classes that started
256 with full stops.
257
258 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
259 captured substring, but while checking the failing first alternative,
260 substring 1 is temporarily captured. If the output vector supplied to
261 pcre_exec() was not big enough for this capture, the yield of the function
262 was still zero ("insufficient space for captured substrings"). This cannot
263 be totally fixed without adding another stack variable, which seems a lot
264 of expense for a edge case. However, I have improved the situation in cases
265 such as /(a)(b)x|abc/ matched against "abc", where the return code
266 indicates that fewer than the maximum number of slots in the ovector have
267 been set.
268
269 3. Related to (2) above: when there are more back references in a pattern than
270 slots in the output vector, pcre_exec() uses temporary memory during
271 matching, and copies in the captures as far as possible afterwards. It was
272 using the entire output vector, but this conflicts with the specification
273 that only 2/3 is used for passing back captured substrings. Now it uses
274 only the first 2/3, for compatibility. This is, of course, another edge
275 case.
276
277 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
278 main code base, and can be used by building with --enable-jit. When this is
279 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
280 runtime --no-jit option is given.
281
282 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
283 ovector, the return from the function was zero, implying that there were
284 other matches that did not fit. The correct "exactly full" value is now
285 returned.
286
287 6. If a subpattern that was called recursively or as a subroutine contained
288 (*PRUNE) or any other control that caused it to give a non-standard return,
289 invalid errors such as "Error -26 (nested recursion at the same subject
290 position)" or even infinite loops could occur.
291
292 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
293 computing the minimum length on reaching *ACCEPT, and so ended up with the
294 wrong value of 1 rather than 0. Further investigation indicates that
295 computing a minimum subject length in the presence of *ACCEPT is difficult
296 (think back references, subroutine calls), and so I have changed the code
297 so that no minimum is registered for a pattern that contains *ACCEPT.
298
299 8. If (*THEN) was present in the first (true) branch of a conditional group,
300 it was not handled as intended. [But see 16 below.]
301
302 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
303 Sheri Pierce.
304
305 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
306 the first byte in a match must be "a".
307
308 11. Change 17 for 8.13 increased the recursion depth for patterns like
309 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
310 pattern contains any instances of (*THEN). If it does not, the old
311 optimizations are restored. It would be nice to do this on a per-group
312 basis, but at the moment that is not feasible.
313
314 12. In some environments, the output of pcretest -C is CRLF terminated. This
315 broke RunTest's code that checks for the link size. A single white space
316 character after the value is now allowed for.
317
318 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
319 For "fr", it uses the Windows-specific input and output files.
320
321 14. If (*THEN) appeared in a group that was called recursively or as a
322 subroutine, it did not work as intended. [But see next item.]
323
324 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
325 pattern fragments (but not containing any | characters). If A and B are
326 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
327 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
328 D. In other words, Perl considers parentheses that do not contain any |
329 characters to be part of a surrounding alternative, whereas PCRE was
330 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
331 differently. PCRE now behaves in the same way as Perl, except in the case
332 of subroutine/recursion calls such as (?1) which have in any case always
333 been different (but PCRE had them first :-).
334
335 16. Related to 15 above: Perl does not treat the | in a conditional group as
336 creating alternatives. Such a group is treated in the same way as an
337 ordinary group without any | characters when processing (*THEN). PCRE has
338 been changed to match Perl's behaviour.
339
340 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
341 RunGrepTest script failed.
342
343 18. Change 22 for version 13 caused atomic groups to use more stack. This is
344 inevitable for groups that contain captures, but it can lead to a lot of
345 stack use in large patterns. The old behaviour has been restored for atomic
346 groups that do not contain any capturing parentheses.
347
348 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
349 suppress the check for a minimum subject length at run time. (If it was
350 given to pcre_exec() or pcre_dfa_exec() it did work.)
351
352 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
353 fail to work when decoding hex characters in data strings in EBCDIC
354 environments.
355
356 21. It appears that in at least one Mac OS environment, the isxdigit() function
357 is implemented as a macro that evaluates to its argument more than once,
358 contravening the C 90 Standard (I haven't checked a later standard). There
359 was an instance in pcretest which caused it to go wrong when processing
360 \x{...} escapes in subject strings. The has been rewritten to avoid using
361 things like p++ in the argument of isxdigit().
362
363
364 Version 8.13 16-Aug-2011
365 ------------------------
366
367 1. The Unicode data tables have been updated to Unicode 6.0.0.
368
369 2. Two minor typos in pcre_internal.h have been fixed.
370
371 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
372 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
373 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
374
375 4. There were a number of related bugs in the code for matching backrefences
376 caselessly in UTF-8 mode when codes for the characters concerned were
377 different numbers of bytes. For example, U+023A and U+2C65 are an upper
378 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
379 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
380 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
381 2-byte code at the end of the subject (it thought there wasn't enough data
382 left).
383
384 5. Comprehensive information about what went wrong is now returned by
385 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
386 as the output vector has at least 2 elements. The offset of the start of
387 the failing character and a reason code are placed in the vector.
388
389 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
390 now returned is for the first byte of the failing character, instead of the
391 last byte inspected. This is an incompatible change, but I hope it is small
392 enough not to be a problem. It makes the returned offset consistent with
393 pcre_exec() and pcre_dfa_exec().
394
395 7. pcretest now gives a text phrase as well as the error number when
396 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
397 failure, the offset and reason code are output.
398
399 8. When \R was used with a maximizing quantifier it failed to skip backwards
400 over a \r\n pair if the subsequent match failed. Instead, it just skipped
401 back over a single character (\n). This seems wrong (because it treated the
402 two characters as a single entity when going forwards), conflicts with the
403 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
404 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
405 has been changed.
406
407 9. Some internal refactoring has changed the processing so that the handling
408 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
409 time (the PCRE_DOTALL option was changed this way some time ago: version
410 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
411 which was always a bit of a fudge. It also means that there is one less
412 argument for the match() function, which reduces its stack requirements
413 slightly. This change also fixes an incompatibility with Perl: the pattern
414 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
415
416 10. More internal refactoring has drastically reduced the number of recursive
417 calls to match() for possessively repeated groups such as (abc)++ when
418 using pcre_exec().
419
420 11. While implementing 10, a number of bugs in the handling of groups were
421 discovered and fixed:
422
423 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
424 (a|)*(?1) gave a compile-time internal error.
425 ((a|)+)+ did not notice that the outer group could match an empty string.
426 (^a|^)+ was not marked as anchored.
427 (.*a|.*)+ was not marked as matching at start or after a newline.
428
429 12. Yet more internal refactoring has removed another argument from the match()
430 function. Special calls to this function are now indicated by setting a
431 value in a variable in the "match data" data block.
432
433 13. Be more explicit in pcre_study() instead of relying on "default" for
434 opcodes that mean there is no starting character; this means that when new
435 ones are added and accidentally left out of pcre_study(), testing should
436 pick them up.
437
438 14. The -s option of pcretest has been documented for ages as being an old
439 synonym of -m (show memory usage). I have changed it to mean "force study
440 for every regex", that is, assume /S for every regex. This is similar to -i
441 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
442 using it. It makes it easier to run collections of tests with and without
443 study enabled, and thereby test pcre_study() more easily. All the standard
444 tests are now run with and without -s (but some patterns can be marked as
445 "never study" - see 20 below).
446
447 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
448 restoration of the capturing data to the outer values was not happening
449 correctly.
450
451 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
452 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
453 pattern had matched an empty string, and so incorrectly returned a no
454 match.
455
456 17. There was optimizing code for the last branch of non-capturing parentheses,
457 and also for the obeyed branch of a conditional subexpression, which used
458 tail recursion to cut down on stack usage. Unfortunately, now that there is
459 the possibility of (*THEN) occurring in these branches, tail recursion is
460 no longer possible because the return has to be checked for (*THEN). These
461 two optimizations have therefore been removed. [But see 8.20/11 above.]
462
463 18. If a pattern containing \R was studied, it was assumed that \R always
464 matched two bytes, thus causing the minimum subject length to be
465 incorrectly computed because \R can also match just one byte.
466
467 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
468 was incorrectly computed.
469
470 20. If /S is present twice on a test pattern in pcretest input, it now
471 *disables* studying, thereby overriding the use of -s on the command line
472 (see 14 above). This is necessary for one or two tests to keep the output
473 identical in both cases.
474
475 21. When (*ACCEPT) was used in an assertion that matched an empty string and
476 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
477
478 22. When an atomic group that contained a capturing parenthesis was
479 successfully matched, but the branch in which it appeared failed, the
480 capturing was not being forgotten if a higher numbered group was later
481 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
482 group 1 to "a", when in fact it should be unset. This applied to multi-
483 branched capturing and non-capturing groups, repeated or not, and also to
484 positive assertions (capturing in negative assertions does not happen
485 in PCRE) and also to nested atomic groups.
486
487 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
488 subject after a captured substring, to make it easier to tell which of a
489 number of identical substrings has been captured.
490
491 24. The way atomic groups are processed by pcre_exec() has been changed so that
492 if they are repeated, backtracking one repetition now resets captured
493 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
494 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
495 "aaa". Previously, it would have been "a". As part of this code
496 refactoring, the way recursive calls are handled has also been changed.
497
498 25. If an assertion condition captured any substrings, they were not passed
499 back unless some other capturing happened later. For example, if
500 (?(?=(a))a) was matched against "a", no capturing was returned.
501
502 26. When studying a pattern that contained subroutine calls or assertions,
503 the code for finding the minimum length of a possible match was handling
504 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
505 group 1 called group 2 while simultaneously a separate group 2 called group
506 1). A stack overflow occurred in this case. I have fixed this by limiting
507 the recursion depth to 10.
508
509 27. Updated RunTest.bat in the distribution to the version supplied by Tom
510 Fortmann. This supports explicit test numbers on the command line, and has
511 argument validation and error reporting.
512
513 28. An instance of \X with an unlimited repeat could fail if at any point the
514 first character it looked at was a mark character.
515
516 29. Some minor code refactoring concerning Unicode properties and scripts
517 should reduce the stack requirement of match() slightly.
518
519 30. Added the '=' option to pcretest to check the setting of unused capturing
520 slots at the end of the pattern, which are documented as being -1, but are
521 not included in the return count.
522
523 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
524 compiled something random. Now it gives a compile-time error (as does
525 Perl).
526
527 32. A *MARK encountered during the processing of a positive assertion is now
528 recorded and passed back (compatible with Perl).
529
530 33. If --only-matching or --colour was set on a pcregrep call whose pattern
531 had alternative anchored branches, the search for a second match in a line
532 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
533 matched the line "0102" twice. The same bug affected patterns that started
534 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
535 twice.
536
537 34. Previously, PCRE did not allow quantification of assertions. However, Perl
538 does, and because of capturing effects, quantifying parenthesized
539 assertions may at times be useful. Quantifiers are now allowed for
540 parenthesized assertions.
541
542 35. A minor code tidy in pcre_compile() when checking options for \R usage.
543
544 36. \g was being checked for fancy things in a character class, when it should
545 just be a literal "g".
546
547 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
548 appearance of a nested POSIX class supersedes an apparent external class.
549 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
550 unescaped square brackets may also appear as part of class names. For
551 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
552 more like Perl. (But see 8.20/1 above.)
553
554 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
555 was because it thought it was \N{name}, which is not supported).
556
557 39. Add minix to OS list not supporting the -S option in pcretest.
558
559 40. PCRE tries to detect cases of infinite recursion at compile time, but it
560 cannot analyze patterns in sufficient detail to catch mutual recursions
561 such as ((?1))((?2)). There is now a runtime test that gives an error if a
562 subgroup is called recursively as a subpattern for a second time at the
563 same position in the subject string. In previous releases this might have
564 been caught by the recursion limit, or it might have run out of stack.
565
566 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
567 happen only once. PCRE was, however incorrectly giving a compile time error
568 "recursive call could loop indefinitely" because it cannot analyze the
569 pattern in sufficient detail. The compile time test no longer happens when
570 PCRE is compiling a conditional subpattern, but actual runaway loops are
571 now caught at runtime (see 40 above).
572
573 42. It seems that Perl allows any characters other than a closing parenthesis
574 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
575 has been changed to be the same.
576
577 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
578 as not to get warnings when autogen.sh is called. Also changed
579 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
580
581 44. To help people who use pcregrep to scan files containing exceedingly long
582 lines, the following changes have been made:
583
584 (a) The default value of the buffer size parameter has been increased from
585 8K to 20K. (The actual buffer used is three times this size.)
586
587 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
588 PCRE is built.
589
590 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
591 to be set at run time.
592
593 (d) Numerical values in pcregrep options can be followed by K or M, for
594 example --buffer-size=50K.
595
596 (e) If a line being scanned overflows pcregrep's buffer, an error is now
597 given and the return code is set to 2.
598
599 45. Add a pointer to the latest mark to the callout data block.
600
601 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
602 partial match of an empty string instead of no match. This was specific to
603 the use of ".".
604
605 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
606 complete match instead of a partial match. This bug was dependent on both
607 the PCRE_UTF8 and PCRE_DOTALL options being set.
608
609 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
610 starting byte set, because \b was not being ignored.
611
612
613 Version 8.12 15-Jan-2011
614 ------------------------
615
616 1. Fixed some typos in the markup of the man pages, and wrote a script that
617 checks for such things as part of the documentation building process.
618
619 2. On a big-endian 64-bit system, pcregrep did not correctly process the
620 --match-limit and --recursion-limit options (added for 8.11). In
621 particular, this made one of the standard tests fail. (The integer value
622 went into the wrong half of a long int.)
623
624 3. If the --colour option was given to pcregrep with -v (invert match), it
625 did strange things, either producing crazy output, or crashing. It should,
626 of course, ignore a request for colour when reporting lines that do not
627 match.
628
629 4. Another pcregrep bug caused similar problems if --colour was specified with
630 -M (multiline) and the pattern match finished with a line ending.
631
632 5. In pcregrep, when a pattern that ended with a literal newline sequence was
633 matched in multiline mode, the following line was shown as part of the
634 match. This seems wrong, so I have changed it.
635
636 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
637 the check for further matches in the same line (so they could be coloured)
638 to overrun the end of the current line. If another match was found, it was
639 incorrectly shown (and then shown again when found in the next line).
640
641 7. If pcregrep was compiled under Windows, there was a reference to the
642 function pcregrep_exit() before it was defined. I am assuming this was
643 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
644 reported by a user. I've moved the definition above the reference.
645
646
647 Version 8.11 10-Dec-2010
648 ------------------------
649
650 1. (*THEN) was not working properly if there were untried alternatives prior
651 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
652 backtracked to try for "b" instead of moving to the next alternative branch
653 at the same level (in this case, to look for "c"). The Perl documentation
654 is clear that when (*THEN) is backtracked onto, it goes to the "next
655 alternative in the innermost enclosing group".
656
657 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
658 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
659 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
660 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
661 (*THEN).
662
663 3. If \s appeared in a character class, it removed the VT character from
664 the class, even if it had been included by some previous item, for example
665 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
666 of \s, but is part of the POSIX "space" class.)
667
668 4. A partial match never returns an empty string (because you can always
669 match an empty string at the end of the subject); however the checking for
670 an empty string was starting at the "start of match" point. This has been
671 changed to the "earliest inspected character" point, because the returned
672 data for a partial match starts at this character. This means that, for
673 example, /(?<=abc)def/ gives a partial match for the subject "abc"
674 (previously it gave "no match").
675
676 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
677 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
678 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
679 has an implication that the given string is incomplete (because a partial
680 match is preferred over a full match). For this reason, these items now
681 give a partial match in this situation. [Aside: previously, the one case
682 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
683 match rather than a full match, which was wrong by the old rules, but is
684 now correct.]
685
686 6. There was a bug in the handling of #-introduced comments, recognized when
687 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
688 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
689 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
690 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
691 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
692 places in pcre_compile().
693
694 7. Related to (6) above, when pcre_compile() was skipping #-introduced
695 comments when looking ahead for named forward references to subpatterns,
696 the only newline sequence it recognized was NL. It now handles newlines
697 according to the set newline convention.
698
699 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
700 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
701 cater for a lack of strerror(). These oversights have been fixed.
702
703 9. Added --match-limit and --recursion-limit to pcregrep.
704
705 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
706
707 11. When the -o option was used, pcregrep was setting a return code of 1, even
708 when matches were found, and --line-buffered was not being honoured.
709
710 12. Added an optional parentheses number to the -o and --only-matching options
711 of pcregrep.
712
713 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
714 can match an empty string. The code to do it in pcretest and pcredemo
715 needed fixing:
716
717 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
718 only one byte after an empty string match just before CRLF (this case
719 just got forgotten; "any" and "anycrlf" were OK).
720
721 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
722 mode when an empty string match preceded an ASCII character followed by
723 a non-ASCII character. (The code for advancing by one character rather
724 than one byte was nonsense.)
725
726 (c) The pcredemo.c sample program did not have any code at all to handle
727 the cases when CRLF is a valid newline sequence.
728
729 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
730 as a starting offset was within the subject string. There is now a new
731 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
732 negative or greater than the length of the string. In order to test this,
733 pcretest is extended to allow the setting of negative starting offsets.
734
735 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
736 starting offset points to the beginning of a UTF-8 character was
737 unnecessarily clumsy. I tidied it up.
738
739 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
740 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
741
742 17. Nobody had reported that the --include_dir option, which was added in
743 release 7.7 should have been called --include-dir (hyphen, not underscore)
744 for compatibility with GNU grep. I have changed it to --include-dir, but
745 left --include_dir as an undocumented synonym, and the same for
746 --exclude-dir, though that is not available in GNU grep, at least as of
747 release 2.5.4.
748
749 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
750 characters from a string of bytes) have been redefined so as not to use
751 loops, in order to improve performance in some environments. At the same
752 time, I abstracted some of the common code into auxiliary macros to save
753 repetition (this should not affect the compiled code).
754
755 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
756 compile-time error is now given if \c is not followed by an ASCII
757 character, that is, a byte less than 128. (In EBCDIC mode, the code is
758 different, and any byte value is allowed.)
759
760 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
761 START_OPTIMIZE option, which is now allowed at compile time - but just
762 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
763 to pcregrep and other applications that have no direct access to PCRE
764 options. The new /Y option in pcretest sets this option when calling
765 pcre_compile().
766
767 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
768 back references. Groups containing recursive back references were forced to
769 be atomic by that change, but in the case of named groups, the amount of
770 memory required was incorrectly computed, leading to "Failed: internal
771 error: code overflow". This has been fixed.
772
773 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
774 pcretest.c, to avoid build problems in some Borland environments.
775
776
777 Version 8.10 25-Jun-2010
778 ------------------------
779
780 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
781 THEN.
782
783 2. (*ACCEPT) was not working when inside an atomic group.
784
785 3. Inside a character class, \B is treated as a literal by default, but
786 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
787 causes the error). The code is unchanged, but I tidied the documentation.
788
789 4. Inside a character class, PCRE always treated \R and \X as literals,
790 whereas Perl faults them if its -w option is set. I have changed PCRE so
791 that it faults them when PCRE_EXTRA is set.
792
793 5. Added support for \N, which always matches any character other than
794 newline. (It is the same as "." when PCRE_DOTALL is not set.)
795
796 6. When compiling pcregrep with newer versions of gcc which may have
797 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
798 declared with attribute warn_unused_result" were given. Just casting the
799 result to (void) does not stop the warnings; a more elaborate fudge is
800 needed. I've used a macro to implement this.
801
802 7. Minor change to pcretest.c to avoid a compiler warning.
803
804 8. Added four artifical Unicode properties to help with an option to make
805 \s etc use properties (see next item). The new properties are: Xan
806 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
807
808 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
809 use Unicode properties. (*UCP) at the start of a pattern can be used to set
810 this option. Modified pcretest to add /W to test this facility. Added
811 REG_UCP to make it available via the POSIX interface.
812
813 10. Added --line-buffered to pcregrep.
814
815 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
816 studied, and the match started with a letter with a code point greater than
817 127 whose first byte was different to the first byte of the other case of
818 the letter, the other case of this starting letter was not recognized
819 (#976).
820
821 12. If a pattern that was studied started with a repeated Unicode property
822 test, for example, \p{Nd}+, there was the theoretical possibility of
823 setting up an incorrect bitmap of starting bytes, but fortunately it could
824 not have actually happened in practice until change 8 above was made (it
825 added property types that matched character-matching opcodes).
826
827 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
828 possible starting bytes for non-anchored patterns.
829
830 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
831 \R, and also a number of cases that involve Unicode properties, both
832 explicit and implicit when PCRE_UCP is set.
833
834 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
835 input, it could crash or give wrong results if characters with values
836 greater than 0xc0 were present in the subject string. (Detail: it assumed
837 UTF-8 input when processing these items.)
838
839 16. Added a lot of (int) casts to avoid compiler warnings in systems where
840 size_t is 64-bit (#991).
841
842 17. Added a check for running out of memory when PCRE is compiled with
843 --disable-stack-for-recursion (#990).
844
845 18. If the last data line in a file for pcretest does not have a newline on
846 the end, a newline was missing in the output.
847
848 19. The default pcre_chartables.c file recognizes only ASCII characters (values
849 less than 128) in its various bitmaps. However, there is a facility for
850 generating tables according to the current locale when PCRE is compiled. It
851 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
852 characters, are recognized by isspace() and therefore were getting set in
853 these tables, and indeed these tables seem to approximate to ISO 8859. This
854 caused a problem in UTF-8 mode when pcre_study() was used to create a list
855 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
856 which of course cannot start UTF-8 characters. I have changed the code so
857 that only real ASCII characters (less than 128) and the correct starting
858 bytes for UTF-8 encodings are set for characters greater than 127 when in
859 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
860 altogether.)
861
862 20. Added the /T option to pcretest so as to be able to run tests with non-
863 standard character tables, thus making it possible to include the tests
864 used for 19 above in the standard set of tests.
865
866 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
867 reference to a subpattern the other side of a comment that contains an
868 opening parenthesis caused either an internal compiling error, or a
869 reference to the wrong subpattern.
870
871
872 Version 8.02 19-Mar-2010
873 ------------------------
874
875 1. The Unicode data tables have been updated to Unicode 5.2.0.
876
877 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
878 configured.
879
880 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
881 original author of that file, following a query about its status.
882
883 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
884 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
885
886 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
887 quantifier applied to a forward-referencing subroutine call, could compile
888 incorrect code or give the error "internal error: previously-checked
889 referenced subpattern not found".
890
891 6. Both MS Visual Studio and Symbian OS have problems with initializing
892 variables to point to external functions. For these systems, therefore,
893 pcre_malloc etc. are now initialized to local functions that call the
894 relevant global functions.
895
896 7. There were two entries missing in the vectors called coptable and poptable
897 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
898 I've fixed the data, and added a kludgy way of testing at compile time that
899 the lengths are correct (equal to the number of opcodes).
900
901 8. Following on from 7, I added a similar kludge to check the length of the
902 eint vector in pcreposix.c.
903
904 9. Error texts for pcre_compile() are held as one long string to avoid too
905 much relocation at load time. To find a text, the string is searched,
906 counting zeros. There was no check for running off the end of the string,
907 which could happen if a new error number was added without updating the
908 string.
909
910 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
911
912 11. \K was not working if it appeared in an atomic group or in a group that
913 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
914 \K is "not well defined" if used in an assertion. PCRE now accepts it if
915 the assertion is positive, but not if it is negative.
916
917 12. Change 11 fortuitously reduced the size of the stack frame used in the
918 "match()" function of pcre_exec.c by one pointer. Forthcoming
919 implementation of support for (*MARK) will need an extra pointer on the
920 stack; I have reserved it now, so that the stack frame size does not
921 decrease.
922
923 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
924 item in branch that calls a recursion is a subroutine call - as in the
925 second branch in the above example - was incorrectly given the compile-
926 time error "recursive call could loop indefinitely" because pcre_compile()
927 was not correctly checking the subroutine for matching a non-empty string.
928
929 14. The checks for overrunning compiling workspace could trigger after an
930 overrun had occurred. This is a "should never occur" error, but it can be
931 triggered by pathological patterns such as hundreds of nested parentheses.
932 The checks now trigger 100 bytes before the end of the workspace.
933
934 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
935
936
937 Version 8.01 19-Jan-2010
938 ------------------------
939
940 1. If a pattern contained a conditional subpattern with only one branch (in
941 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
942 computed the wrong minimum data length (which is of course zero for such
943 subpatterns). This could cause incorrect "no match" results.
944
945 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
946 the pattern is reset in the first branch, pcre_compile() failed with
947 "internal error: code overflow at offset...". This happened only when
948 the reset was to the original external option setting. (An optimization
949 abstracts leading options settings into an external setting, which was the
950 cause of this.)
951
952 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
953 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
954 assertion pattern did not match (meaning that the assertion was true), it
955 was incorrectly treated as false if the SKIP had been reached during the
956 matching. This also applied to assertions used as conditions.
957
958 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
959 assertion subpattern, including such a pattern used as a condition,
960 unpredictable results occurred, instead of the error return
961 PCRE_ERROR_DFA_UITEM.
962
963 5. The C++ GlobalReplace function was not working like Perl for the special
964 situation when an empty string is matched. It now does the fancy magic
965 stuff that is necessary.
966
967 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
968 removed. (These were left over from very, very early versions of PCRE.)
969
970 7. Some cosmetic changes to the code to make life easier when compiling it
971 as part of something else:
972
973 (a) Change DEBUG to PCRE_DEBUG.
974
975 (b) In pcre_compile(), rename the member of the "branch_chain" structure
976 called "current" as "current_branch", to prevent a collision with the
977 Linux macro when compiled as a kernel module.
978
979 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
980 prevent a collision with the Linux macro when compiled as a kernel
981 module.
982
983 8. In pcre_compile() there are some checks for integer overflows that used to
984 cast potentially large values to (double). This has been changed to that
985 when building, a check for int64_t is made, and if it is found, it is used
986 instead, thus avoiding the use of floating point arithmetic. (There is no
987 other use of FP in PCRE.) If int64_t is not found, the fallback is to
988 double.
989
990 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
991 2005 (difference between two addresses compared to an unsigned value).
992
993 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
994 custom one, because of the following reported problem in Windows:
995
996 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
997 under Win32.
998 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
999 therefore missing the function definition.
1000 - The compiler thus generates a "C" signature for the test function.
1001 - The linker fails to find the "C" function.
1002 - PCRE fails to configure if asked to do so against libbz2.
1003
1004 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1005 messages were output:
1006
1007 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1008 rerunning libtoolize, to keep the correct libtool macros in-tree.
1009 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1010
1011 I have done both of these things.
1012
1013 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1014 most of the time, it *can* run out if it is given a pattern that contains a
1015 runaway infinite recursion. I updated the discussion in the pcrestack man
1016 page.
1017
1018 13. Now that we have gone to the x.xx style of version numbers, the minor
1019 version may start with zero. Using 08 or 09 is a bad idea because users
1020 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1021 interpreted as invalid octal numbers. I've updated the previous comment in
1022 configure.ac, and also added a check that gives an error if 08 or 09 are
1023 used.
1024
1025 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1026 causing partial matching to fail when the end of the subject matched \W
1027 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1028
1029 15. There were some discrepancies between the declarations in pcre_internal.h
1030 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1031 their definitions. The declarations used "const uschar *" and the
1032 definitions used USPTR. Even though USPTR is normally defined as "const
1033 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1034 reported that: "This difference in casting confuses some C++ compilers, for
1035 example, SunCC recognizes above declarations as different functions and
1036 generates broken code for hbpcre." I have changed the declarations to use
1037 USPTR.
1038
1039 16. GNU libtool is named differently on some systems. The autogen.sh script now
1040 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1041 (FreeBSD).
1042
1043 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1044 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1045 comment: "Figure out how to create a longlong from a string: strtoll and
1046 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1047 instance, but it only takes 2 args instead of 3!"
1048
1049 18. A subtle bug concerned with back references has been fixed by a change of
1050 specification, with a corresponding code fix. A pattern such as
1051 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1052 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1053 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1054 same bug. Such groups have to be quantified to be useful, or contained
1055 inside another quantified group. (If there's no repetition, the reference
1056 can never match.) The problem arises because, having left the group and
1057 moved on to the rest of the pattern, a later failure that backtracks into
1058 the group uses the captured value from the final iteration of the group
1059 rather than the correct earlier one. I have fixed this in PCRE by forcing
1060 any group that contains a reference to itself to be an atomic group; that
1061 is, there cannot be any backtracking into it once it has completed. This is
1062 similar to recursive and subroutine calls.
1063
1064
1065 Version 8.00 19-Oct-09
1066 ----------------------
1067
1068 1. The table for translating pcre_compile() error codes into POSIX error codes
1069 was out-of-date, and there was no check on the pcre_compile() error code
1070 being within the table. This could lead to an OK return being given in
1071 error.
1072
1073 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1074 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1075 in a Windows environment.
1076
1077 3. The pcregrep --count option prints the count for each file even when it is
1078 zero, as does GNU grep. However, pcregrep was also printing all files when
1079 --files-with-matches was added. Now, when both options are given, it prints
1080 counts only for those files that have at least one match. (GNU grep just
1081 prints the file name in this circumstance, but including the count seems
1082 more useful - otherwise, why use --count?) Also ensured that the
1083 combination -clh just lists non-zero counts, with no names.
1084
1085 4. The long form of the pcregrep -F option was incorrectly implemented as
1086 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1087 but it seems right to fix it, and I didn't think it was worth preserving
1088 the old behaviour.
1089
1090 5. The command line items --regex=pattern and --regexp=pattern were not
1091 recognized by pcregrep, which required --regex pattern or --regexp pattern
1092 (with a space rather than an '='). The man page documented the '=' forms,
1093 which are compatible with GNU grep; these now work.
1094
1095 6. No libpcreposix.pc file was created for pkg-config; there was just
1096 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1097
1098 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1099 when UCP support is not needed, by modifying the Python script that
1100 generates it from Unicode data files. This should not matter if the module
1101 is correctly used as a library, but I received one complaint about 50K of
1102 unwanted data. My guess is that the person linked everything into his
1103 program rather than using a library. Anyway, it does no harm.
1104
1105 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1106 was a minimum greater than 1 for a wide character in a possessive
1107 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1108 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1109 character. Chaos in the form of incorrect output or a compiling loop could
1110 result.
1111
1112 9. The restrictions on what a pattern can contain when partial matching is
1113 requested for pcre_exec() have been removed. All patterns can now be
1114 partially matched by this function. In addition, if there are at least two
1115 slots in the offset vector, the offset of the earliest inspected character
1116 for the match and the offset of the end of the subject are set in them when
1117 PCRE_ERROR_PARTIAL is returned.
1118
1119 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1120 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1121 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1122 and may be more useful for multi-segment matching.
1123
1124 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1125 used to be given if ever the end of the subject was reached; now it is
1126 given only if matching could not proceed because another character was
1127 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1128 string "Z", which now yields "no match" instead of "partial match". In the
1129 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1130 final character ended with (*FAIL).
1131
1132 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1133 if the pattern had a "must contain" character that was already found in the
1134 earlier partial match, unless partial matching was again requested. For
1135 example, with the pattern /dog.(body)?/, the "must contain" character is
1136 "g". If the first part-match was for the string "dog", restarting with
1137 "sbody" failed. This bug has been fixed.
1138
1139 13. The string returned by pcre_dfa_exec() after a partial match has been
1140 changed so that it starts at the first inspected character rather than the
1141 first character of the match. This makes a difference only if the pattern
1142 starts with a lookbehind assertion or \b or \B (\K is not supported by
1143 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1144 matching functions compatible, and I think it's the right thing to do.
1145
1146 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1147 so that the demonstration program is easily available in environments where
1148 PCRE has not been installed from source.
1149
1150 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1151 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1152 library.
1153
1154 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1155 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1156 is not the first non-POSIX option to be added. Clearly some people find
1157 these options useful.
1158
1159 17. If a caller to the POSIX matching function regexec() passes a non-zero
1160 value for nmatch with a NULL value for pmatch, the value of
1161 nmatch is forced to zero.
1162
1163 18. RunGrepTest did not have a test for the availability of the -u option of
1164 the diff command, as RunTest does. It now checks in the same way as
1165 RunTest, and also checks for the -b option.
1166
1167 19. If an odd number of negated classes containing just a single character
1168 interposed, within parentheses, between a forward reference to a named
1169 subpattern and the definition of the subpattern, compilation crashed with
1170 an internal error, complaining that it could not find the referenced
1171 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1172 [The bug was that it was starting one character too far in when skipping
1173 over the character class, thus treating the ] as data rather than
1174 terminating the class. This meant it could skip too much.]
1175
1176 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1177 /g option in pcretest when the pattern contains \K, which makes it possible
1178 to have an empty string match not at the start, even when the pattern is
1179 anchored. Updated pcretest and pcredemo to use this option.
1180
1181 21. If the maximum number of capturing subpatterns in a recursion was greater
1182 than the maximum at the outer level, the higher number was returned, but
1183 with unset values at the outer level. The correct (outer level) value is
1184 now given.
1185
1186 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1187 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1188 make it do so. The string so far is captured, making this feature
1189 compatible with Perl.
1190
1191 23. The tests have been re-organized, adding tests 11 and 12, to make it
1192 possible to check the Perl 5.10 features against Perl 5.10.
1193
1194 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1195 pattern matches a fixed length string. PCRE did not allow this; now it
1196 does. Neither allows recursion.
1197
1198 25. I finally figured out how to implement a request to provide the minimum
1199 length of subject string that was needed in order to match a given pattern.
1200 (It was back references and recursion that I had previously got hung up
1201 on.) This code has now been added to pcre_study(); it finds a lower bound
1202 to the length of subject needed. It is not necessarily the greatest lower
1203 bound, but using it to avoid searching strings that are too short does give
1204 some useful speed-ups. The value is available to calling programs via
1205 pcre_fullinfo().
1206
1207 26. While implementing 25, I discovered to my embarrassment that pcretest had
1208 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1209 study optimizations had never been tested with that matching function.
1210 Oops. What is worse, even when it was passed study data, there was a bug in
1211 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1212 were also very few tests of studied patterns with pcre_dfa_exec().
1213
1214 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1215 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1216 on the other side of the coin, they are no longer allowed to have different
1217 names, because these cannot be distinguished in PCRE, and this has caused
1218 confusion. (This is a difference from Perl.)
1219
1220 28. When duplicate subpattern names are present (necessarily with different
1221 numbers, as required by 27 above), and a test is made by name in a
1222 conditional pattern, either for a subpattern having been matched, or for
1223 recursion in such a pattern, all the associated numbered subpatterns are
1224 tested, and the overall condition is true if the condition is true for any
1225 one of them. This is the way Perl works, and is also more like the way
1226 testing by number works.
1227
1228
1229 Version 7.9 11-Apr-09
1230 ---------------------
1231
1232 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1233 (pcretest), all targets were linked against these libraries. This included
1234 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1235 libraries. This caused unwanted dependencies to be created. This problem
1236 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1237 pcretest is linked with readline.
1238
1239 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1240 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1241 moved outside it again, because FALSE and TRUE are already defined in AIX,
1242 but BOOL is not.
1243
1244 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1245 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1246
1247 4. The pcregrep documentation said spaces were inserted as well as colons (or
1248 hyphens) following file names and line numbers when outputting matching
1249 lines. This is not true; no spaces are inserted. I have also clarified the
1250 wording for the --colour (or --color) option.
1251
1252 5. In pcregrep, when --colour was used with -o, the list of matching strings
1253 was not coloured; this is different to GNU grep, so I have changed it to be
1254 the same.
1255
1256 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1257 each matching line was coloured. Now it goes on to look for further matches
1258 of any of the test patterns, which is the same behaviour as GNU grep.
1259
1260 7. A pattern that could match an empty string could cause pcregrep to loop; it
1261 doesn't make sense to accept an empty string match in pcregrep, so I have
1262 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1263 seems to be how GNU grep behaves.
1264
1265 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1266 start or after a newline", because the conditional assertion was not being
1267 correctly handled. The rule now is that both the assertion and what follows
1268 in the first alternative must satisfy the test.
1269
1270 9. If auto-callout was enabled in a pattern with a conditional group whose
1271 condition was an assertion, PCRE could crash during matching, both with
1272 pcre_exec() and pcre_dfa_exec().
1273
1274 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1275 used for matching.
1276
1277 11. Unicode property support in character classes was not working for
1278 characters (bytes) greater than 127 when not in UTF-8 mode.
1279
1280 12. Added the -M command line option to pcretest.
1281
1282 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1283
1284 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1285
1286 16. Added comments and documentation about mis-use of no_arg in the C++
1287 wrapper.
1288
1289 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1290 from Martin Jerabek that uses macro names for all relevant character and
1291 string constants.
1292
1293 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1294 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1295 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1296 these, but not everybody uses configure.
1297
1298 19. A conditional group that had only one branch was not being correctly
1299 recognized as an item that could match an empty string. This meant that an
1300 enclosing group might also not be so recognized, causing infinite looping
1301 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1302 with the subject "ab", where knowledge that the repeated group can match
1303 nothing is needed in order to break the loop.
1304
1305 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1306 exec(), but without supplying a callout function, matching went wrong.
1307
1308 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1309 leak if the size of the offset vector was greater than 30. When the vector
1310 is smaller, the saved offsets during recursion go onto a local stack
1311 vector, but for larger vectors malloc() is used. It was failing to free
1312 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1313 error, in fact).
1314
1315 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1316 heapframe that is used only when UTF-8 support is enabled. This caused no
1317 problem, but was untidy.
1318
1319 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1320 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1321 included within another project.
1322
1323 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1324 slightly modified by me:
1325
1326 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1327 not building pcregrep.
1328
1329 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1330 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1331
1332 25. Forward references, both numeric and by name, in patterns that made use of
1333 duplicate group numbers, could behave incorrectly or give incorrect errors,
1334 because when scanning forward to find the reference group, PCRE was not
1335 taking into account the duplicate group numbers. A pattern such as
1336 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1337
1338 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1339 the feature of a custom pointer more persuasive (as requested by a user).
1340
1341 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1342 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1343 pre-defined.
1344
1345 28. Added support for (*UTF8) at the start of a pattern.
1346
1347 29. Arrange for flags added by the "release type" setting in CMake to be shown
1348 in the configuration summary.
1349
1350
1351 Version 7.8 05-Sep-08
1352 ---------------------
1353
1354 1. Replaced UCP searching code with optimized version as implemented for Ad
1355 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1356 stage table and inline lookup instead of a function, giving speed ups of 2
1357 to 5 times on some simple patterns that I tested. Permission was given to
1358 distribute the MultiStage2.py script that generates the tables (it's not in
1359 the tarball, but is in the Subversion repository).
1360
1361 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1362 scripts.
1363
1364 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1365 a group with a zero qualifier. The result of the study could be incorrect,
1366 or the function might crash, depending on the pattern.
1367
1368 4. Caseless matching was not working for non-ASCII characters in back
1369 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1370 It now works when Unicode Property Support is available.
1371
1372 5. In pcretest, an escape such as \x{de} in the data was always generating
1373 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1374 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1375 truncation.
1376
1377 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1378
1379 7. Added two (int) casts to pcregrep when printing the difference of two
1380 pointers, in case they are 64-bit values.
1381
1382 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1383 test 2 if it fails.
1384
1385 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1386 and a #define of that name to empty if it is not externally set. This is to
1387 allow users of MSVC to set it if necessary.
1388
1389 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1390 the convenience functions in the pcre_get.c source file.
1391
1392 11. An option change at the start of a pattern that had top-level alternatives
1393 could cause overwriting and/or a crash. This command provoked a crash in
1394 some environments:
1395
1396 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1397
1398 This potential security problem was recorded as CVE-2008-2371.
1399
1400 12. For a pattern where the match had to start at the beginning or immediately
1401 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1402 pcre_dfa_exec() could read past the end of the passed subject if there was
1403 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1404 pcretest so that it places the subject at the end of its malloc-ed buffer.
1405
1406 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1407 exec() might read past the end of the data buffer in UTF-8 mode.
1408
1409 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1410 the data contained the byte 0x85 as part of a UTF-8 character within its
1411 first line. This applied both to normal and DFA matching.
1412
1413 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1414 /^[^d]*?$/8 failed to match "abc".
1415
1416 16. Added a missing copyright notice to pcrecpp_internal.h.
1417
1418 17. Make it more clear in the documentation that values returned from
1419 pcre_exec() in ovector are byte offsets, not character counts.
1420
1421 18. Tidied a few places to stop certain compilers from issuing warnings.
1422
1423 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1424 supplied by Stefan Weber. I made a further small update for 7.8 because
1425 there is a change of source arrangements: the pcre_searchfuncs.c module is
1426 replaced by pcre_ucd.c.
1427
1428
1429 Version 7.7 07-May-08
1430 ---------------------
1431
1432 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1433 a string to a long long, pretend we don't even have a long long." This is
1434 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1435
1436 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1437 pre-7.6 versions, which defined a global no_arg variable instead of putting
1438 it in the RE class. (See also #8 below.)
1439
1440 3. Remove a line of dead code, identified by coverity and reported by Nuno
1441 Lopes.
1442
1443 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1444
1445 (1) The include/exclude patterns were being applied to the whole pathnames
1446 of files, instead of just to the final components.
1447
1448 (2) If there was more than one level of directory, the subdirectories were
1449 skipped unless they satisfied the include/exclude conditions. This is
1450 inconsistent with GNU grep (and could even be seen as contrary to the
1451 pcregrep specification - which I improved to make it absolutely clear).
1452 The action now is always to scan all levels of directory, and just
1453 apply the include/exclude patterns to regular files.
1454
1455 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1456 --exclude_dir in the tests to avoid scanning .svn directories.
1457
1458 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1459 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1460 doesn't support NULs in patterns.
1461
1462 7. Added some missing "const"s to declarations of static tables in
1463 pcre_compile.c and pcre_dfa_exec.c.
1464
1465 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1466 caused by fix #2 above. (Subsequently also a second patch to fix the
1467 first patch. And a third patch - this was a messy problem.)
1468
1469 9. Applied Craig's patch to remove the use of push_back().
1470
1471 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1472 matching function regexec().
1473
1474 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1475 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1476 references. PCRE supports relative numbers with this syntax (I don't think
1477 Oniguruma does).
1478
1479 12. Previously, a group with a zero repeat such as (...){0} was completely
1480 omitted from the compiled regex. However, this means that if the group
1481 was called as a subroutine from elsewhere in the pattern, things went wrong
1482 (an internal error was given). Such groups are now left in the compiled
1483 pattern, with a new opcode that causes them to be skipped at execution
1484 time.
1485
1486 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1487 to the way PCRE behaves:
1488
1489 (a) A lone ] character is dis-allowed (Perl treats it as data).
1490
1491 (b) A back reference to an unmatched subpattern matches an empty string
1492 (Perl fails the current match path).
1493
1494 (c) A data ] in a character class must be notated as \] because if the
1495 first data character in a class is ], it defines an empty class. (In
1496 Perl it is not possible to have an empty class.) The empty class []
1497 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1498 The negative empty class [^] matches any one character, independently
1499 of the DOTALL setting.
1500
1501 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1502 non-existent subpattern following a character class starting with ']' and
1503 containing () gave an internal compiling error instead of "reference to
1504 non-existent subpattern". Fortunately, when the pattern did exist, the
1505 compiled code was correct. (When scanning forwards to check for the
1506 existencd of the subpattern, it was treating the data ']' as terminating
1507 the class, so got the count wrong. When actually compiling, the reference
1508 was subsequently set up correctly.)
1509
1510 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1511 it was being rejected as not supported by pcre_dfa_exec(), even though
1512 other assertions are supported. I have made pcre_dfa_exec() support
1513 (*FAIL).
1514
1515 16. The implementation of 13c above involved the invention of a new opcode,
1516 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1517 cannot be changed at match time, I realized I could make a small
1518 improvement to matching performance by compiling OP_ALLANY instead of
1519 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1520 on the OP_ANY path.
1521
1522 17. Compiling pcretest on Windows with readline support failed without the
1523 following two fixes: (1) Make the unistd.h include conditional on
1524 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1525
1526 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1527 ncurses library to be included for pcretest when ReadLine support is
1528 requested, but also to allow for it to be overridden. This patch came from
1529 Daniel Bergström.
1530
1531 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1532 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1533 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1534 spotting this.
1535
1536
1537 Version 7.6 28-Jan-08
1538 ---------------------
1539
1540 1. A character class containing a very large number of characters with
1541 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1542 overflow.
1543
1544 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1545 HAVE_LONG_LONG is not defined.
1546
1547 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1548 bring them up to date and include new features. This patch includes:
1549
1550 - Fixed PH's badly added libz and libbz2 support.
1551 - Fixed a problem with static linking.
1552 - Added pcredemo. [But later removed - see 7 below.]
1553 - Fixed dftables problem and added an option.
1554 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1555 HAVE_LONG_LONG.
1556 - Added readline support for pcretest.
1557 - Added an listing of the option settings after cmake has run.
1558
1559 4. A user submitted a patch to Makefile that makes it easy to create
1560 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1561 Makefile.am that cause it to include this special target, without
1562 affecting anything else. Note that the same mingw target plus all
1563 the other distribution libraries and programs are now supported
1564 when configuring with CMake (see 6 below) instead of with
1565 Configure/Make.
1566
1567 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1568 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1569 exported in the Windows port". It has not yet been confirmed that the patch
1570 solves the problem, but it does no harm.
1571
1572 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1573 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1574 with CMake, and also correct the comment about stack recursion.
1575
1576 7. Remove the automatic building of pcredemo from the ./configure system and
1577 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1578 of a program that users should build themselves after PCRE is installed, so
1579 building it automatically is not really right. What is more, it gave
1580 trouble in some build environments.
1581
1582 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1583
1584
1585 Version 7.5 10-Jan-08
1586 ---------------------
1587
1588 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1589 values in parens when parsing an RE using the C++ wrapper."
1590
1591 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1592 Characters greater than 255 were excluded from the class instead of being
1593 included.
1594
1595 3. The same bug as (2) above applied to negated POSIX classes such as
1596 [:^space:].
1597
1598 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1599 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1600 I have changed it.
1601
1602 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1603 first named subpattern) and a construct such as (?&a) would reference the
1604 first named subpattern whose name started with "a" (in other words, the
1605 length check was missing). Both these problems are fixed. "Subpattern name
1606 expected" is now given for (?&) (a zero-length name), and this patch also
1607 makes it give the same error for \k'' (previously it complained that that
1608 was a reference to a non-existent subpattern).
1609
1610 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1611 this is right because (?- can be followed by option settings as well as by
1612 digits. I have, however, made the messages clearer.
1613
1614 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1615 than the number used in the conditional) now cause a compile-time error.
1616 This is actually not compatible with Perl, which accepts such patterns, but
1617 treats the conditional as always being FALSE (as PCRE used to), but it
1618 seems to me that giving a diagnostic is better.
1619
1620 8. Change "alphameric" to the more common word "alphanumeric" in comments
1621 and messages.
1622
1623 9. Fix two occurrences of "backslash" in comments that should have been
1624 "backspace".
1625
1626 10. Remove two redundant lines of code that can never be obeyed (their function
1627 was moved elsewhere).
1628
1629 11. The program that makes PCRE's Unicode character property table had a bug
1630 which caused it to generate incorrect table entries for sequences of
1631 characters that have the same character type, but are in different scripts.
1632 It amalgamated them into a single range, with the script of the first of
1633 them. In other words, some characters were in the wrong script. There were
1634 thirteen such cases, affecting characters in the following ranges:
1635
1636 U+002b0 - U+002c1
1637 U+0060c - U+0060d
1638 U+0061e - U+00612
1639 U+0064b - U+0065e
1640 U+0074d - U+0076d
1641 U+01800 - U+01805
1642 U+01d00 - U+01d77
1643 U+01d9b - U+01dbf
1644 U+0200b - U+0200f
1645 U+030fc - U+030fe
1646 U+03260 - U+0327f
1647 U+0fb46 - U+0fbb1
1648 U+10450 - U+1049d
1649
1650 12. The -o option (show only the matching part of a line) for pcregrep was not
1651 compatible with GNU grep in that, if there was more than one match in a
1652 line, it showed only the first of them. It now behaves in the same way as
1653 GNU grep.
1654
1655 13. If the -o and -v options were combined for pcregrep, it printed a blank
1656 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1657 does the same. The return code can be used to tell if there were any
1658 non-matching lines.
1659
1660 14. Added --file-offsets and --line-offsets to pcregrep.
1661
1662 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1663 infinitely looping recursion. The bug was that positive lookaheads were not
1664 being skipped when checking for a possible empty match (negative lookaheads
1665 and both kinds of lookbehind were skipped).
1666
1667 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1668 inclusion of <windows.h> to before rather than after the definition of
1669 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1670
1671 17. Specifying a possessive quantifier with a specific limit for a Unicode
1672 character property caused pcre_compile() to compile bad code, which led at
1673 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1674 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1675 caused the error; without that there was no problem.
1676
1677 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1678
1679 19. Added --enable-pcretest-libreadline.
1680
1681 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1682 RE::GlobalReplace(). As a result, the number of replacements returned was
1683 double what it should be. I removed one of the increments, but Craig sent a
1684 later patch that removed the other one (the right fix) and added unit tests
1685 that check the return values (which was not done before).
1686
1687 21. Several CMake things:
1688
1689 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1690 the names libpcre and libpcreposix, not just pcre and pcreposix.
1691
1692 (2) The above change means that pcretest and pcregrep are now correctly
1693 linked with the newly-built libraries, not previously installed ones.
1694
1695 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1696
1697 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1698 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1699 UTF-8 newline character). The key issue is that the pattern starts .*;
1700 this means that the match must be either at the beginning, or after a
1701 newline. The bug was in the code for advancing after a failed match and
1702 checking that the new position followed a newline. It was not taking
1703 account of UTF-8 characters correctly.
1704
1705 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1706 character classes. PCRE was not treating the sequence [:...:] as a
1707 character class unless the ... were all letters. Perl, however, seems to
1708 allow any characters between [: and :], though of course it rejects as
1709 unknown any "names" that contain non-letters, because all the known class
1710 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1711 for example, whereas PCRE did not - it did not recognize a POSIX character
1712 class. This seemed a bit dangerous, so the code has been changed to be
1713 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1714 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1715 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1716 Perl does, and where it didn't before.
1717
1718 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1719 Windows environments %n is disabled by default.
1720
1721
1722 Version 7.4 21-Sep-07
1723 ---------------------
1724
1725 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1726 means that a class such as [\s] counted as "explicit reference to CR or
1727 LF". That isn't really right - the whole point of the change was to try to
1728 help when there was an actual mention of one of the two characters. So now
1729 the change happens only if \r or \n (or a literal CR or LF) character is
1730 encountered.
1731
1732 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1733 of both had grown to the point where there were only 3 bits left.
1734 Fortunately, there was spare space in the data structure, and so I have
1735 moved the internal flags into a new 16-bit field to free up more option
1736 bits.
1737
1738 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1739 but did not set the internal JCHANGED flag - either of these is enough to
1740 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1741 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1742 start sets both bits.
1743
1744 4. Added options (at build time, compile time, exec time) to change \R from
1745 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1746
1747 5. doc/pcresyntax.html was missing from the distribution.
1748
1749 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1750 compatibility, even though it is no longer used.
1751
1752 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1753 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1754 windows.h file is present (where different names are used). [This was
1755 reversed later after testing - see 16 below.]
1756
1757 8. Changed all #include <config.h> to #include "config.h". There were also
1758 some further <pcre.h> cases that I changed to "pcre.h".
1759
1760 9. When pcregrep was used with the --colour option, it missed the line ending
1761 sequence off the lines that it output.
1762
1763 10. It was pointed out to me that arrays of string pointers cause lots of
1764 relocations when a shared library is dynamically loaded. A technique of
1765 using a single long string with a table of offsets can drastically reduce
1766 these. I have refactored PCRE in four places to do this. The result is
1767 dramatic:
1768
1769 Originally: 290
1770 After changing UCP table: 187
1771 After changing error message table: 43
1772 After changing table of "verbs" 36
1773 After changing table of Posix names 22
1774
1775 Thanks to the folks working on Gregex for glib for this insight.
1776
1777 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1778 unicode-properties was also set.
1779
1780 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1781
1782 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1783 checked only for CRLF.
1784
1785 14. Added casts to pcretest.c to avoid compiler warnings.
1786
1787 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1788
1789 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1790 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1791 entirely. This removes changes made in 7 above.
1792
1793 17. The CMake files have been updated, and there is now more information about
1794 building with CMake in the NON-UNIX-USE document.
1795
1796
1797 Version 7.3 28-Aug-07
1798 ---------------------
1799
1800 1. In the rejigging of the build system that eventually resulted in 7.1, the
1801 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1802 brackets there is not right, since it causes compilers to look for an
1803 installed pcre.h, not the version that is in the source that is being
1804 compiled (which of course may be different). I have changed it back to:
1805
1806 #include "pcre.h"
1807
1808 I have a vague recollection that the change was concerned with compiling in
1809 different directories, but in the new build system, that is taken care of
1810 by the VPATH setting the Makefile.
1811
1812 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1813 when the subject happened to end in the byte 0x85 (e.g. if the last
1814 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1815 characters but of course it shouldn't be taken as a newline when it is part
1816 of another character. The bug was that, for an unlimited repeat of . in
1817 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1818 characters when looking for a newline.
1819
1820 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1821
1822 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1823 in debug output.
1824
1825 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1826 long printing in the pcrecpp unittest when running under MinGW.
1827
1828 6. ESC_K was left out of the EBCDIC table.
1829
1830 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1831 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1832 limit also applies to "virtual nesting" when a pattern is recursive, and in
1833 this case 1000 isn't so big. I have been able to remove this limit at the
1834 expense of backing off one optimization in certain circumstances. Normally,
1835 when pcre_exec() would call its internal match() function recursively and
1836 immediately return the result unconditionally, it uses a "tail recursion"
1837 feature to save stack. However, when a subpattern that can match an empty
1838 string has an unlimited repetition quantifier, it no longer makes this
1839 optimization. That gives it a stack frame in which to save the data for
1840 checking that an empty string has been matched. Previously this was taken
1841 from the 1000-entry workspace that had been reserved. So now there is no
1842 explicit limit, but more stack is used.
1843
1844 8. Applied Daniel's patches to solve problems with the import/export magic
1845 syntax that is required for Windows, and which was going wrong for the
1846 pcreposix and pcrecpp parts of the library. These were overlooked when this
1847 problem was solved for the main library.
1848
1849 9. There were some crude static tests to avoid integer overflow when computing
1850 the size of patterns that contain repeated groups with explicit upper
1851 limits. As the maximum quantifier is 65535, the maximum group length was
1852 set at 30,000 so that the product of these two numbers did not overflow a
1853 32-bit integer. However, it turns out that people want to use groups that
1854 are longer than 30,000 bytes (though not repeat them that many times).
1855 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1856 made it possible to implement the integer overflow checks in a much more
1857 dynamic way, which I have now done. The artificial limitation on group
1858 length has been removed - we now have only the limit on the total length of
1859 the compiled pattern, which depends on the LINK_SIZE setting.
1860
1861 10. Fixed a bug in the documentation for get/copy named substring when
1862 duplicate names are permitted. If none of the named substrings are set, the
1863 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1864 empty string.
1865
1866 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1867 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1868 because the ] is interpreted as the first data character and the
1869 terminating ] is not found. PCRE has been made compatible with Perl in this
1870 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1871 cause memory overwriting.
1872
1873 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1874 string has been matched (to stop an infinite loop). It was not recognizing
1875 a conditional subpattern that could match an empty string if that
1876 subpattern was within another subpattern. For example, it looped when
1877 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1878 condition was not nested. This bug has been fixed.
1879
1880 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1881 past the start of the subject in the presence of bytes with the top bit
1882 set, for example "\x8aBCD".
1883
1884 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1885 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1886
1887 14. Optimized (?!) to (*FAIL).
1888
1889 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1890 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1891 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1892 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1893 does: it's just the validity check that is more restrictive.
1894
1895 16. Inserted checks for integer overflows during escape sequence (backslash)
1896 processing, and also fixed erroneous offset values for syntax errors during
1897 backslash processing.
1898
1899 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1900 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1901
1902 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1903 caused an overrun.
1904
1905 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1906 something other than just ASCII characters) inside a group that had an
1907 unlimited repeat caused a loop at compile time (while checking to see
1908 whether the group could match an empty string).
1909
1910 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1911 [\P{Any}] did so. (Error in the code for printing property names.)
1912
1913 21. An orphan \E inside a character class could cause a crash.
1914
1915 22. A repeated capturing bracket such as (A)? could cause a wild memory
1916 reference during compilation.
1917
1918 23. There are several functions in pcre_compile() that scan along a compiled
1919 expression for various reasons (e.g. to see if it's fixed length for look
1920 behind). There were bugs in these functions when a repeated \p or \P was
1921 present in the pattern. These operators have additional parameters compared
1922 with \d, etc, and these were not being taken into account when moving along
1923 the compiled data. Specifically:
1924
1925 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1926 length.
1927
1928 (b) An item such as \pL+ within a repeated group could cause crashes or
1929 loops.
1930
1931 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1932 "reference to non-existent subpattern" error.
1933
1934 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1935
1936 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1937 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1938
1939 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1940
1941 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1942 character were causing crashes (broken optimization).
1943
1944 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1945 \p or \P) caused a compile-time loop.
1946
1947 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1948 break. For example, the unstudied pattern [\r\n]A does not match the string
1949 "\r\nA" because change 7.0/46 below moves the current point on by two
1950 characters after failing to match at the start. However, the pattern \nA
1951 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1952 the same is true. There doesn't seem any very clean way out of this, but
1953 what I have chosen to do makes the common cases work: PCRE now takes note
1954 of whether there can be an explicit match for \r or \n anywhere in the
1955 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1956 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1957 pattern has explicit CR or LF references.
1958
1959 29. Added (*CR) etc for changing newline setting at start of pattern.
1960
1961
1962 Version 7.2 19-Jun-07
1963 ---------------------
1964
1965 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1966 which is apparently normally available under Windows.
1967
1968 2. Re-jig the pcregrep tests with different newline settings in an attempt
1969 to make them independent of the local environment's newline setting.
1970
1971 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1972
1973 4. Some of the "internals" tests were previously cut out when the link size
1974 was not 2, because the output contained actual offsets. The recent new
1975 "Z" feature of pcretest means that these can be cut out, making the tests
1976 usable with all link sizes.
1977
1978 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1979 stack recursion. This gives a massive performance boost under BSD, but just
1980 a small improvement under Linux. However, it saves one field in the frame
1981 in all cases.
1982
1983 6. Added more features from the forthcoming Perl 5.10:
1984
1985 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1986 recursion call. It refers to the nth most recently opened parentheses.
1987
1988 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1989 to be opened parentheses.
1990
1991 (c) Conditions that refer to capturing parentheses can be specified
1992 relatively, for example, (?(-2)... or (?(+3)...
1993
1994 (d) \K resets the start of the current match so that everything before
1995 is not part of it.
1996
1997 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1998
1999 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2000 reference syntax.
2001
2002 (g) (?| introduces a group in which the numbering of parentheses in each
2003 alternative starts with the same number.
2004
2005 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2006
2007 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2008 PCRE_INFO_JCHANGED.
2009
2010 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2011 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2012 for detecting groups that can match an empty string.
2013
2014 9. A pattern with a very large number of alternatives (more than several
2015 hundred) was running out of internal workspace during the pre-compile
2016 phase, where pcre_compile() figures out how much memory will be needed. A
2017 bit of new cunning has reduced the workspace needed for groups with
2018 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2019 workspace instead of running out of the 4096 that are available.
2020
2021 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2022
2023 11. Applied patch from Google to remove an optimization that didn't quite work.
2024 The report of the bug said:
2025
2026 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2027 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2028 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2029
2030 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2031 it matched the wrong number of bytes.
2032
2033
2034 Version 7.1 24-Apr-07
2035 ---------------------
2036
2037 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2038 that is more "standard", making use of automake and other Autotools. There
2039 is some re-arrangement of the files and adjustment of comments consequent
2040 on this.
2041
2042 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2043 for recursive directory scanning broke on some systems because the files
2044 are not scanned in any specific order and on different systems the order
2045 was different. A call to "sort" has been inserted into RunGrepTest for the
2046 approprate test as a short-term fix. In the longer term there may be an
2047 alternative.
2048
2049 3. I had an email from Eric Raymond about problems translating some of PCRE's
2050 man pages to HTML (despite the fact that I distribute HTML pages, some
2051 people do their own conversions for various reasons). The problems
2052 concerned the use of low-level troff macros .br and .in. I have therefore
2053 removed all such uses from the man pages (some were redundant, some could
2054 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2055 HTML has been updated to handle .nf/.fi and to complain if it encounters
2056 .br or .in.
2057
2058 4. Updated comments in configure.ac that get placed in config.h.in and also
2059 arranged for config.h to be included in the distribution, with the name
2060 config.h.generic, for the benefit of those who have to compile without
2061 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2062
2063 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2064 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2065 makevp.bat for the current PCRE, using the additional files
2066 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2067
2068 6. A Windows user reported a minor discrepancy with test 2, which turned out
2069 to be caused by a trailing space on an input line that had got lost in his
2070 copy. The trailing space was an accident, so I've just removed it.
2071
2072 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2073 that is needed.
2074
2075 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2076 as "const" (a) because they are and (b) because it helps the PHP
2077 maintainers who have recently made a script to detect big data structures
2078 in the php code that should be moved to the .rodata section. I remembered
2079 to update Builducptable as well, so it won't revert if ucptable.h is ever
2080 re-created.
2081
2082 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2083 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2084 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2085 support is not required. This saves 1.5-2K of code, which is important in
2086 some applications.
2087
2088 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2089 so as not to refer to the tables, even though these functions will never be
2090 called when UTF-8 support is disabled. Otherwise there are problems with a
2091 shared library.
2092
2093 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2094
2095 (a) It was defining its arguments as char * instead of void *.
2096
2097 (b) It was assuming that all moves were upwards in memory; this was true
2098 a long time ago when I wrote it, but is no longer the case.
2099
2100 The emulated memove() is provided for those environments that have neither
2101 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2102 is clearly not the case, as these two bugs were recently reported.
2103
2104 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2105 and Detrail to create the HTML documentation, the .txt form of the man
2106 pages, and it removes trailing spaces from listed files. It also creates
2107 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2108 case, it wraps all the #defines with #ifndefs. This script should be run
2109 before "make dist".
2110
2111 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2112 with Unicode property support.
2113
2114 (a) For a maximizing quantifier, if the two different cases of the
2115 character were of different lengths in their UTF-8 codings (there are
2116 some cases like this - I found 11), and the matching function had to
2117 back up over a mixture of the two cases, it incorrectly assumed they
2118 were both the same length.
2119
2120 (b) When PCRE was configured to use the heap rather than the stack for
2121 recursion during matching, it was not correctly preserving the data for
2122 the other case of a UTF-8 character when checking ahead for a match
2123 while processing a minimizing repeat. If the check also involved
2124 matching a wide character, but failed, corruption could cause an
2125 erroneous result when trying to check for a repeat of the original
2126 character.
2127
2128 13. Some tidying changes to the testing mechanism:
2129
2130 (a) The RunTest script now detects the internal link size and whether there
2131 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2132 values substituted by "configure". (The RunGrepTest script already did
2133 this for UTF-8.) The configure.ac script no longer substitutes the
2134 relevant variables.
2135
2136 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2137 with length and offset values. This means that the output is different
2138 for different internal link sizes. Test 2 is skipped for link sizes
2139 other than 2 because of this, bypassing the problem. Unfortunately,
2140 there was also a test in test 3 (the locale tests) that used /B and
2141 failed for link sizes other than 2. Rather than cut the whole test out,
2142 I have added a new /Z option to pcretest that replaces the length and
2143 offset values with spaces. This is now used to make test 3 independent
2144 of link size. (Test 2 will be tidied up later.)
2145
2146 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2147 segmentation fault instead of returning the appropriate error message.
2148
2149 15. In multiline mode when the newline sequence was set to "any", the pattern
2150 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2151 This doesn't seem right; it now treats the CRLF combination as the line
2152 ending, and so does not match in that case. It's only a pattern such as ^$
2153 that would hit this one: something like ^ABC$ would have failed after \r
2154 and then tried again after \r\n.
2155
2156 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2157 in an attempt to make files that differ only in their line terminators
2158 compare equal. This works on Linux.
2159
2160 17. Under certain error circumstances pcregrep might try to free random memory
2161 as it exited. This is now fixed, thanks to valgrind.
2162
2163 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2164 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2165 was because its rules for how to advance for /g after matching an empty
2166 string at the end of a line did not allow for this case. They now check for
2167 it specially.
2168
2169 20. pcretest is supposed to handle patterns and data of any length, by
2170 extending its buffers when necessary. It was getting this wrong when the
2171 buffer for a data line had to be extended.
2172
2173 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2174 CRLF as a newline sequence.
2175
2176 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2177 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2178 I have nevertheless tidied it up.
2179
2180 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2181
2182 24. Added a man page for pcre-config.
2183
2184
2185 Version 7.0 19-Dec-06
2186 ---------------------
2187
2188 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2189 moving to gcc 4.1.1.
2190
2191 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2192 sys/time.h, which is documented as needed for this function. It doesn't
2193 seem to matter on Linux, but it showed up on some releases of OS X.
2194
2195 3. It seems that there are systems where bytes whose values are greater than
2196 127 match isprint() in the "C" locale. The "C" locale should be the
2197 default when a C program starts up. In most systems, only ASCII printing
2198 characters match isprint(). This difference caused the output from pcretest
2199 to vary, making some of the tests fail. I have changed pcretest so that:
2200
2201 (a) When it is outputting text in the compiled version of a pattern, bytes
2202 other than 32-126 are always shown as hex escapes.
2203
2204 (b) When it is outputting text that is a matched part of a subject string,
2205 it does the same, unless a different locale has been set for the match
2206 (using the /L modifier). In this case, it uses isprint() to decide.
2207
2208 4. Fixed a major bug that caused incorrect computation of the amount of memory
2209 required for a compiled pattern when options that changed within the
2210 pattern affected the logic of the preliminary scan that determines the
2211 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2212 that the computed length was too small. The symptoms of this bug were
2213 either the PCRE error "internal error: code overflow" from pcre_compile(),
2214 or a glibc crash with a message such as "pcretest: free(): invalid next
2215 size (fast)". Examples of patterns that provoked this bug (shown in
2216 pcretest format) are:
2217
2218 /(?-x: )/x
2219 /(?x)(?-x: \s*#\s*)/
2220 /((?i)[\x{c0}])/8
2221 /(?i:[\x{c0}])/8
2222
2223 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2224 is now done differently.
2225
2226 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2227 wrapper classes; (b) implement a new function in the C++ scanner that is
2228 more efficient than the old way of doing things because it avoids levels of
2229 recursion in the regex matching; (c) add a paragraph to the documentation
2230 for the FullMatch() function.
2231
2232 6. The escape sequence \n was being treated as whatever was defined as
2233 "newline". Not only was this contrary to the documentation, which states
2234 that \n is character 10 (hex 0A), but it also went horribly wrong when
2235 "newline" was defined as CRLF. This has been fixed.
2236
2237 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2238 was being set to -1 for the "end of line" case (supposedly a value that no
2239 character can have). Though this value is never used (the check for end of
2240 line is "zero bytes in current character"), it caused compiler complaints.
2241 I've changed it to 0xffffffff.
2242
2243 8. In pcre_version.c, the version string was being built by a sequence of
2244 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2245 string (as it is for production releases) called a macro with an empty
2246 argument. The C standard says the result of this is undefined. The gcc
2247 compiler treats it as an empty string (which was what was wanted) but it is
2248 reported that Visual C gives an error. The source has been hacked around to
2249 avoid this problem.
2250
2251 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2252 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2253 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2254 of them did).
2255
2256 10. Originally, pcretest opened its input and output without "b"; then I was
2257 told that "b" was needed in some environments, so it was added for release
2258 5.0 to both the input and output. (It makes no difference on Unix-like
2259 systems.) Later I was told that it is wrong for the input on Windows. I've
2260 now abstracted the modes into two macros, to make it easier to fiddle with
2261 them, and removed "b" from the input mode under Windows.
2262
2263 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2264
2265 12. Added -help and --help to pcretest as an official way of being reminded
2266 of the options.
2267
2268 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2269 and pcrecpp.cc because they annoy compilers at high warning levels.
2270
2271 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2272
2273 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2274 scripts are not C programs :-) and which was not noticed because it works
2275 on Linux.
2276
2277 16. pcretest is supposed to handle any length of pattern and data line (as one
2278 line or as a continued sequence of lines) by extending its input buffer if
2279 necessary. This feature was broken for very long pattern lines, leading to
2280 a string of junk being passed to pcre_compile() if the pattern was longer
2281 than about 50K.
2282
2283 17. I have done a major re-factoring of the way pcre_compile() computes the
2284 amount of memory needed for a compiled pattern. Previously, there was code
2285 that made a preliminary scan of the pattern in order to do this. That was
2286 OK when PCRE was new, but as the facilities have expanded, it has become
2287 harder and harder to keep it in step with the real compile phase, and there
2288 have been a number of bugs (see for example, 4 above). I have now found a
2289 cunning way of running the real compile function in a "fake" mode that
2290 enables it to compute how much memory it would need, while actually only
2291 ever using a few hundred bytes of working memory and without too many
2292 tests of the mode. This should make future maintenance and development
2293 easier. A side effect of this work is that the limit of 200 on the nesting
2294 depth of parentheses has been removed (though this was never a serious
2295 limitation, I suspect). However, there is a downside: pcre_compile() now
2296 runs more slowly than before (30% or more, depending on the pattern). I
2297 hope this isn't a big issue. There is no effect on runtime performance.
2298
2299 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2300 newline (only possible for the last line of a file) and it was a
2301 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2302
2303 19. Added additional timing features to pcretest. (1) The -tm option now times
2304 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2305 separate command line item, by a number that specifies the number of
2306 repeats to use when timing. The default is 50000; this gives better
2307 precision, but takes uncomfortably long for very large patterns.
2308
2309 20. Extended pcre_study() to be more clever in cases where a branch of a
2310 subpattern has no definite first character. For example, (a*|b*)[cd] would
2311 previously give no result from pcre_study(). Now it recognizes that the
2312 first character must be a, b, c, or d.
2313
2314 21. There was an incorrect error "recursive call could loop indefinitely" if
2315 a subpattern (or the entire pattern) that was being tested for matching an
2316 empty string contained only one non-empty item after a nested subpattern.
2317 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2318 incorrectly, because the \d was being skipped in the check.
2319
2320 22. The pcretest program now has a new pattern option /B and a command line
2321 option -b, which is equivalent to adding /B to every pattern. This causes
2322 it to show the compiled bytecode, without the additional information that
2323 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2324 is the same as /B/I).
2325
2326 23. A new optimization is now able automatically to treat some sequences such
2327 as a*b as a*+b. More specifically, if something simple (such as a character
2328 or a simple class like \d) has an unlimited quantifier, and is followed by
2329 something that cannot possibly match the quantified thing, the quantifier
2330 is automatically "possessified".
2331
2332 24. A recursive reference to a subpattern whose number was greater than 39
2333 went wrong under certain circumstances in UTF-8 mode. This bug could also
2334 have affected the operation of pcre_study().
2335
2336 25. Realized that a little bit of performance could be had by replacing
2337 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2338
2339 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2340
2341 27. Possessive quantifiers such as a++ were previously implemented by turning
2342 them into atomic groups such as ($>a+). Now they have their own opcodes,
2343 which improves performance. This includes the automatically created ones
2344 from 23 above.
2345
2346 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2347 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2348 the first matched character to be a colon. This applied both to named and
2349 numbered groups.
2350
2351 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2352
2353 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2354 building PCRE on Windows easier, so I have included it in the distribution.
2355
2356 31. There is now a check in pcretest against a ridiculously large number being
2357 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2358 loop, the loop is abandoned.
2359
2360 32. Forward references to subpatterns in conditions such as (?(2)...) where
2361 subpattern 2 is defined later cause pcre_compile() to search forwards in
2362 the pattern for the relevant set of parentheses. This search went wrong
2363 when there were unescaped parentheses in a character class, parentheses
2364 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2365
2366 33. "Subroutine" calls and backreferences were previously restricted to
2367 referencing subpatterns earlier in the regex. This restriction has now
2368 been removed.
2369
2370 34. Added a number of extra features that are going to be in Perl 5.10. On the
2371 whole, these are just syntactic alternatives for features that PCRE had
2372 previously implemented using the Python syntax or my own invention. The
2373 other formats are all retained for compatibility.
2374
2375 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2376 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2377 also .NET compatible.
2378
2379 (b) A recursion or subroutine call to a named group can now be defined as
2380 (?&name) as well as (?P>name).
2381
2382 (c) A backreference to a named group can now be defined as \k<name> or
2383 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2384 5.10, are also .NET compatible.
2385
2386 (d) A conditional reference to a named group can now use the syntax
2387 (?(<name>) or (?('name') as well as (?(name).
2388
2389 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2390 groups (named and numbered) that are never evaluated inline, but can be
2391 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2392 is always false. There may be only one alternative in such a group.
2393
2394 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2395 as the simple (?(R). The condition is true only if the most recent
2396 recursion is that of the given number or name. It does not search out
2397 through the entire recursion stack.
2398
2399 (g) The escape \gN or \g{N} has been added, where N is a positive or
2400 negative number, specifying an absolute or relative reference.
2401
2402 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2403 some "unreachable code" warnings.
2404
2405 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2406 things, this adds five new scripts.
2407
2408 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2409 There were also incompatibilities regarding the handling of \Q..\E inside
2410 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2411 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2412
2413 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2414 matches an empty string, and forcibly breaks the loop. There were bugs in
2415 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2416 against aaaa the result was just "a" rather than "aaaa", for example. Two
2417 separate and independent bugs (that affected different cases) have been
2418 fixed.
2419
2420 39. Refactored the code to abolish the use of different opcodes for small
2421 capturing bracket numbers. This is a tidy that I avoided doing when I
2422 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2423 The new approach is not only tidier, it makes it possible to reduce the
2424 memory needed to fix the previous bug (38).
2425
2426 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2427 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2428 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2429 mode.
2430
2431 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2432 report.
2433
2434 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2435 copy construction and assignment in the C++ wrapper.
2436
2437 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2438 couple of bugs that could have given wrong results in the "--newline=crlf"
2439 case.
2440
2441 44. Added a number of casts and did some reorganization of signed/unsigned int
2442 variables following suggestions from Dair Grant. Also renamed the variable
2443 "this" as "item" because it is a C++ keyword.
2444
2445 45. Arranged for dftables to add
2446
2447 #include "pcre_internal.h"
2448
2449 to pcre_chartables.c because without it, gcc 4.x may remove the array
2450 definition from the final binary if PCRE is built into a static library and
2451 dead code stripping is activated.
2452
2453 46. For an unanchored pattern, if a match attempt fails at the start of a
2454 newline sequence, and the newline setting is CRLF or ANY, and the next two
2455 characters are CRLF, advance by two characters instead of one.
2456
2457
2458 Version 6.7 04-Jul-06
2459 ---------------------
2460
2461 1. In order to handle tests when input lines are enormously long, pcretest has
2462 been re-factored so that it automatically extends its buffers when
2463 necessary. The code is crude, but this _is_ just a test program. The
2464 default size has been increased from 32K to 50K.
2465
2466 2. The code in pcre_study() was using the value of the re argument before
2467 testing it for NULL. (Of course, in any sensible call of the function, it
2468 won't be NULL.)
2469
2470 3. The memmove() emulation function in pcre_internal.h, which is used on
2471 systems that lack both memmove() and bcopy() - that is, hardly ever -
2472 was missing a "static" storage class specifier.
2473
2474 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2475 containing an extended class (one that cannot be represented by a bitmap
2476 because it contains high-valued characters or Unicode property items, e.g.
2477 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2478 pattern, but PCRE should not loop if you do not (it no longer does).
2479 [Detail: two cases were found: (a) a repeated subpattern containing an
2480 extended class; (b) a recursive reference to a subpattern that followed a
2481 previous extended class. It wasn't skipping over the extended class
2482 correctly when UTF-8 mode was not set.]
2483
2484 5. A negated single-character class was not being recognized as fixed-length
2485 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2486 compile error "lookbehind assertion is not fixed length".
2487
2488 6. The RunPerlTest auxiliary script was showing an unexpected difference
2489 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2490 write a Perl script that can interpret lines of an input file either as
2491 byte characters or as UTF-8, which is what "perltest" was being required to
2492 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2493 can't do is switch easily at run time between having the "use utf8;" pragma
2494 or not. In the end, I fudged it by using the RunPerlTest script to insert
2495 "use utf8;" explicitly for the UTF-8 tests.
2496
2497 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2498 the end of the subject string, contrary to the documentation and to what
2499 Perl does. This was true of both matching functions. Now it matches only at
2500 the start of the subject and immediately after *internal* newlines.
2501
2502 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2503 a pointer to an int instead of a pointer to an unsigned long int. This
2504 caused problems on 64-bit systems.
2505
2506 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2507 instance of the 'standard' template library not being so standard".
2508
2509 10. There was no check on the number of named subpatterns nor the maximum
2510 length of a subpattern name. The product of these values is used to compute
2511 the size of the memory block for a compiled pattern. By supplying a very
2512 long subpattern name and a large number of named subpatterns, the size
2513 computation could be caused to overflow. This is now prevented by limiting
2514 the length of names to 32 characters, and the number of named subpatterns
2515 to 10,000.
2516
2517 11. Subpatterns that are repeated with specific counts have to be replicated in
2518 the compiled pattern. The size of memory for this was computed from the
2519 length of the subpattern and the repeat count. The latter is limited to
2520 65535, but there was no limit on the former, meaning that integer overflow
2521 could in principle occur. The compiled length of a repeated subpattern is
2522 now limited to 30,000 bytes in order to prevent this.
2523
2524 12. Added the optional facility to have named substrings with the same name.
2525
2526 13. Added the ability to use a named substring as a condition, using the
2527 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2528 are numbers (not recommended). Forward references are permitted.
2529
2530 14. Added forward references in named backreferences (if you see what I mean).
2531
2532 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2533 pattern could run off the end of the subject. For example, the pattern
2534 "(?s)(.{1,5})"8 did this with the subject "ab".
2535
2536 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2537 PCRE_CASELESS was set when matching characters that were quantified with ?
2538 or *.
2539
2540 17. A character class other than a single negated character that had a minimum
2541 but no maximum quantifier - for example [ab]{6,} - was not handled
2542 correctly by pce_dfa_exec(). It would match only one character.
2543
2544 18. A valid (though odd) pattern that looked like a POSIX character
2545 class but used an invalid character after [ (for example [[,abc,]]) caused
2546 pcre_compile() to give the error "Failed: internal error: code overflow" or
2547 in some cases to crash with a glibc free() error. This could even happen if
2548 the pattern terminated after [[ but there just happened to be a sequence of
2549 letters, a binary zero, and a closing ] in the memory that followed.
2550
2551 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2552 over the years. Originally (before any Unicode support), just the bottom 8
2553 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2554 output from "man perlunicode" includes this:
2555
2556 The regular expression compiler produces polymorphic opcodes. That
2557 is, the pattern adapts to the data and automatically switches to
2558 the Unicode character scheme when presented with Unicode data--or
2559 instead uses a traditional byte scheme when presented with byte
2560 data.
2561
2562 Sadly, a wide octal escape does not cause a switch, and in a string with
2563 no other multibyte characters, these octal escapes are treated as before.
2564 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2565 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2566 Unicode string.
2567
2568 I have not perpetrated such confusion in PCRE. Up till now, it took just
2569 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2570 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2571 translate to the appropriate multibyte character.
2572
2573 29. Applied some refactoring to reduce the number of warnings from Microsoft
2574 and Borland compilers. This has included removing the fudge introduced
2575 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2576 a warning about an unused variable.
2577
2578 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2579 characters since release 4.0, because Perl (from release 5.004) does not.
2580 [Or at least, is documented not to: some releases seem to be in conflict
2581 with the documentation.] However, when a pattern was studied with
2582 pcre_study() and all its branches started with \s, PCRE still included VT
2583 as a possible starting character. Of course, this did no harm; it just
2584 caused an unnecessary match attempt.
2585
2586 22. Removed a now-redundant internal flag bit that recorded the fact that case
2587 dependency changed within the pattern. This was once needed for "required
2588 byte" processing, but is no longer used. This recovers a now-scarce options
2589 bit. Also moved the least significant internal flag bit to the most-
2590 significant bit of the word, which was not previously used (hangover from
2591 the days when it was an int rather than a uint) to free up another bit for
2592 the future.
2593
2594 23. Added support for CRLF line endings as well as CR and LF. As well as the
2595 default being selectable at build time, it can now be changed at runtime
2596 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2597 specify that it is scanning data with non-default line endings.
2598
2599 24. Changed the definition of CXXLINK to make it agree with the definition of
2600 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2601
2602 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2603 recursions. This makes a big different to stack usage for some patterns.
2604
2605 26. If a subpattern containing a named recursion or subroutine reference such
2606 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2607 the space required for the compiled pattern went wrong and gave too small a
2608 value. Depending on the environment, this could lead to "Failed: internal
2609 error: code overflow at offset 49" or "glibc detected double free or
2610 corruption" errors.
2611
2612 27. Applied patches from Google (a) to support the new newline modes and (b) to
2613 advance over multibyte UTF-8 characters in GlobalReplace.
2614
2615 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2616 difference for some implementation of PCRE in some Windows version.
2617
2618 29. Added some extra testing facilities to pcretest:
2619
2620 \q<number> in a data line sets the "match limit" value
2621 \Q<number> in a data line sets the "match recursion limt" value
2622 -S <number> sets the stack size, where <number> is in megabytes
2623
2624 The -S option isn't available for Windows.
2625
2626
2627 Version 6.6 06-Feb-06
2628 ---------------------
2629
2630 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2631 in pcreposix.h. I have copied the definition from pcre.h.
2632
2633 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2634 because pcre.h is no longer a built file.
2635
2636 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2637 not normally included in the compiled code.
2638
2639
2640 Version 6.5 01-Feb-06
2641 ---------------------
2642
2643 1. When using the partial match feature with pcre_dfa_exec(), it was not
2644 anchoring the second and subsequent partial matches at the new starting
2645 point. This could lead to incorrect results. For example, with the pattern
2646 /1234/, partially matching against "123" and then "a4" gave a match.
2647
2648 2. Changes to pcregrep:
2649
2650 (a) All non-match returns from pcre_exec() were being treated as failures
2651 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2652 error message is output. Some extra information is given for the
2653 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2654 probably the only errors that are likely to be caused by users (by
2655 specifying a regex that has nested indefinite repeats, for instance).
2656 If there are more than 20 of these errors, pcregrep is abandoned.
2657
2658 (b) A binary zero was treated as data while matching, but terminated the
2659 output line if it was written out. This has been fixed: binary zeroes
2660 are now no different to any other data bytes.
2661
2662 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2663 used to set a locale for matching. The --locale=xxxx long option has
2664 been added (no short equivalent) to specify a locale explicitly on the
2665 pcregrep command, overriding the environment variables.
2666
2667 (d) When -B was used with -n, some line numbers in the output were one less
2668 than they should have been.
2669
2670 (e) Added the -o (--only-matching) option.
2671
2672 (f) If -A or -C was used with -c (count only), some lines of context were
2673 accidentally printed for the final match.
2674
2675 (g) Added the -H (--with-filename) option.
2676
2677 (h) The combination of options -rh failed to suppress file names for files
2678 that were found from directory arguments.
2679
2680 (i) Added the -D (--devices) and -d (--directories) options.
2681
2682 (j) Added the -F (--fixed-strings) option.
2683
2684 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2685
2686 (l) Added the --colo(u)r option.
2687
2688 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2689 is not present by default.
2690
2691 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2692 items such as (?R) or (?1), when the recursion could match a number of
2693 alternatives. If it matched one of the alternatives, but subsequently,
2694 outside the recursion, there was a failure, the code tried to back up into
2695 the recursion. However, because of the way PCRE is implemented, this is not
2696 possible, and the result was an incorrect result from the match.
2697
2698 In order to prevent this happening, the specification of recursion has
2699 been changed so that all such subpatterns are automatically treated as
2700 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2701
2702 4. I had overlooked the fact that, in some locales, there are characters for
2703 which isalpha() is true but neither isupper() nor islower() are true. In
2704 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2705 and ordfeminine) are like this. This affected the treatment of \w and \W
2706 when they appeared in character classes, but not when they appeared outside
2707 a character class. The bit map for "word" characters is now created
2708 separately from the results of isalnum() instead of just taking it from the
2709 upper, lower, and digit maps. (Plus the underscore character, of course.)
2710
2711 5. The above bug also affected the handling of POSIX character classes such as
2712 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2713 permanent tables. Instead, the bit maps for such a class were previously
2714 created as the appropriate unions of the upper, lower, and digit bitmaps.
2715 Now they are created by subtraction from the [[:word:]] class, which has
2716 its own bitmap.
2717
2718 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2719 It is created by subtracting the vertical space characters (\x09, \x0a,
2720 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2721 subtraction was done in the overall bitmap for a character class, meaning
2722 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2723 be recognized. This bug has been fixed.
2724
2725 7. Patches from the folks at Google:
2726
2727 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2728 real life, but is still worth protecting against".
2729
2730 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2731 regular expressions".
2732
2733 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2734 have it.
2735
2736 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2737 "configure" and the latter not, in order to fix a problem somebody had
2738 with compiling the Arg class on HP-UX.
2739
2740 (e) Improve the error-handling of the C++ wrapper a little bit.
2741
2742 (f) New tests for checking recursion limiting.
2743
2744 8. The pcre_memmove() function, which is used only if the environment does not
2745 have a standard memmove() function (and is therefore rarely compiled),
2746 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2747 returning a result (though PCRE never actually uses the result).
2748
2749 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2750 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2751 returned instead of calling malloc() with an overflowing number that would
2752 most likely cause subsequent chaos.
2753
2754 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2755
2756 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2757 with this option is matched, the nmatch and pmatch options of regexec() are
2758 ignored.
2759
2760 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2761 provided in case anyone wants to the the POSIX interface with UTF-8
2762 strings.
2763
2764 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2765 C++ linking (needed for some HP-UX environments).
2766
2767 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2768 (unused parameter) and in the pcre_printint() function (omitted "default"
2769 switch label when the default is to do nothing).
2770
2771 15. Added some code to make it possible, when PCRE is compiled as a C++
2772 library, to replace subject pointers for pcre_exec() with a smart pointer
2773 class, thus making it possible to process discontinuous strings.
2774
2775 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2776 much the same function. They were added by different people who were trying
2777 to make PCRE easy to compile on non-Unix systems. It has been suggested
2778 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2779 for compiling on Windows systems. I have therefore replaced it with
2780 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2781 defaults to "extern" for C or "extern C" for C++, which works fine on
2782 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2783 SCOPE with something explicit in config.h. In addition:
2784
2785 (a) pcreposix.h still had just "extern" instead of either of these macros;
2786 I have replaced it with PCRE_DATA_SCOPE.
2787
2788 (b) Functions such as _pcre_xclass(), which are internal to the library,
2789 but external in the C sense, all had PCRE_EXPORT in their definitions.
2790 This is apparently wrong for the Windows case, so I have removed it.
2791 (It makes no difference on Unix-like systems.)
2792
2793 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2794 of recursive calls to match(). This is different to MATCH_LIMIT because
2795 that limits the total number of calls to match(), not all of which increase
2796 the depth of recursion. Limiting the recursion depth limits the amount of
2797 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2798 when PCRE is compiled, and changed at run time. A patch from Google adds
2799 this functionality to the C++ interface.
2800
2801 18. Changes to the handling of Unicode character properties:
2802
2803 (a) Updated the table to Unicode 4.1.0.
2804
2805 (b) Recognize characters that are not in the table as "Cn" (undefined).
2806
2807 (c) I revised the way the table is implemented to a much improved format
2808 which includes recognition of ranges. It now supports the ranges that
2809 are defined in UnicodeData.txt, and it also amalgamates other
2810 characters into ranges. This has reduced the number of entries in the
2811 table from around 16,000 to around 3,000, thus reducing its size
2812 considerably. I realized I did not need to use a tree structure after
2813 all - a binary chop search is just as efficient. Having reduced the
2814 number of entries, I extended their size from 6 bytes to 8 bytes to
2815 allow for more data.
2816
2817 (d) Added support for Unicode script names via properties such as \p{Han}.
2818
2819 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2820 matching that character.
2821
2822 20. When matching a repeated Unicode property with a minimum greater than zero,
2823 (for example \pL{2,}), PCRE could look past the end of the subject if it
2824 reached it while seeking the minimum number of characters. This could
2825 happen only if some of the characters were more than one byte long, because
2826 there is a check for at least the minimum number of bytes.
2827
2828 21. Refactored the implementation of \p and \P so as to be more general, to
2829 allow for more different types of property in future. This has changed the
2830 compiled form incompatibly. Anybody with saved compiled patterns that use
2831 \p or \P will have to recompile them.
2832
2833 22. Added "Any" and "L&" to the supported property types.
2834
2835 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2836 but give a compile time error if the value is greater than 0xff.
2837
2838 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2839 accidentally not being installed or uninstalled.
2840
2841 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2842 made were to insert the current release number. This seemed silly, because
2843 it made things harder for people building PCRE on systems that don't run
2844 "configure". I have turned pcre.h into a distributed file, no longer built
2845 by "configure", with the version identification directly included. There is
2846 no longer a pcre.h.in file.
2847
2848 However, this change necessitated a change to the pcre-config script as
2849 well. It is built from pcre-config.in, and one of the substitutions was the
2850 release number. I have updated configure.ac so that ./configure now finds
2851 the release number by grepping pcre.h.
2852
2853 26. Added the ability to run the tests under valgrind.
2854
2855
2856 Version 6.4 05-Sep-05
2857 ---------------------
2858
2859 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2860 "--" to be printed when multiple files were scanned, even when none of the
2861 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2862 consider it to be a bug, and have restored the previous behaviour.
2863
2864 2. A couple of code tidies to get rid of compiler warnings.
2865
2866 3. The pcretest program used to cheat by referring to symbols in the library
2867 whose names begin with _pcre_. These are internal symbols that are not
2868 really supposed to be visible externally, and in some environments it is
2869 possible to suppress them. The cheating is now confined to including
2870 certain files from the library's source, which is a bit cleaner.
2871
2872 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2873 file's purpose clearer.
2874
2875 5. Reorganized pcre_ucp_findchar().
2876
2877
2878 Version 6.3 15-Aug-05
2879 ---------------------
2880
2881 1. The file libpcre.pc.in did not have general read permission in the tarball.
2882
2883 2. There were some problems when building without C++ support:
2884
2885 (a) If C++ support was not built, "make install" and "make test" still
2886 tried to test it.
2887
2888 (b) There were problems when the value of CXX was explicitly set. Some
2889 changes have been made to try to fix these, and ...
2890
2891 (c) --disable-cpp can now be used to explicitly disable C++ support.
2892
2893 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2894 backslash in a target when C++ was disabled. This confuses some
2895 versions of "make", apparently. Using an intermediate variable solves
2896 this. (Same for CPP_LOBJ.)
2897
2898 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2899 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2900 necessary on certain architectures.
2901
2902 4. Added a setting of -export-symbols-regex to the link command to remove
2903 those symbols that are exported in the C sense, but actually are local
2904 within the library, and not documented. Their names all begin with
2905 "_pcre_". This is not a perfect job, because (a) we have to except some
2906 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2907 available (and never for static libraries). I have made a note to try to
2908 find a way round (a) in the future.
2909
2910
2911 Version 6.2 01-Aug-05
2912 ---------------------
2913
2914 1. There was no test for integer overflow of quantifier values. A construction
2915 such as {1111111111111111} would give undefined results. What is worse, if
2916 a minimum quantifier for a parenthesized subpattern overflowed and became
2917 negative, the calculation of the memory size went wrong. This could have
2918 led to memory overwriting.
2919
2920 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2921
2922 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2923 operating environments where this matters.
2924
2925 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2926 PCRE options from within the C++ wrapper.
2927
2928 5. Named capturing subpatterns were not being correctly counted when a pattern
2929 was compiled. This caused two problems: (a) If there were more than 100
2930 such subpatterns, the calculation of the memory needed for the whole
2931 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2932 back references of the form \12, where the number was greater than 9, were
2933 not recognized as back references, even though there were sufficient
2934 previous subpatterns.
2935
2936 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2937 versions of gcc, e.g. 2.95.4.
2938
2939
2940 Version 6.1 21-Jun-05
2941 ---------------------
2942
2943 1. There was one reference to the variable "posix" in pcretest.c that was not
2944 surrounded by "#if !defined NOPOSIX".
2945
2946 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2947 the cross-check on the old pcre_info() function, for the benefit of the
2948 cut-down version of PCRE that is currently imported into Exim.
2949
2950 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2951 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2952 patterns that start (?i) but I don't think that matters. The use of (?i) is
2953 just an example; this all applies to the other options as well.
2954
2955 4. Since libtool seems to echo the compile commands it is issuing, the output
2956 from "make" can be reduced a bit by putting "@" in front of each libtool
2957 compile command.
2958
2959 5. Patch from the folks at Google for configure.in to be a bit more thorough
2960 in checking for a suitable C++ installation before trying to compile the
2961 C++ stuff. This should fix a reported problem when a compiler was present,
2962 but no suitable headers.
2963
2964 6. The man pages all had just "PCRE" as their title. I have changed them to
2965 be the relevant file name. I have also arranged that these names are
2966 retained in the file doc/pcre.txt, which is a concatenation in text format
2967 of all the man pages except the little individual ones for each function.
2968
2969 7. The NON-UNIX-USE file had not been updated for the different set of source
2970 files that come with release 6. I also added a few comments about the C++
2971 wrapper.
2972
2973
2974 Version 6.0 07-Jun-05
2975 ---------------------
2976
2977 1. Some minor internal re-organization to help with my DFA experiments.
2978
2979 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2980 didn't matter for the library itself when fully configured, but did matter
2981 when compiling without UCP support, or within Exim, where the ucp files are
2982 not imported.
2983
2984 3. Refactoring of the library code to split up the various functions into
2985 different source modules. The addition of the new DFA matching code (see
2986 below) to a single monolithic source would have made it really too
2987 unwieldy, quite apart from causing all the code to be include in a
2988 statically linked application, when only some functions are used. This is
2989 relevant even without the DFA addition now that patterns can be compiled in
2990 one application and matched in another.
2991
2992 The downside of splitting up is that there have to be some external
2993 functions and data tables that are used internally in different modules of
2994 the library but which are not part of the API. These have all had their
2995 names changed to start with "_pcre_" so that they are unlikely to clash
2996 with other external names.
2997
2998 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2999 a different (DFA) algorithm. Although it is slower than the original
3000 function, it does have some advantages for certain types of matching
3001 problem.
3002
3003 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3004 including restarting after a partial match.
3005
3006 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3007 defined when compiling for Windows was sent to me. I have put it into the
3008 code, though I have no means of testing or verifying it.
3009
3010 7. Added the pcre_refcount() auxiliary function.
3011
3012 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3013 match before or at the first newline in the subject string. In pcretest,
3014 the /f option on a pattern can be used to set this.
3015
3016 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3017 would behave wrongly. This has been present in PCRE since release 4.0.
3018
3019 10. A number of changes to the pcregrep command:
3020
3021 (a) Refactored how -x works; insert ^(...)$ instead of setting
3022 PCRE_ANCHORED and checking the length, in preparation for adding
3023 something similar for -w.
3024
3025 (b) Added the -w (match as a word) option.
3026
3027 (c) Refactored the way lines are read and buffered so as to have more
3028 than one at a time available.
3029
3030 (d) Implemented a pcregrep test script.
3031
3032 (e) Added the -M (multiline match) option. This allows patterns to match
3033 over several lines of the subject. The buffering ensures that at least
3034 8K, or the rest of the document (whichever is the shorter) is available
3035 for matching (and similarly the previous 8K for lookbehind assertions).
3036
3037 (f) Changed the --help output so that it now says
3038
3039 -w, --word-regex(p)
3040
3041 instead of two lines, one with "regex" and the other with "regexp"
3042 because that confused at least one person since the short forms are the
3043 same. (This required a bit of code, as the output is generated
3044 automatically from a table. It wasn't just a text change.)
3045
3046 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3047 option but starts with a hyphen. Could be a pattern or a path name
3048 starting with a hyphen, for instance.
3049
3050 (h) "-" can be given as a file name to represent stdin.
3051
3052 (i) When file names are being printed, "(standard input)" is used for
3053 the standard input, for compatibility with GNU grep. Previously
3054 "<stdin>" was used.
3055
3056 (j) The option --label=xxx can be used to supply a name to be used for
3057 stdin when file names are being printed. There is no short form.
3058
3059 (k) Re-factored the options decoding logic because we are going to add
3060 two more options that take data. Such options can now be given in four
3061 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3062
3063 (l) Added the -A, -B, and -C options for requesting that lines of context
3064 around matches be printed.
3065
3066 (m) Added the -L option to print the names of files that do not contain
3067 any matching lines, that is, the complement of -l.
3068
3069 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3070 continue to scan other files.
3071
3072 (o) The -s option was incorrectly implemented. For compatibility with other
3073 greps, it now suppresses the error message for a non-existent or non-
3074 accessible file (but not the return code). There is a new option called
3075 -q that suppresses the output of matching lines, which was what -s was
3076 previously doing.
3077
3078 (p) Added --include and --exclude options to specify files for inclusion
3079 and exclusion when recursing.
3080
3081 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3082 Hopefully, it now does.
3083
3084 12. Missing cast in pcre_study().
3085
3086 13. Added an "uninstall" target to the makefile.
3087
3088 14. Replaced "extern" in the function prototypes in Makefile.in with
3089 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3090 world, but is set differently for Windows.
3091
3092 15. Added a second compiling function called pcre_compile2(). The only
3093 difference is that it has an extra argument, which is a pointer to an
3094 integer error code. When there is a compile-time failure, this is set
3095 non-zero, in addition to the error test pointer being set to point to an
3096 error message. The new argument may be NULL if no error number is required
3097 (but then you may as well call pcre_compile(), which is now just a
3098 wrapper). This facility is provided because some applications need a
3099 numeric error indication, but it has also enabled me to tidy up the way
3100 compile-time errors are handled in the POSIX wrapper.
3101
3102 16. Added VPATH=.libs to the makefile; this should help when building with one
3103 prefix path and installing with another. (Or so I'm told by someone who
3104 knows more about this stuff than I do.)
3105
3106 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3107 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3108 match everything, including newlines. This is not POSIX-compatible, but
3109 somebody wanted the feature. From pcretest it can be activated by using
3110 both the P and the s flags.
3111
3112 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3113
3114 19. libpcre.pc was being incorrectly installed as executable.
3115
3116 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3117 it now also looks for '\r' so that it will work unmodified on Windows.
3118
3119 21. Added Google's contributed C++ wrapper to the distribution.
3120
3121 22. Added some untidy missing memory free() calls in pcretest, to keep
3122 Electric Fence happy when testing.
3123
3124
3125
3126 Version 5.0 13-Sep-04
3127 ---------------------
3128
3129 1. Internal change: literal characters are no longer packed up into items
3130 containing multiple characters in a single byte-string. Each character
3131 is now matched using a separate opcode. However, there may be more than one
3132 byte in the character in UTF-8 mode.
3133
3134 2. The pcre_callout_block structure has two new fields: pattern_position and
3135 next_item_length. These contain the offset in the pattern to the next match
3136 item, and its length, respectively.
3137
3138 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3139 insertion of callouts before each pattern item. Added the /C option to
3140 pcretest to make use of this.
3141
3142 4. On the advice of a Windows user, the lines
3143
3144 #if defined(_WIN32) || defined(WIN32)
3145 _setmode( _fileno( stdout ), 0x8000 );
3146 #endif /* defined(_WIN32) || defined(WIN32) */
3147
3148 have been added to the source of pcretest. This apparently does useful
3149 magic in relation to line terminators.
3150
3151 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3152 for the benefit of those environments where the "b" makes a difference.
3153
3154 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3155 to know about it. I have put a hack into configure.in that adds in code
3156 to set GCC=yes if CC=icc. This seems to end up at a point in the
3157 generated configure script that is early enough to affect the setting of
3158 compiler options, which is what is needed, but I have no means of testing
3159 whether it really works. (The user who reported this had patched the
3160 generated configure script, which of course I cannot do.)
3161
3162 LATER: After change 22 below (new libtool files), the configure script
3163 seems to know about icc (and also ecc). Therefore, I have commented out
3164 this hack in configure.in.
3165
3166 7. Added support for pkg-config (2 patches were sent in).
3167
3168 8. Negated POSIX character classes that used a combination of internal tables
3169 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3170 [[:^ascii]]. Typically, they would match almost any characters. The other
3171 POSIX classes were not broken in this way.
3172
3173 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3174 to find the match, as PCRE was deluded into thinking that the match had to
3175 start at the start point or following a newline. The same bug applied to
3176 patterns with negative forward assertions or any backward assertions
3177 preceding ".*" at the start, unless the pattern required a fixed first
3178 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3179
3180 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3181 starting at the last subject character, bytes beyond the end of the subject
3182 string were read.
3183
3184 11. Renamed the variable "class" as "classbits" to make life easier for C++
3185 users. (Previously there was a macro definition, but it apparently wasn't
3186 enough.)
3187
3188 12. Added the new field "tables" to the extra data so that tables can be passed
3189 in at exec time, or the internal tables can be re-selected. This allows
3190 a compiled regex to be saved and re-used at a later time by a different
3191 program that might have everything at different addresses.
3192
3193 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3194 -R library as well as a -L library.
3195
3196 14. The debugging options of pcretest (-d on the command line or D on a
3197 pattern) showed incorrect output for anything following an extended class
3198 that contained multibyte characters and which was followed by a quantifier.
3199
3200 15. Added optional support for general category Unicode character properties
3201 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3202 support. It adds about 90K to the size of the library. The meanings of the
3203 inbuilt class escapes such as \d and \s have NOT been changed.
3204
3205 16. Updated pcredemo.c to include calls to free() to release the memory for the
3206 compiled pattern.
3207
3208 17. The generated file chartables.c was being created in the source directory
3209 instead of in the building directory. This caused the build to fail if the
3210 source directory was different from the building directory, and was
3211 read-only.
3212
3213 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3214 file. No doubt somebody will tell me if they don't make sense... Also added
3215 Dan Mooney's comments about building on OpenVMS.
3216
3217 19. Added support for partial matching via the PCRE_PARTIAL option for
3218 pcre_exec() and the \P data escape in pcretest.
3219
3220 20. Extended pcretest with 3 new pattern features:
3221
3222 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3223 write the compiled pattern to the file whose name is "rest-of-line".
3224 This is a straight binary dump of the data, with the saved pointer to
3225 the character tables forced to be NULL. The study data, if any, is
3226 written too. After writing, pcretest reads a new pattern.
3227
3228 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3229 compiled pattern from the given file. There must not be any
3230 occurrences of "<" in the file name (pretty unlikely); if there are,
3231 pcretest will instead treat the initial "<" as a pattern delimiter.
3232 After reading in the pattern, pcretest goes on to read data lines as
3233 usual.
3234
3235 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3236 and 16-bit fields in a compiled pattern, to simulate a pattern that
3237 was compiled on a host of opposite endianness.
3238
3239 21. The pcre-exec() function can now cope with patterns that were compiled on
3240 hosts of opposite endianness, with this restriction:
3241
3242 As for any compiled expression that is saved and used later, the tables
3243 pointer field cannot be preserved; the extra_data field in the arguments
3244 to pcre_exec() should be used to pass in a tables address if a value
3245 other than the default internal tables were used at compile time.
3246
3247 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3248 now diagnosed as an error. Previously, most of the time, a negative number
3249 would have been treated as zero, but if in addition "ovector" was passed as
3250 NULL, a crash could occur.
3251
3252 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3253 new versions from the libtool 1.5 distribution (the last one is a copy of
3254 a file called libtool.m4). This seems to have fixed the need to patch
3255 "configure" to support Darwin 1.3 (which I used to do). However, I still
3256 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3257 workstation).
3258
3259 24. Changed the PCRE licence to be the more standard "BSD" licence.
3260
3261
3262 Version 4.5 01-Dec-03
3263 ---------------------
3264
3265 1. There has been some re-arrangement of the code for the match() function so
3266 that it can be compiled in a version that does not call itself recursively.
3267 Instead, it keeps those local variables that need separate instances for
3268 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3269 needs to "recurse". Keeping track of where control must go is done by means
3270 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3271 hide most of the details from the main code, and operates only if
3272 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3273 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3274 operating.
3275
3276 To make it easier for callers to provide specially tailored get/free
3277 functions for this usage, two new functions, pcre_stack_malloc, and
3278 pcre_stack_free, are used. They are always called in strict stacking order,
3279 and the size of block requested is always the same.
3280
3281 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3282 PCRE has been compiled to use the stack or the heap for recursion. The
3283 -C option of pcretest uses this to show which version is compiled.
3284
3285 A new data escape \S, is added to pcretest; it causes the amounts of store
3286 obtained and freed by both kinds of malloc/free at match time to be added
3287 to the output.
3288
3289 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3290 what's available on my current Linux desktop machine.
3291
3292 3. When matching a UTF-8 string, the test for a valid string at the start has
3293 been extended. If start_offset is not zero, PCRE now checks that it points
3294 to a byte that is the start of a UTF-8 character. If not, it returns
3295 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3296 this is necessary because there may be backward assertions in the pattern.
3297 When matching the same subject several times, it may save resources to use
3298 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3299
3300 4. The code for checking the validity of UTF-8 strings has been tightened so
3301 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3302 containing "overlong sequences".
3303
3304 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3305 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3306 should have been "&", but it just so happened that all the cases this let
3307 through by mistake were picked up later in the function.
3308
3309 6. I had used a variable called "isblank" - this is a C99 function, causing
3310 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3311
3312 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3313 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3314 so the version is shown; (c) stop "make test" echoing "./RunTest".
3315
3316 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3317
3318 9. The prototype for memmove() for systems that don't have it was using
3319 size_t, but the inclusion of the header that defines size_t was later. I've
3320 moved the #includes for the C headers earlier to avoid this.
3321
3322 10. Added some adjustments to the code to make it easier to compiler on certain
3323 special systems:
3324
3325 (a) Some "const" qualifiers were missing.
3326 (b) Added the macro EXPORT before all exported functions; by default this
3327 is defined to be empty.
3328 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3329 that it reads its output file name as an argument instead of writing
3330 to the standard output and assuming this can be redirected.
3331
3332 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3333 class containing characters with values greater than 255, PCRE compilation
3334 went into a loop.
3335
3336 12. A recursive reference to a subpattern that was within another subpattern
3337 that had a minimum quantifier of zero caused PCRE to crash. For example,
3338 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3339 recursion. If the recursively-called subpattern itself had a zero repeat,
3340 that was OK.
3341
3342 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3343 buffer into which it was copied (for escape processing) was still set at
3344 1024, so long lines caused crashes.
3345
3346 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3347 "internal error: code overflow...". This applied to any character class
3348 that was followed by a possessive quantifier.
3349
3350 15. Modified the Makefile to add libpcre.la as a prerequisite for
3351 libpcreposix.la because I was told this is needed for a parallel build to
3352 work.
3353
3354 16. If a pattern that contained .* following optional items at the start was
3355 studied, the wrong optimizing data was generated, leading to matching
3356 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3357 matching string must start with a or b or c. The correct conclusion for
3358 this pattern is that a match can start with any character.
3359
3360
3361 Version 4.4 13-Aug-03
3362 ---------------------
3363
3364 1. In UTF-8 mode, a character class containing characters with values between
3365 127 and 255 was not handled correctly if the compiled pattern was studied.
3366 In fixing this, I have also improved the studying algorithm for such
3367 classes (slightly).
3368
3369 2. Three internal functions had redundant arguments passed to them. Removal
3370 might give a very teeny performance improvement.
3371
3372 3. Documentation bug: the value of the capture_top field in a callout is *one
3373 more than* the number of the hightest numbered captured substring.
3374
3375 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3376 in incorrectly linking with a previously installed version. They now link
3377 explicitly with libpcre.la.
3378
3379 5. configure.in no longer needs to recognize Cygwin specially.
3380
3381 6. A problem in pcre.in for Windows platforms is fixed.
3382
3383 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3384 pcretest, it used to include the size of the study block as part of its
3385 output. Unfortunately, the structure contains a field that has a different
3386 size on different hardware architectures. This meant that the tests that
3387 showed this size failed. As the block is currently always of a fixed size,
3388 this information isn't actually particularly useful in pcretest output, so
3389 I have just removed it.
3390
3391 8. Three pre-processor statements accidentally did not start in column 1.
3392 Sadly, there are *still* compilers around that complain, even though
3393 standard C has not required this for well over a decade. Sigh.
3394
3395 9. In pcretest, the code for checking callouts passed small integers in the
3396 callout_data field, which is a void * field. However, some picky compilers
3397 complained about the casts involved for this on 64-bit systems. Now
3398 pcretest passes the address of the small integer instead, which should get
3399 rid of the warnings.
3400
3401 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3402 both compile and run time, and gives an error if an invalid UTF-8 sequence
3403 is found. There is a option for disabling this check in cases where the
3404 string is known to be correct and/or the maximum performance is wanted.
3405
3406 11. In response to a bug report, I changed one line in Makefile.in from
3407
3408 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3409 to
3410 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3411
3412 to look similar to other lines, but I have no way of telling whether this
3413 is the right thing to do, as I do not use Windows. No doubt I'll get told
3414 if it's wrong...
3415
3416
3417 Version 4.3 21-May-03
3418 ---------------------
3419
3420 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3421 Makefile.
3422
3423 2. Some refactoring to improve the quality of the code:
3424
3425 (i) The utf8_table... variables are now declared "const".
3426
3427 (ii) The code for \cx, which used the "case flipping" table to upper case
3428 lower case letters, now just substracts 32. This is ASCII-specific,
3429 but the whole concept of \cx is ASCII-specific, so it seems
3430 reasonable.
3431
3432 (iii) PCRE was using its character types table to recognize decimal and
3433 hexadecimal digits in the pattern. This is silly, because it handles
3434 only 0-9, a-f, and A-F, but the character types table is locale-
3435 specific, which means strange things might happen. A private
3436 table is now used for this - though it costs 256 bytes, a table is
3437 much faster than multiple explicit tests. Of course, the standard
3438 character types table is still used for matching digits in subject
3439 strings against \d.
3440
3441 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3442 ending in _t are). So I've renamed it as ESC_tee.
3443
3444 3. The first argument for regexec() in the POSIX wrapper should have been
3445 defined as "const".
3446
3447 4. Changed pcretest to use malloc() for its buffers so that they can be
3448 Electric Fenced for debugging.
3449
3450 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3451 to read one or more bytes before the start of the subject string. Often this
3452 had no effect on PCRE's behaviour, but in some circumstances it could
3453 provoke a segmentation fault.
3454
3455 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3456 to try to read one or more bytes before the start of the subject string.
3457
3458 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3459 UTF-8 support could misbehave in various ways if the subject string
3460 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3461 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3462 back over UTF-8 characters.)
3463
3464
3465 Version 4.2 14-Apr-03
3466 ---------------------
3467
3468 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3469
3470 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3471 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3472 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3473 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3474 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3475 and BUILD_EXEEXT
3476 Note: automatic setting of the BUILD variables is not yet working
3477 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3478 compile-time but not at link-time
3479 [LINK]: use for linking executables only
3480 make different versions for Windows and non-Windows
3481 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3482 libraries
3483 [LINK_FOR_BUILD]: new variable
3484 [OBJEXT]: use throughout
3485 [EXEEXT]: use throughout
3486 <winshared>: new target
3487 <wininstall>: new target
3488 <dftables.o>: use native compiler
3489 <dftables>: use native linker
3490 <install>: handle Windows platform correctly
3491 <clean>: ditto
3492 <check>: ditto
3493 copy DLL to top builddir before testing
3494
3495 As part of these changes, -no-undefined was removed again. This was reported
3496 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3497 in any case.
3498
3499 3. Some tidies to get rid of compiler warnings:
3500
3501 . In the match_data structure, match_limit was an unsigned long int, whereas
3502 match_call_count was an int. I've made them both unsigned long ints.
3503
3504 . In pcretest the fact that a const uschar * doesn't automatically cast to
3505 a void * provoked a warning.
3506
3507 . Turning on some more compiler warnings threw up some "shadow" variables
3508 and a few more missing casts.
3509
3510 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3511 option, a class that contained a single character with a value between 128
3512 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3513
3514 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3515 option, a class that contained several characters, but with at least one
3516 whose value was between 128 and 255 caused PCRE to crash.
3517
3518
3519 Version 4.1 12-Mar-03
3520 ---------------------
3521
3522 1. Compiling with gcc -pedantic found a couple of places where casts were
3523 needed, and a string in dftables.c that was longer than standard compilers are
3524 required to support.
3525
3526 2. Compiling with Sun's compiler found a few more places where the code could
3527 be tidied up in order to avoid warnings.
3528
3529 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3530 first of these names is deprecated in the latest Autoconf in favour of the name
3531 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3532 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3533 analogy I have changed it to CFLAGS_FOR_BUILD.
3534
3535 4. Added -no-undefined to the linking command in the Makefile, because this is
3536 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3537 linking step for the pcreposix library.
3538
3539 5. PCRE was failing to diagnose the case of two named groups with the same
3540 name.
3541
3542 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3543 literal character that is needed in the subject for a match, and scans along to
3544 ensure that it is present before embarking on the full matching process. This
3545 saves time in cases of nested unlimited repeats that are never going to match.
3546 Problem: the scan can take a lot of time if the subject is very long (e.g.
3547 megabytes), thus penalizing straightforward matches. It is now done only if the
3548 amount of subject to be scanned is less than 1000 bytes.
3549
3550 7. A lesser problem with the same optimization is that it was recording the
3551 first character of an anchored pattern as "needed", thus provoking a search
3552 right along the subject, even when the first match of the pattern was going to
3553 fail. The "needed" character is now not set for anchored patterns, unless it
3554 follows something in the pattern that is of non-fixed length. Thus, it still
3555 fulfils its original purpose of finding quick non-matches in cases of nested
3556 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3557
3558
3559 Version 4.0 17-Feb-03
3560 ---------------------
3561
3562 1. If a comment in an extended regex that started immediately after a meta-item
3563 extended to the end of string, PCRE compiled incorrect data. This could lead to
3564 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3565
3566 2. Moved to autoconf 2.53 and libtool 1.4.2.
3567
3568 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3569 the special perltest8 script is no longer needed - all the tests can be run
3570 from a single perltest script.
3571
3572 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3573 by \s. It has now been removed in PCRE. This means it isn't recognized as
3574 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3575 class [:space:] *does* include VT, thereby creating a mess.
3576
3577 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3578 space and tab.
3579
3580 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3581 its new features into the main test script, reducing the number of scripts.
3582
3583 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3584 were backward compatible, and made the (?i) apply to the whole pattern, as if
3585 /i were given. Now it behaves more logically, and applies the option setting
3586 only to what follows. PCRE has been changed to follow suit. However, if it
3587 finds options settings right at the start of the pattern, it extracts them into
3588 the global options, as before. Thus, they show up in the info data.
3589
3590 8. Added support for the \Q...\E escape sequence. Characters in between are
3591 treated as literals. This is slightly different from Perl in that $ and @ are
3592 also handled as literals inside the quotes. In Perl, they will cause variable
3593 interpolation. Note the following examples:
3594
3595 Pattern PCRE matches Perl matches
3596
3597 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3598 \Qabc\$xyz\E abc\$xyz abc\$xyz
3599 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3600
3601 For compatibility with Perl, \Q...\E sequences are recognized inside character
3602 classes as well as outside them.
3603
3604 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3605 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3606 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3607 signed/unsigned warnings.
3608
3609 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3610 option for pcretest, so I've replaced it by a simple function that does just
3611 that job.
3612
3613 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3614 "pcregrep -".
3615
3616 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3617 Java package. This provides some syntactic sugar for simple cases of what my
3618 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3619 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3620 item, you can use this simplified notation. Note that only makes sense with
3621 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3622 greediness, whatever the setting of the PCRE_UNGREEDY option.
3623
3624 13. A change of greediness default within a pattern was not taking effect at
3625 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3626 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3627 was abstracted outside.
3628
3629 14. PCRE now supports the \G assertion. It is true when the current matching
3630 position is at the start point of the match. This differs from \A when the
3631 starting offset is non-zero. Used with the /g option of pcretest (or similar
3632 code), it works in the same way as it does for Perl's /g option. If all
3633 alternatives of a regex begin with \G, the expression is anchored to the start
3634 match position, and the "anchored" flag is set in the compiled expression.
3635
3636 15. Some bugs concerning the handling of certain option changes within patterns
3637 have been fixed. These applied to options other than (?ims). For example,
3638 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3639 been the other way round. Some of this was related to change 7 above.
3640
3641 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3642 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3643 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3644 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3645
3646 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3647 mode. Unlike ".", it always matches newline, whatever the setting of
3648 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3649 assertions. Perl allows it, but it doesn't (in general) work because it can't
3650 calculate the length of the lookbehind. At least, that's the case for Perl
3651 5.8.0 - I've been told they are going to document that it doesn't work in
3652 future.
3653
3654 18. Added an error diagnosis for escapes that PCRE does not support: these are
3655 \L, \l, \N, \P, \p, \U, \u, and \X.
3656
3657 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3658 reading past the end of the pattern in cases such as /[abcd/.
3659
3660 20. PCRE was getting more memory than necessary for patterns with classes that
3661 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3662
3663 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3664 compiling PCRE for use with Virtual Pascal.
3665
3666 22. Small fix to the Makefile to make it work properly if the build is done
3667 outside the source tree.
3668
3669 23. Added a new extension: a condition to go with recursion. If a conditional
3670 subpattern starts with (?(R) the "true" branch is used if recursion has
3671 happened, whereas the "false" branch is used only at the top level.
3672
3673 24. When there was a very long string of literal characters (over 255 bytes
3674 without UTF support, over 250 bytes with UTF support), the computation of how
3675 much memory was required could be incorrect, leading to segfaults or other
3676 strange effects.
3677
3678 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3679 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3680 there was a subsequent back reference to those brackets. This meant that, for
3681 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3682 possible to check for precisely this case. All we can do is abandon the
3683 optimization if .* occurs inside capturing brackets when there are any back
3684 references whatsoever. (See below for a better fix that came later.)
3685
3686 26. The handling of the optimization for finding the first character of a
3687 non-anchored pattern, and for finding a character that is required later in the
3688 match were failing in some cases. This didn't break the matching; it just
3689 failed to optimize when it could. The way this is done has been re-implemented.
3690
3691 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3692
3693 28. Added a new feature that provides some of the functionality that Perl
3694 provides with (?{...}). The facility is termed a "callout". The way it is done
3695 in PCRE is for the caller to provide an optional function, by setting
3696 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3697 global variable. By default it is unset, which disables all calling out. To get
3698 the function called, the regex must include (?C) at appropriate points. This
3699 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3700 This provides a means of identifying different callout points. When PCRE
3701 reaches such a point in the regex, if pcre_callout has been set, the external
3702 function is called. It is provided with data in a structure called
3703 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3704 matching continues; if it returns a non-zero value, the match at the current
3705 point fails. However, backtracking will occur if possible. [This was changed
3706 later and other features added - see item 49 below.]
3707
3708 29. pcretest is upgraded to test the callout functionality. It provides a
3709 callout function that displays information. By default, it shows the start of
3710 the match and the current position in the text. There are some new data escapes
3711 to vary what happens:
3712
3713 \C+ in addition, show current contents of captured substrings
3714 \C- do not supply a callout function
3715 \C!n return 1 when callout number n is reached
3716 \C!n!m return 1 when callout number n is reached for the mth time
3717
3718 30. If pcregrep was called with the -l option and just a single file name, it
3719 output "<stdin>" if a match was found, instead of the file name.
3720
3721 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3722 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3723 pcre_exec(). This saves a malloc/free per call. The default value of
3724 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3725 when configuring.
3726
3727 32. The default maximum size of a compiled pattern is 64K. There have been a
3728 few cases of people hitting this limit. The code now uses macros to handle the
3729 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3730 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3731 configuring. Tests 2 and 5 work only with 2-byte links because they output
3732 debugging information about compiled patterns.
3733
3734 33. Internal code re-arrangements:
3735
3736 (a) Moved the debugging function for printing out a compiled regex into
3737 its own source file (printint.c) and used #include to pull it into
3738 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3739 separate copies.
3740
3741 (b) Defined the list of op-code names for debugging as a macro in
3742 internal.h so that it is next to the definition of the opcodes.
3743
3744 (c) Defined a table of op-code lengths for simpler skipping along compiled
3745 code. This is again a macro in internal.h so that it is next to the
3746 definition of the opcodes.
3747
3748 34. Added support for recursive calls to individual subpatterns, along the
3749 lines of Robin Houston's patch (but implemented somewhat differently).
3750
3751 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3752 allow it to read and process whole directories in Win32. This code was
3753 contributed by Lionel Fourquaux; it has not been tested by me.
3754
3755 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3756 used to name a group. Names consist of alphanumerics and underscores, and must
3757 be unique. Back references use the syntax (?P=name) and recursive calls use
3758 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3759 numbers. The function pcre_fullinfo() can be used after compilation to extract
3760 a name/number map. There are three relevant calls:
3761
3762 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3763 PCRE_INFO_NAMECOUNT yields the number of entries
3764 PCRE_INFO_NAMETABLE yields a pointer to the map.
3765
3766 The map is a vector of fixed-size entries. The size of each entry depends on
3767 the length of the longest name used. The first two bytes of each entry are the
3768 group number, most significant byte first. There follows the corresponding
3769 name, zero terminated. The names are in alphabetical order.
3770
3771 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3772 case instead of 255. Making it the same both with and without UTF-8 support
3773 means that the same test output works with both.
3774
3775 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3776 calling malloc() with a zero argument.
3777
3778 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3779 optimization. I've improved things by keeping a bitmap of backreferences with
3780 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3781 fact referenced, the optimization can be applied. It is unlikely that a
3782 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3783 the match to follow \n) will appear inside brackets with a number greater than
3784 31, but if it does, any back reference > 31 suppresses the optimization.
3785
3786 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3787 of disabling numbered capturing parentheses. Any opening parenthesis that is
3788 not followed by ? behaves as if it were followed by ?: but named parentheses
3789 can still be used for capturing (and they will acquire numbers in the usual
3790 way).
3791
3792 41. Redesigned the return codes from the match() function into yes/no/error so
3793 that errors can be passed back from deep inside the nested calls. A malloc
3794 failure while inside a recursive subpattern call now causes the
3795 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3796
3797 42. It is now possible to set a limit on the number of times the match()
3798 function is called in a call to pcre_exec(). This facility makes it possible to
3799 limit the amount of recursion and backtracking, though not in a directly
3800 obvious way, because the match() function is used in a number of different
3801 circumstances. The count starts from zero for each position in the subject
3802 string (for non-anchored patterns). The default limit is, for compatibility, a
3803 large number, namely 10 000 000. You can change this in two ways:
3804
3805 (a) When configuring PCRE before making, you can use --with-match-limit=n
3806 to set a default value for the compiled library.
3807
3808 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3809 a different value is set. See 45 below.
3810
3811 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3812
3813 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3814 of things that can be changed at compile time. The first argument specifies
3815 what is wanted and the second points to where the information is to be placed.
3816 The current list of available information is:
3817
3818 PCRE_CONFIG_UTF8
3819
3820 The output is an integer that is set to one if UTF-8 support is available;
3821 otherwise it is set to zero.
3822
3823 PCRE_CONFIG_NEWLINE
3824
3825 The output is an integer that it set to the value of the code that is used for
3826 newline. It is either LF (10) or CR (13).
3827
3828 PCRE_CONFIG_LINK_SIZE
3829
3830 The output is an integer that contains the number of bytes used for internal
3831 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3832
3833 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3834
3835 The output is an integer that contains the threshold above which the POSIX
3836 interface uses malloc() for output vectors. See item 31 above.
3837
3838 PCRE_CONFIG_MATCH_LIMIT
3839
3840 The output is an unsigned integer that contains the default limit of the number
3841 of match() calls in a pcre_exec() execution. See 42 above.
3842
3843 44. pcretest has been upgraded by the addition of the -C option. This causes it
3844 to extract all the available output from the new pcre_config() function, and to
3845 output it. The program then exits immediately.
3846
3847 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3848 order to support additional features. One way would have been to define
3849 pcre_exec2() (for example) with extra arguments, but this would not have been
3850 extensible, and would also have required all calls to the original function to
3851 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3852 is used for passing in "extra" data from pcre_study().
3853
3854 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3855 contains the following fields:
3856
3857 flags a bitmap indicating which of the following fields are set
3858 study_data opaque data from pcre_study()
3859 match_limit a way of specifying a limit on match() calls for a specific
3860 call to pcre_exec()
3861 callout_data data for callouts (see 49 below)
3862
3863 The flag bits are also defined in pcre.h, and are
3864
3865 PCRE_EXTRA_STUDY_DATA
3866 PCRE_EXTRA_MATCH_LIMIT
3867 PCRE_EXTRA_CALLOUT_DATA
3868
3869 The pcre_study() function now returns one of these new pcre_extra blocks, with
3870 the actual study data pointed to by the study_data field, and the
3871 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3872 before. That is, this change is entirely upwards-compatible and requires no
3873 change to existing code.
3874
3875 If you want to pass in additional data to pcre_exec(), you can either place it
3876 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3877 block.
3878
3879 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3880 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3881 times with different match limits, until it finds the minimum value needed for
3882 pcre_exec() to complete. The value is then output. This can be instructive; for
3883 most simple matches the number is quite small, but for pathological cases it
3884 gets very large very quickly.
3885
3886 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3887 returns the size of the data block pointed to by the study_data field in a
3888 pcre_extra block, that is, the value that was passed as the argument to
3889 pcre_malloc() when PCRE was getting memory in which to place the information
3890 created by pcre_study(). The fourth argument should point to a size_t variable.
3891 pcretest has been extended so that this information is shown after a successful
3892 pcre_study() call when information about the compiled regex is being displayed.
3893
3894 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3895 because what follows is always an absolute path. (Later: it turns out that this
3896 is more than cosmetic for MinGW, because it doesn't like empty path
3897 components.)
3898
3899 49. Some changes have been made to the callout feature (see 28 above):
3900
3901 (i) A callout function now has three choices for what it returns:
3902
3903 0 => success, carry on matching
3904 > 0 => failure at this point, but backtrack if possible
3905 < 0 => serious error, return this value from pcre_exec()
3906
3907 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3908 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3909 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3910 use by callout functions. It will never be used by PCRE itself.
3911
3912 (ii) The pcre_extra structure (see 45 above) has a void * field called
3913 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3914 pcre_callout_block structure has a field of the same name. The contents of
3915 the field passed in the pcre_extra structure are passed to the callout
3916 function in the corresponding field in the callout block. This makes it
3917 easier to use the same callout-containing regex from multiple threads. For
3918 testing, the pcretest program has a new data escape
3919
3920 \C*n pass the number n (may be negative) as callout_data
3921
3922 If the callout function in pcretest receives a non-zero value as
3923 callout_data, it returns that value.
3924
3925 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3926 there were some redundant $(CFLAGS) in commands that are now specified as
3927 $(LINK), which already includes $(CFLAGS).
3928
3929 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3930 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3931 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3932 one-byte characters throughout. Note that case-insensitive matching applies
3933 only to characters whose values are less than 256. PCRE doesn't support the
3934 notion of cases for higher-valued characters.
3935
3936 (i) A character class whose characters are all within 0-255 is handled as
3937 a bit map, and the map is inverted for negative classes. Previously, a
3938 character > 255 always failed to match such a class; however it should
3939 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3940
3941 (ii) A negated character class with a single character < 255 is coded as
3942 "not this character" (OP_NOT). This wasn't working properly when the test
3943 character was multibyte, either singly or repeated.
3944
3945 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3946 mode, for example: \x{100}{2,3}.
3947
3948 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3949 singly or repeated) now correctly test multibyte characters. However,
3950 PCRE doesn't recognize any characters with values greater than 255 as
3951 digits, spaces, or word characters. Such characters always match \D, \S,
3952 and \W, and never match \d, \s, or \w.
3953
3954 (v) Classes may now contain characters and character ranges with values
3955 greater than 255. For example: [ab\x{100}-\x{400}].
3956
3957 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3958 PCRE in UTF-8 mode.
3959
3960 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3961 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3962 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3963 value.)
3964
3965 53. The single man page has become too large. I have therefore split it up into
3966 a number of separate man pages. These also give rise to individual HTML pages;
3967 these are now put in a separate directory, and there is an index.html page that
3968 lists them all. Some hyperlinking between the pages has been installed.
3969
3970 54. Added convenience functions for handling named capturing parentheses.
3971
3972 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3973 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3974 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3975 are faulted.
3976
3977 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3978 calling configure. These values are used when compiling the dftables.c program
3979 which is run to generate the source of the default character tables. They
3980 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3981 you will need to set these values.
3982
3983 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3984
3985
3986 Version 3.9 02-Jan-02
3987 ---------------------
3988
3989 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3990
3991 2. If --disable-static was given, the building process failed when trying to
3992 build pcretest and pcregrep. (For some reason it was using libtool to compile
3993 them, which is not right, as they aren't part of the library.)
3994
3995
3996 Version 3.8 18-Dec-01
3997 ---------------------
3998
3999 1. The experimental UTF-8 code was completely screwed up. It was packing the
4000 bytes in the wrong order. How dumb can you get?
4001
4002
4003 Version 3.7 29-Oct-01
4004 ---------------------
4005
4006 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4007 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4008 this didn't happen under Solaris 8, where I normally test things.
4009
4010 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4011 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4012 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4013 this fix an hour or so after the initial 3.7 release.)
4014
4015
4016 Version 3.6 23-Oct-01
4017 ---------------------
4018
4019 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4020 offsets passed as NULL with zero offset count.
4021
4022 2. The config.guess and config.sub files had not been updated when I moved to
4023 the latest autoconf.
4024
4025
4026 Version 3.5 15-Aug-01
4027 ---------------------
4028
4029 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4030 had been forgotten.
4031
4032 2. By using declared but undefined structures, we can avoid using "void"
4033 definitions in pcre.h while keeping the internal definitions of the structures
4034 private.
4035
4036 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4037 user point of view, this means that both static and shared libraries are built
4038 by default, but this can be individually controlled. More of the work of
4039 handling this static/shared cases is now inside libtool instead of PCRE's make
4040 file.
4041
4042 4. The pcretest utility is now installed along with pcregrep because it is
4043 useful for users (to test regexs) and by doing this, it automatically gets
4044 relinked by libtool. The documentation has been turned into a man page, so
4045 there are now .1, .txt, and .html versions in /doc.
4046
4047 5. Upgrades to pcregrep:
4048 (i) Added long-form option names like gnu grep.
4049 (ii) Added --help to list all options with an explanatory phrase.
4050 (iii) Added -r, --recursive to recurse into sub-directories.
4051 (iv) Added -f, --file to read patterns from a file.
4052
4053 6. pcre_exec() was referring to its "code" argument before testing that
4054 argument for NULL (and giving an error if it was NULL).
4055
4056 7. Upgraded Makefile.in to allow for compiling in a different directory from
4057 the source directory.
4058
4059 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4060 options bits, the pointer it was passed was to an int instead of to an unsigned
4061 long int. This mattered only on 64-bit systems.
4062
4063 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4064 generated) instead of pcre.in, which it its source. Also made the same change
4065 in several of the .c files.
4066
4067 10. A new release of gcc defines printf() as a macro, which broke pcretest
4068 because it had an ifdef in the middle of a string argument for printf(). Fixed
4069 by using separate calls to printf().
4070
4071 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4072 script, to force use of CR or LF instead of \n in the source. On non-Unix
4073 systems, the value can be set in config.h.
4074
4075 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4076 absolute limit. Changed the text of the error message to make this clear, and
4077 likewise updated the man page.
4078
4079 13. The limit of 99 on the number of capturing subpatterns has been removed.
4080 The new limit is 65535, which I hope will not be a "real" limit.
4081
4082
4083 Version 3.4 22-Aug-00
4084 ---------------------
4085
4086 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4087
4088 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4089
4090
4091 Version 3.3 01-Aug-00
4092 ---------------------
4093
4094 1. If an octal character was given, but the value was greater than \377, it
4095 was not getting masked to the least significant bits, as documented. This could
4096 lead to crashes in some systems.
4097
4098 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4099 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4100
4101 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4102 These just pass their arguments on to (pcre_free)(), but they are provided
4103 because some uses of PCRE bind it to non-C systems that can call its functions,
4104 but cannot call free() or pcre_free() directly.
4105
4106 4. Add "make test" as a synonym for "make check". Corrected some comments in
4107 the Makefile.
4108
4109 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4110 Makefile.
4111
4112 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4113 command called pgrep for grepping around the active processes.
4114
4115 7. Added the beginnings of support for UTF-8 character strings.
4116
4117 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4118 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4119 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4120 out for the ar command.)
4121
4122
4123 Version 3.2 12-May-00
4124 ---------------------
4125
4126 This is purely a bug fixing release.
4127
4128 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4129 of ZA. This was just one example of several cases that could provoke this bug,
4130 which was introduced by change 9 of version 2.00. The code for breaking
4131 infinite loops after an iteration that matches an empty string was't working
4132 correctly.
4133
4134 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4135 when matched against abbab (for example). After matching an empty string, it
4136 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4137 caused it to match further down the string than it should.
4138
4139 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4140 was there because it doesn't seem to be needed, and it causes trouble on some
4141 systems, as it is not a Standard C header. It has been removed.
4142
4143 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4144 were reported on the Macintosh. The changes were from
4145
4146 while ((c = *(++ptr)) != 0 && c != '\n');
4147 to
4148 while ((c = *(++ptr)) != 0 && c != '\n') ;
4149
4150 Totally extraordinary, but if that's what it takes...
4151
4152 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4153 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4154 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4155 assumes the way PCRE uses memmove() (always moving upwards).
4156
4157 6. PCRE is being used in one environment where strchr() is not available. There
4158 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4159 faster code anyway.
4160
4161
4162 Version 3.1 09-Feb-00
4163 ---------------------
4164
4165 The only change in this release is the fixing of some bugs in Makefile.in for
4166 the "install" target:
4167
4168 (1) It was failing to install pcreposix.h.
4169
4170 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4171
4172
4173 Version 3.0 01-Feb-00
4174 ---------------------
4175
4176 1. Add support for the /+ modifier to perltest (to output $` like it does in
4177 pcretest).
4178
4179 2. Add support for the /g modifier to perltest.
4180
4181 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4182 matches null strings.
4183
4184 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4185 pattern. Perl treats empty patterns specially - it reuses the most recent
4186 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4187 effect.
4188
4189 5. The POSIX interface was broken in that it was just handing over the POSIX
4190 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4191 required a bigger vector, with some working space on the end. This means that
4192 the POSIX wrapper now has to get and free some memory, and copy the results.
4193
4194 6. Added some simple autoconf support, placing the test data and the
4195 documentation in separate directories, re-organizing some of the
4196 information files, and making it build pcre-config (a GNU standard). Also added
4197 libtool support for building PCRE as a shared library, which is now the
4198 default.
4199
4200 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4201 09 are not valid octal constants. Single digits will be used for minor values
4202 less than 10.
4203
4204 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4205 existing programs that set these in the POSIX interface can use PCRE without
4206 modification.
4207
4208 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4209 return all that pcre_info() returns, plus additional data. The pcre_info()
4210 function is retained for compatibility, but is considered to be obsolete.
4211
4212 10. Added experimental recursion feature (?R) to handle one common case that
4213 Perl 5.6 will be able to do with (?p{...}).
4214
4215 11. Added support for POSIX character classes like [:alpha:], which Perl is
4216 adopting.
4217
4218
4219 Version 2.08 31-Aug-99
4220 ----------------------
4221
4222 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4223 trying to match at the startoffset position, but instead was moving forward to
4224 the next newline as if a previous match had failed.
4225
4226 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4227 and could get into a loop if a null string was matched other than at the start
4228 of the subject.
4229
4230 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4231 be distinguished at compile time, and for completeness also added PCRE_DATE.
4232
4233 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4234 in GnuWin32 environments.
4235
4236
4237 Version 2.07 29-Jul-99
4238 ----------------------
4239
4240 1. The documentation is now supplied in plain text form and HTML as well as in
4241 the form of man page sources.
4242
4243 2. C++ compilers don't like assigning (void *) values to other pointer types.
4244 In particular this affects malloc(). Although there is no problem in Standard
4245 C, I've put in casts to keep C++ compilers happy.
4246
4247 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4248 should be (const char *).
4249
4250 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4251 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4252 However, I haven't made this a standard facility. The documentation doesn't
4253 mention it, and the Makefile doesn't support it.
4254
4255 5. The Makefile now contains an "install" target, with editable destinations at
4256 the top of the file. The pcretest program is not installed.
4257
4258 6. pgrep -V now gives the PCRE version number and date.
4259
4260 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4261 causing the entire string to be ignored, instead of just the last character.
4262
4263 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4264 non-matching string, it can take a very, very long time, even for strings of
4265 quite modest length, because of the nested recursion. PCRE now does better in
4266 some of these cases. It does this by remembering the last required literal
4267 character in the pattern, and pre-searching the subject to ensure it is present
4268 before running the real match. In other words, it applies a heuristic to detect
4269 some types of certain failure quickly, and in the above example, if presented
4270 with a string that has no trailing " it gives "no match" very quickly.
4271
4272 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4273 other alternatives are tried instead.
4274
4275
4276 Version 2.06 09-Jun-99
4277 ----------------------
4278
4279 1. Change pcretest's output for amount of store used to show just the code
4280 space, because the remainder (the data block) varies in size between 32-bit and
4281 64-bit systems.
4282
4283 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4284 start matching at. This allows lookbehinds to work when searching for multiple
4285 occurrences in a string.
4286
4287 3. Added additional options to pcretest for testing multiple occurrences:
4288
4289 /+ outputs the rest of the string that follows a match
4290 /g loops for multiple occurrences, using the new startoffset argument
4291 /G loops for multiple occurrences by passing an incremented pointer
4292
4293 4. PCRE wasn't doing the "first character" optimization for patterns starting
4294 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4295 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4296 the letter 'x'. On long subject strings, this gives a significant speed-up.
4297
4298
4299 Version 2.05 21-Apr-99
4300 ----------------------
4301
4302 1. Changed the type of magic_number from int to long int so that it works
4303 properly on 16-bit systems.
4304
4305 2. Fixed a bug which caused patterns starting with .* not to work correctly
4306 when the subject string contained newline characters. PCRE was assuming
4307 anchoring for such patterns in all cases, which is not correct because .* will
4308 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4309 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4310 must be retried after every newline in the subject.
4311
4312
4313 Version 2.04 18-Feb-99
4314 ----------------------
4315
4316 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4317 computation of the store needed to hold the pattern was incorrect (too large).
4318 If such patterns were nested a few deep, this could multiply and become a real
4319 problem.
4320
4321 2. Added /M option to pcretest to show the memory requirement of a specific
4322 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4323
4324 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4325 compiled in such a way that the backtracking after subsequent failure was
4326 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4327 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4328
4329
4330 Version 2.03 02-Feb-99
4331 ----------------------
4332
4333 1. Fixed typo and small mistake in man page.
4334
4335 2. Added 4th condition (GPL supersedes if conflict) and created separate
4336 LICENCE file containing the conditions.
4337
4338 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4339 Perl, that is the internal \ allows the delimiter to be included in the
4340 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4341 the final delimiter, add \ to the end of the pattern (to test the error).
4342
4343 4. Added the convenience functions for extracting substrings after a successful
4344 match. Updated pcretest to make it able to test these functions.
4345
4346
4347 Version 2.02 14-Jan-99
4348 ----------------------
4349
4350 1. Initialized the working variables associated with each extraction so that
4351 their saving and restoring doesn't refer to uninitialized store.
4352
4353 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4354 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4355 fix the problem.
4356
4357 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4358 calls, and wasn't printing the correct value for compiling calls. Increased the
4359 default value of LOOPREPEAT, and the number of significant figures in the
4360 times.
4361
4362 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4363
4364 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4365 a building problem on Windows NT with a FAT file system.
4366
4367
4368 Version 2.01 21-Oct-98
4369 ----------------------
4370
4371 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4372 to character tables built by pcre_maketables() in the current locale. If NULL
4373 is passed, the default tables are used.
4374
4375
4376 Version 2.00 24-Sep-98
4377 ----------------------
4378
4379 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4380 it any more.
4381
4382 2. Allow quantification of (?>) groups, and make it work correctly.
4383
4384 3. The first character computation wasn't working for (?>) groups.
4385
4386 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4387 end of the subject) and add 5.005's \z, which really does match only at the
4388 very end of the subject.
4389
4390 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4391
4392 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4393 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4394 localized options. All options to pcre_study() were also removed.
4395
4396 7. Add other new features from 5.005:
4397
4398 $(?<= positive lookbehind
4399 $(?<! negative lookbehind
4400 (?imsx-imsx) added the unsetting capability
4401 such a setting is global if at outer level; local otherwise
4402 (?imsx-imsx:) non-capturing groups with option setting
4403 (?(cond)re|re) conditional pattern matching
4404
4405 A backreference to itself in a repeated group matches the previous
4406 captured string.
4407
4408 8. General tidying up of studying (both automatic and via "study")
4409 consequential on the addition of new assertions.
4410
4411 9. As in 5.005, unlimited repeated groups that could match an empty substring
4412 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4413 runtime if any iteration does actually match an empty substring.
4414
4415 10. Include the RunTest script in the distribution.
4416
4417 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4418 discrepancies, some of which were old and were also with respect to 5.004. They
4419 have now been fixed.
4420
4421
4422 Version 1.09 28-Apr-98
4423 ----------------------
4424
4425 1. A negated single character class followed by a quantifier with a minimum
4426 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4427 program crashes, or just wrong answers. This did not apply to negated classes
4428 containing more than one character, or to minima other than one.
4429
4430
4431 Version 1.08 27-Mar-98
4432 ----------------------
4433
4434 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4435
4436 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4437 latter must appear before anything that relies on it in the pattern.
4438
4439
4440 Version 1.07 16-Feb-98
4441 ----------------------
4442
4443 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4444 repeat of a potentially empty string).
4445
4446
4447 Version 1.06 23-Jan-98
4448 ----------------------
4449
4450 1. Added Markus Oberhumer's little patches for C++.
4451
4452 2. Literal strings longer than 255 characters were broken.
4453
4454
4455 Version 1.05 23-Dec-97
4456 ----------------------
4457
4458 1. Negated character classes containing more than one character were failing if
4459 PCRE_CASELESS was set at run time.
4460
4461
4462 Version 1.04 19-Dec-97
4463 ----------------------
4464
4465 1. Corrected the man page, where some "const" qualifiers had been omitted.
4466
4467 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4468 input syntax.
4469
4470 3. Fixed memory leak which occurred when a regex with back references was
4471 matched with an offsets vector that wasn't big enough. The temporary memory
4472 that is used in this case wasn't being freed if the match failed.
4473
4474 4. Tidied pcretest to ensure it frees memory that it gets.
4475
4476 5. Temporary memory was being obtained in the case where the passed offsets
4477 vector was exactly big enough.
4478
4479 6. Corrected definition of offsetof() from change 5 below.
4480
4481 7. I had screwed up change 6 below and broken the rules for the use of
4482 setjmp(). Now fixed.
4483
4484
4485 Version 1.03 18-Dec-97
4486 ----------------------
4487
4488 1. A erroneous regex with a missing opening parenthesis was correctly
4489 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4490 on some systems.
4491
4492 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4493 it was reported that one broken compiler failed on the former because "code" is
4494 also an independent variable.
4495
4496 3. The erroneous regex a[]b caused an array overrun reference.
4497
4498 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4499 fail on data ending with that character. (It was going on too far, and checking
4500 the next character, typically a binary zero.) This was specific to the
4501 optimized code for single-character negative classes.
4502
4503 5. Added a contributed patch from the TIN world which does the following:
4504
4505 + Add an undef for memmove, in case the the system defines a macro for it.
4506
4507 + Add a definition of offsetof(), in case there isn't one. (I don't know
4508 the reason behind this - offsetof() is part of the ANSI standard - but
4509 it does no harm).
4510
4511 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4512 most of the places where whitespace preceded '#'. I have given up and
4513 allowed the remaining 2 cases to be at the margin.
4514
4515 + Rename some variables in pcre to eliminate shadowing. This seems very
4516 pedantic, but does no harm, of course.
4517
4518 6. Moved the call to setjmp() into its own function, to get rid of warnings
4519 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4520
4521 7. Constructs such as \d{8,} were compiling into the equivalent of
4522 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4523 outcome, but in this particular case used more store than had been allocated,
4524 which caused the bug to be discovered because it threw up an internal error.
4525
4526 8. The debugging code in both pcre and pcretest for outputting the compiled
4527 form of a regex was going wrong in the case of back references followed by
4528 curly-bracketed repeats.
4529
4530
4531 Version 1.02 12-Dec-97
4532 ----------------------
4533
4534 1. Typos in pcre.3 and comments in the source fixed.
4535
4536 2. Applied a contributed patch to get rid of places where it used to remove
4537 'const' from variables, and fixed some signed/unsigned and uninitialized
4538 variable warnings.
4539
4540 3. Added the "runtest" target to Makefile.
4541
4542 4. Set default compiler flag to -O2 rather than just -O.
4543
4544
4545 Version 1.01 19-Nov-97
4546 ----------------------
4547
4548 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4549 like /([ab]*)*/, that is, for classes with more than one character in them.
4550
4551 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4552 as /((?>a*))*/ (a PCRE_EXTRA facility).
4553
4554
4555 Version 1.00 18-Nov-97
4556 ----------------------
4557
4558 1. Added compile-time macros to support systems such as SunOS4 which don't have
4559 memmove() or strerror() but have other things that can be used instead.
4560
4561 2. Arranged that "make clean" removes the executables.
4562
4563
4564 Version 0.99 27-Oct-97
4565 ----------------------
4566
4567 1. Fixed bug in code for optimizing classes with only one character. It was
4568 initializing a 32-byte map regardless, which could cause it to run off the end
4569 of the memory it had got.
4570
4571 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4572
4573
4574 Version 0.98 22-Oct-97
4575 ----------------------
4576
4577 1. Fixed bug in code for handling temporary memory usage when there are more
4578 back references than supplied space in the ovector. This could cause segfaults.
4579
4580
4581 Version 0.97 21-Oct-97
4582 ----------------------
4583
4584 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4585
4586 2. Optimized negated single characters not to use a bit map.
4587
4588 3. Brought error texts together as macro definitions; clarified some of them;
4589 fixed one that was wrong - it said "range out of order" when it meant "invalid
4590 escape sequence".
4591
4592 4. Changed some char * arguments to const char *.
4593
4594 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4595
4596 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4597 pcretest.
4598
4599
4600 Version 0.96 16-Oct-97
4601 ----------------------
4602
4603 1. Added a simple "pgrep" utility to the distribution.
4604
4605 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4606 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4607 where "ddd" means "one or more decimal digits".
4608
4609 3. Fixed serious bug. If a pattern had a back reference, but the call to
4610 pcre_exec() didn't supply a large enough ovector to record the related
4611 identifying subpattern, the match always failed. PCRE now remembers the number
4612 of the largest back reference, and gets some temporary memory in which to save
4613 the offsets during matching if necessary, in order to ensure that
4614 backreferences always work.
4615
4616 4. Increased the compatibility with Perl in a number of ways:
4617
4618 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4619 to request this handling. The option can be set at compile or exec time.
4620
4621 (b) $ matches before a terminating newline by default; an option
4622 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4623 mode). The option can be set at compile or exec time.
4624
4625 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4626 the same as Perl's. If the decimal number it represents is less than 10
4627 or there aren't that many previous left capturing parentheses, an octal
4628 escape is read. Inside a character class, it's always an octal escape,
4629 even if it is a single digit.
4630
4631 (d) An escaped but undefined alphabetic character is taken as a literal,
4632 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4633 escapes.
4634
4635 (e) {0} is now permitted. (The previous item is removed from the compiled
4636 pattern).
4637
4638 5. Changed all the names of code files so that the basic parts are no longer
4639 than 10 characters, and abolished the teeny "globals.c" file.
4640
4641 6. Changed the handling of character classes; they are now done with a 32-byte
4642 bit map always.
4643
4644 7. Added the -d and /D options to pcretest to make it possible to look at the
4645 internals of compilation without having to recompile pcre.
4646
4647
4648 Version 0.95 23-Sep-97
4649 ----------------------
4650
4651 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4652 \x20 at the start of a run of normal characters. These were being treated as
4653 real characters, instead of the source characters being re-checked.
4654
4655
4656 Version 0.94 18-Sep-97
4657 ----------------------
4658
4659 1. The functions are now thread-safe, with the caveat that the global variables
4660 containing pointers to malloc() and free() or alternative functions are the
4661 same for all threads.
4662
4663 2. Get pcre_study() to generate a bitmap of initial characters for non-
4664 anchored patterns when this is possible, and use it if passed to pcre_exec().
4665
4666
4667 Version 0.93 15-Sep-97
4668 ----------------------
4669
4670 1. /(b)|(:+)/ was computing an incorrect first character.
4671
4672 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4673 but not actually doing anything yet.
4674
4675 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4676 as Perl does (e.g. [-az] or [az-]).
4677
4678 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4679 all possible positions.
4680
4681 5. Split up into different modules to avoid including unneeded functions in a
4682 compiled binary. However, compile and exec are still in one module. The "study"
4683 function is split off.
4684
4685 6. The character tables are now in a separate module whose source is generated
4686 by an auxiliary program - but can then be edited by hand if required. There are
4687 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4688 toupper() in the code.
4689
4690 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4691 make them global. Abolish the function for setting them, as the caller can now
4692 set them directly.
4693
4694
4695 Version 0.92 11-Sep-97
4696 ----------------------
4697
4698 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4699 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4700
4701 2. Caseless matching was not working in character classes if the characters in
4702 the pattern were in upper case.
4703
4704 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4705
4706 4. Make PCRE_ANCHORED public and accept as a compile option.
4707
4708 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4709 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4710 pass them.
4711
4712 6. Give an error if bad option bits passed at compile or run time.
4713
4714 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4715 pcretest to cause it to pass that flag.
4716
4717 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4718 options, and the first character, if set.
4719
4720 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4721
4722
4723 Version 0.91 10-Sep-97
4724 ----------------------
4725
4726 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4727 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4728
4729 2. PCRE was looping on encountering an indefinitely repeated back reference to
4730 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4731 Perl does - treats the match as successful.
4732
4733 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5